diff --git a/StructurePlan.txt b/StructurePlan.txt index 9240bb2..5377bab 100644 --- a/StructurePlan.txt +++ b/StructurePlan.txt @@ -1,11 +1,14 @@ 1 From AST (with clang lib) generate: -generated.cpp generated.hpp with proper members -Have a cash system which will partially generate things +Add it all to global context and expose it per module. Having as multiple .cpp and .hpp files does feel significantly better performance wise as it would allow to not recompile everything when something is changed and make it modular. In addition we would be able to remove need for registry if generated.hpp file would hold a reference to metadata. The downside is having a lot of files. BUT THREAAADING CAN BE IMPLEMENTED TO MAKE IT PARALLEL. +Add supported list of attributes to our precompiler step, which will give warning based on if it is in there already or not. + +LLVM + clang; parse AST myself + 2 During compilation populate unordered map with type - members Different types and info we need to keep track of: @@ -16,6 +19,7 @@ Different types and info we need to keep track of: 4 Functions: -//-, const, static, virtual and co, pointer to function, 5 Template instances: our concept thing, member functions and variables, allignment, size, name with parameters, 6 Enums: underlying type, scoped?, values + 7 unions? If we implement inheritance reflected_member - we will incur 8*virtual_func_num hidden bytes @@ -25,3 +29,13 @@ Need to figure out which modifiers exactly can be used in the future. Need to add a way to manually add reflections. buffer_string,dynamic string + +Need to include namespace into strings + +For choices: Hash function, hash structure, compile vs runtime containers + +For choices, where to store runtime reflection data - BSS vs Heap vs Stack + +Names are made from recursive parents. No runtime reflected namespaces objects. + +Had multiple parsing iterations. diff --git a/VisualStudio22-All.bat b/VisualStudio22-All.bat index d316591..91b6d8b 100644 --- a/VisualStudio22-All.bat +++ b/VisualStudio22-All.bat @@ -1 +1,2 @@ -rythe-build-system\premake5 --file=rythe-build-system/premake5.lua vs2022 --workspace-name=reflection-experiments \ No newline at end of file +rythe-build-system\premake5 --file=rythe-build-system/premake5.lua vs2022 --workspace-name=reflection-experiments +pause \ No newline at end of file diff --git a/applications/reflection-experiments/.rythe_project b/applications/reflection-experiments/.rythe_project index 3409d83..b93a687 100644 --- a/applications/reflection-experiments/.rythe_project +++ b/applications/reflection-experiments/.rythe_project @@ -2,16 +2,9 @@ local fs = dofile("filesystem.lua") local project = { additional_types = { "test" }, - dependencies = { "public rythe/rythe-standard-library", "private third_party/clang-c:header-only" }, + dependencies = { "public rythe/rythe-standard-library", "private third_party/clang-c:header-only", + "private third_party/xxhash", "public rythe/reflection-core"}, additional_link_targets = { "libclang" }, } -function project.init(proj, context) - local targetDir = fs.sanitize(fs.parentPath(_MAIN_SCRIPT_DIR) .. "/build/" .. _ACTION) .. fs.getPathSeperator() - local srcDir = fs.sanitize(fs.parentPath(_MAIN_SCRIPT_DIR) .. "/" .. proj.location .. "/src/test.hpp") - proj.post_build = { "xcopy \"" .. srcDir .. "\" \"" .. targetDir .. "\" /i /r /y /s /c" } - - return proj -end - return project \ No newline at end of file diff --git a/applications/reflection-experiments/src/notes.cpp b/applications/reflection-experiments/notes.cpp similarity index 100% rename from applications/reflection-experiments/src/notes.cpp rename to applications/reflection-experiments/notes.cpp diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_class.cpp b/applications/reflection-experiments/src/reflection_containers/reflected_class.cpp deleted file mode 100644 index fc13f1d..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_class.cpp +++ /dev/null @@ -1 +0,0 @@ -#include "reflected_class.h" diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_class.h b/applications/reflection-experiments/src/reflection_containers/reflected_class.h deleted file mode 100644 index 8a704a3..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_class.h +++ /dev/null @@ -1,6 +0,0 @@ -#pragma once - -struct reflected_class -{ - -}; diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_enum.h b/applications/reflection-experiments/src/reflection_containers/reflected_enum.h deleted file mode 100644 index 6e4b221..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_enum.h +++ /dev/null @@ -1,7 +0,0 @@ -#pragma once -#include "access_modifier.h" - -struct reflected_enum -{ - -}; diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_function.cpp b/applications/reflection-experiments/src/reflection_containers/reflected_function.cpp deleted file mode 100644 index 6c120e3..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_function.cpp +++ /dev/null @@ -1 +0,0 @@ -#include "reflected_function.h" diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_function.h b/applications/reflection-experiments/src/reflection_containers/reflected_function.h deleted file mode 100644 index b1ec84c..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_function.h +++ /dev/null @@ -1,34 +0,0 @@ -#pragma once -#include -#include "../reflection_properties/access_modifier.h" -#include "../reflection_id/reflection_id.h" -#include "rsl/impl/containers/hashed_string.hpp" - -namespace rythe::reflection_containers { - struct reflected_function { - rsl::hashed_string name; - rsl::dynamic_string current_namespace; - reflection_properties::acess_modifier access_modifier; - rsl::dynamic_string attributes[]; - - bool is_static; - bool is_const; - - void* memory_reference; - - reflection_id return_type_id; - rsl::dynamic_array parameters; - - template - reflected_function() { - } - - const reflection_id& get_return_type() const { - return return_type_id; - } - - void* get_value_ptr(void* obj) { - return reinterpret_cast(obj); - } - }; -} \ No newline at end of file diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_global_variable.h b/applications/reflection-experiments/src/reflection_containers/reflected_global_variable.h deleted file mode 100644 index 97df22d..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_global_variable.h +++ /dev/null @@ -1,6 +0,0 @@ -#pragma once - -struct reflected_global_variable -{ - -}; diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_template.h b/applications/reflection-experiments/src/reflection_containers/reflected_template.h deleted file mode 100644 index 68c612f..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_template.h +++ /dev/null @@ -1,6 +0,0 @@ -#pragma once - -struct reflected_template -{ - -}; diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_variable.cpp b/applications/reflection-experiments/src/reflection_containers/reflected_variable.cpp deleted file mode 100644 index f61a30a..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_variable.cpp +++ /dev/null @@ -1 +0,0 @@ -#include "reflected_variable.h" diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_variable.h b/applications/reflection-experiments/src/reflection_containers/reflected_variable.h deleted file mode 100644 index 2095791..0000000 --- a/applications/reflection-experiments/src/reflection_containers/reflected_variable.h +++ /dev/null @@ -1,40 +0,0 @@ -#pragma once -#include -#include "../reflection_properties/access_modifier.h" -#include "../reflection_id/reflection_id.h" -#include "rsl/impl/containers/hashed_string.hpp" - -namespace rythe::reflection_containers { - - struct reflected_variable { - rsl::hashed_string name; - rsl::dynamic_string current_namespace; - reflection_properties::acess_modifier access_modifier; - rsl::dynamic_string attributes[]; - - bool is_static; - bool is_const; - - bool is_array; - int array_size; - - size_t offset; - size_t size; - uint16_t alignment; - - reflection_id type_id; - - template - reflected_variable() - : size(sizeof(T)) { - } - - const reflection_id& GetType() const { - return type_id; - } - - void* GetValuePtr(void* obj) { - return reinterpret_cast(obj) + offset; - } - }; -} diff --git a/applications/reflection-experiments/src/reflection_id/reflection_id.cpp b/applications/reflection-experiments/src/reflection_id/reflection_id.cpp deleted file mode 100644 index 93d7129..0000000 --- a/applications/reflection-experiments/src/reflection_id/reflection_id.cpp +++ /dev/null @@ -1 +0,0 @@ -#include "reflection_id.h" diff --git a/applications/reflection-experiments/src/reflection_id/reflection_id.h b/applications/reflection-experiments/src/reflection_id/reflection_id.h deleted file mode 100644 index 1d74d67..0000000 --- a/applications/reflection-experiments/src/reflection_id/reflection_id.h +++ /dev/null @@ -1,5 +0,0 @@ -#pragma once - -struct reflection_id { - -}; diff --git a/applications/reflection-experiments/src/reflection_properties/access_modifier.h b/applications/reflection-experiments/src/reflection_properties/access_modifier.h deleted file mode 100644 index d42ee5d..0000000 --- a/applications/reflection-experiments/src/reflection_properties/access_modifier.h +++ /dev/null @@ -1,9 +0,0 @@ -#pragma once - -namespace reflection_properties { - enum acess_modifier { - private_access, - protected_access, - public_access - }; -} \ No newline at end of file diff --git a/applications/reflection-experiments/src/source.cpp b/applications/reflection-experiments/src/source.cpp index 2b8c6ac..4c5a64e 100644 --- a/applications/reflection-experiments/src/source.cpp +++ b/applications/reflection-experiments/src/source.cpp @@ -1,4 +1,34 @@ +#include -int main() { +#include + +#include "impl/reflection_parsers/ast_source_parser.h" +#include "impl/reflection_context/reflection_context.h" +#include "impl/reflection_context/reflection_registration_registry.h" +//#include "target/test_generated.hpp" + +int main(int, char* argv[]) +{ + rsl::cli_parser cmdl(argv); + + // any particular reason for unordered_set? + std::unordered_set folders; + + for(auto& param : cmdl.params("fpath")) + { + std::cout << '\t' << param.first << " : " << param.second << '\n'; + folders.insert(param.second); + } + + auto parser = reflection_parsers::ast_source_parser(); + + parser.generate_reflection_files(folders); + reflection_registration_registry::instance().clear(); + //register_reflection_file_0(); + reflection_registration_registry::instance().run_all(); + // lala + reflection_context::instance().print_classes(); + + return 0; } diff --git a/applications/reflection-experiments/src/target/test.hpp b/applications/reflection-experiments/src/target/test.hpp new file mode 100644 index 0000000..87fc1a7 --- /dev/null +++ b/applications/reflection-experiments/src/target/test.hpp @@ -0,0 +1,34 @@ +class very_cool_target_class +{ +public: + int int_variable; + + float variable; + + void void_function(int int_parameter); + + class nested_class + { + public: + nested_class() : value(0) {} + + void set_value(int v) { value = v; } + + int get_value() const { return value; } + private: + int value; + }; + +private: + char char_variable; +}; + +class target_class2 +{ +public: + int int_variable2; + + void void_function2(int int_parameter); +private: + char char_variable2; +}; \ No newline at end of file diff --git a/applications/reflection-experiments/src/target/test_generated.hpp b/applications/reflection-experiments/src/target/test_generated.hpp new file mode 100644 index 0000000..87f64a8 --- /dev/null +++ b/applications/reflection-experiments/src/target/test_generated.hpp @@ -0,0 +1,72 @@ +#include "../runtime_reflection_containers/runtime_reflected_class.h" +#include "../runtime_reflection_containers/runtime_reflected_function.h" +#include "../runtime_reflection_containers/runtime_reflected_variable.h" +#include "../reflection_id/reflection_id.h" +#include "../reflection_context/reflection_registration_registry.h" +#include "../reflection_context/reflection_context.h" +void register_reflection_file_0() +{ + runtime_reflected_class class_14985410343605238495; + class_14985410343605238495.id = reflection_id(1644627251862529318ULL, 13483286640053009272ULL, 14985410343605238495ULL); + class_14985410343605238495.type_spelling = rsl::dynamic_string::from_string_length("very_cool_target_class"); + + runtime_reflected_variable variable_6501503083828029313; + variable_6501503083828029313.id = reflection_id(10933754376540684078ULL, 16350452351519940193ULL, 6501503083828029313ULL); + variable_6501503083828029313.offset = 0; + variable_6501503083828029313.type_spelling = rsl::dynamic_string::from_string_length("int"); + class_14985410343605238495.add_variable(std::move(variable_6501503083828029313)); + + runtime_reflected_variable variable_126163732700301715; + variable_126163732700301715.id = reflection_id(5770744143431503129ULL, 12913254394975708779ULL, 126163732700301715ULL); + variable_126163732700301715.offset = 4; + variable_126163732700301715.type_spelling = rsl::dynamic_string::from_string_length("float"); + class_14985410343605238495.add_variable(std::move(variable_126163732700301715)); + + runtime_reflected_variable variable_17337086075296791468; + variable_17337086075296791468.id = reflection_id(15779825311949492597ULL, 9221959329564397894ULL, 17337086075296791468ULL); + variable_17337086075296791468.offset = 8; + variable_17337086075296791468.type_spelling = rsl::dynamic_string::from_string_length("char"); + class_14985410343605238495.add_variable(std::move(variable_17337086075296791468)); + + runtime_reflected_class class_3168970719907517523; + class_3168970719907517523.id = reflection_id(8627741348740311923ULL, 10691473971882747459ULL, 3168970719907517523ULL); + class_3168970719907517523.type_spelling = rsl::dynamic_string::from_string_length("nested_class"); + + runtime_reflected_variable variable_9809713342160364025; + variable_9809713342160364025.id = reflection_id(785375445157317512ULL, 16350452351519940193ULL, 9809713342160364025ULL); + variable_9809713342160364025.offset = 0; + variable_9809713342160364025.type_spelling = rsl::dynamic_string::from_string_length("int"); + class_3168970719907517523.add_variable(std::move(variable_9809713342160364025)); + + class_14985410343605238495.add_class(std::move(class_3168970719907517523)); + + reflection_context::instance().add_class(std::move(class_14985410343605238495)); + + runtime_reflected_class class_862151746601595251; + class_862151746601595251.id = reflection_id(14404923345981690111ULL, 31932944099034707ULL, 862151746601595251ULL); + class_862151746601595251.type_spelling = rsl::dynamic_string::from_string_length("target_class2"); + + runtime_reflected_variable variable_13153249853553726320; + variable_13153249853553726320.id = reflection_id(4281340778281841492ULL, 16350452351519940193ULL, 13153249853553726320ULL); + variable_13153249853553726320.offset = 0; + variable_13153249853553726320.type_spelling = rsl::dynamic_string::from_string_length("int"); + class_862151746601595251.add_variable(std::move(variable_13153249853553726320)); + + runtime_reflected_variable variable_6526241590453974317; + variable_6526241590453974317.id = reflection_id(10234543199726084352ULL, 9221959329564397894ULL, 6526241590453974317ULL); + variable_6526241590453974317.offset = 4; + variable_6526241590453974317.type_spelling = rsl::dynamic_string::from_string_length("char"); + class_862151746601595251.add_variable(std::move(variable_6526241590453974317)); + + reflection_context::instance().add_class(std::move(class_862151746601595251)); + +}; + +struct reflection_file_registration_helper +{ + reflection_file_registration_helper() + { + reflection_registration_registry::instance().add(®ister_reflection_file_0); + } +}; +static reflection_file_registration_helper registration_instance; \ No newline at end of file diff --git a/applications/reflection-experiments/src/test.hpp b/applications/reflection-experiments/src/test.hpp deleted file mode 100644 index 8be4a80..0000000 --- a/applications/reflection-experiments/src/test.hpp +++ /dev/null @@ -1,47 +0,0 @@ -#define RYTHE_STRINGIFY(x) #x - -namespace rsl -{ - struct custom_attribute{}; -} - -#define RYTHE_DEFINE_ATTRIBUTE_WITH_PARAMS(Attrib, ...) clang::annotate(RYTHE_STRINGIFY(attributes::Attrib(__VA_ARGS__))) -#define RYTHE_DEFINE_ATTRIBUTE(Attrib) clang::annotate(RYTHE_STRINGIFY(attributes::Attrib())) - - -namespace attributes -{ - struct my_attrib : public rsl::custom_attribute - { - float f; - int i; - my_attrib(float flt, int integ) : f(flt), i(integ) {} - }; -} - -#define my_attrib(...) RYTHE_DEFINE_ATTRIBUTE_WITH_PARAMS(my_attrib, __VA_ARGS__) -#define my_other_attrib RYTHE_DEFINE_ATTRIBUTE(my_other_attrib) - -namespace rsl -{ - struct [[my_attrib(0.5f, 2)]] MyStruct - { - int value; - [[my_other_attrib]] bool b; - - float my_func([[maybe_unused]] float f) { return 0.4f; } - void my_other_func() {} - }; - - struct OtherStruct - { - struct ChildStruct{}; - float f; - }; -} - -struct GlobalNamespaceStruct -{ - bool bb; - unsigned uInt; -}; \ No newline at end of file diff --git a/applications/reflection-experiments/third_party/blake3/blake3.c b/applications/reflection-experiments/third_party/blake3/blake3.c new file mode 100644 index 0000000..74fc2e7 --- /dev/null +++ b/applications/reflection-experiments/third_party/blake3/blake3.c @@ -0,0 +1,650 @@ +#include +#include +#include + +#include "blake3.h" +#include "blake3_impl.h" + +const char *blake3_version(void) { return BLAKE3_VERSION_STRING; } + +INLINE void chunk_state_init(blake3_chunk_state *self, const uint32_t key[8], + uint8_t flags) { + memcpy(self->cv, key, BLAKE3_KEY_LEN); + self->chunk_counter = 0; + memset(self->buf, 0, BLAKE3_BLOCK_LEN); + self->buf_len = 0; + self->blocks_compressed = 0; + self->flags = flags; +} + +INLINE void chunk_state_reset(blake3_chunk_state *self, const uint32_t key[8], + uint64_t chunk_counter) { + memcpy(self->cv, key, BLAKE3_KEY_LEN); + self->chunk_counter = chunk_counter; + self->blocks_compressed = 0; + memset(self->buf, 0, BLAKE3_BLOCK_LEN); + self->buf_len = 0; +} + +INLINE size_t chunk_state_len(const blake3_chunk_state *self) { + return (BLAKE3_BLOCK_LEN * (size_t)self->blocks_compressed) + + ((size_t)self->buf_len); +} + +INLINE size_t chunk_state_fill_buf(blake3_chunk_state *self, + const uint8_t *input, size_t input_len) { + size_t take = BLAKE3_BLOCK_LEN - ((size_t)self->buf_len); + if (take > input_len) { + take = input_len; + } + uint8_t *dest = self->buf + ((size_t)self->buf_len); + memcpy(dest, input, take); + self->buf_len += (uint8_t)take; + return take; +} + +INLINE uint8_t chunk_state_maybe_start_flag(const blake3_chunk_state *self) { + if (self->blocks_compressed == 0) { + return CHUNK_START; + } else { + return 0; + } +} + +typedef struct { + uint32_t input_cv[8]; + uint64_t counter; + uint8_t block[BLAKE3_BLOCK_LEN]; + uint8_t block_len; + uint8_t flags; +} output_t; + +INLINE output_t make_output(const uint32_t input_cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, + uint8_t flags) { + output_t ret; + memcpy(ret.input_cv, input_cv, 32); + memcpy(ret.block, block, BLAKE3_BLOCK_LEN); + ret.block_len = block_len; + ret.counter = counter; + ret.flags = flags; + return ret; +} + +// Chaining values within a given chunk (specifically the compress_in_place +// interface) are represented as words. This avoids unnecessary bytes<->words +// conversion overhead in the portable implementation. However, the hash_many +// interface handles both user input and parent node blocks, so it accepts +// bytes. For that reason, chaining values in the CV stack are represented as +// bytes. +INLINE void output_chaining_value(const output_t *self, uint8_t cv[32]) { + uint32_t cv_words[8]; + memcpy(cv_words, self->input_cv, 32); + blake3_compress_in_place(cv_words, self->block, self->block_len, + self->counter, self->flags); + store_cv_words(cv, cv_words); +} + +INLINE void output_root_bytes(const output_t *self, uint64_t seek, uint8_t *out, + size_t out_len) { + if (out_len == 0) { + return; + } + uint64_t output_block_counter = seek / 64; + size_t offset_within_block = seek % 64; + uint8_t wide_buf[64]; + if(offset_within_block) { + blake3_compress_xof(self->input_cv, self->block, self->block_len, output_block_counter, self->flags | ROOT, wide_buf); + const size_t available_bytes = 64 - offset_within_block; + const size_t bytes = out_len > available_bytes ? available_bytes : out_len; + memcpy(out, wide_buf + offset_within_block, bytes); + out += bytes; + out_len -= bytes; + output_block_counter += 1; + } + if(out_len / 64) { + blake3_xof_many(self->input_cv, self->block, self->block_len, output_block_counter, self->flags | ROOT, out, out_len / 64); + } + output_block_counter += out_len / 64; + out += out_len & -64; + out_len -= out_len & -64; + if(out_len) { + blake3_compress_xof(self->input_cv, self->block, self->block_len, output_block_counter, self->flags | ROOT, wide_buf); + memcpy(out, wide_buf, out_len); + } +} + +INLINE void chunk_state_update(blake3_chunk_state *self, const uint8_t *input, + size_t input_len) { + if (self->buf_len > 0) { + size_t take = chunk_state_fill_buf(self, input, input_len); + input += take; + input_len -= take; + if (input_len > 0) { + blake3_compress_in_place( + self->cv, self->buf, BLAKE3_BLOCK_LEN, self->chunk_counter, + self->flags | chunk_state_maybe_start_flag(self)); + self->blocks_compressed += 1; + self->buf_len = 0; + memset(self->buf, 0, BLAKE3_BLOCK_LEN); + } + } + + while (input_len > BLAKE3_BLOCK_LEN) { + blake3_compress_in_place(self->cv, input, BLAKE3_BLOCK_LEN, + self->chunk_counter, + self->flags | chunk_state_maybe_start_flag(self)); + self->blocks_compressed += 1; + input += BLAKE3_BLOCK_LEN; + input_len -= BLAKE3_BLOCK_LEN; + } + + chunk_state_fill_buf(self, input, input_len); +} + +INLINE output_t chunk_state_output(const blake3_chunk_state *self) { + uint8_t block_flags = + self->flags | chunk_state_maybe_start_flag(self) | CHUNK_END; + return make_output(self->cv, self->buf, self->buf_len, self->chunk_counter, + block_flags); +} + +INLINE output_t parent_output(const uint8_t block[BLAKE3_BLOCK_LEN], + const uint32_t key[8], uint8_t flags) { + return make_output(key, block, BLAKE3_BLOCK_LEN, 0, flags | PARENT); +} + +// Given some input larger than one chunk, return the number of bytes that +// should go in the left subtree. This is the largest power-of-2 number of +// chunks that leaves at least 1 byte for the right subtree. +INLINE size_t left_subtree_len(size_t input_len) { + // Subtract 1 to reserve at least one byte for the right side. input_len + // should always be greater than BLAKE3_CHUNK_LEN. + size_t full_chunks = (input_len - 1) / BLAKE3_CHUNK_LEN; + return round_down_to_power_of_2(full_chunks) * BLAKE3_CHUNK_LEN; +} + +// Use SIMD parallelism to hash up to MAX_SIMD_DEGREE chunks at the same time +// on a single thread. Write out the chunk chaining values and return the +// number of chunks hashed. These chunks are never the root and never empty; +// those cases use a different codepath. +INLINE size_t compress_chunks_parallel(const uint8_t *input, size_t input_len, + const uint32_t key[8], + uint64_t chunk_counter, uint8_t flags, + uint8_t *out) { +#if defined(BLAKE3_TESTING) + assert(0 < input_len); + assert(input_len <= MAX_SIMD_DEGREE * BLAKE3_CHUNK_LEN); +#endif + + const uint8_t *chunks_array[MAX_SIMD_DEGREE]; + size_t input_position = 0; + size_t chunks_array_len = 0; + while (input_len - input_position >= BLAKE3_CHUNK_LEN) { + chunks_array[chunks_array_len] = &input[input_position]; + input_position += BLAKE3_CHUNK_LEN; + chunks_array_len += 1; + } + + blake3_hash_many(chunks_array, chunks_array_len, + BLAKE3_CHUNK_LEN / BLAKE3_BLOCK_LEN, key, chunk_counter, + true, flags, CHUNK_START, CHUNK_END, out); + + // Hash the remaining partial chunk, if there is one. Note that the empty + // chunk (meaning the empty message) is a different codepath. + if (input_len > input_position) { + uint64_t counter = chunk_counter + (uint64_t)chunks_array_len; + blake3_chunk_state chunk_state; + chunk_state_init(&chunk_state, key, flags); + chunk_state.chunk_counter = counter; + chunk_state_update(&chunk_state, &input[input_position], + input_len - input_position); + output_t output = chunk_state_output(&chunk_state); + output_chaining_value(&output, &out[chunks_array_len * BLAKE3_OUT_LEN]); + return chunks_array_len + 1; + } else { + return chunks_array_len; + } +} + +// Use SIMD parallelism to hash up to MAX_SIMD_DEGREE parents at the same time +// on a single thread. Write out the parent chaining values and return the +// number of parents hashed. (If there's an odd input chaining value left over, +// return it as an additional output.) These parents are never the root and +// never empty; those cases use a different codepath. +INLINE size_t compress_parents_parallel(const uint8_t *child_chaining_values, + size_t num_chaining_values, + const uint32_t key[8], uint8_t flags, + uint8_t *out) { +#if defined(BLAKE3_TESTING) + assert(2 <= num_chaining_values); + assert(num_chaining_values <= 2 * MAX_SIMD_DEGREE_OR_2); +#endif + + const uint8_t *parents_array[MAX_SIMD_DEGREE_OR_2]; + size_t parents_array_len = 0; + while (num_chaining_values - (2 * parents_array_len) >= 2) { + parents_array[parents_array_len] = + &child_chaining_values[2 * parents_array_len * BLAKE3_OUT_LEN]; + parents_array_len += 1; + } + + blake3_hash_many(parents_array, parents_array_len, 1, key, + 0, // Parents always use counter 0. + false, flags | PARENT, + 0, // Parents have no start flags. + 0, // Parents have no end flags. + out); + + // If there's an odd child left over, it becomes an output. + if (num_chaining_values > 2 * parents_array_len) { + memcpy(&out[parents_array_len * BLAKE3_OUT_LEN], + &child_chaining_values[2 * parents_array_len * BLAKE3_OUT_LEN], + BLAKE3_OUT_LEN); + return parents_array_len + 1; + } else { + return parents_array_len; + } +} + +// The wide helper function returns (writes out) an array of chaining values +// and returns the length of that array. The number of chaining values returned +// is the dynamically detected SIMD degree, at most MAX_SIMD_DEGREE. Or fewer, +// if the input is shorter than that many chunks. The reason for maintaining a +// wide array of chaining values going back up the tree, is to allow the +// implementation to hash as many parents in parallel as possible. +// +// As a special case when the SIMD degree is 1, this function will still return +// at least 2 outputs. This guarantees that this function doesn't perform the +// root compression. (If it did, it would use the wrong flags, and also we +// wouldn't be able to implement extendable output.) Note that this function is +// not used when the whole input is only 1 chunk long; that's a different +// codepath. +// +// Why not just have the caller split the input on the first update(), instead +// of implementing this special rule? Because we don't want to limit SIMD or +// multi-threading parallelism for that update(). +size_t blake3_compress_subtree_wide(const uint8_t *input, size_t input_len, + const uint32_t key[8], + uint64_t chunk_counter, uint8_t flags, + uint8_t *out, bool use_tbb) { + // Note that the single chunk case does *not* bump the SIMD degree up to 2 + // when it is 1. If this implementation adds multi-threading in the future, + // this gives us the option of multi-threading even the 2-chunk case, which + // can help performance on smaller platforms. + if (input_len <= blake3_simd_degree() * BLAKE3_CHUNK_LEN) { + return compress_chunks_parallel(input, input_len, key, chunk_counter, flags, + out); + } + + // With more than simd_degree chunks, we need to recurse. Start by dividing + // the input into left and right subtrees. (Note that this is only optimal + // as long as the SIMD degree is a power of 2. If we ever get a SIMD degree + // of 3 or something, we'll need a more complicated strategy.) + size_t left_input_len = left_subtree_len(input_len); + size_t right_input_len = input_len - left_input_len; + const uint8_t *right_input = &input[left_input_len]; + uint64_t right_chunk_counter = + chunk_counter + (uint64_t)(left_input_len / BLAKE3_CHUNK_LEN); + + // Make space for the child outputs. Here we use MAX_SIMD_DEGREE_OR_2 to + // account for the special case of returning 2 outputs when the SIMD degree + // is 1. + uint8_t cv_array[2 * MAX_SIMD_DEGREE_OR_2 * BLAKE3_OUT_LEN]; + size_t degree = blake3_simd_degree(); + if (left_input_len > BLAKE3_CHUNK_LEN && degree == 1) { + // The special case: We always use a degree of at least two, to make + // sure there are two outputs. Except, as noted above, at the chunk + // level, where we allow degree=1. (Note that the 1-chunk-input case is + // a different codepath.) + degree = 2; + } + uint8_t *right_cvs = &cv_array[degree * BLAKE3_OUT_LEN]; + + // Recurse! + size_t left_n = -1; + size_t right_n = -1; + +#if defined(BLAKE3_USE_TBB) + blake3_compress_subtree_wide_join_tbb( + key, flags, use_tbb, + // left-hand side + input, left_input_len, chunk_counter, cv_array, &left_n, + // right-hand side + right_input, right_input_len, right_chunk_counter, right_cvs, &right_n); +#else + left_n = blake3_compress_subtree_wide( + input, left_input_len, key, chunk_counter, flags, cv_array, use_tbb); + right_n = blake3_compress_subtree_wide(right_input, right_input_len, key, + right_chunk_counter, flags, right_cvs, + use_tbb); +#endif // BLAKE3_USE_TBB + + // The special case again. If simd_degree=1, then we'll have left_n=1 and + // right_n=1. Rather than compressing them into a single output, return + // them directly, to make sure we always have at least two outputs. + if (left_n == 1) { + memcpy(out, cv_array, 2 * BLAKE3_OUT_LEN); + return 2; + } + + // Otherwise, do one layer of parent node compression. + size_t num_chaining_values = left_n + right_n; + return compress_parents_parallel(cv_array, num_chaining_values, key, flags, + out); +} + +// Hash a subtree with compress_subtree_wide(), and then condense the resulting +// list of chaining values down to a single parent node. Don't compress that +// last parent node, however. Instead, return its message bytes (the +// concatenated chaining values of its children). This is necessary when the +// first call to update() supplies a complete subtree, because the topmost +// parent node of that subtree could end up being the root. It's also necessary +// for extended output in the general case. +// +// As with compress_subtree_wide(), this function is not used on inputs of 1 +// chunk or less. That's a different codepath. +INLINE void +compress_subtree_to_parent_node(const uint8_t *input, size_t input_len, + const uint32_t key[8], uint64_t chunk_counter, + uint8_t flags, uint8_t out[2 * BLAKE3_OUT_LEN], + bool use_tbb) { +#if defined(BLAKE3_TESTING) + assert(input_len > BLAKE3_CHUNK_LEN); +#endif + + uint8_t cv_array[MAX_SIMD_DEGREE_OR_2 * BLAKE3_OUT_LEN]; + size_t num_cvs = blake3_compress_subtree_wide(input, input_len, key, + chunk_counter, flags, cv_array, use_tbb); + assert(num_cvs <= MAX_SIMD_DEGREE_OR_2); + // The following loop never executes when MAX_SIMD_DEGREE_OR_2 is 2, because + // as we just asserted, num_cvs will always be <=2 in that case. But GCC + // (particularly GCC 8.5) can't tell that it never executes, and if NDEBUG is + // set then it emits incorrect warnings here. We tried a few different + // hacks to silence these, but in the end our hacks just produced different + // warnings (see https://github.com/BLAKE3-team/BLAKE3/pull/380). Out of + // desperation, we ifdef out this entire loop when we know it's not needed. +#if MAX_SIMD_DEGREE_OR_2 > 2 + // If MAX_SIMD_DEGREE_OR_2 is greater than 2 and there's enough input, + // compress_subtree_wide() returns more than 2 chaining values. Condense + // them into 2 by forming parent nodes repeatedly. + uint8_t out_array[MAX_SIMD_DEGREE_OR_2 * BLAKE3_OUT_LEN / 2]; + while (num_cvs > 2) { + num_cvs = + compress_parents_parallel(cv_array, num_cvs, key, flags, out_array); + memcpy(cv_array, out_array, num_cvs * BLAKE3_OUT_LEN); + } +#endif + memcpy(out, cv_array, 2 * BLAKE3_OUT_LEN); +} + +INLINE void hasher_init_base(blake3_hasher *self, const uint32_t key[8], + uint8_t flags) { + memcpy(self->key, key, BLAKE3_KEY_LEN); + chunk_state_init(&self->chunk, key, flags); + self->cv_stack_len = 0; +} + +void blake3_hasher_init(blake3_hasher *self) { hasher_init_base(self, IV, 0); } + +void blake3_hasher_init_keyed(blake3_hasher *self, + const uint8_t key[BLAKE3_KEY_LEN]) { + uint32_t key_words[8]; + load_key_words(key, key_words); + hasher_init_base(self, key_words, KEYED_HASH); +} + +void blake3_hasher_init_derive_key_raw(blake3_hasher *self, const void *context, + size_t context_len) { + blake3_hasher context_hasher; + hasher_init_base(&context_hasher, IV, DERIVE_KEY_CONTEXT); + blake3_hasher_update(&context_hasher, context, context_len); + uint8_t context_key[BLAKE3_KEY_LEN]; + blake3_hasher_finalize(&context_hasher, context_key, BLAKE3_KEY_LEN); + uint32_t context_key_words[8]; + load_key_words(context_key, context_key_words); + hasher_init_base(self, context_key_words, DERIVE_KEY_MATERIAL); +} + +void blake3_hasher_init_derive_key(blake3_hasher *self, const char *context) { + blake3_hasher_init_derive_key_raw(self, context, strlen(context)); +} + +// As described in hasher_push_cv() below, we do "lazy merging", delaying +// merges until right before the next CV is about to be added. This is +// different from the reference implementation. Another difference is that we +// aren't always merging 1 chunk at a time. Instead, each CV might represent +// any power-of-two number of chunks, as long as the smaller-above-larger stack +// order is maintained. Instead of the "count the trailing 0-bits" algorithm +// described in the spec, we use a "count the total number of 1-bits" variant +// that doesn't require us to retain the subtree size of the CV on top of the +// stack. The principle is the same: each CV that should remain in the stack is +// represented by a 1-bit in the total number of chunks (or bytes) so far. +INLINE void hasher_merge_cv_stack(blake3_hasher *self, uint64_t total_len) { + size_t post_merge_stack_len = (size_t)popcnt(total_len); + while (self->cv_stack_len > post_merge_stack_len) { + uint8_t *parent_node = + &self->cv_stack[(self->cv_stack_len - 2) * BLAKE3_OUT_LEN]; + output_t output = parent_output(parent_node, self->key, self->chunk.flags); + output_chaining_value(&output, parent_node); + self->cv_stack_len -= 1; + } +} + +// In reference_impl.rs, we merge the new CV with existing CVs from the stack +// before pushing it. We can do that because we know more input is coming, so +// we know none of the merges are root. +// +// This setting is different. We want to feed as much input as possible to +// compress_subtree_wide(), without setting aside anything for the chunk_state. +// If the user gives us 64 KiB, we want to parallelize over all 64 KiB at once +// as a single subtree, if at all possible. +// +// This leads to two problems: +// 1) This 64 KiB input might be the only call that ever gets made to update. +// In this case, the root node of the 64 KiB subtree would be the root node +// of the whole tree, and it would need to be ROOT finalized. We can't +// compress it until we know. +// 2) This 64 KiB input might complete a larger tree, whose root node is +// similarly going to be the root of the whole tree. For example, maybe +// we have 196 KiB (that is, 128 + 64) hashed so far. We can't compress the +// node at the root of the 256 KiB subtree until we know how to finalize it. +// +// The second problem is solved with "lazy merging". That is, when we're about +// to add a CV to the stack, we don't merge it with anything first, as the +// reference impl does. Instead we do merges using the *previous* CV that was +// added, which is sitting on top of the stack, and we put the new CV +// (unmerged) on top of the stack afterwards. This guarantees that we never +// merge the root node until finalize(). +// +// Solving the first problem requires an additional tool, +// compress_subtree_to_parent_node(). That function always returns the top +// *two* chaining values of the subtree it's compressing. We then do lazy +// merging with each of them separately, so that the second CV will always +// remain unmerged. (That also helps us support extendable output when we're +// hashing an input all-at-once.) +INLINE void hasher_push_cv(blake3_hasher *self, uint8_t new_cv[BLAKE3_OUT_LEN], + uint64_t chunk_counter) { + hasher_merge_cv_stack(self, chunk_counter); + memcpy(&self->cv_stack[self->cv_stack_len * BLAKE3_OUT_LEN], new_cv, + BLAKE3_OUT_LEN); + self->cv_stack_len += 1; +} + +INLINE void blake3_hasher_update_base(blake3_hasher *self, const void *input, + size_t input_len, bool use_tbb) { + // Explicitly checking for zero avoids causing UB by passing a null pointer + // to memcpy. This comes up in practice with things like: + // std::vector v; + // blake3_hasher_update(&hasher, v.data(), v.size()); + if (input_len == 0) { + return; + } + + const uint8_t *input_bytes = (const uint8_t *)input; + + // If we have some partial chunk bytes in the internal chunk_state, we need + // to finish that chunk first. + if (chunk_state_len(&self->chunk) > 0) { + size_t take = BLAKE3_CHUNK_LEN - chunk_state_len(&self->chunk); + if (take > input_len) { + take = input_len; + } + chunk_state_update(&self->chunk, input_bytes, take); + input_bytes += take; + input_len -= take; + // If we've filled the current chunk and there's more coming, finalize this + // chunk and proceed. In this case we know it's not the root. + if (input_len > 0) { + output_t output = chunk_state_output(&self->chunk); + uint8_t chunk_cv[32]; + output_chaining_value(&output, chunk_cv); + hasher_push_cv(self, chunk_cv, self->chunk.chunk_counter); + chunk_state_reset(&self->chunk, self->key, self->chunk.chunk_counter + 1); + } else { + return; + } + } + + // Now the chunk_state is clear, and we have more input. If there's more than + // a single chunk (so, definitely not the root chunk), hash the largest whole + // subtree we can, with the full benefits of SIMD (and maybe in the future, + // multi-threading) parallelism. Two restrictions: + // - The subtree has to be a power-of-2 number of chunks. Only subtrees along + // the right edge can be incomplete, and we don't know where the right edge + // is going to be until we get to finalize(). + // - The subtree must evenly divide the total number of chunks up until this + // point (if total is not 0). If the current incomplete subtree is only + // waiting for 1 more chunk, we can't hash a subtree of 4 chunks. We have + // to complete the current subtree first. + // Because we might need to break up the input to form powers of 2, or to + // evenly divide what we already have, this part runs in a loop. + while (input_len > BLAKE3_CHUNK_LEN) { + size_t subtree_len = round_down_to_power_of_2(input_len); + uint64_t count_so_far = self->chunk.chunk_counter * BLAKE3_CHUNK_LEN; + // Shrink the subtree_len until it evenly divides the count so far. We know + // that subtree_len itself is a power of 2, so we can use a bitmasking + // trick instead of an actual remainder operation. (Note that if the caller + // consistently passes power-of-2 inputs of the same size, as is hopefully + // typical, this loop condition will always fail, and subtree_len will + // always be the full length of the input.) + // + // An aside: We don't have to shrink subtree_len quite this much. For + // example, if count_so_far is 1, we could pass 2 chunks to + // compress_subtree_to_parent_node. Since we'll get 2 CVs back, we'll still + // get the right answer in the end, and we might get to use 2-way SIMD + // parallelism. The problem with this optimization, is that it gets us + // stuck always hashing 2 chunks. The total number of chunks will remain + // odd, and we'll never graduate to higher degrees of parallelism. See + // https://github.com/BLAKE3-team/BLAKE3/issues/69. + while ((((uint64_t)(subtree_len - 1)) & count_so_far) != 0) { + subtree_len /= 2; + } + // The shrunken subtree_len might now be 1 chunk long. If so, hash that one + // chunk by itself. Otherwise, compress the subtree into a pair of CVs. + uint64_t subtree_chunks = subtree_len / BLAKE3_CHUNK_LEN; + if (subtree_len <= BLAKE3_CHUNK_LEN) { + blake3_chunk_state chunk_state; + chunk_state_init(&chunk_state, self->key, self->chunk.flags); + chunk_state.chunk_counter = self->chunk.chunk_counter; + chunk_state_update(&chunk_state, input_bytes, subtree_len); + output_t output = chunk_state_output(&chunk_state); + uint8_t cv[BLAKE3_OUT_LEN]; + output_chaining_value(&output, cv); + hasher_push_cv(self, cv, chunk_state.chunk_counter); + } else { + // This is the high-performance happy path, though getting here depends + // on the caller giving us a long enough input. + uint8_t cv_pair[2 * BLAKE3_OUT_LEN]; + compress_subtree_to_parent_node(input_bytes, subtree_len, self->key, + self->chunk.chunk_counter, + self->chunk.flags, cv_pair, use_tbb); + hasher_push_cv(self, cv_pair, self->chunk.chunk_counter); + hasher_push_cv(self, &cv_pair[BLAKE3_OUT_LEN], + self->chunk.chunk_counter + (subtree_chunks / 2)); + } + self->chunk.chunk_counter += subtree_chunks; + input_bytes += subtree_len; + input_len -= subtree_len; + } + + // If there's any remaining input less than a full chunk, add it to the chunk + // state. In that case, also do a final merge loop to make sure the subtree + // stack doesn't contain any unmerged pairs. The remaining input means we + // know these merges are non-root. This merge loop isn't strictly necessary + // here, because hasher_push_chunk_cv already does its own merge loop, but it + // simplifies blake3_hasher_finalize below. + if (input_len > 0) { + chunk_state_update(&self->chunk, input_bytes, input_len); + hasher_merge_cv_stack(self, self->chunk.chunk_counter); + } +} + +void blake3_hasher_update(blake3_hasher *self, const void *input, + size_t input_len) { + bool use_tbb = false; + blake3_hasher_update_base(self, input, input_len, use_tbb); +} + +#if defined(BLAKE3_USE_TBB) +void blake3_hasher_update_tbb(blake3_hasher *self, const void *input, + size_t input_len) { + bool use_tbb = true; + blake3_hasher_update_base(self, input, input_len, use_tbb); +} +#endif // BLAKE3_USE_TBB + +void blake3_hasher_finalize(const blake3_hasher *self, uint8_t *out, + size_t out_len) { + blake3_hasher_finalize_seek(self, 0, out, out_len); +} + +void blake3_hasher_finalize_seek(const blake3_hasher *self, uint64_t seek, + uint8_t *out, size_t out_len) { + // Explicitly checking for zero avoids causing UB by passing a null pointer + // to memcpy. This comes up in practice with things like: + // std::vector v; + // blake3_hasher_finalize(&hasher, v.data(), v.size()); + if (out_len == 0) { + return; + } + + // If the subtree stack is empty, then the current chunk is the root. + if (self->cv_stack_len == 0) { + output_t output = chunk_state_output(&self->chunk); + output_root_bytes(&output, seek, out, out_len); + return; + } + // If there are any bytes in the chunk state, finalize that chunk and do a + // roll-up merge between that chunk hash and every subtree in the stack. In + // this case, the extra merge loop at the end of blake3_hasher_update + // guarantees that none of the subtrees in the stack need to be merged with + // each other first. Otherwise, if there are no bytes in the chunk state, + // then the top of the stack is a chunk hash, and we start the merge from + // that. + output_t output; + size_t cvs_remaining; + if (chunk_state_len(&self->chunk) > 0) { + cvs_remaining = self->cv_stack_len; + output = chunk_state_output(&self->chunk); + } else { + // There are always at least 2 CVs in the stack in this case. + cvs_remaining = self->cv_stack_len - 2; + output = parent_output(&self->cv_stack[cvs_remaining * 32], self->key, + self->chunk.flags); + } + while (cvs_remaining > 0) { + cvs_remaining -= 1; + uint8_t parent_block[BLAKE3_BLOCK_LEN]; + memcpy(parent_block, &self->cv_stack[cvs_remaining * 32], 32); + output_chaining_value(&output, &parent_block[32]); + output = parent_output(parent_block, self->key, self->chunk.flags); + } + output_root_bytes(&output, seek, out, out_len); +} + +void blake3_hasher_reset(blake3_hasher *self) { + chunk_state_reset(&self->chunk, self->key, 0); + self->cv_stack_len = 0; +} diff --git a/applications/reflection-experiments/third_party/blake3/blake3.h b/applications/reflection-experiments/third_party/blake3/blake3.h new file mode 100644 index 0000000..0893295 --- /dev/null +++ b/applications/reflection-experiments/third_party/blake3/blake3.h @@ -0,0 +1,86 @@ +#ifndef BLAKE3_H +#define BLAKE3_H + +#include +#include + +#if !defined(BLAKE3_API) +# if defined(_WIN32) || defined(__CYGWIN__) +# if defined(BLAKE3_DLL) +# if defined(BLAKE3_DLL_EXPORTS) +# define BLAKE3_API __declspec(dllexport) +# else +# define BLAKE3_API __declspec(dllimport) +# endif +# define BLAKE3_PRIVATE +# else +# define BLAKE3_API +# define BLAKE3_PRIVATE +# endif +# elif __GNUC__ >= 4 +# define BLAKE3_API __attribute__((visibility("default"))) +# define BLAKE3_PRIVATE __attribute__((visibility("hidden"))) +# else +# define BLAKE3_API +# define BLAKE3_PRIVATE +# endif +#endif + +#ifdef __cplusplus +extern "C" { +#endif + +#define BLAKE3_VERSION_STRING "1.8.2" +#define BLAKE3_KEY_LEN 32 +#define BLAKE3_OUT_LEN 32 +#define BLAKE3_BLOCK_LEN 64 +#define BLAKE3_CHUNK_LEN 1024 +#define BLAKE3_MAX_DEPTH 54 + +// This struct is a private implementation detail. It has to be here because +// it's part of blake3_hasher below. +typedef struct { + uint32_t cv[8]; + uint64_t chunk_counter; + uint8_t buf[BLAKE3_BLOCK_LEN]; + uint8_t buf_len; + uint8_t blocks_compressed; + uint8_t flags; +} blake3_chunk_state; + +typedef struct { + uint32_t key[8]; + blake3_chunk_state chunk; + uint8_t cv_stack_len; + // The stack size is MAX_DEPTH + 1 because we do lazy merging. For example, + // with 7 chunks, we have 3 entries in the stack. Adding an 8th chunk + // requires a 4th entry, rather than merging everything down to 1, because we + // don't know whether more input is coming. This is different from how the + // reference implementation does things. + uint8_t cv_stack[(BLAKE3_MAX_DEPTH + 1) * BLAKE3_OUT_LEN]; +} blake3_hasher; + +BLAKE3_API const char *blake3_version(void); +BLAKE3_API void blake3_hasher_init(blake3_hasher *self); +BLAKE3_API void blake3_hasher_init_keyed(blake3_hasher *self, + const uint8_t key[BLAKE3_KEY_LEN]); +BLAKE3_API void blake3_hasher_init_derive_key(blake3_hasher *self, const char *context); +BLAKE3_API void blake3_hasher_init_derive_key_raw(blake3_hasher *self, const void *context, + size_t context_len); +BLAKE3_API void blake3_hasher_update(blake3_hasher *self, const void *input, + size_t input_len); +#if defined(BLAKE3_USE_TBB) +BLAKE3_API void blake3_hasher_update_tbb(blake3_hasher *self, const void *input, + size_t input_len); +#endif // BLAKE3_USE_TBB +BLAKE3_API void blake3_hasher_finalize(const blake3_hasher *self, uint8_t *out, + size_t out_len); +BLAKE3_API void blake3_hasher_finalize_seek(const blake3_hasher *self, uint64_t seek, + uint8_t *out, size_t out_len); +BLAKE3_API void blake3_hasher_reset(blake3_hasher *self); + +#ifdef __cplusplus +} +#endif + +#endif /* BLAKE3_H */ diff --git a/applications/reflection-experiments/third_party/blake3/blake3_dispatch.c b/applications/reflection-experiments/third_party/blake3/blake3_dispatch.c new file mode 100644 index 0000000..52db905 --- /dev/null +++ b/applications/reflection-experiments/third_party/blake3/blake3_dispatch.c @@ -0,0 +1,332 @@ +#include +#include +#include + +#include "blake3_impl.h" + +#if defined(_MSC_VER) +#include +#endif + +#if defined(IS_X86) +#if defined(_MSC_VER) +#include +#elif defined(__GNUC__) +#include +#else +#undef IS_X86 /* Unimplemented! */ +#endif +#endif + +#if !defined(BLAKE3_ATOMICS) +#if defined(__has_include) +#if __has_include() && !defined(_MSC_VER) +#define BLAKE3_ATOMICS 1 +#else +#define BLAKE3_ATOMICS 0 +#endif /* __has_include() && !defined(_MSC_VER) */ +#else +#define BLAKE3_ATOMICS 0 +#endif /* defined(__has_include) */ +#endif /* BLAKE3_ATOMICS */ + +#if BLAKE3_ATOMICS +#define ATOMIC_INT _Atomic int +#define ATOMIC_LOAD(x) x +#define ATOMIC_STORE(x, y) x = y +#elif defined(_MSC_VER) +#define ATOMIC_INT LONG +#define ATOMIC_LOAD(x) InterlockedOr(&x, 0) +#define ATOMIC_STORE(x, y) InterlockedExchange(&x, y) +#else +#define ATOMIC_INT int +#define ATOMIC_LOAD(x) x +#define ATOMIC_STORE(x, y) x = y +#endif + +#define MAYBE_UNUSED(x) (void)((x)) + +#if defined(IS_X86) +static uint64_t xgetbv(void) { +#if defined(_MSC_VER) + return _xgetbv(0); +#else + uint32_t eax = 0, edx = 0; + __asm__ __volatile__("xgetbv\n" : "=a"(eax), "=d"(edx) : "c"(0)); + return ((uint64_t)edx << 32) | eax; +#endif +} + +static void cpuid(uint32_t out[4], uint32_t id) { +#if defined(_MSC_VER) + __cpuid((int *)out, id); +#elif defined(__i386__) || defined(_M_IX86) + __asm__ __volatile__("movl %%ebx, %1\n" + "cpuid\n" + "xchgl %1, %%ebx\n" + : "=a"(out[0]), "=r"(out[1]), "=c"(out[2]), "=d"(out[3]) + : "a"(id)); +#else + __asm__ __volatile__("cpuid\n" + : "=a"(out[0]), "=b"(out[1]), "=c"(out[2]), "=d"(out[3]) + : "a"(id)); +#endif +} + +static void cpuidex(uint32_t out[4], uint32_t id, uint32_t sid) { +#if defined(_MSC_VER) + __cpuidex((int *)out, id, sid); +#elif defined(__i386__) || defined(_M_IX86) + __asm__ __volatile__("movl %%ebx, %1\n" + "cpuid\n" + "xchgl %1, %%ebx\n" + : "=a"(out[0]), "=r"(out[1]), "=c"(out[2]), "=d"(out[3]) + : "a"(id), "c"(sid)); +#else + __asm__ __volatile__("cpuid\n" + : "=a"(out[0]), "=b"(out[1]), "=c"(out[2]), "=d"(out[3]) + : "a"(id), "c"(sid)); +#endif +} + +#endif + +enum cpu_feature { + SSE2 = 1 << 0, + SSSE3 = 1 << 1, + SSE41 = 1 << 2, + AVX = 1 << 3, + AVX2 = 1 << 4, + AVX512F = 1 << 5, + AVX512VL = 1 << 6, + /* ... */ + UNDEFINED = 1 << 30 +}; + +#if !defined(BLAKE3_TESTING) +static /* Allow the variable to be controlled manually for testing */ +#endif + ATOMIC_INT g_cpu_features = UNDEFINED; + +#if !defined(BLAKE3_TESTING) +static +#endif + enum cpu_feature + get_cpu_features(void) { + + /* If TSAN detects a data race here, try compiling with -DBLAKE3_ATOMICS=1 */ + enum cpu_feature features = ATOMIC_LOAD(g_cpu_features); + if (features != UNDEFINED) { + return features; + } else { +#if defined(IS_X86) + uint32_t regs[4] = {0}; + uint32_t *eax = ®s[0], *ebx = ®s[1], *ecx = ®s[2], *edx = ®s[3]; + (void)edx; + features = 0; + cpuid(regs, 0); + const int max_id = *eax; + cpuid(regs, 1); +#if defined(__amd64__) || defined(_M_X64) + features |= SSE2; +#else + if (*edx & (1UL << 26)) + features |= SSE2; +#endif + if (*ecx & (1UL << 9)) + features |= SSSE3; + if (*ecx & (1UL << 19)) + features |= SSE41; + + if (*ecx & (1UL << 27)) { // OSXSAVE + const uint64_t mask = xgetbv(); + if ((mask & 6) == 6) { // SSE and AVX states + if (*ecx & (1UL << 28)) + features |= AVX; + if (max_id >= 7) { + cpuidex(regs, 7, 0); + if (*ebx & (1UL << 5)) + features |= AVX2; + if ((mask & 224) == 224) { // Opmask, ZMM_Hi256, Hi16_Zmm + if (*ebx & (1UL << 31)) + features |= AVX512VL; + if (*ebx & (1UL << 16)) + features |= AVX512F; + } + } + } + } + ATOMIC_STORE(g_cpu_features, features); + return features; +#else + /* How to detect NEON? */ + return 0; +#endif + } +} + +void blake3_compress_in_place(uint32_t cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, + uint8_t flags) { +#if defined(IS_X86) + const enum cpu_feature features = get_cpu_features(); + MAYBE_UNUSED(features); +#if !defined(BLAKE3_NO_AVX512) + if (features & AVX512VL) { + blake3_compress_in_place_avx512(cv, block, block_len, counter, flags); + return; + } +#endif +#if !defined(BLAKE3_NO_SSE41) + if (features & SSE41) { + blake3_compress_in_place_sse41(cv, block, block_len, counter, flags); + return; + } +#endif +#if !defined(BLAKE3_NO_SSE2) + if (features & SSE2) { + blake3_compress_in_place_sse2(cv, block, block_len, counter, flags); + return; + } +#endif +#endif + blake3_compress_in_place_portable(cv, block, block_len, counter, flags); +} + +void blake3_compress_xof(const uint32_t cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, uint8_t flags, + uint8_t out[64]) { +#if defined(IS_X86) + const enum cpu_feature features = get_cpu_features(); + MAYBE_UNUSED(features); +#if !defined(BLAKE3_NO_AVX512) + if (features & AVX512VL) { + blake3_compress_xof_avx512(cv, block, block_len, counter, flags, out); + return; + } +#endif +#if !defined(BLAKE3_NO_SSE41) + if (features & SSE41) { + blake3_compress_xof_sse41(cv, block, block_len, counter, flags, out); + return; + } +#endif +#if !defined(BLAKE3_NO_SSE2) + if (features & SSE2) { + blake3_compress_xof_sse2(cv, block, block_len, counter, flags, out); + return; + } +#endif +#endif + blake3_compress_xof_portable(cv, block, block_len, counter, flags, out); +} + + +void blake3_xof_many(const uint32_t cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, uint8_t flags, + uint8_t out[64], size_t outblocks) { + if (outblocks == 0) { + // The current assembly implementation always outputs at least 1 block. + return; + } +#if defined(IS_X86) + const enum cpu_feature features = get_cpu_features(); + MAYBE_UNUSED(features); +#if !defined(_WIN32) && !defined(BLAKE3_NO_AVX512) + if (features & AVX512VL) { + blake3_xof_many_avx512(cv, block, block_len, counter, flags, out, outblocks); + return; + } +#endif +#endif + for(size_t i = 0; i < outblocks; ++i) { + blake3_compress_xof(cv, block, block_len, counter + i, flags, out + 64*i); + } +} + +void blake3_hash_many(const uint8_t *const *inputs, size_t num_inputs, + size_t blocks, const uint32_t key[8], uint64_t counter, + bool increment_counter, uint8_t flags, + uint8_t flags_start, uint8_t flags_end, uint8_t *out) { +#if defined(IS_X86) + const enum cpu_feature features = get_cpu_features(); + MAYBE_UNUSED(features); +#if !defined(BLAKE3_NO_AVX512) + if ((features & (AVX512F|AVX512VL)) == (AVX512F|AVX512VL)) { + blake3_hash_many_avx512(inputs, num_inputs, blocks, key, counter, + increment_counter, flags, flags_start, flags_end, + out); + return; + } +#endif +#if !defined(BLAKE3_NO_AVX2) + if (features & AVX2) { + blake3_hash_many_avx2(inputs, num_inputs, blocks, key, counter, + increment_counter, flags, flags_start, flags_end, + out); + return; + } +#endif +#if !defined(BLAKE3_NO_SSE41) + if (features & SSE41) { + blake3_hash_many_sse41(inputs, num_inputs, blocks, key, counter, + increment_counter, flags, flags_start, flags_end, + out); + return; + } +#endif +#if !defined(BLAKE3_NO_SSE2) + if (features & SSE2) { + blake3_hash_many_sse2(inputs, num_inputs, blocks, key, counter, + increment_counter, flags, flags_start, flags_end, + out); + return; + } +#endif +#endif + +#if BLAKE3_USE_NEON == 1 + blake3_hash_many_neon(inputs, num_inputs, blocks, key, counter, + increment_counter, flags, flags_start, flags_end, out); + return; +#endif + + blake3_hash_many_portable(inputs, num_inputs, blocks, key, counter, + increment_counter, flags, flags_start, flags_end, + out); +} + +// The dynamically detected SIMD degree of the current platform. +size_t blake3_simd_degree(void) { +#if defined(IS_X86) + const enum cpu_feature features = get_cpu_features(); + MAYBE_UNUSED(features); +#if !defined(BLAKE3_NO_AVX512) + if ((features & (AVX512F|AVX512VL)) == (AVX512F|AVX512VL)) { + return 16; + } +#endif +#if !defined(BLAKE3_NO_AVX2) + if (features & AVX2) { + return 8; + } +#endif +#if !defined(BLAKE3_NO_SSE41) + if (features & SSE41) { + return 4; + } +#endif +#if !defined(BLAKE3_NO_SSE2) + if (features & SSE2) { + return 4; + } +#endif +#endif +#if BLAKE3_USE_NEON == 1 + return 4; +#endif + return 1; +} diff --git a/applications/reflection-experiments/third_party/blake3/blake3_portable.c b/applications/reflection-experiments/third_party/blake3/blake3_portable.c new file mode 100644 index 0000000..062dd1b --- /dev/null +++ b/applications/reflection-experiments/third_party/blake3/blake3_portable.c @@ -0,0 +1,160 @@ +#include "blake3_impl.h" +#include + +INLINE uint32_t rotr32(uint32_t w, uint32_t c) { + return (w >> c) | (w << (32 - c)); +} + +INLINE void g(uint32_t *state, size_t a, size_t b, size_t c, size_t d, + uint32_t x, uint32_t y) { + state[a] = state[a] + state[b] + x; + state[d] = rotr32(state[d] ^ state[a], 16); + state[c] = state[c] + state[d]; + state[b] = rotr32(state[b] ^ state[c], 12); + state[a] = state[a] + state[b] + y; + state[d] = rotr32(state[d] ^ state[a], 8); + state[c] = state[c] + state[d]; + state[b] = rotr32(state[b] ^ state[c], 7); +} + +INLINE void round_fn(uint32_t state[16], const uint32_t *msg, size_t round) { + // Select the message schedule based on the round. + const uint8_t *schedule = MSG_SCHEDULE[round]; + + // Mix the columns. + g(state, 0, 4, 8, 12, msg[schedule[0]], msg[schedule[1]]); + g(state, 1, 5, 9, 13, msg[schedule[2]], msg[schedule[3]]); + g(state, 2, 6, 10, 14, msg[schedule[4]], msg[schedule[5]]); + g(state, 3, 7, 11, 15, msg[schedule[6]], msg[schedule[7]]); + + // Mix the rows. + g(state, 0, 5, 10, 15, msg[schedule[8]], msg[schedule[9]]); + g(state, 1, 6, 11, 12, msg[schedule[10]], msg[schedule[11]]); + g(state, 2, 7, 8, 13, msg[schedule[12]], msg[schedule[13]]); + g(state, 3, 4, 9, 14, msg[schedule[14]], msg[schedule[15]]); +} + +INLINE void compress_pre(uint32_t state[16], const uint32_t cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, uint8_t flags) { + uint32_t block_words[16]; + block_words[0] = load32(block + 4 * 0); + block_words[1] = load32(block + 4 * 1); + block_words[2] = load32(block + 4 * 2); + block_words[3] = load32(block + 4 * 3); + block_words[4] = load32(block + 4 * 4); + block_words[5] = load32(block + 4 * 5); + block_words[6] = load32(block + 4 * 6); + block_words[7] = load32(block + 4 * 7); + block_words[8] = load32(block + 4 * 8); + block_words[9] = load32(block + 4 * 9); + block_words[10] = load32(block + 4 * 10); + block_words[11] = load32(block + 4 * 11); + block_words[12] = load32(block + 4 * 12); + block_words[13] = load32(block + 4 * 13); + block_words[14] = load32(block + 4 * 14); + block_words[15] = load32(block + 4 * 15); + + state[0] = cv[0]; + state[1] = cv[1]; + state[2] = cv[2]; + state[3] = cv[3]; + state[4] = cv[4]; + state[5] = cv[5]; + state[6] = cv[6]; + state[7] = cv[7]; + state[8] = IV[0]; + state[9] = IV[1]; + state[10] = IV[2]; + state[11] = IV[3]; + state[12] = counter_low(counter); + state[13] = counter_high(counter); + state[14] = (uint32_t)block_len; + state[15] = (uint32_t)flags; + + round_fn(state, &block_words[0], 0); + round_fn(state, &block_words[0], 1); + round_fn(state, &block_words[0], 2); + round_fn(state, &block_words[0], 3); + round_fn(state, &block_words[0], 4); + round_fn(state, &block_words[0], 5); + round_fn(state, &block_words[0], 6); +} + +void blake3_compress_in_place_portable(uint32_t cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, + uint8_t flags) { + uint32_t state[16]; + compress_pre(state, cv, block, block_len, counter, flags); + cv[0] = state[0] ^ state[8]; + cv[1] = state[1] ^ state[9]; + cv[2] = state[2] ^ state[10]; + cv[3] = state[3] ^ state[11]; + cv[4] = state[4] ^ state[12]; + cv[5] = state[5] ^ state[13]; + cv[6] = state[6] ^ state[14]; + cv[7] = state[7] ^ state[15]; +} + +void blake3_compress_xof_portable(const uint32_t cv[8], + const uint8_t block[BLAKE3_BLOCK_LEN], + uint8_t block_len, uint64_t counter, + uint8_t flags, uint8_t out[64]) { + uint32_t state[16]; + compress_pre(state, cv, block, block_len, counter, flags); + + store32(&out[0 * 4], state[0] ^ state[8]); + store32(&out[1 * 4], state[1] ^ state[9]); + store32(&out[2 * 4], state[2] ^ state[10]); + store32(&out[3 * 4], state[3] ^ state[11]); + store32(&out[4 * 4], state[4] ^ state[12]); + store32(&out[5 * 4], state[5] ^ state[13]); + store32(&out[6 * 4], state[6] ^ state[14]); + store32(&out[7 * 4], state[7] ^ state[15]); + store32(&out[8 * 4], state[8] ^ cv[0]); + store32(&out[9 * 4], state[9] ^ cv[1]); + store32(&out[10 * 4], state[10] ^ cv[2]); + store32(&out[11 * 4], state[11] ^ cv[3]); + store32(&out[12 * 4], state[12] ^ cv[4]); + store32(&out[13 * 4], state[13] ^ cv[5]); + store32(&out[14 * 4], state[14] ^ cv[6]); + store32(&out[15 * 4], state[15] ^ cv[7]); +} + +INLINE void hash_one_portable(const uint8_t *input, size_t blocks, + const uint32_t key[8], uint64_t counter, + uint8_t flags, uint8_t flags_start, + uint8_t flags_end, uint8_t out[BLAKE3_OUT_LEN]) { + uint32_t cv[8]; + memcpy(cv, key, BLAKE3_KEY_LEN); + uint8_t block_flags = flags | flags_start; + while (blocks > 0) { + if (blocks == 1) { + block_flags |= flags_end; + } + blake3_compress_in_place_portable(cv, input, BLAKE3_BLOCK_LEN, counter, + block_flags); + input = &input[BLAKE3_BLOCK_LEN]; + blocks -= 1; + block_flags = flags; + } + store_cv_words(out, cv); +} + +void blake3_hash_many_portable(const uint8_t *const *inputs, size_t num_inputs, + size_t blocks, const uint32_t key[8], + uint64_t counter, bool increment_counter, + uint8_t flags, uint8_t flags_start, + uint8_t flags_end, uint8_t *out) { + while (num_inputs > 0) { + hash_one_portable(inputs[0], blocks, key, counter, flags, flags_start, + flags_end, out); + if (increment_counter) { + counter += 1; + } + inputs += 1; + num_inputs -= 1; + out = &out[BLAKE3_OUT_LEN]; + } +} diff --git a/applications/reflection-experiments/third_party/lib/DELETE_ME (here you can place .lib files).txt b/applications/reflection-experiments/third_party/lib/DELETE_ME (here you can place .lib files).txt deleted file mode 100644 index e69de29..0000000 diff --git a/applications/reflection-experiments/third_party/lib/clangAST.lib b/applications/reflection-experiments/third_party/lib/clangAST.lib new file mode 100644 index 0000000..da4a9cb Binary files /dev/null and b/applications/reflection-experiments/third_party/lib/clangAST.lib differ diff --git a/applications/reflection-experiments/third_party/lib/clangASTMatchers.lib b/applications/reflection-experiments/third_party/lib/clangASTMatchers.lib new file mode 100644 index 0000000..4255461 Binary files /dev/null and b/applications/reflection-experiments/third_party/lib/clangASTMatchers.lib differ diff --git a/applications/reflection-experiments/third_party/xxhash/xxhash.c b/applications/reflection-experiments/third_party/xxhash/xxhash.c new file mode 100644 index 0000000..e60cc37 --- /dev/null +++ b/applications/reflection-experiments/third_party/xxhash/xxhash.c @@ -0,0 +1,42 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Copyright (C) 2012-2023 Yann Collet + * + * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following disclaimer + * in the documentation and/or other materials provided with the + * distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * You can contact the author at: + * - xxHash homepage: https://www.xxhash.com + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +/* + * xxhash.c instantiates functions defined in xxhash.h + */ + +#define XXH_STATIC_LINKING_ONLY /* access advanced declarations */ +#define XXH_IMPLEMENTATION /* access definitions */ + +#include "xxhash.h" diff --git a/applications/reflection-experiments/third_party/xxhash/xxhash.h b/applications/reflection-experiments/third_party/xxhash/xxhash.h new file mode 100644 index 0000000..78fc2e8 --- /dev/null +++ b/applications/reflection-experiments/third_party/xxhash/xxhash.h @@ -0,0 +1,7238 @@ +/* + * xxHash - Extremely Fast Hash algorithm + * Header File + * Copyright (C) 2012-2023 Yann Collet + * + * BSD 2-Clause License (https://www.opensource.org/licenses/bsd-license.php) + * + * Redistribution and use in source and binary forms, with or without + * modification, are permitted provided that the following conditions are + * met: + * + * * Redistributions of source code must retain the above copyright + * notice, this list of conditions and the following disclaimer. + * * Redistributions in binary form must reproduce the above + * copyright notice, this list of conditions and the following disclaimer + * in the documentation and/or other materials provided with the + * distribution. + * + * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS + * "AS IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT + * LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR + * A PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT + * OWNER OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, + * SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT + * LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, + * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY + * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT + * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE + * OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE. + * + * You can contact the author at: + * - xxHash homepage: https://www.xxhash.com + * - xxHash source repository: https://github.com/Cyan4973/xxHash + */ + +/*! + * @mainpage xxHash + * + * xxHash is an extremely fast non-cryptographic hash algorithm, working at RAM speed + * limits. + * + * It is proposed in four flavors, in three families: + * 1. @ref XXH32_family + * - Classic 32-bit hash function. Simple, compact, and runs on almost all + * 32-bit and 64-bit systems. + * 2. @ref XXH64_family + * - Classic 64-bit adaptation of XXH32. Just as simple, and runs well on most + * 64-bit systems (but _not_ 32-bit systems). + * 3. @ref XXH3_family + * - Modern 64-bit and 128-bit hash function family which features improved + * strength and performance across the board, especially on smaller data. + * It benefits greatly from SIMD and 64-bit without requiring it. + * + * Benchmarks + * --- + * The reference system uses an Intel i7-9700K CPU, and runs Ubuntu x64 20.04. + * The open source benchmark program is compiled with clang v10.0 using -O3 flag. + * + * | Hash Name | ISA ext | Width | Large Data Speed | Small Data Velocity | + * | -------------------- | ------- | ----: | ---------------: | ------------------: | + * | XXH3_64bits() | @b AVX2 | 64 | 59.4 GB/s | 133.1 | + * | MeowHash | AES-NI | 128 | 58.2 GB/s | 52.5 | + * | XXH3_128bits() | @b AVX2 | 128 | 57.9 GB/s | 118.1 | + * | CLHash | PCLMUL | 64 | 37.1 GB/s | 58.1 | + * | XXH3_64bits() | @b SSE2 | 64 | 31.5 GB/s | 133.1 | + * | XXH3_128bits() | @b SSE2 | 128 | 29.6 GB/s | 118.1 | + * | RAM sequential read | | N/A | 28.0 GB/s | N/A | + * | ahash | AES-NI | 64 | 22.5 GB/s | 107.2 | + * | City64 | | 64 | 22.0 GB/s | 76.6 | + * | T1ha2 | | 64 | 22.0 GB/s | 99.0 | + * | City128 | | 128 | 21.7 GB/s | 57.7 | + * | FarmHash | AES-NI | 64 | 21.3 GB/s | 71.9 | + * | XXH64() | | 64 | 19.4 GB/s | 71.0 | + * | SpookyHash | | 64 | 19.3 GB/s | 53.2 | + * | Mum | | 64 | 18.0 GB/s | 67.0 | + * | CRC32C | SSE4.2 | 32 | 13.0 GB/s | 57.9 | + * | XXH32() | | 32 | 9.7 GB/s | 71.9 | + * | City32 | | 32 | 9.1 GB/s | 66.0 | + * | Blake3* | @b AVX2 | 256 | 4.4 GB/s | 8.1 | + * | Murmur3 | | 32 | 3.9 GB/s | 56.1 | + * | SipHash* | | 64 | 3.0 GB/s | 43.2 | + * | Blake3* | @b SSE2 | 256 | 2.4 GB/s | 8.1 | + * | HighwayHash | | 64 | 1.4 GB/s | 6.0 | + * | FNV64 | | 64 | 1.2 GB/s | 62.7 | + * | Blake2* | | 256 | 1.1 GB/s | 5.1 | + * | SHA1* | | 160 | 0.8 GB/s | 5.6 | + * | MD5* | | 128 | 0.6 GB/s | 7.8 | + * @note + * - Hashes which require a specific ISA extension are noted. SSE2 is also noted, + * even though it is mandatory on x64. + * - Hashes with an asterisk are cryptographic. Note that MD5 is non-cryptographic + * by modern standards. + * - Small data velocity is a rough average of algorithm's efficiency for small + * data. For more accurate information, see the wiki. + * - More benchmarks and strength tests are found on the wiki: + * https://github.com/Cyan4973/xxHash/wiki + * + * Usage + * ------ + * All xxHash variants use a similar API. Changing the algorithm is a trivial + * substitution. + * + * @pre + * For functions which take an input and length parameter, the following + * requirements are assumed: + * - The range from [`input`, `input + length`) is valid, readable memory. + * - The only exception is if the `length` is `0`, `input` may be `NULL`. + * - For C++, the objects must have the *TriviallyCopyable* property, as the + * functions access bytes directly as if it was an array of `unsigned char`. + * + * @anchor single_shot_example + * **Single Shot** + * + * These functions are stateless functions which hash a contiguous block of memory, + * immediately returning the result. They are the easiest and usually the fastest + * option. + * + * XXH32(), XXH64(), XXH3_64bits(), XXH3_128bits() + * + * @code{.c} + * #include + * #include "xxhash.h" + * + * // Example for a function which hashes a null terminated string with XXH32(). + * XXH32_hash_t hash_string(const char* string, XXH32_hash_t seed) + * { + * // NULL pointers are only valid if the length is zero + * size_t length = (string == NULL) ? 0 : strlen(string); + * return XXH32(string, length, seed); + * } + * @endcode + * + * + * @anchor streaming_example + * **Streaming** + * + * These groups of functions allow incremental hashing of unknown size, even + * more than what would fit in a size_t. + * + * XXH32_reset(), XXH64_reset(), XXH3_64bits_reset(), XXH3_128bits_reset() + * + * @code{.c} + * #include + * #include + * #include "xxhash.h" + * // Example for a function which hashes a FILE incrementally with XXH3_64bits(). + * XXH64_hash_t hashFile(FILE* f) + * { + * // Allocate a state struct. Do not just use malloc() or new. + * XXH3_state_t* state = XXH3_createState(); + * assert(state != NULL && "Out of memory!"); + * // Reset the state to start a new hashing session. + * XXH3_64bits_reset(state); + * char buffer[4096]; + * size_t count; + * // Read the file in chunks + * while ((count = fread(buffer, 1, sizeof(buffer), f)) != 0) { + * // Run update() as many times as necessary to process the data + * XXH3_64bits_update(state, buffer, count); + * } + * // Retrieve the finalized hash. This will not change the state. + * XXH64_hash_t result = XXH3_64bits_digest(state); + * // Free the state. Do not use free(). + * XXH3_freeState(state); + * return result; + * } + * @endcode + * + * Streaming functions generate the xxHash value from an incremental input. + * This method is slower than single-call functions, due to state management. + * For small inputs, prefer `XXH32()` and `XXH64()`, which are better optimized. + * + * An XXH state must first be allocated using `XXH*_createState()`. + * + * Start a new hash by initializing the state with a seed using `XXH*_reset()`. + * + * Then, feed the hash state by calling `XXH*_update()` as many times as necessary. + * + * The function returns an error code, with 0 meaning OK, and any other value + * meaning there is an error. + * + * Finally, a hash value can be produced anytime, by using `XXH*_digest()`. + * This function returns the nn-bits hash as an int or long long. + * + * It's still possible to continue inserting input into the hash state after a + * digest, and generate new hash values later on by invoking `XXH*_digest()`. + * + * When done, release the state using `XXH*_freeState()`. + * + * + * @anchor canonical_representation_example + * **Canonical Representation** + * + * The default return values from XXH functions are unsigned 32, 64 and 128 bit + * integers. + * This the simplest and fastest format for further post-processing. + * + * However, this leaves open the question of what is the order on the byte level, + * since little and big endian conventions will store the same number differently. + * + * The canonical representation settles this issue by mandating big-endian + * convention, the same convention as human-readable numbers (large digits first). + * + * When writing hash values to storage, sending them over a network, or printing + * them, it's highly recommended to use the canonical representation to ensure + * portability across a wider range of systems, present and future. + * + * The following functions allow transformation of hash values to and from + * canonical format. + * + * XXH32_canonicalFromHash(), XXH32_hashFromCanonical(), + * XXH64_canonicalFromHash(), XXH64_hashFromCanonical(), + * XXH128_canonicalFromHash(), XXH128_hashFromCanonical(), + * + * @code{.c} + * #include + * #include "xxhash.h" + * + * // Example for a function which prints XXH32_hash_t in human readable format + * void printXxh32(XXH32_hash_t hash) + * { + * XXH32_canonical_t cano; + * XXH32_canonicalFromHash(&cano, hash); + * size_t i; + * for(i = 0; i < sizeof(cano.digest); ++i) { + * printf("%02x", cano.digest[i]); + * } + * printf("\n"); + * } + * + * // Example for a function which converts XXH32_canonical_t to XXH32_hash_t + * XXH32_hash_t convertCanonicalToXxh32(XXH32_canonical_t cano) + * { + * XXH32_hash_t hash = XXH32_hashFromCanonical(&cano); + * return hash; + * } + * @endcode + * + * + * @file xxhash.h + * xxHash prototypes and implementation + */ + +#if defined (__cplusplus) +extern "C" { +#endif + +/* **************************** + * INLINE mode + ******************************/ +/*! + * @defgroup public Public API + * Contains details on the public xxHash functions. + * @{ + */ +#ifdef XXH_DOXYGEN +/*! + * @brief Gives access to internal state declaration, required for static allocation. + * + * Incompatible with dynamic linking, due to risks of ABI changes. + * + * Usage: + * @code{.c} + * #define XXH_STATIC_LINKING_ONLY + * #include "xxhash.h" + * @endcode + */ +# define XXH_STATIC_LINKING_ONLY +/* Do not undef XXH_STATIC_LINKING_ONLY for Doxygen */ + +/*! + * @brief Gives access to internal definitions. + * + * Usage: + * @code{.c} + * #define XXH_STATIC_LINKING_ONLY + * #define XXH_IMPLEMENTATION + * #include "xxhash.h" + * @endcode + */ +# define XXH_IMPLEMENTATION +/* Do not undef XXH_IMPLEMENTATION for Doxygen */ + +/*! + * @brief Exposes the implementation and marks all functions as `inline`. + * + * Use these build macros to inline xxhash into the target unit. + * Inlining improves performance on small inputs, especially when the length is + * expressed as a compile-time constant: + * + * https://fastcompression.blogspot.com/2018/03/xxhash-for-small-keys-impressive-power.html + * + * It also keeps xxHash symbols private to the unit, so they are not exported. + * + * Usage: + * @code{.c} + * #define XXH_INLINE_ALL + * #include "xxhash.h" + * @endcode + * Do not compile and link xxhash.o as a separate object, as it is not useful. + */ +# define XXH_INLINE_ALL +# undef XXH_INLINE_ALL +/*! + * @brief Exposes the implementation without marking functions as inline. + */ +# define XXH_PRIVATE_API +# undef XXH_PRIVATE_API +/*! + * @brief Emulate a namespace by transparently prefixing all symbols. + * + * If you want to include _and expose_ xxHash functions from within your own + * library, but also want to avoid symbol collisions with other libraries which + * may also include xxHash, you can use @ref XXH_NAMESPACE to automatically prefix + * any public symbol from xxhash library with the value of @ref XXH_NAMESPACE + * (therefore, avoid empty or numeric values). + * + * Note that no change is required within the calling program as long as it + * includes `xxhash.h`: Regular symbol names will be automatically translated + * by this header. + */ +# define XXH_NAMESPACE /* YOUR NAME HERE */ +# undef XXH_NAMESPACE +#endif + +#if (defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API)) \ + && !defined(XXH_INLINE_ALL_31684351384) + /* this section should be traversed only once */ +# define XXH_INLINE_ALL_31684351384 + /* give access to the advanced API, required to compile implementations */ +# undef XXH_STATIC_LINKING_ONLY /* avoid macro redef */ +# define XXH_STATIC_LINKING_ONLY + /* make all functions private */ +# undef XXH_PUBLIC_API +# if defined(__GNUC__) +# define XXH_PUBLIC_API static __inline __attribute__((__unused__)) +# elif defined (__cplusplus) || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) +# define XXH_PUBLIC_API static inline +# elif defined(_MSC_VER) +# define XXH_PUBLIC_API static __inline +# else + /* note: this version may generate warnings for unused static functions */ +# define XXH_PUBLIC_API static +# endif + + /* + * This part deals with the special case where a unit wants to inline xxHash, + * but "xxhash.h" has previously been included without XXH_INLINE_ALL, + * such as part of some previously included *.h header file. + * Without further action, the new include would just be ignored, + * and functions would effectively _not_ be inlined (silent failure). + * The following macros solve this situation by prefixing all inlined names, + * avoiding naming collision with previous inclusions. + */ + /* Before that, we unconditionally #undef all symbols, + * in case they were already defined with XXH_NAMESPACE. + * They will then be redefined for XXH_INLINE_ALL + */ +# undef XXH_versionNumber + /* XXH32 */ +# undef XXH32 +# undef XXH32_createState +# undef XXH32_freeState +# undef XXH32_reset +# undef XXH32_update +# undef XXH32_digest +# undef XXH32_copyState +# undef XXH32_canonicalFromHash +# undef XXH32_hashFromCanonical + /* XXH64 */ +# undef XXH64 +# undef XXH64_createState +# undef XXH64_freeState +# undef XXH64_reset +# undef XXH64_update +# undef XXH64_digest +# undef XXH64_copyState +# undef XXH64_canonicalFromHash +# undef XXH64_hashFromCanonical + /* XXH3_64bits */ +# undef XXH3_64bits +# undef XXH3_64bits_withSecret +# undef XXH3_64bits_withSeed +# undef XXH3_64bits_withSecretandSeed +# undef XXH3_createState +# undef XXH3_freeState +# undef XXH3_copyState +# undef XXH3_64bits_reset +# undef XXH3_64bits_reset_withSeed +# undef XXH3_64bits_reset_withSecret +# undef XXH3_64bits_update +# undef XXH3_64bits_digest +# undef XXH3_generateSecret + /* XXH3_128bits */ +# undef XXH128 +# undef XXH3_128bits +# undef XXH3_128bits_withSeed +# undef XXH3_128bits_withSecret +# undef XXH3_128bits_reset +# undef XXH3_128bits_reset_withSeed +# undef XXH3_128bits_reset_withSecret +# undef XXH3_128bits_reset_withSecretandSeed +# undef XXH3_128bits_update +# undef XXH3_128bits_digest +# undef XXH128_isEqual +# undef XXH128_cmp +# undef XXH128_canonicalFromHash +# undef XXH128_hashFromCanonical + /* Finally, free the namespace itself */ +# undef XXH_NAMESPACE + + /* employ the namespace for XXH_INLINE_ALL */ +# define XXH_NAMESPACE XXH_INLINE_ + /* + * Some identifiers (enums, type names) are not symbols, + * but they must nonetheless be renamed to avoid redeclaration. + * Alternative solution: do not redeclare them. + * However, this requires some #ifdefs, and has a more dispersed impact. + * Meanwhile, renaming can be achieved in a single place. + */ +# define XXH_IPREF(Id) XXH_NAMESPACE ## Id +# define XXH_OK XXH_IPREF(XXH_OK) +# define XXH_ERROR XXH_IPREF(XXH_ERROR) +# define XXH_errorcode XXH_IPREF(XXH_errorcode) +# define XXH32_canonical_t XXH_IPREF(XXH32_canonical_t) +# define XXH64_canonical_t XXH_IPREF(XXH64_canonical_t) +# define XXH128_canonical_t XXH_IPREF(XXH128_canonical_t) +# define XXH32_state_s XXH_IPREF(XXH32_state_s) +# define XXH32_state_t XXH_IPREF(XXH32_state_t) +# define XXH64_state_s XXH_IPREF(XXH64_state_s) +# define XXH64_state_t XXH_IPREF(XXH64_state_t) +# define XXH3_state_s XXH_IPREF(XXH3_state_s) +# define XXH3_state_t XXH_IPREF(XXH3_state_t) +# define XXH128_hash_t XXH_IPREF(XXH128_hash_t) + /* Ensure the header is parsed again, even if it was previously included */ +# undef XXHASH_H_5627135585666179 +# undef XXHASH_H_STATIC_13879238742 +#endif /* XXH_INLINE_ALL || XXH_PRIVATE_API */ + +/* **************************************************************** + * Stable API + *****************************************************************/ +#ifndef XXHASH_H_5627135585666179 +#define XXHASH_H_5627135585666179 1 + +/*! @brief Marks a global symbol. */ +#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) +# if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) +# ifdef XXH_EXPORT +# define XXH_PUBLIC_API __declspec(dllexport) +# elif XXH_IMPORT +# define XXH_PUBLIC_API __declspec(dllimport) +# endif +# else +# define XXH_PUBLIC_API /* do nothing */ +# endif +#endif + +#ifdef XXH_NAMESPACE +# define XXH_CAT(A,B) A##B +# define XXH_NAME2(A,B) XXH_CAT(A,B) +# define XXH_versionNumber XXH_NAME2(XXH_NAMESPACE, XXH_versionNumber) +/* XXH32 */ +# define XXH32 XXH_NAME2(XXH_NAMESPACE, XXH32) +# define XXH32_createState XXH_NAME2(XXH_NAMESPACE, XXH32_createState) +# define XXH32_freeState XXH_NAME2(XXH_NAMESPACE, XXH32_freeState) +# define XXH32_reset XXH_NAME2(XXH_NAMESPACE, XXH32_reset) +# define XXH32_update XXH_NAME2(XXH_NAMESPACE, XXH32_update) +# define XXH32_digest XXH_NAME2(XXH_NAMESPACE, XXH32_digest) +# define XXH32_copyState XXH_NAME2(XXH_NAMESPACE, XXH32_copyState) +# define XXH32_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH32_canonicalFromHash) +# define XXH32_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH32_hashFromCanonical) +/* XXH64 */ +# define XXH64 XXH_NAME2(XXH_NAMESPACE, XXH64) +# define XXH64_createState XXH_NAME2(XXH_NAMESPACE, XXH64_createState) +# define XXH64_freeState XXH_NAME2(XXH_NAMESPACE, XXH64_freeState) +# define XXH64_reset XXH_NAME2(XXH_NAMESPACE, XXH64_reset) +# define XXH64_update XXH_NAME2(XXH_NAMESPACE, XXH64_update) +# define XXH64_digest XXH_NAME2(XXH_NAMESPACE, XXH64_digest) +# define XXH64_copyState XXH_NAME2(XXH_NAMESPACE, XXH64_copyState) +# define XXH64_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH64_canonicalFromHash) +# define XXH64_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH64_hashFromCanonical) +/* XXH3_64bits */ +# define XXH3_64bits XXH_NAME2(XXH_NAMESPACE, XXH3_64bits) +# define XXH3_64bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecret) +# define XXH3_64bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSeed) +# define XXH3_64bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_withSecretandSeed) +# define XXH3_createState XXH_NAME2(XXH_NAMESPACE, XXH3_createState) +# define XXH3_freeState XXH_NAME2(XXH_NAMESPACE, XXH3_freeState) +# define XXH3_copyState XXH_NAME2(XXH_NAMESPACE, XXH3_copyState) +# define XXH3_64bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset) +# define XXH3_64bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSeed) +# define XXH3_64bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecret) +# define XXH3_64bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_reset_withSecretandSeed) +# define XXH3_64bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_update) +# define XXH3_64bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_64bits_digest) +# define XXH3_generateSecret XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret) +# define XXH3_generateSecret_fromSeed XXH_NAME2(XXH_NAMESPACE, XXH3_generateSecret_fromSeed) +/* XXH3_128bits */ +# define XXH128 XXH_NAME2(XXH_NAMESPACE, XXH128) +# define XXH3_128bits XXH_NAME2(XXH_NAMESPACE, XXH3_128bits) +# define XXH3_128bits_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSeed) +# define XXH3_128bits_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecret) +# define XXH3_128bits_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_withSecretandSeed) +# define XXH3_128bits_reset XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset) +# define XXH3_128bits_reset_withSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSeed) +# define XXH3_128bits_reset_withSecret XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecret) +# define XXH3_128bits_reset_withSecretandSeed XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_reset_withSecretandSeed) +# define XXH3_128bits_update XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_update) +# define XXH3_128bits_digest XXH_NAME2(XXH_NAMESPACE, XXH3_128bits_digest) +# define XXH128_isEqual XXH_NAME2(XXH_NAMESPACE, XXH128_isEqual) +# define XXH128_cmp XXH_NAME2(XXH_NAMESPACE, XXH128_cmp) +# define XXH128_canonicalFromHash XXH_NAME2(XXH_NAMESPACE, XXH128_canonicalFromHash) +# define XXH128_hashFromCanonical XXH_NAME2(XXH_NAMESPACE, XXH128_hashFromCanonical) +#endif + + +/* ************************************* +* Compiler specifics +***************************************/ + +/* specific declaration modes for Windows */ +#if !defined(XXH_INLINE_ALL) && !defined(XXH_PRIVATE_API) +# if defined(_WIN32) && defined(_MSC_VER) && (defined(XXH_IMPORT) || defined(XXH_EXPORT)) +# ifdef XXH_EXPORT +# define XXH_PUBLIC_API __declspec(dllexport) +# elif XXH_IMPORT +# define XXH_PUBLIC_API __declspec(dllimport) +# endif +# else +# define XXH_PUBLIC_API /* do nothing */ +# endif +#endif + +#if defined (__GNUC__) +# define XXH_CONSTF __attribute__((__const__)) +# define XXH_PUREF __attribute__((__pure__)) +# define XXH_MALLOCF __attribute__((__malloc__)) +#else +# define XXH_CONSTF /* disable */ +# define XXH_PUREF +# define XXH_MALLOCF +#endif + +/* ************************************* +* Version +***************************************/ +#define XXH_VERSION_MAJOR 0 +#define XXH_VERSION_MINOR 8 +#define XXH_VERSION_RELEASE 3 +/*! @brief Version number, encoded as two digits each */ +#define XXH_VERSION_NUMBER (XXH_VERSION_MAJOR *100*100 + XXH_VERSION_MINOR *100 + XXH_VERSION_RELEASE) + +/*! + * @brief Obtains the xxHash version. + * + * This is mostly useful when xxHash is compiled as a shared library, + * since the returned value comes from the library, as opposed to header file. + * + * @return @ref XXH_VERSION_NUMBER of the invoked library. + */ +XXH_PUBLIC_API XXH_CONSTF unsigned XXH_versionNumber (void); + + +/* **************************** +* Common basic types +******************************/ +#include /* size_t */ +/*! + * @brief Exit code for the streaming API. + */ +typedef enum { + XXH_OK = 0, /*!< OK */ + XXH_ERROR /*!< Error */ +} XXH_errorcode; + + +/*-********************************************************************** +* 32-bit hash +************************************************************************/ +#if defined(XXH_DOXYGEN) /* Don't show include */ +/*! + * @brief An unsigned 32-bit integer. + * + * Not necessarily defined to `uint32_t` but functionally equivalent. + */ +typedef uint32_t XXH32_hash_t; + +#elif !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint32_t XXH32_hash_t; + +#else +# include +# if UINT_MAX == 0xFFFFFFFFUL + typedef unsigned int XXH32_hash_t; +# elif ULONG_MAX == 0xFFFFFFFFUL + typedef unsigned long XXH32_hash_t; +# else +# error "unsupported platform: need a 32-bit type" +# endif +#endif + +/*! + * @} + * + * @defgroup XXH32_family XXH32 family + * @ingroup public + * Contains functions used in the classic 32-bit xxHash algorithm. + * + * @note + * XXH32 is useful for older platforms, with no or poor 64-bit performance. + * Note that the @ref XXH3_family provides competitive speed for both 32-bit + * and 64-bit systems, and offers true 64/128 bit hash results. + * + * @see @ref XXH64_family, @ref XXH3_family : Other xxHash families + * @see @ref XXH32_impl for implementation details + * @{ + */ + +/*! + * @brief Calculates the 32-bit hash of @p input using xxHash32. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 32-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 32-bit xxHash32 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32 (const void* input, size_t length, XXH32_hash_t seed); + +#ifndef XXH_NO_STREAM +/*! + * @typedef struct XXH32_state_s XXH32_state_t + * @brief The opaque state struct for the XXH32 streaming API. + * + * @see XXH32_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH32_state_s XXH32_state_t; + +/*! + * @brief Allocates an @ref XXH32_state_t. + * + * @return An allocated pointer of @ref XXH32_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH32_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_MALLOCF XXH32_state_t* XXH32_createState(void); +/*! + * @brief Frees an @ref XXH32_state_t. + * + * @param statePtr A pointer to an @ref XXH32_state_t allocated with @ref XXH32_createState(). + * + * @return @ref XXH_OK. + * + * @note @p statePtr must be allocated with XXH32_createState(). + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr); +/*! + * @brief Copies one @ref XXH32_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dst_state, const XXH32_state_t* src_state); + +/*! + * @brief Resets an @ref XXH32_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 32-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note This function resets and seeds a state. Call it before @ref XXH32_update(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH32_reset (XXH32_state_t* statePtr, XXH32_hash_t seed); + +/*! + * @brief Consumes a block of @p input to an @ref XXH32_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH32_update (XXH32_state_t* statePtr, const void* input, size_t length); + +/*! + * @brief Returns the calculated hash value from an @ref XXH32_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated 32-bit xxHash32 value from that state. + * + * @note + * Calling XXH32_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_digest (const XXH32_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! + * @brief Canonical (big endian) representation of @ref XXH32_hash_t. + */ +typedef struct { + unsigned char digest[4]; /*!< Hash bytes, big endian */ +} XXH32_canonical_t; + +/*! + * @brief Converts an @ref XXH32_hash_t to a big endian @ref XXH32_canonical_t. + * + * @param dst The @ref XXH32_canonical_t pointer to be stored to. + * @param hash The @ref XXH32_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash); + +/*! + * @brief Converts an @ref XXH32_canonical_t to a native @ref XXH32_hash_t. + * + * @param src The @ref XXH32_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src); + + +/*! @cond Doxygen ignores this part */ +#ifdef __has_attribute +# define XXH_HAS_ATTRIBUTE(x) __has_attribute(x) +#else +# define XXH_HAS_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * C23 __STDC_VERSION__ number hasn't been specified yet. For now + * leave as `201711L` (C17 + 1). + * TODO: Update to correct value when its been specified. + */ +#define XXH_C23_VN 201711L +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* C-language Attributes are added in C23. */ +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) && defined(__has_c_attribute) +# define XXH_HAS_C_ATTRIBUTE(x) __has_c_attribute(x) +#else +# define XXH_HAS_C_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +#if defined(__cplusplus) && defined(__has_cpp_attribute) +# define XXH_HAS_CPP_ATTRIBUTE(x) __has_cpp_attribute(x) +#else +# define XXH_HAS_CPP_ATTRIBUTE(x) 0 +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * Define XXH_FALLTHROUGH macro for annotating switch case with the 'fallthrough' attribute + * introduced in CPP17 and C23. + * CPP17 : https://en.cppreference.com/w/cpp/language/attributes/fallthrough + * C23 : https://en.cppreference.com/w/c/language/attributes/fallthrough + */ +#if XXH_HAS_C_ATTRIBUTE(fallthrough) || XXH_HAS_CPP_ATTRIBUTE(fallthrough) +# define XXH_FALLTHROUGH [[fallthrough]] +#elif XXH_HAS_ATTRIBUTE(__fallthrough__) +# define XXH_FALLTHROUGH __attribute__ ((__fallthrough__)) +#else +# define XXH_FALLTHROUGH /* fallthrough */ +#endif +/*! @endcond */ + +/*! @cond Doxygen ignores this part */ +/* + * Define XXH_NOESCAPE for annotated pointers in public API. + * https://clang.llvm.org/docs/AttributeReference.html#noescape + * As of writing this, only supported by clang. + */ +#if XXH_HAS_ATTRIBUTE(noescape) +# define XXH_NOESCAPE __attribute__((__noescape__)) +#else +# define XXH_NOESCAPE +#endif +/*! @endcond */ + + +/*! + * @} + * @ingroup public + * @{ + */ + +#ifndef XXH_NO_LONG_LONG +/*-********************************************************************** +* 64-bit hash +************************************************************************/ +#if defined(XXH_DOXYGEN) /* don't include */ +/*! + * @brief An unsigned 64-bit integer. + * + * Not necessarily defined to `uint64_t` but functionally equivalent. + */ +typedef uint64_t XXH64_hash_t; +#elif !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint64_t XXH64_hash_t; +#else +# include +# if defined(__LP64__) && ULONG_MAX == 0xFFFFFFFFFFFFFFFFULL + /* LP64 ABI says uint64_t is unsigned long */ + typedef unsigned long XXH64_hash_t; +# else + /* the following type must have a width of 64-bit */ + typedef unsigned long long XXH64_hash_t; +# endif +#endif + +/*! + * @} + * + * @defgroup XXH64_family XXH64 family + * @ingroup public + * @{ + * Contains functions used in the classic 64-bit xxHash algorithm. + * + * @note + * XXH3 provides competitive speed for both 32-bit and 64-bit systems, + * and offers true 64/128 bit hash results. + * It provides better speed for systems with vector processing capabilities. + */ + +/*! + * @brief Calculates the 64-bit hash of @p input using xxHash64. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 64-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit xxHash64 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/*! + * @brief The opaque state struct for the XXH64 streaming API. + * + * @see XXH64_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH64_state_s XXH64_state_t; /* incomplete type */ + +/*! + * @brief Allocates an @ref XXH64_state_t. + * + * @return An allocated pointer of @ref XXH64_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH64_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_MALLOCF XXH64_state_t* XXH64_createState(void); + +/*! + * @brief Frees an @ref XXH64_state_t. + * + * @param statePtr A pointer to an @ref XXH64_state_t allocated with @ref XXH64_createState(). + * + * @return @ref XXH_OK. + * + * @note @p statePtr must be allocated with XXH64_createState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr); + +/*! + * @brief Copies one @ref XXH64_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dst_state, const XXH64_state_t* src_state); + +/*! + * @brief Resets an @ref XXH64_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note This function resets and seeds a state. Call it before @ref XXH64_update(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_reset (XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed); + +/*! + * @brief Consumes a block of @p input to an @ref XXH64_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH64_update (XXH_NOESCAPE XXH64_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated hash value from an @ref XXH64_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated 64-bit xxHash64 value from that state. + * + * @note + * Calling XXH64_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_digest (XXH_NOESCAPE const XXH64_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ +/******* Canonical representation *******/ + +/*! + * @brief Canonical (big endian) representation of @ref XXH64_hash_t. + */ +typedef struct { unsigned char digest[sizeof(XXH64_hash_t)]; } XXH64_canonical_t; + +/*! + * @brief Converts an @ref XXH64_hash_t to a big endian @ref XXH64_canonical_t. + * + * @param dst The @ref XXH64_canonical_t pointer to be stored to. + * @param hash The @ref XXH64_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash); + +/*! + * @brief Converts an @ref XXH64_canonical_t to a native @ref XXH64_hash_t. + * + * @param src The @ref XXH64_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src); + +#ifndef XXH_NO_XXH3 + +/*! + * @} + * ************************************************************************ + * @defgroup XXH3_family XXH3 family + * @ingroup public + * @{ + * + * XXH3 is a more recent hash algorithm featuring: + * - Improved speed for both small and large inputs + * - True 64-bit and 128-bit outputs + * - SIMD acceleration + * - Improved 32-bit viability + * + * Speed analysis methodology is explained here: + * + * https://fastcompression.blogspot.com/2019/03/presenting-xxh3.html + * + * Compared to XXH64, expect XXH3 to run approximately + * ~2x faster on large inputs and >3x faster on small ones, + * exact differences vary depending on platform. + * + * XXH3's speed benefits greatly from SIMD and 64-bit arithmetic, + * but does not require it. + * Most 32-bit and 64-bit targets that can run XXH32 smoothly can run XXH3 + * at competitive speeds, even without vector support. Further details are + * explained in the implementation. + * + * XXH3 has a fast scalar implementation, but it also includes accelerated SIMD + * implementations for many common platforms: + * - AVX512 + * - AVX2 + * - SSE2 + * - ARM NEON + * - WebAssembly SIMD128 + * - POWER8 VSX + * - s390x ZVector + * This can be controlled via the @ref XXH_VECTOR macro, but it automatically + * selects the best version according to predefined macros. For the x86 family, an + * automatic runtime dispatcher is included separately in @ref xxh_x86dispatch.c. + * + * XXH3 implementation is portable: + * it has a generic C90 formulation that can be compiled on any platform, + * all implementations generate exactly the same hash value on all platforms. + * Starting from v0.8.0, it's also labelled "stable", meaning that + * any future version will also generate the same hash value. + * + * XXH3 offers 2 variants, _64bits and _128bits. + * + * When only 64 bits are needed, prefer invoking the _64bits variant, as it + * reduces the amount of mixing, resulting in faster speed on small inputs. + * It's also generally simpler to manipulate a scalar return type than a struct. + * + * The API supports one-shot hashing, streaming mode, and custom secrets. + */ + +/*! + * @ingroup tuning + * @brief Possible values for @ref XXH_VECTOR. + * + * Unless set explicitly, determined automatically. + */ +# define XXH_SCALAR 0 /*!< Portable scalar version */ +# define XXH_SSE2 1 /*!< SSE2 for Pentium 4, Opteron, all x86_64. */ +# define XXH_AVX2 2 /*!< AVX2 for Haswell and Bulldozer */ +# define XXH_AVX512 3 /*!< AVX512 for Skylake and Icelake */ +# define XXH_NEON 4 /*!< NEON for most ARMv7-A, all AArch64, and WASM SIMD128 */ +# define XXH_VSX 5 /*!< VSX and ZVector for POWER8/z13 (64-bit) */ +# define XXH_SVE 6 /*!< SVE for some ARMv8-A and ARMv9-A */ +# define XXH_LSX 7 /*!< LSX (128-bit SIMD) for LoongArch64 */ + + +/*-********************************************************************** +* XXH3 64-bit variant +************************************************************************/ + +/*! + * @brief Calculates 64-bit unseeded variant of XXH3 hash of @p input. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @note + * This is equivalent to @ref XXH3_64bits_withSeed() with a seed of `0`, however + * it may have slightly better performance due to constant propagation of the + * defaults. + * + * @see + * XXH3_64bits_withSeed(), XXH3_64bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Calculates 64-bit seeded variant of XXH3 hash of @p input. + * + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @note + * seed == 0 produces the same results as @ref XXH3_64bits(). + * + * This variant generates a custom secret on the fly based on default secret + * altered using the @p seed value. + * + * While this operation is decently fast, note that it's not completely free. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed); + +/*! + * The bare minimum size for a custom secret. + * + * @see + * XXH3_64bits_withSecret(), XXH3_64bits_reset_withSecret(), + * XXH3_128bits_withSecret(), XXH3_128bits_reset_withSecret(). + */ +#define XXH3_SECRET_SIZE_MIN 136 + +/*! + * @brief Calculates 64-bit variant of XXH3 with a custom "secret". + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @return The calculated 64-bit XXH3 hash value. + * + * @pre + * The memory between @p data and @p data + @p len must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p data may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * It's possible to provide any blob of bytes as a "secret" to generate the hash. + * This makes it more difficult for an external actor to prepare an intentional collision. + * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). + * However, the quality of the secret impacts the dispersion of the hash algorithm. + * Therefore, the secret _must_ look like a bunch of random bytes. + * Avoid "trivial" or structured data such as repeated sequences or a text document. + * Whenever in doubt about the "randomness" of the blob of bytes, + * consider employing @ref XXH3_generateSecret() instead (see below). + * It will generate a proper high entropy secret derived from the blob of bytes. + * Another advantage of using XXH3_generateSecret() is that + * it guarantees that all bits within the initial blob of bytes + * will impact every bit of the output. + * This is not necessarily the case when using the blob of bytes directly + * because, when hashing _small_ inputs, only a portion of the secret is employed. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); + + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/* + * Streaming requires state maintenance. + * This operation costs memory and CPU. + * As a consequence, streaming is slower than one-shot hashing. + * For better performance, prefer one-shot functions whenever applicable. + */ + +/*! + * @brief The opaque state struct for the XXH3 streaming API. + * + * @see XXH3_state_s for details. + * @see @ref streaming_example "Streaming Example" + */ +typedef struct XXH3_state_s XXH3_state_t; +XXH_PUBLIC_API XXH_MALLOCF XXH3_state_t* XXH3_createState(void); +XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr); + +/*! + * @brief Copies one @ref XXH3_state_t to another. + * + * @param dst_state The state to copy to. + * @param src_state The state to copy from. + * @pre + * @p dst_state and @p src_state must not be `NULL` and must not overlap. + */ +XXH_PUBLIC_API void XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state); + +/*! + * @brief Resets an @ref XXH3_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret with default parameters. + * - Call this function before @ref XXH3_64bits_update(). + * - Digest will be equivalent to `XXH3_64bits()`. + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); + +/*! + * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret from `seed`. + * - Call this function before @ref XXH3_64bits_update(). + * - Digest will be equivalent to `XXH3_64bits_withSeed()`. + * + * @see @ref streaming_example "Streaming Example" + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); + +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * `secret` is referenced, it _must outlive_ the hash streaming session. + * + * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, + * and the quality of produced hash values depends on secret's entropy + * (secret's content should look like a bunch of random bytes). + * When in doubt about the randomness of a candidate `secret`, + * consider employing `XXH3_generateSecret()` instead (see below). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); + +/*! + * @brief Consumes a block of @p input to an @ref XXH3_state_t. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * @pre + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note Call this to incrementally consume blocks of data. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_64bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated XXH3 64-bit hash value from an @ref XXH3_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated XXH3 64-bit hash value from that state. + * + * @note + * Calling XXH3_64bits_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/* note : canonical representation of XXH3 is the same as XXH64 + * since they both produce XXH64_hash_t values */ + + +/*-********************************************************************** +* XXH3 128-bit variant +************************************************************************/ + +/*! + * @brief The return value from 128-bit hashes. + * + * Stored in little endian order, although the fields themselves are in native + * endianness. + */ +typedef struct { + XXH64_hash_t low64; /*!< `value & 0xFFFFFFFFFFFFFFFF` */ + XXH64_hash_t high64; /*!< `value >> 64` */ +} XXH128_hash_t; + +/*! + * @brief Calculates 128-bit unseeded variant of XXH3 of @p data. + * + * @param data The block of data to be hashed, at least @p length bytes in size. + * @param len The length of @p data, in bytes. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * The 128-bit variant of XXH3 has more strength, but it has a bit of overhead + * for shorter inputs. + * + * This is equivalent to @ref XXH3_128bits_withSeed() with a seed of `0`, however + * it may have slightly better performance due to constant propagation of the + * defaults. + * + * @see XXH3_128bits_withSeed(), XXH3_128bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* data, size_t len); +/*! @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The block of data to be hashed, at least @p length bytes in size. + * @param len The length of @p data, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * @note + * seed == 0 produces the same results as @ref XXH3_64bits(). + * + * This variant generates a custom secret on the fly based on default secret + * altered using the @p seed value. + * + * While this operation is decently fast, note that it's not completely free. + * + * @see XXH3_128bits(), XXH3_128bits_withSecret(): other seeding variants + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSeed(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); +/*! + * @brief Calculates 128-bit variant of XXH3 with a custom "secret". + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @return The calculated 128-bit variant of XXH3 value. + * + * It's possible to provide any blob of bytes as a "secret" to generate the hash. + * This makes it more difficult for an external actor to prepare an intentional collision. + * The main condition is that @p secretSize *must* be large enough (>= @ref XXH3_SECRET_SIZE_MIN). + * However, the quality of the secret impacts the dispersion of the hash algorithm. + * Therefore, the secret _must_ look like a bunch of random bytes. + * Avoid "trivial" or structured data such as repeated sequences or a text document. + * Whenever in doubt about the "randomness" of the blob of bytes, + * consider employing @ref XXH3_generateSecret() instead (see below). + * It will generate a proper high entropy secret derived from the blob of bytes. + * Another advantage of using XXH3_generateSecret() is that + * it guarantees that all bits within the initial blob of bytes + * will impact every bit of the output. + * This is not necessarily the case when using the blob of bytes directly + * because, when hashing _small_ inputs, only a portion of the secret is employed. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_withSecret(XXH_NOESCAPE const void* data, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize); + +/******* Streaming *******/ +#ifndef XXH_NO_STREAM +/* + * Streaming requires state maintenance. + * This operation costs memory and CPU. + * As a consequence, streaming is slower than one-shot hashing. + * For better performance, prefer one-shot functions whenever applicable. + * + * XXH3_128bits uses the same XXH3_state_t as XXH3_64bits(). + * Use already declared XXH3_createState() and XXH3_freeState(). + * + * All reset and streaming functions have same meaning as their 64-bit counterpart. + */ + +/*! + * @brief Resets an @ref XXH3_state_t to begin a new hash. + * + * @param statePtr The state struct to reset. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret with default parameters. + * - Call it before @ref XXH3_128bits_update(). + * - Digest will be equivalent to `XXH3_128bits()`. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr); + +/*! + * @brief Resets an @ref XXH3_state_t with 64-bit seed to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * - This function resets `statePtr` and generate a secret from `seed`. + * - Call it before @ref XXH3_128bits_update(). + * - Digest will be equivalent to `XXH3_128bits_withSeed()`. + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed); +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr The state struct to reset. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * `secret` is referenced, it _must outlive_ the hash streaming session. + * Similar to one-shot API, `secretSize` must be >= @ref XXH3_SECRET_SIZE_MIN, + * and the quality of produced hash values depends on secret's entropy + * (secret's content should look like a bunch of random bytes). + * When in doubt about the randomness of a candidate `secret`, + * consider employing `XXH3_generateSecret()` instead (see below). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize); + +/*! + * @brief Consumes a block of @p input to an @ref XXH3_state_t. + * + * Call this to incrementally consume blocks of data. + * + * @param statePtr The state struct to update. + * @param input The block of data to be hashed, at least @p length bytes in size. + * @param length The length of @p input, in bytes. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @note + * The memory between @p input and @p input + @p length must be valid, + * readable, contiguous memory. However, if @p length is `0`, @p input may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + */ +XXH_PUBLIC_API XXH_errorcode XXH3_128bits_update (XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* input, size_t length); + +/*! + * @brief Returns the calculated XXH3 128-bit hash value from an @ref XXH3_state_t. + * + * @param statePtr The state struct to calculate the hash from. + * + * @pre + * @p statePtr must not be `NULL`. + * + * @return The calculated XXH3 128-bit hash value from that state. + * + * @note + * Calling XXH3_128bits_digest() will not affect @p statePtr, so you can update, + * digest, and update again. + * + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* statePtr); +#endif /* !XXH_NO_STREAM */ + +/* Following helper functions make it possible to compare XXH128_hast_t values. + * Since XXH128_hash_t is a structure, this capability is not offered by the language. + * Note: For better performance, these functions can be inlined using XXH_INLINE_ALL */ + +/*! + * @brief Check equality of two XXH128_hash_t values + * + * @param h1 The 128-bit hash value. + * @param h2 Another 128-bit hash value. + * + * @return `1` if `h1` and `h2` are equal. + * @return `0` if they are not. + */ +XXH_PUBLIC_API XXH_PUREF int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2); + +/*! + * @brief Compares two @ref XXH128_hash_t + * + * This comparator is compatible with stdlib's `qsort()`/`bsearch()`. + * + * @param h128_1 Left-hand side value + * @param h128_2 Right-hand side value + * + * @return >0 if @p h128_1 > @p h128_2 + * @return =0 if @p h128_1 == @p h128_2 + * @return <0 if @p h128_1 < @p h128_2 + */ +XXH_PUBLIC_API XXH_PUREF int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2); + + +/******* Canonical representation *******/ +typedef struct { unsigned char digest[sizeof(XXH128_hash_t)]; } XXH128_canonical_t; + + +/*! + * @brief Converts an @ref XXH128_hash_t to a big endian @ref XXH128_canonical_t. + * + * @param dst The @ref XXH128_canonical_t pointer to be stored to. + * @param hash The @ref XXH128_hash_t to be converted. + * + * @pre + * @p dst must not be `NULL`. + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API void XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash); + +/*! + * @brief Converts an @ref XXH128_canonical_t to a native @ref XXH128_hash_t. + * + * @param src The @ref XXH128_canonical_t to convert. + * + * @pre + * @p src must not be `NULL`. + * + * @return The converted hash. + * @see @ref canonical_representation_example "Canonical Representation Example" + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src); + + +#endif /* !XXH_NO_XXH3 */ +#endif /* XXH_NO_LONG_LONG */ + +/*! + * @} + */ +#endif /* XXHASH_H_5627135585666179 */ + + + +#if defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) +#define XXHASH_H_STATIC_13879238742 +/* **************************************************************************** + * This section contains declarations which are not guaranteed to remain stable. + * They may change in future versions, becoming incompatible with a different + * version of the library. + * These declarations should only be used with static linking. + * Never use them in association with dynamic linking! + ***************************************************************************** */ + +/* + * These definitions are only present to allow static allocation + * of XXH states, on stack or in a struct, for example. + * Never **ever** access their members directly. + */ + +/*! + * @internal + * @brief Structure for XXH32 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is + * an opaque type. This allows fields to safely be changed. + * + * Typedef'd to @ref XXH32_state_t. + * Do not access the members of this struct directly. + * @see XXH64_state_s, XXH3_state_s + */ +struct XXH32_state_s { + XXH32_hash_t total_len_32; /*!< Total length hashed, modulo 2^32 */ + XXH32_hash_t large_len; /*!< Whether the hash is >= 16 (handles @ref total_len_32 overflow) */ + XXH32_hash_t acc[4]; /*!< Accumulator lanes */ + unsigned char buffer[16]; /*!< Internal buffer for partial reads. */ + XXH32_hash_t bufferedSize; /*!< Amount of data in @ref buffer */ + XXH32_hash_t reserved; /*!< Reserved field. Do not read nor write to it. */ +}; /* typedef'd to XXH32_state_t */ + + +#ifndef XXH_NO_LONG_LONG /* defined when there is no 64-bit support */ + +/*! + * @internal + * @brief Structure for XXH64 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. Otherwise it is + * an opaque type. This allows fields to safely be changed. + * + * Typedef'd to @ref XXH64_state_t. + * Do not access the members of this struct directly. + * @see XXH32_state_s, XXH3_state_s + */ +struct XXH64_state_s { + XXH64_hash_t total_len; /*!< Total length hashed. This is always 64-bit. */ + XXH64_hash_t acc[4]; /*!< Accumulator lanes */ + unsigned char buffer[32]; /*!< Internal buffer for partial reads.. */ + XXH32_hash_t bufferedSize; /*!< Amount of data in @ref buffer */ + XXH32_hash_t reserved32; /*!< Reserved field, needed for padding anyways*/ + XXH64_hash_t reserved64; /*!< Reserved field. Do not read or write to it. */ +}; /* typedef'd to XXH64_state_t */ + +#ifndef XXH_NO_XXH3 + +#if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* >= C11 */ +# define XXH_ALIGN(n) _Alignas(n) +#elif defined(__cplusplus) && (__cplusplus >= 201103L) /* >= C++11 */ +/* In C++ alignas() is a keyword */ +# define XXH_ALIGN(n) alignas(n) +#elif defined(__GNUC__) +# define XXH_ALIGN(n) __attribute__ ((aligned(n))) +#elif defined(_MSC_VER) +# define XXH_ALIGN(n) __declspec(align(n)) +#else +# define XXH_ALIGN(n) /* disabled */ +#endif + +/* Old GCC versions only accept the attribute after the type in structures. */ +#if !(defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L)) /* C11+ */ \ + && ! (defined(__cplusplus) && (__cplusplus >= 201103L)) /* >= C++11 */ \ + && defined(__GNUC__) +# define XXH_ALIGN_MEMBER(align, type) type XXH_ALIGN(align) +#else +# define XXH_ALIGN_MEMBER(align, type) XXH_ALIGN(align) type +#endif + +/*! + * @brief The size of the internal XXH3 buffer. + * + * This is the optimal update size for incremental hashing. + * + * @see XXH3_64b_update(), XXH3_128b_update(). + */ +#define XXH3_INTERNALBUFFER_SIZE 256 + +/*! + * @internal + * @brief Default size of the secret buffer (and @ref XXH3_kSecret). + * + * This is the size used in @ref XXH3_kSecret and the seeded functions. + * + * Not to be confused with @ref XXH3_SECRET_SIZE_MIN. + */ +#define XXH3_SECRET_DEFAULT_SIZE 192 + +/*! + * @internal + * @brief Structure for XXH3 streaming API. + * + * @note This is only defined when @ref XXH_STATIC_LINKING_ONLY, + * @ref XXH_INLINE_ALL, or @ref XXH_IMPLEMENTATION is defined. + * Otherwise it is an opaque type. + * Never use this definition in combination with dynamic library. + * This allows fields to safely be changed in the future. + * + * @note ** This structure has a strict alignment requirement of 64 bytes!! ** + * Do not allocate this with `malloc()` or `new`, + * it will not be sufficiently aligned. + * Use @ref XXH3_createState() and @ref XXH3_freeState(), or stack allocation. + * + * Typedef'd to @ref XXH3_state_t. + * Do never access the members of this struct directly. + * + * @see XXH3_INITSTATE() for stack initialization. + * @see XXH3_createState(), XXH3_freeState(). + * @see XXH32_state_s, XXH64_state_s + */ +struct XXH3_state_s { + XXH_ALIGN_MEMBER(64, XXH64_hash_t acc[8]); + /*!< The 8 accumulators. See @ref XXH32_state_s::v and @ref XXH64_state_s::v */ + XXH_ALIGN_MEMBER(64, unsigned char customSecret[XXH3_SECRET_DEFAULT_SIZE]); + /*!< Used to store a custom secret generated from a seed. */ + XXH_ALIGN_MEMBER(64, unsigned char buffer[XXH3_INTERNALBUFFER_SIZE]); + /*!< The internal buffer. @see XXH32_state_s::mem32 */ + XXH32_hash_t bufferedSize; + /*!< The amount of memory in @ref buffer, @see XXH32_state_s::memsize */ + XXH32_hash_t useSeed; + /*!< Reserved field. Needed for padding on 64-bit. */ + size_t nbStripesSoFar; + /*!< Number or stripes processed. */ + XXH64_hash_t totalLen; + /*!< Total length hashed. 64-bit even on 32-bit targets. */ + size_t nbStripesPerBlock; + /*!< Number of stripes per block. */ + size_t secretLimit; + /*!< Size of @ref customSecret or @ref extSecret */ + XXH64_hash_t seed; + /*!< Seed for _withSeed variants. Must be zero otherwise, @see XXH3_INITSTATE() */ + XXH64_hash_t reserved64; + /*!< Reserved field. */ + const unsigned char* extSecret; + /*!< Reference to an external secret for the _withSecret variants, NULL + * for other variants. */ + /* note: there may be some padding at the end due to alignment on 64 bytes */ +}; /* typedef'd to XXH3_state_t */ + +#undef XXH_ALIGN_MEMBER + +/*! + * @brief Initializes a stack-allocated `XXH3_state_s`. + * + * When the @ref XXH3_state_t structure is merely emplaced on stack, + * it should be initialized with XXH3_INITSTATE() or a memset() + * in case its first reset uses XXH3_NNbits_reset_withSeed(). + * This init can be omitted if the first reset uses default or _withSecret mode. + * This operation isn't necessary when the state is created with XXH3_createState(). + * Note that this doesn't prepare the state for a streaming operation, + * it's still necessary to use XXH3_NNbits_reset*() afterwards. + */ +#define XXH3_INITSTATE(XXH3_state_ptr) \ + do { \ + XXH3_state_t* tmp_xxh3_state_ptr = (XXH3_state_ptr); \ + tmp_xxh3_state_ptr->seed = 0; \ + tmp_xxh3_state_ptr->extSecret = NULL; \ + } while(0) + + +/*! + * @brief Calculates the 128-bit hash of @p data using XXH3. + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param seed The 64-bit seed to alter the hash's output predictably. + * + * @pre + * The memory between @p data and @p data + @p len must be valid, + * readable, contiguous memory. However, if @p len is `0`, @p data may be + * `NULL`. In C++, this also must be *TriviallyCopyable*. + * + * @return The calculated 128-bit XXH3 value. + * + * @see @ref single_shot_example "Single Shot Example" for an example. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t XXH128(XXH_NOESCAPE const void* data, size_t len, XXH64_hash_t seed); + + +/* === Experimental API === */ +/* Symbols defined below must be considered tied to a specific library version. */ + +/*! + * @brief Derive a high-entropy secret from any user-defined content, named customSeed. + * + * @param secretBuffer A writable buffer for derived high-entropy secret data. + * @param secretSize Size of secretBuffer, in bytes. Must be >= XXH3_SECRET_SIZE_MIN. + * @param customSeed A user-defined content. + * @param customSeedSize Size of customSeed, in bytes. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * The generated secret can be used in combination with `*_withSecret()` functions. + * The `_withSecret()` variants are useful to provide a higher level of protection + * than 64-bit seed, as it becomes much more difficult for an external actor to + * guess how to impact the calculation logic. + * + * The function accepts as input a custom seed of any length and any content, + * and derives from it a high-entropy secret of length @p secretSize into an + * already allocated buffer @p secretBuffer. + * + * The generated secret can then be used with any `*_withSecret()` variant. + * The functions @ref XXH3_128bits_withSecret(), @ref XXH3_64bits_withSecret(), + * @ref XXH3_128bits_reset_withSecret() and @ref XXH3_64bits_reset_withSecret() + * are part of this list. They all accept a `secret` parameter + * which must be large enough for implementation reasons (>= @ref XXH3_SECRET_SIZE_MIN) + * _and_ feature very high entropy (consist of random-looking bytes). + * These conditions can be a high bar to meet, so @ref XXH3_generateSecret() can + * be employed to ensure proper quality. + * + * @p customSeed can be anything. It can have any size, even small ones, + * and its content can be anything, even "poor entropy" sources such as a bunch + * of zeroes. The resulting `secret` will nonetheless provide all required qualities. + * + * @pre + * - @p secretSize must be >= @ref XXH3_SECRET_SIZE_MIN + * - When @p customSeedSize > 0, supplying NULL as customSeed is undefined behavior. + * + * Example code: + * @code{.c} + * #include + * #include + * #include + * #define XXH_STATIC_LINKING_ONLY // expose unstable API + * #include "xxhash.h" + * // Hashes argv[2] using the entropy from argv[1]. + * int main(int argc, char* argv[]) + * { + * char secret[XXH3_SECRET_SIZE_MIN]; + * if (argv != 3) { return 1; } + * XXH3_generateSecret(secret, sizeof(secret), argv[1], strlen(argv[1])); + * XXH64_hash_t h = XXH3_64bits_withSecret( + * argv[2], strlen(argv[2]), + * secret, sizeof(secret) + * ); + * printf("%016llx\n", (unsigned long long) h); + * } + * @endcode + */ +XXH_PUBLIC_API XXH_errorcode XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize); + +/*! + * @brief Generate the same secret as the _withSeed() variants. + * + * @param secretBuffer A writable buffer of @ref XXH3_SECRET_DEFAULT_SIZE bytes + * @param seed The 64-bit seed to alter the hash result predictably. + * + * The generated secret can be used in combination with + *`*_withSecret()` and `_withSecretandSeed()` variants. + * + * Example C++ `std::string` hash class: + * @code{.cpp} + * #include + * #define XXH_STATIC_LINKING_ONLY // expose unstable API + * #include "xxhash.h" + * // Slow, seeds each time + * class HashSlow { + * XXH64_hash_t seed; + * public: + * HashSlow(XXH64_hash_t s) : seed{s} {} + * size_t operator()(const std::string& x) const { + * return size_t{XXH3_64bits_withSeed(x.c_str(), x.length(), seed)}; + * } + * }; + * // Fast, caches the seeded secret for future uses. + * class HashFast { + * unsigned char secret[XXH3_SECRET_DEFAULT_SIZE]; + * public: + * HashFast(XXH64_hash_t s) { + * XXH3_generateSecret_fromSeed(secret, seed); + * } + * size_t operator()(const std::string& x) const { + * return size_t{ + * XXH3_64bits_withSecret(x.c_str(), x.length(), secret, sizeof(secret)) + * }; + * } + * }; + * @endcode + */ +XXH_PUBLIC_API void XXH3_generateSecret_fromSeed(XXH_NOESCAPE void* secretBuffer, XXH64_hash_t seed); + +/*! + * @brief Maximum size of "short" key in bytes. + */ +#define XXH3_MIDSIZE_MAX 240 + +/*! + * @brief Calculates 64/128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The block of data to be hashed, at least @p len bytes in size. + * @param len The length of @p data, in bytes. + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed The 64-bit seed to alter the hash result predictably. + * + * These variants generate hash values using either: + * - @p seed for "short" keys (< @ref XXH3_MIDSIZE_MAX = 240 bytes) + * - @p secret for "large" keys (>= @ref XXH3_MIDSIZE_MAX). + * + * This generally benefits speed, compared to `_withSeed()` or `_withSecret()`. + * `_withSeed()` has to generate the secret on the fly for "large" keys. + * It's fast, but can be perceptible for "not so large" keys (< 1 KB). + * `_withSecret()` has to generate the masks on the fly for "small" keys, + * which requires more instructions than _withSeed() variants. + * Therefore, _withSecretandSeed variant combines the best of both worlds. + * + * When @p secret has been generated by XXH3_generateSecret_fromSeed(), + * this variant produces *exactly* the same results as `_withSeed()` variant, + * hence offering only a pure speed benefit on "large" input, + * by skipping the need to regenerate the secret for every large input. + * + * Another usage scenario is to hash the secret to a 64-bit hash value, + * for example with XXH3_64bits(), which then becomes the seed, + * and then employ both the seed and the secret in _withSecretandSeed(). + * On top of speed, an added benefit is that each bit in the secret + * has a 50% chance to swap each bit in the output, via its impact to the seed. + * + * This is not guaranteed when using the secret directly in "small data" scenarios, + * because only portions of the secret are employed for small data. + */ +XXH_PUBLIC_API XXH_PUREF XXH64_hash_t +XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* data, size_t len, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed); + +/*! + * @brief Calculates 128-bit seeded variant of XXH3 hash of @p data. + * + * @param data The memory segment to be hashed, at least @p len bytes in size. + * @param length The length of @p data, in bytes. + * @param secret The secret used to alter hash result predictably. + * @param secretSize The length of @p secret, in bytes (must be >= XXH3_SECRET_SIZE_MIN) + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(): contract is the same. + */ +XXH_PUBLIC_API XXH_PUREF XXH128_hash_t +XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +#ifndef XXH_NO_STREAM +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(). Contract is identical. + */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +/*! + * @brief Resets an @ref XXH3_state_t with secret data to begin a new hash. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * @param secret The secret data. + * @param secretSize The length of @p secret, in bytes. + * @param seed64 The 64-bit seed to alter the hash result predictably. + * + * @return @ref XXH_OK on success. + * @return @ref XXH_ERROR on failure. + * + * @see XXH3_64bits_withSecretandSeed(). Contract is identical. + * + * Note: there was a bug in an earlier version of this function (<= v0.8.2) + * that would make it generate an incorrect hash value + * when @p seed == 0 and @p length < XXH3_MIDSIZE_MAX + * and @p secret is different from XXH3_generateSecret_fromSeed(). + * As stated in the contract, the correct hash result must be + * the same as XXH3_128bits_withSeed() when @p length <= XXH3_MIDSIZE_MAX. + * Results generated by this older version are wrong, hence not comparable. + */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, + XXH_NOESCAPE const void* secret, size_t secretSize, + XXH64_hash_t seed64); + +#endif /* !XXH_NO_STREAM */ + +#endif /* !XXH_NO_XXH3 */ +#endif /* XXH_NO_LONG_LONG */ +#if defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) +# define XXH_IMPLEMENTATION +#endif + +#endif /* defined(XXH_STATIC_LINKING_ONLY) && !defined(XXHASH_H_STATIC_13879238742) */ + + +/* ======================================================================== */ +/* ======================================================================== */ +/* ======================================================================== */ + + +/*-********************************************************************** + * xxHash implementation + *-********************************************************************** + * xxHash's implementation used to be hosted inside xxhash.c. + * + * However, inlining requires implementation to be visible to the compiler, + * hence be included alongside the header. + * Previously, implementation was hosted inside xxhash.c, + * which was then #included when inlining was activated. + * This construction created issues with a few build and install systems, + * as it required xxhash.c to be stored in /include directory. + * + * xxHash implementation is now directly integrated within xxhash.h. + * As a consequence, xxhash.c is no longer needed in /include. + * + * xxhash.c is still available and is still useful. + * In a "normal" setup, when xxhash is not inlined, + * xxhash.h only exposes the prototypes and public symbols, + * while xxhash.c can be built into an object file xxhash.o + * which can then be linked into the final binary. + ************************************************************************/ + +#if ( defined(XXH_INLINE_ALL) || defined(XXH_PRIVATE_API) \ + || defined(XXH_IMPLEMENTATION) ) && !defined(XXH_IMPLEM_13a8737387) +# define XXH_IMPLEM_13a8737387 + +/* ************************************* +* Tuning parameters +***************************************/ + +/*! + * @defgroup tuning Tuning parameters + * @{ + * + * Various macros to control xxHash's behavior. + */ +#ifdef XXH_DOXYGEN +/*! + * @brief Define this to disable 64-bit code. + * + * Useful if only using the @ref XXH32_family and you have a strict C90 compiler. + */ +# define XXH_NO_LONG_LONG +# undef XXH_NO_LONG_LONG /* don't actually */ +/*! + * @brief Controls how unaligned memory is accessed. + * + * By default, access to unaligned memory is controlled by `memcpy()`, which is + * safe and portable. + * + * Unfortunately, on some target/compiler combinations, the generated assembly + * is sub-optimal. + * + * The below switch allow selection of a different access method + * in the search for improved performance. + * + * @par Possible options: + * + * - `XXH_FORCE_MEMORY_ACCESS=0` (default): `memcpy` + * @par + * Use `memcpy()`. Safe and portable. Note that most modern compilers will + * eliminate the function call and treat it as an unaligned access. + * + * - `XXH_FORCE_MEMORY_ACCESS=1`: `__attribute__((aligned(1)))` + * @par + * Depends on compiler extensions and is therefore not portable. + * This method is safe _if_ your compiler supports it, + * and *generally* as fast or faster than `memcpy`. + * + * - `XXH_FORCE_MEMORY_ACCESS=2`: Direct cast + * @par + * Casts directly and dereferences. This method doesn't depend on the + * compiler, but it violates the C standard as it directly dereferences an + * unaligned pointer. It can generate buggy code on targets which do not + * support unaligned memory accesses, but in some circumstances, it's the + * only known way to get the most performance. + * + * - `XXH_FORCE_MEMORY_ACCESS=3`: Byteshift + * @par + * Also portable. This can generate the best code on old compilers which don't + * inline small `memcpy()` calls, and it might also be faster on big-endian + * systems which lack a native byteswap instruction. However, some compilers + * will emit literal byteshifts even if the target supports unaligned access. + * + * + * @warning + * Methods 1 and 2 rely on implementation-defined behavior. Use these with + * care, as what works on one compiler/platform/optimization level may cause + * another to read garbage data or even crash. + * + * See https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html for details. + * + * Prefer these methods in priority order (0 > 3 > 1 > 2) + */ +# define XXH_FORCE_MEMORY_ACCESS 0 + +/*! + * @def XXH_SIZE_OPT + * @brief Controls how much xxHash optimizes for size. + * + * xxHash, when compiled, tends to result in a rather large binary size. This + * is mostly due to heavy usage to forced inlining and constant folding of the + * @ref XXH3_family to increase performance. + * + * However, some developers prefer size over speed. This option can + * significantly reduce the size of the generated code. When using the `-Os` + * or `-Oz` options on GCC or Clang, this is defined to 1 by default, + * otherwise it is defined to 0. + * + * Most of these size optimizations can be controlled manually. + * + * This is a number from 0-2. + * - `XXH_SIZE_OPT` == 0: Default. xxHash makes no size optimizations. Speed + * comes first. + * - `XXH_SIZE_OPT` == 1: Default for `-Os` and `-Oz`. xxHash is more + * conservative and disables hacks that increase code size. It implies the + * options @ref XXH_NO_INLINE_HINTS == 1, @ref XXH_FORCE_ALIGN_CHECK == 0, + * and @ref XXH3_NEON_LANES == 8 if they are not already defined. + * - `XXH_SIZE_OPT` == 2: xxHash tries to make itself as small as possible. + * Performance may cry. For example, the single shot functions just use the + * streaming API. + */ +# define XXH_SIZE_OPT 0 + +/*! + * @def XXH_FORCE_ALIGN_CHECK + * @brief If defined to non-zero, adds a special path for aligned inputs (XXH32() + * and XXH64() only). + * + * This is an important performance trick for architectures without decent + * unaligned memory access performance. + * + * It checks for input alignment, and when conditions are met, uses a "fast + * path" employing direct 32-bit/64-bit reads, resulting in _dramatically + * faster_ read speed. + * + * The check costs one initial branch per hash, which is generally negligible, + * but not zero. + * + * Moreover, it's not useful to generate an additional code path if memory + * access uses the same instruction for both aligned and unaligned + * addresses (e.g. x86 and aarch64). + * + * In these cases, the alignment check can be removed by setting this macro to 0. + * Then the code will always use unaligned memory access. + * Align check is automatically disabled on x86, x64, ARM64, and some ARM chips + * which are platforms known to offer good unaligned memory accesses performance. + * + * It is also disabled by default when @ref XXH_SIZE_OPT >= 1. + * + * This option does not affect XXH3 (only XXH32 and XXH64). + */ +# define XXH_FORCE_ALIGN_CHECK 0 + +/*! + * @def XXH_NO_INLINE_HINTS + * @brief When non-zero, sets all functions to `static`. + * + * By default, xxHash tries to force the compiler to inline almost all internal + * functions. + * + * This can usually improve performance due to reduced jumping and improved + * constant folding, but significantly increases the size of the binary which + * might not be favorable. + * + * Additionally, sometimes the forced inlining can be detrimental to performance, + * depending on the architecture. + * + * XXH_NO_INLINE_HINTS marks all internal functions as static, giving the + * compiler full control on whether to inline or not. + * + * When not optimizing (-O0), using `-fno-inline` with GCC or Clang, or if + * @ref XXH_SIZE_OPT >= 1, this will automatically be defined. + */ +# define XXH_NO_INLINE_HINTS 0 + +/*! + * @def XXH3_INLINE_SECRET + * @brief Determines whether to inline the XXH3 withSecret code. + * + * When the secret size is known, the compiler can improve the performance + * of XXH3_64bits_withSecret() and XXH3_128bits_withSecret(). + * + * However, if the secret size is not known, it doesn't have any benefit. This + * happens when xxHash is compiled into a global symbol. Therefore, if + * @ref XXH_INLINE_ALL is *not* defined, this will be defined to 0. + * + * Additionally, this defaults to 0 on GCC 12+, which has an issue with function pointers + * that are *sometimes* force inline on -Og, and it is impossible to automatically + * detect this optimization level. + */ +# define XXH3_INLINE_SECRET 0 + +/*! + * @def XXH32_ENDJMP + * @brief Whether to use a jump for `XXH32_finalize`. + * + * For performance, `XXH32_finalize` uses multiple branches in the finalizer. + * This is generally preferable for performance, + * but depending on exact architecture, a jmp may be preferable. + * + * This setting is only possibly making a difference for very small inputs. + */ +# define XXH32_ENDJMP 0 + +/*! + * @internal + * @brief Redefines old internal names. + * + * For compatibility with code that uses xxHash's internals before the names + * were changed to improve namespacing. There is no other reason to use this. + */ +# define XXH_OLD_NAMES +# undef XXH_OLD_NAMES /* don't actually use, it is ugly. */ + +/*! + * @def XXH_NO_STREAM + * @brief Disables the streaming API. + * + * When xxHash is not inlined and the streaming functions are not used, disabling + * the streaming functions can improve code size significantly, especially with + * the @ref XXH3_family which tends to make constant folded copies of itself. + */ +# define XXH_NO_STREAM +# undef XXH_NO_STREAM /* don't actually */ +#endif /* XXH_DOXYGEN */ +/*! + * @} + */ + +#ifndef XXH_FORCE_MEMORY_ACCESS /* can be defined externally, on command line for example */ + /* prefer __packed__ structures (method 1) for GCC + * < ARMv7 with unaligned access (e.g. Raspbian armhf) still uses byte shifting, so we use memcpy + * which for some reason does unaligned loads. */ +# if defined(__GNUC__) && !(defined(__ARM_ARCH) && __ARM_ARCH < 7 && defined(__ARM_FEATURE_UNALIGNED)) +# define XXH_FORCE_MEMORY_ACCESS 1 +# endif +#endif + +#ifndef XXH_SIZE_OPT + /* default to 1 for -Os or -Oz */ +# if (defined(__GNUC__) || defined(__clang__)) && defined(__OPTIMIZE_SIZE__) +# define XXH_SIZE_OPT 1 +# else +# define XXH_SIZE_OPT 0 +# endif +#endif + +#ifndef XXH_FORCE_ALIGN_CHECK /* can be defined externally */ + /* don't check on sizeopt, x86, aarch64, or arm when unaligned access is available */ +# if XXH_SIZE_OPT >= 1 || \ + defined(__i386) || defined(__x86_64__) || defined(__aarch64__) || defined(__ARM_FEATURE_UNALIGNED) \ + || defined(_M_IX86) || defined(_M_X64) || defined(_M_ARM64) || defined(_M_ARM) /* visual */ +# define XXH_FORCE_ALIGN_CHECK 0 +# else +# define XXH_FORCE_ALIGN_CHECK 1 +# endif +#endif + +#ifndef XXH_NO_INLINE_HINTS +# if XXH_SIZE_OPT >= 1 || defined(__NO_INLINE__) /* -O0, -fno-inline */ +# define XXH_NO_INLINE_HINTS 1 +# else +# define XXH_NO_INLINE_HINTS 0 +# endif +#endif + +#ifndef XXH3_INLINE_SECRET +# if (defined(__GNUC__) && !defined(__clang__) && __GNUC__ >= 12) \ + || !defined(XXH_INLINE_ALL) +# define XXH3_INLINE_SECRET 0 +# else +# define XXH3_INLINE_SECRET 1 +# endif +#endif + +#ifndef XXH32_ENDJMP +/* generally preferable for performance */ +# define XXH32_ENDJMP 0 +#endif + +/*! + * @defgroup impl Implementation + * @{ + */ + + +/* ************************************* +* Includes & Memory related functions +***************************************/ +#if defined(XXH_NO_STREAM) +/* nothing */ +#elif defined(XXH_NO_STDLIB) + +/* When requesting to disable any mention of stdlib, + * the library loses the ability to invoked malloc / free. + * In practice, it means that functions like `XXH*_createState()` + * will always fail, and return NULL. + * This flag is useful in situations where + * xxhash.h is integrated into some kernel, embedded or limited environment + * without access to dynamic allocation. + */ + +static XXH_CONSTF void* XXH_malloc(size_t s) { (void)s; return NULL; } +static void XXH_free(void* p) { (void)p; } + +#else + +/* + * Modify the local functions below should you wish to use + * different memory routines for malloc() and free() + */ +#include + +/*! + * @internal + * @brief Modify this function to use a different routine than malloc(). + */ +static XXH_MALLOCF void* XXH_malloc(size_t s) { return malloc(s); } + +/*! + * @internal + * @brief Modify this function to use a different routine than free(). + */ +static void XXH_free(void* p) { free(p); } + +#endif /* XXH_NO_STDLIB */ + +#include + +/*! + * @internal + * @brief Modify this function to use a different routine than memcpy(). + */ +static void* XXH_memcpy(void* dest, const void* src, size_t size) +{ + return memcpy(dest,src,size); +} + +#include /* ULLONG_MAX */ + + +/* ************************************* +* Compiler Specific Options +***************************************/ +#ifdef _MSC_VER /* Visual Studio warning fix */ +# pragma warning(disable : 4127) /* disable: C4127: conditional expression is constant */ +#endif + +#if XXH_NO_INLINE_HINTS /* disable inlining hints */ +# if defined(__GNUC__) || defined(__clang__) +# define XXH_FORCE_INLINE static __attribute__((__unused__)) +# else +# define XXH_FORCE_INLINE static +# endif +# define XXH_NO_INLINE static +/* enable inlining hints */ +#elif defined(__GNUC__) || defined(__clang__) +# define XXH_FORCE_INLINE static __inline__ __attribute__((__always_inline__, __unused__)) +# define XXH_NO_INLINE static __attribute__((__noinline__)) +#elif defined(_MSC_VER) /* Visual Studio */ +# define XXH_FORCE_INLINE static __forceinline +# define XXH_NO_INLINE static __declspec(noinline) +#elif defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L)) /* C99 */ +# define XXH_FORCE_INLINE static inline +# define XXH_NO_INLINE static +#else +# define XXH_FORCE_INLINE static +# define XXH_NO_INLINE static +#endif + +#if defined(XXH_INLINE_ALL) +# define XXH_STATIC XXH_FORCE_INLINE +#else +# define XXH_STATIC static +#endif + +#if XXH3_INLINE_SECRET +# define XXH3_WITH_SECRET_INLINE XXH_FORCE_INLINE +#else +# define XXH3_WITH_SECRET_INLINE XXH_NO_INLINE +#endif + +#if ((defined(sun) || defined(__sun)) && __cplusplus) /* Solaris includes __STDC_VERSION__ with C++. Tested with GCC 5.5 */ +# define XXH_RESTRICT /* disable */ +#elif defined (__STDC_VERSION__) && __STDC_VERSION__ >= 199901L /* >= C99 */ +# define XXH_RESTRICT restrict +#elif (defined (__GNUC__) && ((__GNUC__ > 3) || (__GNUC__ == 3 && __GNUC_MINOR__ >= 1))) \ + || (defined (__clang__)) \ + || (defined (_MSC_VER) && (_MSC_VER >= 1400)) \ + || (defined (__INTEL_COMPILER) && (__INTEL_COMPILER >= 1300)) +/* + * There are a LOT more compilers that recognize __restrict but this + * covers the major ones. + */ +# define XXH_RESTRICT __restrict +#else +# define XXH_RESTRICT /* disable */ +#endif + +/* ************************************* +* Debug +***************************************/ +/*! + * @ingroup tuning + * @def XXH_DEBUGLEVEL + * @brief Sets the debugging level. + * + * XXH_DEBUGLEVEL is expected to be defined externally, typically via the + * compiler's command line options. The value must be a number. + */ +#ifndef XXH_DEBUGLEVEL +# ifdef DEBUGLEVEL /* backwards compat */ +# define XXH_DEBUGLEVEL DEBUGLEVEL +# else +# define XXH_DEBUGLEVEL 0 +# endif +#endif + +#if (XXH_DEBUGLEVEL>=1) +# include /* note: can still be disabled with NDEBUG */ +# define XXH_ASSERT(c) assert(c) +#else +# if defined(__INTEL_COMPILER) +# define XXH_ASSERT(c) XXH_ASSUME((unsigned char) (c)) +# else +# define XXH_ASSERT(c) XXH_ASSUME(c) +# endif +#endif + +/* note: use after variable declarations */ +#ifndef XXH_STATIC_ASSERT +# if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= 201112L) /* C11 */ +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { _Static_assert((c),m); } while(0) +# elif defined(__cplusplus) && (__cplusplus >= 201103L) /* C++11 */ +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { static_assert((c),m); } while(0) +# else +# define XXH_STATIC_ASSERT_WITH_MESSAGE(c,m) do { struct xxh_sa { char x[(c) ? 1 : -1]; }; } while(0) +# endif +# define XXH_STATIC_ASSERT(c) XXH_STATIC_ASSERT_WITH_MESSAGE((c),#c) +#endif + +/*! + * @internal + * @def XXH_COMPILER_GUARD(var) + * @brief Used to prevent unwanted optimizations for @p var. + * + * It uses an empty GCC inline assembly statement with a register constraint + * which forces @p var into a general purpose register (eg eax, ebx, ecx + * on x86) and marks it as modified. + * + * This is used in a few places to avoid unwanted autovectorization (e.g. + * XXH32_round()). All vectorization we want is explicit via intrinsics, + * and _usually_ isn't wanted elsewhere. + * + * We also use it to prevent unwanted constant folding for AArch64 in + * XXH3_initCustomSecret_scalar(). + */ +#if defined(__GNUC__) || defined(__clang__) +# define XXH_COMPILER_GUARD(var) __asm__("" : "+r" (var)) +#else +# define XXH_COMPILER_GUARD(var) ((void)0) +#endif + +/* Specifically for NEON vectors which use the "w" constraint, on + * Clang. */ +#if defined(__clang__) && defined(__ARM_ARCH) && !defined(__wasm__) +# define XXH_COMPILER_GUARD_CLANG_NEON(var) __asm__("" : "+w" (var)) +#else +# define XXH_COMPILER_GUARD_CLANG_NEON(var) ((void)0) +#endif + +/* ************************************* +* Basic Types +***************************************/ +#if !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) +# ifdef _AIX +# include +# else +# include +# endif + typedef uint8_t xxh_u8; +#else + typedef unsigned char xxh_u8; +#endif +typedef XXH32_hash_t xxh_u32; + +#ifdef XXH_OLD_NAMES +# warning "XXH_OLD_NAMES is planned to be removed starting v0.9. If the program depends on it, consider moving away from it by employing newer type names directly" +# define BYTE xxh_u8 +# define U8 xxh_u8 +# define U32 xxh_u32 +#endif + +/* *** Memory access *** */ + +/*! + * @internal + * @fn xxh_u32 XXH_read32(const void* ptr) + * @brief Reads an unaligned 32-bit integer from @p ptr in native endianness. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit native endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readLE32(const void* ptr) + * @brief Reads an unaligned 32-bit little endian integer from @p ptr. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit little endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readBE32(const void* ptr) + * @brief Reads an unaligned 32-bit big endian integer from @p ptr. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * + * @param ptr The pointer to read from. + * @return The 32-bit big endian integer from the bytes at @p ptr. + */ + +/*! + * @internal + * @fn xxh_u32 XXH_readLE32_align(const void* ptr, XXH_alignment align) + * @brief Like @ref XXH_readLE32(), but has an option for aligned reads. + * + * Affected by @ref XXH_FORCE_MEMORY_ACCESS. + * Note that when @ref XXH_FORCE_ALIGN_CHECK == 0, the @p align parameter is + * always @ref XXH_alignment::XXH_unaligned. + * + * @param ptr The pointer to read from. + * @param align Whether @p ptr is aligned. + * @pre + * If @p align == @ref XXH_alignment::XXH_aligned, @p ptr must be 4 byte + * aligned. + * @return The 32-bit little endian integer from the bytes at @p ptr. + */ + +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) +/* + * Manual byteshift. Best for old compilers which don't inline memcpy. + * We actually directly use XXH_readLE32 and XXH_readBE32. + */ +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) + +/* + * Force direct memory access. Only works on CPU which support unaligned memory + * access in hardware. + */ +static xxh_u32 XXH_read32(const void* memPtr) { return *(const xxh_u32*) memPtr; } + +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) + +/* + * __attribute__((aligned(1))) is supported by gcc and clang. Originally the + * documentation claimed that it only increased the alignment, but actually it + * can decrease it on gcc, clang, and icc: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, + * https://gcc.godbolt.org/z/xYez1j67Y. + */ +#ifdef XXH_OLD_NAMES +typedef union { xxh_u32 u32; } __attribute__((__packed__)) unalign; +#endif +static xxh_u32 XXH_read32(const void* ptr) +{ + typedef __attribute__((__aligned__(1))) xxh_u32 xxh_unalign32; + return *((const xxh_unalign32*)ptr); +} + +#else + +/* + * Portable and safe solution. Generally efficient. + * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html + */ +static xxh_u32 XXH_read32(const void* memPtr) +{ + xxh_u32 val; + XXH_memcpy(&val, memPtr, sizeof(val)); + return val; +} + +#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ + + +/* *** Endianness *** */ + +/*! + * @ingroup tuning + * @def XXH_CPU_LITTLE_ENDIAN + * @brief Whether the target is little endian. + * + * Defined to 1 if the target is little endian, or 0 if it is big endian. + * It can be defined externally, for example on the compiler command line. + * + * If it is not defined, + * a runtime check (which is usually constant folded) is used instead. + * + * @note + * This is not necessarily defined to an integer constant. + * + * @see XXH_isLittleEndian() for the runtime check. + */ +#ifndef XXH_CPU_LITTLE_ENDIAN +/* + * Try to detect endianness automatically, to avoid the nonstandard behavior + * in `XXH_isLittleEndian()` + */ +# if defined(_WIN32) /* Windows is always little endian */ \ + || defined(__LITTLE_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) +# define XXH_CPU_LITTLE_ENDIAN 1 +# elif defined(__BIG_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) +# define XXH_CPU_LITTLE_ENDIAN 0 +# else +/*! + * @internal + * @brief Runtime check for @ref XXH_CPU_LITTLE_ENDIAN. + * + * Most compilers will constant fold this. + */ +static int XXH_isLittleEndian(void) +{ + /* + * Portable and well-defined behavior. + * Don't use static: it is detrimental to performance. + */ + const union { xxh_u32 u; xxh_u8 c[4]; } one = { 1 }; + return one.c[0]; +} +# define XXH_CPU_LITTLE_ENDIAN XXH_isLittleEndian() +# endif +#endif + + + + +/* **************************************** +* Compiler-specific Functions and Macros +******************************************/ +#define XXH_GCC_VERSION (__GNUC__ * 100 + __GNUC_MINOR__) + +#ifdef __has_builtin +# define XXH_HAS_BUILTIN(x) __has_builtin(x) +#else +# define XXH_HAS_BUILTIN(x) 0 +#endif + + + +/* + * C23 and future versions have standard "unreachable()". + * Once it has been implemented reliably we can add it as an + * additional case: + * + * ``` + * #if defined(__STDC_VERSION__) && (__STDC_VERSION__ >= XXH_C23_VN) + * # include + * # ifdef unreachable + * # define XXH_UNREACHABLE() unreachable() + * # endif + * #endif + * ``` + * + * Note C++23 also has std::unreachable() which can be detected + * as follows: + * ``` + * #if defined(__cpp_lib_unreachable) && (__cpp_lib_unreachable >= 202202L) + * # include + * # define XXH_UNREACHABLE() std::unreachable() + * #endif + * ``` + * NB: `__cpp_lib_unreachable` is defined in the `` header. + * We don't use that as including `` in `extern "C"` blocks + * doesn't work on GCC12 + */ + +#if XXH_HAS_BUILTIN(__builtin_unreachable) +# define XXH_UNREACHABLE() __builtin_unreachable() + +#elif defined(_MSC_VER) +# define XXH_UNREACHABLE() __assume(0) + +#else +# define XXH_UNREACHABLE() +#endif + +#if XXH_HAS_BUILTIN(__builtin_assume) +# define XXH_ASSUME(c) __builtin_assume(c) +#else +# define XXH_ASSUME(c) if (!(c)) { XXH_UNREACHABLE(); } +#endif + +/*! + * @internal + * @def XXH_rotl32(x,r) + * @brief 32-bit rotate left. + * + * @param x The 32-bit integer to be rotated. + * @param r The number of bits to rotate. + * @pre + * @p r > 0 && @p r < 32 + * @note + * @p x and @p r may be evaluated multiple times. + * @return The rotated result. + */ +#if !defined(NO_CLANG_BUILTIN) && XXH_HAS_BUILTIN(__builtin_rotateleft32) \ + && XXH_HAS_BUILTIN(__builtin_rotateleft64) +# define XXH_rotl32 __builtin_rotateleft32 +# define XXH_rotl64 __builtin_rotateleft64 +#elif XXH_HAS_BUILTIN(__builtin_stdc_rotate_left) +# define XXH_rotl32 __builtin_stdc_rotate_left +# define XXH_rotl64 __builtin_stdc_rotate_left +/* Note: although _rotl exists for minGW (GCC under windows), performance seems poor */ +#elif defined(_MSC_VER) +# define XXH_rotl32(x,r) _rotl(x,r) +# define XXH_rotl64(x,r) _rotl64(x,r) +#else +# define XXH_rotl32(x,r) (((x) << (r)) | ((x) >> (32 - (r)))) +# define XXH_rotl64(x,r) (((x) << (r)) | ((x) >> (64 - (r)))) +#endif + +/*! + * @internal + * @fn xxh_u32 XXH_swap32(xxh_u32 x) + * @brief A 32-bit byteswap. + * + * @param x The 32-bit integer to byteswap. + * @return @p x, byteswapped. + */ +#if defined(_MSC_VER) /* Visual Studio */ +# define XXH_swap32 _byteswap_ulong +#elif XXH_GCC_VERSION >= 403 +# define XXH_swap32 __builtin_bswap32 +#else +static xxh_u32 XXH_swap32 (xxh_u32 x) +{ + return ((x << 24) & 0xff000000 ) | + ((x << 8) & 0x00ff0000 ) | + ((x >> 8) & 0x0000ff00 ) | + ((x >> 24) & 0x000000ff ); +} +#endif + + +/* *************************** +* Memory reads +*****************************/ + +/*! + * @internal + * @brief Enum to indicate whether a pointer is aligned. + */ +typedef enum { + XXH_aligned, /*!< Aligned */ + XXH_unaligned /*!< Possibly unaligned */ +} XXH_alignment; + +/* + * XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. + * + * This is ideal for older compilers which don't inline memcpy. + */ +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) + +XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[0] + | ((xxh_u32)bytePtr[1] << 8) + | ((xxh_u32)bytePtr[2] << 16) + | ((xxh_u32)bytePtr[3] << 24); +} + +XXH_FORCE_INLINE xxh_u32 XXH_readBE32(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[3] + | ((xxh_u32)bytePtr[2] << 8) + | ((xxh_u32)bytePtr[1] << 16) + | ((xxh_u32)bytePtr[0] << 24); +} + +#else +XXH_FORCE_INLINE xxh_u32 XXH_readLE32(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_read32(ptr) : XXH_swap32(XXH_read32(ptr)); +} + +static xxh_u32 XXH_readBE32(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_swap32(XXH_read32(ptr)) : XXH_read32(ptr); +} +#endif + +XXH_FORCE_INLINE xxh_u32 +XXH_readLE32_align(const void* ptr, XXH_alignment align) +{ + if (align==XXH_unaligned) { + return XXH_readLE32(ptr); + } else { + return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u32*)ptr : XXH_swap32(*(const xxh_u32*)ptr); + } +} + + +/* ************************************* +* Misc +***************************************/ +/*! @ingroup public */ +XXH_PUBLIC_API unsigned XXH_versionNumber (void) { return XXH_VERSION_NUMBER; } + + +/* ******************************************************************* +* 32-bit hash functions +*********************************************************************/ +/*! + * @} + * @defgroup XXH32_impl XXH32 implementation + * @ingroup impl + * + * Details on the XXH32 implementation. + * @{ + */ + /* #define instead of static const, to be used as initializers */ +#define XXH_PRIME32_1 0x9E3779B1U /*!< 0b10011110001101110111100110110001 */ +#define XXH_PRIME32_2 0x85EBCA77U /*!< 0b10000101111010111100101001110111 */ +#define XXH_PRIME32_3 0xC2B2AE3DU /*!< 0b11000010101100101010111000111101 */ +#define XXH_PRIME32_4 0x27D4EB2FU /*!< 0b00100111110101001110101100101111 */ +#define XXH_PRIME32_5 0x165667B1U /*!< 0b00010110010101100110011110110001 */ + +#ifdef XXH_OLD_NAMES +# define PRIME32_1 XXH_PRIME32_1 +# define PRIME32_2 XXH_PRIME32_2 +# define PRIME32_3 XXH_PRIME32_3 +# define PRIME32_4 XXH_PRIME32_4 +# define PRIME32_5 XXH_PRIME32_5 +#endif + +/*! + * @internal + * @brief Normal stripe processing routine. + * + * This shuffles the bits so that any bit from @p input impacts several bits in + * @p acc. + * + * @param acc The accumulator lane. + * @param input The stripe of input to mix. + * @return The mixed accumulator lane. + */ +static xxh_u32 XXH32_round(xxh_u32 acc, xxh_u32 input) +{ + acc += input * XXH_PRIME32_2; + acc = XXH_rotl32(acc, 13); + acc *= XXH_PRIME32_1; +#if (defined(__SSE4_1__) || defined(__aarch64__) || defined(__wasm_simd128__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) + /* + * UGLY HACK: + * A compiler fence is used to prevent GCC and Clang from + * autovectorizing the XXH32 loop (pragmas and attributes don't work for some + * reason) without globally disabling SSE4.1. + * + * The reason we want to avoid vectorization is because despite working on + * 4 integers at a time, there are multiple factors slowing XXH32 down on + * SSE4: + * - There's a ridiculous amount of lag from pmulld (10 cycles of latency on + * newer chips!) making it slightly slower to multiply four integers at + * once compared to four integers independently. Even when pmulld was + * fastest, Sandy/Ivy Bridge, it is still not worth it to go into SSE + * just to multiply unless doing a long operation. + * + * - Four instructions are required to rotate, + * movqda tmp, v // not required with VEX encoding + * pslld tmp, 13 // tmp <<= 13 + * psrld v, 19 // x >>= 19 + * por v, tmp // x |= tmp + * compared to one for scalar: + * roll v, 13 // reliably fast across the board + * shldl v, v, 13 // Sandy Bridge and later prefer this for some reason + * + * - Instruction level parallelism is actually more beneficial here because + * the SIMD actually serializes this operation: While v1 is rotating, v2 + * can load data, while v3 can multiply. SSE forces them to operate + * together. + * + * This is also enabled on AArch64, as Clang is *very aggressive* in vectorizing + * the loop. NEON is only faster on the A53, and with the newer cores, it is less + * than half the speed. + * + * Additionally, this is used on WASM SIMD128 because it JITs to the same + * SIMD instructions and has the same issue. + */ + XXH_COMPILER_GUARD(acc); +#endif + return acc; +} + +/*! + * @internal + * @brief Mixes all bits to finalize the hash. + * + * The final mix ensures that all input bits have a chance to impact any bit in + * the output digest, resulting in an unbiased distribution. + * + * @param hash The hash to avalanche. + * @return The avalanched hash. + */ +static xxh_u32 XXH32_avalanche(xxh_u32 hash) +{ + hash ^= hash >> 15; + hash *= XXH_PRIME32_2; + hash ^= hash >> 13; + hash *= XXH_PRIME32_3; + hash ^= hash >> 16; + return hash; +} + +#define XXH_get32bits(p) XXH_readLE32_align(p, align) + +/*! + * @internal + * @brief Sets up the initial accumulator state for XXH32(). + */ +XXH_FORCE_INLINE void +XXH32_initAccs(xxh_u32 *acc, xxh_u32 seed) +{ + XXH_ASSERT(acc != NULL); + acc[0] = seed + XXH_PRIME32_1 + XXH_PRIME32_2; + acc[1] = seed + XXH_PRIME32_2; + acc[2] = seed + 0; + acc[3] = seed - XXH_PRIME32_1; +} + +/*! + * @internal + * @brief Consumes a block of data for XXH32(). + * + * @return the end input pointer. + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH32_consumeLong( + xxh_u32 *XXH_RESTRICT acc, + xxh_u8 const *XXH_RESTRICT input, + size_t len, + XXH_alignment align +) +{ + const xxh_u8* const bEnd = input + len; + const xxh_u8* const limit = bEnd - 15; + XXH_ASSERT(acc != NULL); + XXH_ASSERT(input != NULL); + XXH_ASSERT(len >= 16); + do { + acc[0] = XXH32_round(acc[0], XXH_get32bits(input)); input += 4; + acc[1] = XXH32_round(acc[1], XXH_get32bits(input)); input += 4; + acc[2] = XXH32_round(acc[2], XXH_get32bits(input)); input += 4; + acc[3] = XXH32_round(acc[3], XXH_get32bits(input)); input += 4; + } while (input < limit); + + return input; +} + +/*! + * @internal + * @brief Merges the accumulator lanes together for XXH32() + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u32 +XXH32_mergeAccs(const xxh_u32 *acc) +{ + XXH_ASSERT(acc != NULL); + return XXH_rotl32(acc[0], 1) + XXH_rotl32(acc[1], 7) + + XXH_rotl32(acc[2], 12) + XXH_rotl32(acc[3], 18); +} + +/*! + * @internal + * @brief Processes the last 0-15 bytes of @p ptr. + * + * There may be up to 15 bytes remaining to consume from the input. + * This final stage will digest them to ensure that all input bytes are present + * in the final mix. + * + * @param hash The hash to finalize. + * @param ptr The pointer to the remaining input. + * @param len The remaining length, modulo 16. + * @param align Whether @p ptr is aligned. + * @return The finalized hash. + * @see XXH64_finalize(). + */ +static XXH_PUREF xxh_u32 +XXH32_finalize(xxh_u32 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) +{ +#define XXH_PROCESS1 do { \ + hash += (*ptr++) * XXH_PRIME32_5; \ + hash = XXH_rotl32(hash, 11) * XXH_PRIME32_1; \ +} while (0) + +#define XXH_PROCESS4 do { \ + hash += XXH_get32bits(ptr) * XXH_PRIME32_3; \ + ptr += 4; \ + hash = XXH_rotl32(hash, 17) * XXH_PRIME32_4; \ +} while (0) + + if (ptr==NULL) XXH_ASSERT(len == 0); + + /* Compact rerolled version; generally faster */ + if (!XXH32_ENDJMP) { + len &= 15; + while (len >= 4) { + XXH_PROCESS4; + len -= 4; + } + while (len > 0) { + XXH_PROCESS1; + --len; + } + return XXH32_avalanche(hash); + } else { + switch(len&15) /* or switch(bEnd - p) */ { + case 12: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 8: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 4: XXH_PROCESS4; + return XXH32_avalanche(hash); + + case 13: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 9: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 5: XXH_PROCESS4; + XXH_PROCESS1; + return XXH32_avalanche(hash); + + case 14: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 10: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 6: XXH_PROCESS4; + XXH_PROCESS1; + XXH_PROCESS1; + return XXH32_avalanche(hash); + + case 15: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 11: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 7: XXH_PROCESS4; + XXH_FALLTHROUGH; /* fallthrough */ + case 3: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 2: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 1: XXH_PROCESS1; + XXH_FALLTHROUGH; /* fallthrough */ + case 0: return XXH32_avalanche(hash); + } + XXH_ASSERT(0); + return hash; /* reaching this point is deemed impossible */ + } +} + +#ifdef XXH_OLD_NAMES +# define PROCESS1 XXH_PROCESS1 +# define PROCESS4 XXH_PROCESS4 +#else +# undef XXH_PROCESS1 +# undef XXH_PROCESS4 +#endif + +/*! + * @internal + * @brief The implementation for @ref XXH32(). + * + * @param input , len , seed Directly passed from @ref XXH32(). + * @param align Whether @p input is aligned. + * @return The calculated hash. + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u32 +XXH32_endian_align(const xxh_u8* input, size_t len, xxh_u32 seed, XXH_alignment align) +{ + xxh_u32 h32; + + if (input==NULL) XXH_ASSERT(len == 0); + + if (len>=16) { + xxh_u32 acc[4]; + XXH32_initAccs(acc, seed); + + input = XXH32_consumeLong(acc, input, len, align); + + h32 = XXH32_mergeAccs(acc); + } else { + h32 = seed + XXH_PRIME32_5; + } + + h32 += (xxh_u32)len; + + return XXH32_finalize(h32, input, len&15, align); +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32 (const void* input, size_t len, XXH32_hash_t seed) +{ +#if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 + /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ + XXH32_state_t state; + XXH32_reset(&state, seed); + XXH32_update(&state, (const xxh_u8*)input, len); + return XXH32_digest(&state); +#else + if (XXH_FORCE_ALIGN_CHECK) { + if ((((size_t)input) & 3) == 0) { /* Input is 4-bytes aligned, leverage the speed benefit */ + return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); + } } + + return XXH32_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); +#endif +} + + + +/******* Hash streaming *******/ +#ifndef XXH_NO_STREAM +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_state_t* XXH32_createState(void) +{ + return (XXH32_state_t*)XXH_malloc(sizeof(XXH32_state_t)); +} +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode XXH32_freeState(XXH32_state_t* statePtr) +{ + XXH_free(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API void XXH32_copyState(XXH32_state_t* dstState, const XXH32_state_t* srcState) +{ + XXH_memcpy(dstState, srcState, sizeof(*dstState)); +} + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode XXH32_reset(XXH32_state_t* statePtr, XXH32_hash_t seed) +{ + XXH_ASSERT(statePtr != NULL); + memset(statePtr, 0, sizeof(*statePtr)); + XXH32_initAccs(statePtr->acc, seed); + return XXH_OK; +} + + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH_errorcode +XXH32_update(XXH32_state_t* state, const void* input, size_t len) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + state->total_len_32 += (XXH32_hash_t)len; + state->large_len |= (XXH32_hash_t)((len>=16) | (state->total_len_32>=16)); + + XXH_ASSERT(state->bufferedSize < sizeof(state->buffer)); + if (len < sizeof(state->buffer) - state->bufferedSize) { /* fill in tmp buffer */ + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + { const xxh_u8* xinput = (const xxh_u8*)input; + const xxh_u8* const bEnd = xinput + len; + + if (state->bufferedSize) { /* non-empty buffer: complete first */ + XXH_memcpy(state->buffer + state->bufferedSize, xinput, sizeof(state->buffer) - state->bufferedSize); + xinput += sizeof(state->buffer) - state->bufferedSize; + /* then process one round */ + (void)XXH32_consumeLong(state->acc, state->buffer, sizeof(state->buffer), XXH_aligned); + state->bufferedSize = 0; + } + + XXH_ASSERT(xinput <= bEnd); + if ((size_t)(bEnd - xinput) >= sizeof(state->buffer)) { + /* Process the remaining data */ + xinput = XXH32_consumeLong(state->acc, xinput, (size_t)(bEnd - xinput), XXH_unaligned); + } + + if (xinput < bEnd) { + /* Copy the leftover to the tmp buffer */ + XXH_memcpy(state->buffer, xinput, (size_t)(bEnd-xinput)); + state->bufferedSize = (unsigned)(bEnd-xinput); + } + } + + return XXH_OK; +} + + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32_digest(const XXH32_state_t* state) +{ + xxh_u32 h32; + + if (state->large_len) { + h32 = XXH32_mergeAccs(state->acc); + } else { + h32 = state->acc[2] /* == seed */ + XXH_PRIME32_5; + } + + h32 += state->total_len_32; + + return XXH32_finalize(h32, state->buffer, state->bufferedSize, XXH_aligned); +} +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API void XXH32_canonicalFromHash(XXH32_canonical_t* dst, XXH32_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH32_canonical_t) == sizeof(XXH32_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap32(hash); + XXH_memcpy(dst, &hash, sizeof(*dst)); +} +/*! @ingroup XXH32_family */ +XXH_PUBLIC_API XXH32_hash_t XXH32_hashFromCanonical(const XXH32_canonical_t* src) +{ + return XXH_readBE32(src); +} + + +#ifndef XXH_NO_LONG_LONG + +/* ******************************************************************* +* 64-bit hash functions +*********************************************************************/ +/*! + * @} + * @ingroup impl + * @{ + */ +/******* Memory access *******/ + +typedef XXH64_hash_t xxh_u64; + +#ifdef XXH_OLD_NAMES +# define U64 xxh_u64 +#endif + +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) +/* + * Manual byteshift. Best for old compilers which don't inline memcpy. + * We actually directly use XXH_readLE64 and XXH_readBE64. + */ +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==2)) + +/* Force direct memory access. Only works on CPU which support unaligned memory access in hardware */ +static xxh_u64 XXH_read64(const void* memPtr) +{ + return *(const xxh_u64*) memPtr; +} + +#elif (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==1)) + +/* + * __attribute__((aligned(1))) is supported by gcc and clang. Originally the + * documentation claimed that it only increased the alignment, but actually it + * can decrease it on gcc, clang, and icc: + * https://gcc.gnu.org/bugzilla/show_bug.cgi?id=69502, + * https://gcc.godbolt.org/z/xYez1j67Y. + */ +#ifdef XXH_OLD_NAMES +typedef union { xxh_u32 u32; xxh_u64 u64; } __attribute__((__packed__)) unalign64; +#endif +static xxh_u64 XXH_read64(const void* ptr) +{ + typedef __attribute__((__aligned__(1))) xxh_u64 xxh_unalign64; + return *((const xxh_unalign64*)ptr); +} + +#else + +/* + * Portable and safe solution. Generally efficient. + * see: https://fastcompression.blogspot.com/2015/08/accessing-unaligned-memory.html + */ +static xxh_u64 XXH_read64(const void* memPtr) +{ + xxh_u64 val; + XXH_memcpy(&val, memPtr, sizeof(val)); + return val; +} + +#endif /* XXH_FORCE_DIRECT_MEMORY_ACCESS */ + +#if defined(_MSC_VER) /* Visual Studio */ +# define XXH_swap64 _byteswap_uint64 +#elif XXH_GCC_VERSION >= 403 +# define XXH_swap64 __builtin_bswap64 +#else +static xxh_u64 XXH_swap64(xxh_u64 x) +{ + return ((x << 56) & 0xff00000000000000ULL) | + ((x << 40) & 0x00ff000000000000ULL) | + ((x << 24) & 0x0000ff0000000000ULL) | + ((x << 8) & 0x000000ff00000000ULL) | + ((x >> 8) & 0x00000000ff000000ULL) | + ((x >> 24) & 0x0000000000ff0000ULL) | + ((x >> 40) & 0x000000000000ff00ULL) | + ((x >> 56) & 0x00000000000000ffULL); +} +#endif + + +/* XXH_FORCE_MEMORY_ACCESS==3 is an endian-independent byteshift load. */ +#if (defined(XXH_FORCE_MEMORY_ACCESS) && (XXH_FORCE_MEMORY_ACCESS==3)) + +XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[0] + | ((xxh_u64)bytePtr[1] << 8) + | ((xxh_u64)bytePtr[2] << 16) + | ((xxh_u64)bytePtr[3] << 24) + | ((xxh_u64)bytePtr[4] << 32) + | ((xxh_u64)bytePtr[5] << 40) + | ((xxh_u64)bytePtr[6] << 48) + | ((xxh_u64)bytePtr[7] << 56); +} + +XXH_FORCE_INLINE xxh_u64 XXH_readBE64(const void* memPtr) +{ + const xxh_u8* bytePtr = (const xxh_u8 *)memPtr; + return bytePtr[7] + | ((xxh_u64)bytePtr[6] << 8) + | ((xxh_u64)bytePtr[5] << 16) + | ((xxh_u64)bytePtr[4] << 24) + | ((xxh_u64)bytePtr[3] << 32) + | ((xxh_u64)bytePtr[2] << 40) + | ((xxh_u64)bytePtr[1] << 48) + | ((xxh_u64)bytePtr[0] << 56); +} + +#else +XXH_FORCE_INLINE xxh_u64 XXH_readLE64(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_read64(ptr) : XXH_swap64(XXH_read64(ptr)); +} + +static xxh_u64 XXH_readBE64(const void* ptr) +{ + return XXH_CPU_LITTLE_ENDIAN ? XXH_swap64(XXH_read64(ptr)) : XXH_read64(ptr); +} +#endif + +XXH_FORCE_INLINE xxh_u64 +XXH_readLE64_align(const void* ptr, XXH_alignment align) +{ + if (align==XXH_unaligned) + return XXH_readLE64(ptr); + else + return XXH_CPU_LITTLE_ENDIAN ? *(const xxh_u64*)ptr : XXH_swap64(*(const xxh_u64*)ptr); +} + + +/******* xxh64 *******/ +/*! + * @} + * @defgroup XXH64_impl XXH64 implementation + * @ingroup impl + * + * Details on the XXH64 implementation. + * @{ + */ +/* #define rather that static const, to be used as initializers */ +#define XXH_PRIME64_1 0x9E3779B185EBCA87ULL /*!< 0b1001111000110111011110011011000110000101111010111100101010000111 */ +#define XXH_PRIME64_2 0xC2B2AE3D27D4EB4FULL /*!< 0b1100001010110010101011100011110100100111110101001110101101001111 */ +#define XXH_PRIME64_3 0x165667B19E3779F9ULL /*!< 0b0001011001010110011001111011000110011110001101110111100111111001 */ +#define XXH_PRIME64_4 0x85EBCA77C2B2AE63ULL /*!< 0b1000010111101011110010100111011111000010101100101010111001100011 */ +#define XXH_PRIME64_5 0x27D4EB2F165667C5ULL /*!< 0b0010011111010100111010110010111100010110010101100110011111000101 */ + +#ifdef XXH_OLD_NAMES +# define PRIME64_1 XXH_PRIME64_1 +# define PRIME64_2 XXH_PRIME64_2 +# define PRIME64_3 XXH_PRIME64_3 +# define PRIME64_4 XXH_PRIME64_4 +# define PRIME64_5 XXH_PRIME64_5 +#endif + +/*! @copydoc XXH32_round */ +static xxh_u64 XXH64_round(xxh_u64 acc, xxh_u64 input) +{ + acc += input * XXH_PRIME64_2; + acc = XXH_rotl64(acc, 31); + acc *= XXH_PRIME64_1; +#if (defined(__AVX512F__)) && !defined(XXH_ENABLE_AUTOVECTORIZE) + /* + * DISABLE AUTOVECTORIZATION: + * A compiler fence is used to prevent GCC and Clang from + * autovectorizing the XXH64 loop (pragmas and attributes don't work for some + * reason) without globally disabling AVX512. + * + * Autovectorization of XXH64 tends to be detrimental, + * though the exact outcome may change depending on exact cpu and compiler version. + * For information, it has been reported as detrimental for Skylake-X, + * but possibly beneficial for Zen4. + * + * The default is to disable auto-vectorization, + * but you can select to enable it instead using `XXH_ENABLE_AUTOVECTORIZE` build variable. + */ + XXH_COMPILER_GUARD(acc); +#endif + return acc; +} + +static xxh_u64 XXH64_mergeRound(xxh_u64 acc, xxh_u64 val) +{ + val = XXH64_round(0, val); + acc ^= val; + acc = acc * XXH_PRIME64_1 + XXH_PRIME64_4; + return acc; +} + +/*! @copydoc XXH32_avalanche */ +static xxh_u64 XXH64_avalanche(xxh_u64 hash) +{ + hash ^= hash >> 33; + hash *= XXH_PRIME64_2; + hash ^= hash >> 29; + hash *= XXH_PRIME64_3; + hash ^= hash >> 32; + return hash; +} + + +#define XXH_get64bits(p) XXH_readLE64_align(p, align) + +/*! + * @internal + * @brief Sets up the initial accumulator state for XXH64(). + */ +XXH_FORCE_INLINE void +XXH64_initAccs(xxh_u64 *acc, xxh_u64 seed) +{ + XXH_ASSERT(acc != NULL); + acc[0] = seed + XXH_PRIME64_1 + XXH_PRIME64_2; + acc[1] = seed + XXH_PRIME64_2; + acc[2] = seed + 0; + acc[3] = seed - XXH_PRIME64_1; +} + +/*! + * @internal + * @brief Consumes a block of data for XXH64(). + * + * @return the end input pointer. + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH64_consumeLong( + xxh_u64 *XXH_RESTRICT acc, + xxh_u8 const *XXH_RESTRICT input, + size_t len, + XXH_alignment align +) +{ + const xxh_u8* const bEnd = input + len; + const xxh_u8* const limit = bEnd - 31; + XXH_ASSERT(acc != NULL); + XXH_ASSERT(input != NULL); + XXH_ASSERT(len >= 32); + do { + /* reroll on 32-bit */ + if (sizeof(void *) < sizeof(xxh_u64)) { + size_t i; + for (i = 0; i < 4; i++) { + acc[i] = XXH64_round(acc[i], XXH_get64bits(input)); + input += 8; + } + } else { + acc[0] = XXH64_round(acc[0], XXH_get64bits(input)); input += 8; + acc[1] = XXH64_round(acc[1], XXH_get64bits(input)); input += 8; + acc[2] = XXH64_round(acc[2], XXH_get64bits(input)); input += 8; + acc[3] = XXH64_round(acc[3], XXH_get64bits(input)); input += 8; + } + } while (input < limit); + + return input; +} + +/*! + * @internal + * @brief Merges the accumulator lanes together for XXH64() + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u64 +XXH64_mergeAccs(const xxh_u64 *acc) +{ + XXH_ASSERT(acc != NULL); + { + xxh_u64 h64 = XXH_rotl64(acc[0], 1) + XXH_rotl64(acc[1], 7) + + XXH_rotl64(acc[2], 12) + XXH_rotl64(acc[3], 18); + /* reroll on 32-bit */ + if (sizeof(void *) < sizeof(xxh_u64)) { + size_t i; + for (i = 0; i < 4; i++) { + h64 = XXH64_mergeRound(h64, acc[i]); + } + } else { + h64 = XXH64_mergeRound(h64, acc[0]); + h64 = XXH64_mergeRound(h64, acc[1]); + h64 = XXH64_mergeRound(h64, acc[2]); + h64 = XXH64_mergeRound(h64, acc[3]); + } + return h64; + } +} + +/*! + * @internal + * @brief Processes the last 0-31 bytes of @p ptr. + * + * There may be up to 31 bytes remaining to consume from the input. + * This final stage will digest them to ensure that all input bytes are present + * in the final mix. + * + * @param hash The hash to finalize. + * @param ptr The pointer to the remaining input. + * @param len The remaining length, modulo 32. + * @param align Whether @p ptr is aligned. + * @return The finalized hash + * @see XXH32_finalize(). + */ +XXH_STATIC XXH_PUREF xxh_u64 +XXH64_finalize(xxh_u64 hash, const xxh_u8* ptr, size_t len, XXH_alignment align) +{ + if (ptr==NULL) XXH_ASSERT(len == 0); + len &= 31; + while (len >= 8) { + xxh_u64 const k1 = XXH64_round(0, XXH_get64bits(ptr)); + ptr += 8; + hash ^= k1; + hash = XXH_rotl64(hash,27) * XXH_PRIME64_1 + XXH_PRIME64_4; + len -= 8; + } + if (len >= 4) { + hash ^= (xxh_u64)(XXH_get32bits(ptr)) * XXH_PRIME64_1; + ptr += 4; + hash = XXH_rotl64(hash, 23) * XXH_PRIME64_2 + XXH_PRIME64_3; + len -= 4; + } + while (len > 0) { + hash ^= (*ptr++) * XXH_PRIME64_5; + hash = XXH_rotl64(hash, 11) * XXH_PRIME64_1; + --len; + } + return XXH64_avalanche(hash); +} + +#ifdef XXH_OLD_NAMES +# define PROCESS1_64 XXH_PROCESS1_64 +# define PROCESS4_64 XXH_PROCESS4_64 +# define PROCESS8_64 XXH_PROCESS8_64 +#else +# undef XXH_PROCESS1_64 +# undef XXH_PROCESS4_64 +# undef XXH_PROCESS8_64 +#endif + +/*! + * @internal + * @brief The implementation for @ref XXH64(). + * + * @param input , len , seed Directly passed from @ref XXH64(). + * @param align Whether @p input is aligned. + * @return The calculated hash. + */ +XXH_FORCE_INLINE XXH_PUREF xxh_u64 +XXH64_endian_align(const xxh_u8* input, size_t len, xxh_u64 seed, XXH_alignment align) +{ + xxh_u64 h64; + if (input==NULL) XXH_ASSERT(len == 0); + + if (len>=32) { /* Process a large block of data */ + xxh_u64 acc[4]; + XXH64_initAccs(acc, seed); + + input = XXH64_consumeLong(acc, input, len, align); + + h64 = XXH64_mergeAccs(acc); + } else { + h64 = seed + XXH_PRIME64_5; + } + + h64 += (xxh_u64) len; + + return XXH64_finalize(h64, input, len, align); +} + + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64 (XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ +#if !defined(XXH_NO_STREAM) && XXH_SIZE_OPT >= 2 + /* Simple version, good for code maintenance, but unfortunately slow for small inputs */ + XXH64_state_t state; + XXH64_reset(&state, seed); + XXH64_update(&state, (const xxh_u8*)input, len); + return XXH64_digest(&state); +#else + if (XXH_FORCE_ALIGN_CHECK) { + if ((((size_t)input) & 7)==0) { /* Input is aligned, let's leverage the speed advantage */ + return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_aligned); + } } + + return XXH64_endian_align((const xxh_u8*)input, len, seed, XXH_unaligned); + +#endif +} + +/******* Hash Streaming *******/ +#ifndef XXH_NO_STREAM +/*! @ingroup XXH64_family*/ +XXH_PUBLIC_API XXH64_state_t* XXH64_createState(void) +{ + return (XXH64_state_t*)XXH_malloc(sizeof(XXH64_state_t)); +} +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode XXH64_freeState(XXH64_state_t* statePtr) +{ + XXH_free(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API void XXH64_copyState(XXH_NOESCAPE XXH64_state_t* dstState, const XXH64_state_t* srcState) +{ + XXH_memcpy(dstState, srcState, sizeof(*dstState)); +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode XXH64_reset(XXH_NOESCAPE XXH64_state_t* statePtr, XXH64_hash_t seed) +{ + XXH_ASSERT(statePtr != NULL); + memset(statePtr, 0, sizeof(*statePtr)); + XXH64_initAccs(statePtr->acc, seed); + return XXH_OK; +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH_errorcode +XXH64_update (XXH_NOESCAPE XXH64_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + state->total_len += len; + + XXH_ASSERT(state->bufferedSize <= sizeof(state->buffer)); + if (len < sizeof(state->buffer) - state->bufferedSize) { /* fill in tmp buffer */ + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + { const xxh_u8* xinput = (const xxh_u8*)input; + const xxh_u8* const bEnd = xinput + len; + + if (state->bufferedSize) { /* non-empty buffer => complete first */ + XXH_memcpy(state->buffer + state->bufferedSize, xinput, sizeof(state->buffer) - state->bufferedSize); + xinput += sizeof(state->buffer) - state->bufferedSize; + /* and process one round */ + (void)XXH64_consumeLong(state->acc, state->buffer, sizeof(state->buffer), XXH_aligned); + state->bufferedSize = 0; + } + + XXH_ASSERT(xinput <= bEnd); + if ((size_t)(bEnd - xinput) >= sizeof(state->buffer)) { + /* Process the remaining data */ + xinput = XXH64_consumeLong(state->acc, xinput, (size_t)(bEnd - xinput), XXH_unaligned); + } + + if (xinput < bEnd) { + /* Copy the leftover to the tmp buffer */ + XXH_memcpy(state->buffer, xinput, (size_t)(bEnd-xinput)); + state->bufferedSize = (unsigned)(bEnd-xinput); + } + } + + return XXH_OK; +} + + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64_digest(XXH_NOESCAPE const XXH64_state_t* state) +{ + xxh_u64 h64; + + if (state->total_len >= 32) { + h64 = XXH64_mergeAccs(state->acc); + } else { + h64 = state->acc[2] /*seed*/ + XXH_PRIME64_5; + } + + h64 += (xxh_u64) state->total_len; + + return XXH64_finalize(h64, state->buffer, (size_t)state->total_len, XXH_aligned); +} +#endif /* !XXH_NO_STREAM */ + +/******* Canonical representation *******/ + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API void XXH64_canonicalFromHash(XXH_NOESCAPE XXH64_canonical_t* dst, XXH64_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH64_canonical_t) == sizeof(XXH64_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) hash = XXH_swap64(hash); + XXH_memcpy(dst, &hash, sizeof(*dst)); +} + +/*! @ingroup XXH64_family */ +XXH_PUBLIC_API XXH64_hash_t XXH64_hashFromCanonical(XXH_NOESCAPE const XXH64_canonical_t* src) +{ + return XXH_readBE64(src); +} + +#ifndef XXH_NO_XXH3 + +/* ********************************************************************* +* XXH3 +* New generation hash designed for speed on small keys and vectorization +************************************************************************ */ +/*! + * @} + * @defgroup XXH3_impl XXH3 implementation + * @ingroup impl + * @{ + */ + +/* === Compiler specifics === */ + + +#if (defined(__GNUC__) && (__GNUC__ >= 3)) \ + || (defined(__INTEL_COMPILER) && (__INTEL_COMPILER >= 800)) \ + || defined(__clang__) +# define XXH_likely(x) __builtin_expect(x, 1) +# define XXH_unlikely(x) __builtin_expect(x, 0) +#else +# define XXH_likely(x) (x) +# define XXH_unlikely(x) (x) +#endif + +#ifndef XXH_HAS_INCLUDE +# ifdef __has_include +/* + * Not defined as XXH_HAS_INCLUDE(x) (function-like) because + * this causes segfaults in Apple Clang 4.2 (on Mac OS X 10.7 Lion) + */ +# define XXH_HAS_INCLUDE __has_include +# else +# define XXH_HAS_INCLUDE(x) 0 +# endif +#endif + +#if defined(__GNUC__) || defined(__clang__) +# if defined(__ARM_FEATURE_SVE) +# include +# endif +# if defined(__ARM_NEON__) || defined(__ARM_NEON) \ + || (defined(_M_ARM) && _M_ARM >= 7) \ + || defined(_M_ARM64) || defined(_M_ARM64EC) \ + || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* WASM SIMD128 via SIMDe */ +# define inline __inline__ /* circumvent a clang bug */ +# include +# undef inline +# elif defined(__AVX2__) +# include +# elif defined(__SSE2__) +# include +# elif defined(__loongarch_sx) +# include +# endif +#endif + +#if defined(_MSC_VER) +# include +#endif + +/* + * One goal of XXH3 is to make it fast on both 32-bit and 64-bit, while + * remaining a true 64-bit/128-bit hash function. + * + * This is done by prioritizing a subset of 64-bit operations that can be + * emulated without too many steps on the average 32-bit machine. + * + * For example, these two lines seem similar, and run equally fast on 64-bit: + * + * xxh_u64 x; + * x ^= (x >> 47); // good + * x ^= (x >> 13); // bad + * + * However, to a 32-bit machine, there is a major difference. + * + * x ^= (x >> 47) looks like this: + * + * x.lo ^= (x.hi >> (47 - 32)); + * + * while x ^= (x >> 13) looks like this: + * + * // note: funnel shifts are not usually cheap. + * x.lo ^= (x.lo >> 13) | (x.hi << (32 - 13)); + * x.hi ^= (x.hi >> 13); + * + * The first one is significantly faster than the second, simply because the + * shift is larger than 32. This means: + * - All the bits we need are in the upper 32 bits, so we can ignore the lower + * 32 bits in the shift. + * - The shift result will always fit in the lower 32 bits, and therefore, + * we can ignore the upper 32 bits in the xor. + * + * Thanks to this optimization, XXH3 only requires these features to be efficient: + * + * - Usable unaligned access + * - A 32-bit or 64-bit ALU + * - If 32-bit, a decent ADC instruction + * - A 32 or 64-bit multiply with a 64-bit result + * - For the 128-bit variant, a decent byteswap helps short inputs. + * + * The first two are already required by XXH32, and almost all 32-bit and 64-bit + * platforms which can run XXH32 can run XXH3 efficiently. + * + * Thumb-1, the classic 16-bit only subset of ARM's instruction set, is one + * notable exception. + * + * First of all, Thumb-1 lacks support for the UMULL instruction which + * performs the important long multiply. This means numerous __aeabi_lmul + * calls. + * + * Second of all, the 8 functional registers are just not enough. + * Setup for __aeabi_lmul, byteshift loads, pointers, and all arithmetic need + * Lo registers, and this shuffling results in thousands more MOVs than A32. + * + * A32 and T32 don't have this limitation. They can access all 14 registers, + * do a 32->64 multiply with UMULL, and the flexible operand allowing free + * shifts is helpful, too. + * + * Therefore, we do a quick sanity check. + * + * If compiling Thumb-1 for a target which supports ARM instructions, we will + * emit a warning, as it is not a "sane" platform to compile for. + * + * Usually, if this happens, it is because of an accident and you probably need + * to specify -march, as you likely meant to compile for a newer architecture. + * + * Credit: large sections of the vectorial and asm source code paths + * have been contributed by @easyaspi314 + */ +#if defined(__thumb__) && !defined(__thumb2__) && defined(__ARM_ARCH_ISA_ARM) +# warning "XXH3 is highly inefficient without ARM or Thumb-2." +#endif + +/* ========================================== + * Vectorization detection + * ========================================== */ + +#ifdef XXH_DOXYGEN +/*! + * @ingroup tuning + * @brief Overrides the vectorization implementation chosen for XXH3. + * + * Can be defined to 0 to disable SIMD or any of the values mentioned in + * @ref XXH_VECTOR_TYPE. + * + * If this is not defined, it uses predefined macros to determine the best + * implementation. + */ +# define XXH_VECTOR XXH_SCALAR +/*! + * @ingroup tuning + * @brief Selects the minimum alignment for XXH3's accumulators. + * + * When using SIMD, this should match the alignment required for said vector + * type, so, for example, 32 for AVX2. + * + * Default: Auto detected. + */ +# define XXH_ACC_ALIGN 8 +#endif + +/* Actual definition */ +#ifndef XXH_DOXYGEN +#endif + +#ifndef XXH_VECTOR /* can be defined on command line */ +# if defined(__ARM_FEATURE_SVE) +# define XXH_VECTOR XXH_SVE +# elif ( \ + defined(__ARM_NEON__) || defined(__ARM_NEON) /* gcc */ \ + || defined(_M_ARM) || defined(_M_ARM64) || defined(_M_ARM64EC) /* msvc */ \ + || (defined(__wasm_simd128__) && XXH_HAS_INCLUDE()) /* wasm simd128 via SIMDe */ \ + ) && ( \ + defined(_WIN32) || defined(__LITTLE_ENDIAN__) /* little endian only */ \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_LITTLE_ENDIAN__) \ + ) +# define XXH_VECTOR XXH_NEON +# elif defined(__AVX512F__) +# define XXH_VECTOR XXH_AVX512 +# elif defined(__AVX2__) +# define XXH_VECTOR XXH_AVX2 +# elif defined(__SSE2__) || defined(_M_AMD64) || defined(_M_X64) || (defined(_M_IX86_FP) && (_M_IX86_FP == 2)) +# define XXH_VECTOR XXH_SSE2 +# elif (defined(__PPC64__) && defined(__POWER8_VECTOR__)) \ + || (defined(__s390x__) && defined(__VEC__)) \ + && defined(__GNUC__) /* TODO: IBM XL */ +# define XXH_VECTOR XXH_VSX +# elif defined(__loongarch_sx) +# define XXH_VECTOR XXH_LSX +# else +# define XXH_VECTOR XXH_SCALAR +# endif +#endif + +/* __ARM_FEATURE_SVE is only supported by GCC & Clang. */ +#if (XXH_VECTOR == XXH_SVE) && !defined(__ARM_FEATURE_SVE) +# ifdef _MSC_VER +# pragma warning(once : 4606) +# else +# warning "__ARM_FEATURE_SVE isn't supported. Use SCALAR instead." +# endif +# undef XXH_VECTOR +# define XXH_VECTOR XXH_SCALAR +#endif + +/* + * Controls the alignment of the accumulator, + * for compatibility with aligned vector loads, which are usually faster. + */ +#ifndef XXH_ACC_ALIGN +# if defined(XXH_X86DISPATCH) +# define XXH_ACC_ALIGN 64 /* for compatibility with avx512 */ +# elif XXH_VECTOR == XXH_SCALAR /* scalar */ +# define XXH_ACC_ALIGN 8 +# elif XXH_VECTOR == XXH_SSE2 /* sse2 */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_AVX2 /* avx2 */ +# define XXH_ACC_ALIGN 32 +# elif XXH_VECTOR == XXH_NEON /* neon */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_VSX /* vsx */ +# define XXH_ACC_ALIGN 16 +# elif XXH_VECTOR == XXH_AVX512 /* avx512 */ +# define XXH_ACC_ALIGN 64 +# elif XXH_VECTOR == XXH_SVE /* sve */ +# define XXH_ACC_ALIGN 64 +# elif XXH_VECTOR == XXH_LSX /* lsx */ +# define XXH_ACC_ALIGN 64 +# endif +#endif + +#if defined(XXH_X86DISPATCH) || XXH_VECTOR == XXH_SSE2 \ + || XXH_VECTOR == XXH_AVX2 || XXH_VECTOR == XXH_AVX512 +# define XXH_SEC_ALIGN XXH_ACC_ALIGN +#elif XXH_VECTOR == XXH_SVE +# define XXH_SEC_ALIGN XXH_ACC_ALIGN +#else +# define XXH_SEC_ALIGN 8 +#endif + +#if defined(__GNUC__) || defined(__clang__) +# define XXH_ALIASING __attribute__((__may_alias__)) +#else +# define XXH_ALIASING /* nothing */ +#endif + +/* + * UGLY HACK: + * GCC usually generates the best code with -O3 for xxHash. + * + * However, when targeting AVX2, it is overzealous in its unrolling resulting + * in code roughly 3/4 the speed of Clang. + * + * There are other issues, such as GCC splitting _mm256_loadu_si256 into + * _mm_loadu_si128 + _mm256_inserti128_si256. This is an optimization which + * only applies to Sandy and Ivy Bridge... which don't even support AVX2. + * + * That is why when compiling the AVX2 version, it is recommended to use either + * -O2 -mavx2 -march=haswell + * or + * -O2 -mavx2 -mno-avx256-split-unaligned-load + * for decent performance, or to use Clang instead. + * + * Fortunately, we can control the first one with a pragma that forces GCC into + * -O2, but the other one we can't control without "failed to inline always + * inline function due to target mismatch" warnings. + */ +#if XXH_VECTOR == XXH_AVX2 /* AVX2 */ \ + && defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ + && defined(__OPTIMIZE__) && XXH_SIZE_OPT <= 0 /* respect -O0 and -Os */ +# pragma GCC push_options +# pragma GCC optimize("-O2") +#endif + +#if XXH_VECTOR == XXH_NEON + +/* + * UGLY HACK: While AArch64 GCC on Linux does not seem to care, on macOS, GCC -O3 + * optimizes out the entire hashLong loop because of the aliasing violation. + * + * However, GCC is also inefficient at load-store optimization with vld1q/vst1q, + * so the only option is to mark it as aliasing. + */ +typedef uint64x2_t xxh_aliasing_uint64x2_t XXH_ALIASING; + +/*! + * @internal + * @brief `vld1q_u64` but faster and alignment-safe. + * + * On AArch64, unaligned access is always safe, but on ARMv7-a, it is only + * *conditionally* safe (`vld1` has an alignment bit like `movdq[ua]` in x86). + * + * GCC for AArch64 sees `vld1q_u8` as an intrinsic instead of a load, so it + * prohibits load-store optimizations. Therefore, a direct dereference is used. + * + * Otherwise, `vld1q_u8` is used with `vreinterpretq_u8_u64` to do a safe + * unaligned load. + */ +#if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) +XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) /* silence -Wcast-align */ +{ + return *(xxh_aliasing_uint64x2_t const *)ptr; +} +#else +XXH_FORCE_INLINE uint64x2_t XXH_vld1q_u64(void const* ptr) +{ + return vreinterpretq_u64_u8(vld1q_u8((uint8_t const*)ptr)); +} +#endif + +/*! + * @internal + * @brief `vmlal_u32` on low and high halves of a vector. + * + * This is a workaround for AArch64 GCC < 11 which implemented arm_neon.h with + * inline assembly and were therefore incapable of merging the `vget_{low, high}_u32` + * with `vmlal_u32`. + */ +#if defined(__aarch64__) && defined(__GNUC__) && !defined(__clang__) && __GNUC__ < 11 +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + /* Inline assembly is the only way */ + __asm__("umlal %0.2d, %1.2s, %2.2s" : "+w" (acc) : "w" (lhs), "w" (rhs)); + return acc; +} +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + /* This intrinsic works as expected */ + return vmlal_high_u32(acc, lhs, rhs); +} +#else +/* Portable intrinsic versions */ +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_low_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + return vmlal_u32(acc, vget_low_u32(lhs), vget_low_u32(rhs)); +} +/*! @copydoc XXH_vmlal_low_u32 + * Assume the compiler converts this to vmlal_high_u32 on aarch64 */ +XXH_FORCE_INLINE uint64x2_t +XXH_vmlal_high_u32(uint64x2_t acc, uint32x4_t lhs, uint32x4_t rhs) +{ + return vmlal_u32(acc, vget_high_u32(lhs), vget_high_u32(rhs)); +} +#endif + +/*! + * @ingroup tuning + * @brief Controls the NEON to scalar ratio for XXH3 + * + * This can be set to 2, 4, 6, or 8. + * + * ARM Cortex CPUs are _very_ sensitive to how their pipelines are used. + * + * For example, the Cortex-A73 can dispatch 3 micro-ops per cycle, but only 2 of those + * can be NEON. If you are only using NEON instructions, you are only using 2/3 of the CPU + * bandwidth. + * + * This is even more noticeable on the more advanced cores like the Cortex-A76 which + * can dispatch 8 micro-ops per cycle, but still only 2 NEON micro-ops at once. + * + * Therefore, to make the most out of the pipeline, it is beneficial to run 6 NEON lanes + * and 2 scalar lanes, which is chosen by default. + * + * This does not apply to Apple processors or 32-bit processors, which run better with + * full NEON. These will default to 8. Additionally, size-optimized builds run 8 lanes. + * + * This change benefits CPUs with large micro-op buffers without negatively affecting + * most other CPUs: + * + * | Chipset | Dispatch type | NEON only | 6:2 hybrid | Diff. | + * |:----------------------|:--------------------|----------:|-----------:|------:| + * | Snapdragon 730 (A76) | 2 NEON/8 micro-ops | 8.8 GB/s | 10.1 GB/s | ~16% | + * | Snapdragon 835 (A73) | 2 NEON/3 micro-ops | 5.1 GB/s | 5.3 GB/s | ~5% | + * | Marvell PXA1928 (A53) | In-order dual-issue | 1.9 GB/s | 1.9 GB/s | 0% | + * | Apple M1 | 4 NEON/8 micro-ops | 37.3 GB/s | 36.1 GB/s | ~-3% | + * + * It also seems to fix some bad codegen on GCC, making it almost as fast as clang. + * + * When using WASM SIMD128, if this is 2 or 6, SIMDe will scalarize 2 of the lanes meaning + * it effectively becomes worse 4. + * + * @see XXH3_accumulate_512_neon() + */ +# ifndef XXH3_NEON_LANES +# if (defined(__aarch64__) || defined(__arm64__) || defined(_M_ARM64) || defined(_M_ARM64EC)) \ + && !defined(__APPLE__) && XXH_SIZE_OPT <= 0 +# define XXH3_NEON_LANES 6 +# else +# define XXH3_NEON_LANES XXH_ACC_NB +# endif +# endif +#endif /* XXH_VECTOR == XXH_NEON */ + +/* + * VSX and Z Vector helpers. + * + * This is very messy, and any pull requests to clean this up are welcome. + * + * There are a lot of problems with supporting VSX and s390x, due to + * inconsistent intrinsics, spotty coverage, and multiple endiannesses. + */ +#if XXH_VECTOR == XXH_VSX +/* Annoyingly, these headers _may_ define three macros: `bool`, `vector`, + * and `pixel`. This is a problem for obvious reasons. + * + * These keywords are unnecessary; the spec literally says they are + * equivalent to `__bool`, `__vector`, and `__pixel` and may be undef'd + * after including the header. + * + * We use pragma push_macro/pop_macro to keep the namespace clean. */ +# pragma push_macro("bool") +# pragma push_macro("vector") +# pragma push_macro("pixel") +/* silence potential macro redefined warnings */ +# undef bool +# undef vector +# undef pixel + +# if defined(__s390x__) +# include +# else +# include +# endif + +/* Restore the original macro values, if applicable. */ +# pragma pop_macro("pixel") +# pragma pop_macro("vector") +# pragma pop_macro("bool") + +typedef __vector unsigned long long xxh_u64x2; +typedef __vector unsigned char xxh_u8x16; +typedef __vector unsigned xxh_u32x4; + +/* + * UGLY HACK: Similar to aarch64 macOS GCC, s390x GCC has the same aliasing issue. + */ +typedef xxh_u64x2 xxh_aliasing_u64x2 XXH_ALIASING; + +# ifndef XXH_VSX_BE +# if defined(__BIG_ENDIAN__) \ + || (defined(__BYTE_ORDER__) && __BYTE_ORDER__ == __ORDER_BIG_ENDIAN__) +# define XXH_VSX_BE 1 +# elif defined(__VEC_ELEMENT_REG_ORDER__) && __VEC_ELEMENT_REG_ORDER__ == __ORDER_BIG_ENDIAN__ +# warning "-maltivec=be is not recommended. Please use native endianness." +# define XXH_VSX_BE 1 +# else +# define XXH_VSX_BE 0 +# endif +# endif /* !defined(XXH_VSX_BE) */ + +# if XXH_VSX_BE +# if defined(__POWER9_VECTOR__) || (defined(__clang__) && defined(__s390x__)) +# define XXH_vec_revb vec_revb +# else +/*! + * A polyfill for POWER9's vec_revb(). + */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_revb(xxh_u64x2 val) +{ + xxh_u8x16 const vByteSwap = { 0x07, 0x06, 0x05, 0x04, 0x03, 0x02, 0x01, 0x00, + 0x0F, 0x0E, 0x0D, 0x0C, 0x0B, 0x0A, 0x09, 0x08 }; + return vec_perm(val, val, vByteSwap); +} +# endif +# endif /* XXH_VSX_BE */ + +/*! + * Performs an unaligned vector load and byte swaps it on big endian. + */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_loadu(const void *ptr) +{ + xxh_u64x2 ret; + XXH_memcpy(&ret, ptr, sizeof(xxh_u64x2)); +# if XXH_VSX_BE + ret = XXH_vec_revb(ret); +# endif + return ret; +} + +/* + * vec_mulo and vec_mule are very problematic intrinsics on PowerPC + * + * These intrinsics weren't added until GCC 8, despite existing for a while, + * and they are endian dependent. Also, their meaning swap depending on version. + * */ +# if defined(__s390x__) + /* s390x is always big endian, no issue on this platform */ +# define XXH_vec_mulo vec_mulo +# define XXH_vec_mule vec_mule +# elif defined(__clang__) && XXH_HAS_BUILTIN(__builtin_altivec_vmuleuw) && !defined(__ibmxl__) +/* Clang has a better way to control this, we can just use the builtin which doesn't swap. */ + /* The IBM XL Compiler (which defined __clang__) only implements the vec_* operations */ +# define XXH_vec_mulo __builtin_altivec_vmulouw +# define XXH_vec_mule __builtin_altivec_vmuleuw +# else +/* gcc needs inline assembly */ +/* Adapted from https://github.com/google/highwayhash/blob/master/highwayhash/hh_vsx.h. */ +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mulo(xxh_u32x4 a, xxh_u32x4 b) +{ + xxh_u64x2 result; + __asm__("vmulouw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); + return result; +} +XXH_FORCE_INLINE xxh_u64x2 XXH_vec_mule(xxh_u32x4 a, xxh_u32x4 b) +{ + xxh_u64x2 result; + __asm__("vmuleuw %0, %1, %2" : "=v" (result) : "v" (a), "v" (b)); + return result; +} +# endif /* XXH_vec_mulo, XXH_vec_mule */ +#endif /* XXH_VECTOR == XXH_VSX */ + +#if XXH_VECTOR == XXH_SVE +#define ACCRND(acc, offset) \ +do { \ + svuint64_t input_vec = svld1_u64(mask, xinput + offset); \ + svuint64_t secret_vec = svld1_u64(mask, xsecret + offset); \ + svuint64_t mixed = sveor_u64_x(mask, secret_vec, input_vec); \ + svuint64_t swapped = svtbl_u64(input_vec, kSwap); \ + svuint64_t mixed_lo = svextw_u64_x(mask, mixed); \ + svuint64_t mixed_hi = svlsr_n_u64_x(mask, mixed, 32); \ + svuint64_t mul = svmad_u64_x(mask, mixed_lo, mixed_hi, swapped); \ + acc = svadd_u64_x(mask, acc, mul); \ +} while (0) +#endif /* XXH_VECTOR == XXH_SVE */ + +/* prefetch + * can be disabled, by declaring XXH_NO_PREFETCH build macro */ +#if defined(XXH_NO_PREFETCH) +# define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ +#else +# if XXH_SIZE_OPT >= 1 +# define XXH_PREFETCH(ptr) (void)(ptr) +# elif defined(_MSC_VER) && (defined(_M_X64) || defined(_M_IX86)) /* _mm_prefetch() not defined outside of x86/x64 */ +# include /* https://msdn.microsoft.com/fr-fr/library/84szxsww(v=vs.90).aspx */ +# define XXH_PREFETCH(ptr) _mm_prefetch((const char*)(ptr), _MM_HINT_T0) +# elif defined(__GNUC__) && ( (__GNUC__ >= 4) || ( (__GNUC__ == 3) && (__GNUC_MINOR__ >= 1) ) ) +# define XXH_PREFETCH(ptr) __builtin_prefetch((ptr), 0 /* rw==read */, 3 /* locality */) +# else +# define XXH_PREFETCH(ptr) (void)(ptr) /* disabled */ +# endif +#endif /* XXH_NO_PREFETCH */ + + +/* ========================================== + * XXH3 default settings + * ========================================== */ + +#define XXH_SECRET_DEFAULT_SIZE 192 /* minimum XXH3_SECRET_SIZE_MIN */ + +#if (XXH_SECRET_DEFAULT_SIZE < XXH3_SECRET_SIZE_MIN) +# error "default keyset is not large enough" +#endif + +/*! Pseudorandom secret taken directly from FARSH. */ +XXH_ALIGN(64) static const xxh_u8 XXH3_kSecret[XXH_SECRET_DEFAULT_SIZE] = { + 0xb8, 0xfe, 0x6c, 0x39, 0x23, 0xa4, 0x4b, 0xbe, 0x7c, 0x01, 0x81, 0x2c, 0xf7, 0x21, 0xad, 0x1c, + 0xde, 0xd4, 0x6d, 0xe9, 0x83, 0x90, 0x97, 0xdb, 0x72, 0x40, 0xa4, 0xa4, 0xb7, 0xb3, 0x67, 0x1f, + 0xcb, 0x79, 0xe6, 0x4e, 0xcc, 0xc0, 0xe5, 0x78, 0x82, 0x5a, 0xd0, 0x7d, 0xcc, 0xff, 0x72, 0x21, + 0xb8, 0x08, 0x46, 0x74, 0xf7, 0x43, 0x24, 0x8e, 0xe0, 0x35, 0x90, 0xe6, 0x81, 0x3a, 0x26, 0x4c, + 0x3c, 0x28, 0x52, 0xbb, 0x91, 0xc3, 0x00, 0xcb, 0x88, 0xd0, 0x65, 0x8b, 0x1b, 0x53, 0x2e, 0xa3, + 0x71, 0x64, 0x48, 0x97, 0xa2, 0x0d, 0xf9, 0x4e, 0x38, 0x19, 0xef, 0x46, 0xa9, 0xde, 0xac, 0xd8, + 0xa8, 0xfa, 0x76, 0x3f, 0xe3, 0x9c, 0x34, 0x3f, 0xf9, 0xdc, 0xbb, 0xc7, 0xc7, 0x0b, 0x4f, 0x1d, + 0x8a, 0x51, 0xe0, 0x4b, 0xcd, 0xb4, 0x59, 0x31, 0xc8, 0x9f, 0x7e, 0xc9, 0xd9, 0x78, 0x73, 0x64, + 0xea, 0xc5, 0xac, 0x83, 0x34, 0xd3, 0xeb, 0xc3, 0xc5, 0x81, 0xa0, 0xff, 0xfa, 0x13, 0x63, 0xeb, + 0x17, 0x0d, 0xdd, 0x51, 0xb7, 0xf0, 0xda, 0x49, 0xd3, 0x16, 0x55, 0x26, 0x29, 0xd4, 0x68, 0x9e, + 0x2b, 0x16, 0xbe, 0x58, 0x7d, 0x47, 0xa1, 0xfc, 0x8f, 0xf8, 0xb8, 0xd1, 0x7a, 0xd0, 0x31, 0xce, + 0x45, 0xcb, 0x3a, 0x8f, 0x95, 0x16, 0x04, 0x28, 0xaf, 0xd7, 0xfb, 0xca, 0xbb, 0x4b, 0x40, 0x7e, +}; + +static const xxh_u64 PRIME_MX1 = 0x165667919E3779F9ULL; /*!< 0b0001011001010110011001111001000110011110001101110111100111111001 */ +static const xxh_u64 PRIME_MX2 = 0x9FB21C651E98DF25ULL; /*!< 0b1001111110110010000111000110010100011110100110001101111100100101 */ + +#ifdef XXH_OLD_NAMES +# define kSecret XXH3_kSecret +#endif + +#ifdef XXH_DOXYGEN +/*! + * @brief Calculates a 32-bit to 64-bit long multiply. + * + * Implemented as a macro. + * + * Wraps `__emulu` on MSVC x86 because it tends to call `__allmul` when it doesn't + * need to (but it shouldn't need to anyways, it is about 7 instructions to do + * a 64x64 multiply...). Since we know that this will _always_ emit `MULL`, we + * use that instead of the normal method. + * + * If you are compiling for platforms like Thumb-1 and don't have a better option, + * you may also want to write your own long multiply routine here. + * + * @param x, y Numbers to be multiplied + * @return 64-bit product of the low 32 bits of @p x and @p y. + */ +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64(xxh_u64 x, xxh_u64 y) +{ + return (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF); +} +#elif defined(_MSC_VER) && defined(_M_IX86) +# define XXH_mult32to64(x, y) __emulu((unsigned)(x), (unsigned)(y)) +#else +/* + * Downcast + upcast is usually better than masking on older compilers like + * GCC 4.2 (especially 32-bit ones), all without affecting newer compilers. + * + * The other method, (x & 0xFFFFFFFF) * (y & 0xFFFFFFFF), will AND both operands + * and perform a full 64x64 multiply -- entirely redundant on 32-bit. + */ +# define XXH_mult32to64(x, y) ((xxh_u64)(xxh_u32)(x) * (xxh_u64)(xxh_u32)(y)) +#endif + +/*! + * @brief Calculates a 64->128-bit long multiply. + * + * Uses `__uint128_t` and `_umul128` if available, otherwise uses a scalar + * version. + * + * @param lhs , rhs The 64-bit integers to be multiplied + * @return The 128-bit result represented in an @ref XXH128_hash_t. + */ +static XXH128_hash_t +XXH_mult64to128(xxh_u64 lhs, xxh_u64 rhs) +{ + /* + * GCC/Clang __uint128_t method. + * + * On most 64-bit targets, GCC and Clang define a __uint128_t type. + * This is usually the best way as it usually uses a native long 64-bit + * multiply, such as MULQ on x86_64 or MUL + UMULH on aarch64. + * + * Usually. + * + * Despite being a 32-bit platform, Clang (and emscripten) define this type + * despite not having the arithmetic for it. This results in a laggy + * compiler builtin call which calculates a full 128-bit multiply. + * In that case it is best to use the portable one. + * https://github.com/Cyan4973/xxHash/issues/211#issuecomment-515575677 + */ +#if (defined(__GNUC__) || defined(__clang__)) && !defined(__wasm__) \ + && defined(__SIZEOF_INT128__) \ + || (defined(_INTEGRAL_MAX_BITS) && _INTEGRAL_MAX_BITS >= 128) + + __uint128_t const product = (__uint128_t)lhs * (__uint128_t)rhs; + XXH128_hash_t r128; + r128.low64 = (xxh_u64)(product); + r128.high64 = (xxh_u64)(product >> 64); + return r128; + + /* + * MSVC for x64's _umul128 method. + * + * xxh_u64 _umul128(xxh_u64 Multiplier, xxh_u64 Multiplicand, xxh_u64 *HighProduct); + * + * This compiles to single operand MUL on x64. + */ +#elif (defined(_M_X64) || defined(_M_IA64)) && !defined(_M_ARM64EC) + +#ifndef _MSC_VER +# pragma intrinsic(_umul128) +#endif + xxh_u64 product_high; + xxh_u64 const product_low = _umul128(lhs, rhs, &product_high); + XXH128_hash_t r128; + r128.low64 = product_low; + r128.high64 = product_high; + return r128; + + /* + * MSVC for ARM64's __umulh method. + * + * This compiles to the same MUL + UMULH as GCC/Clang's __uint128_t method. + */ +#elif defined(_M_ARM64) || defined(_M_ARM64EC) + +#ifndef _MSC_VER +# pragma intrinsic(__umulh) +#endif + XXH128_hash_t r128; + r128.low64 = lhs * rhs; + r128.high64 = __umulh(lhs, rhs); + return r128; + +#else + /* + * Portable scalar method. Optimized for 32-bit and 64-bit ALUs. + * + * This is a fast and simple grade school multiply, which is shown below + * with base 10 arithmetic instead of base 0x100000000. + * + * 9 3 // D2 lhs = 93 + * x 7 5 // D2 rhs = 75 + * ---------- + * 1 5 // D2 lo_lo = (93 % 10) * (75 % 10) = 15 + * 4 5 | // D2 hi_lo = (93 / 10) * (75 % 10) = 45 + * 2 1 | // D2 lo_hi = (93 % 10) * (75 / 10) = 21 + * + 6 3 | | // D2 hi_hi = (93 / 10) * (75 / 10) = 63 + * --------- + * 2 7 | // D2 cross = (15 / 10) + (45 % 10) + 21 = 27 + * + 6 7 | | // D2 upper = (27 / 10) + (45 / 10) + 63 = 67 + * --------- + * 6 9 7 5 // D4 res = (27 * 10) + (15 % 10) + (67 * 100) = 6975 + * + * The reasons for adding the products like this are: + * 1. It avoids manual carry tracking. Just like how + * (9 * 9) + 9 + 9 = 99, the same applies with this for UINT64_MAX. + * This avoids a lot of complexity. + * + * 2. It hints for, and on Clang, compiles to, the powerful UMAAL + * instruction available in ARM's Digital Signal Processing extension + * in 32-bit ARMv6 and later, which is shown below: + * + * void UMAAL(xxh_u32 *RdLo, xxh_u32 *RdHi, xxh_u32 Rn, xxh_u32 Rm) + * { + * xxh_u64 product = (xxh_u64)*RdLo * (xxh_u64)*RdHi + Rn + Rm; + * *RdLo = (xxh_u32)(product & 0xFFFFFFFF); + * *RdHi = (xxh_u32)(product >> 32); + * } + * + * This instruction was designed for efficient long multiplication, and + * allows this to be calculated in only 4 instructions at speeds + * comparable to some 64-bit ALUs. + * + * 3. It isn't terrible on other platforms. Usually this will be a couple + * of 32-bit ADD/ADCs. + */ + + /* First calculate all of the cross products. */ + xxh_u64 const lo_lo = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs & 0xFFFFFFFF); + xxh_u64 const hi_lo = XXH_mult32to64(lhs >> 32, rhs & 0xFFFFFFFF); + xxh_u64 const lo_hi = XXH_mult32to64(lhs & 0xFFFFFFFF, rhs >> 32); + xxh_u64 const hi_hi = XXH_mult32to64(lhs >> 32, rhs >> 32); + + /* Now add the products together. These will never overflow. */ + xxh_u64 const cross = (lo_lo >> 32) + (hi_lo & 0xFFFFFFFF) + lo_hi; + xxh_u64 const upper = (hi_lo >> 32) + (cross >> 32) + hi_hi; + xxh_u64 const lower = (cross << 32) | (lo_lo & 0xFFFFFFFF); + + XXH128_hash_t r128; + r128.low64 = lower; + r128.high64 = upper; + return r128; +#endif +} + +/*! + * @brief Calculates a 64-bit to 128-bit multiply, then XOR folds it. + * + * The reason for the separate function is to prevent passing too many structs + * around by value. This will hopefully inline the multiply, but we don't force it. + * + * @param lhs , rhs The 64-bit integers to multiply + * @return The low 64 bits of the product XOR'd by the high 64 bits. + * @see XXH_mult64to128() + */ +static xxh_u64 +XXH3_mul128_fold64(xxh_u64 lhs, xxh_u64 rhs) +{ + XXH128_hash_t product = XXH_mult64to128(lhs, rhs); + return product.low64 ^ product.high64; +} + +/*! Seems to produce slightly better code on GCC for some reason. */ +XXH_FORCE_INLINE XXH_CONSTF xxh_u64 XXH_xorshift64(xxh_u64 v64, int shift) +{ + XXH_ASSERT(0 <= shift && shift < 64); + return v64 ^ (v64 >> shift); +} + +/* + * This is a fast avalanche stage, + * suitable when input bits are already partially mixed + */ +static XXH64_hash_t XXH3_avalanche(xxh_u64 h64) +{ + h64 = XXH_xorshift64(h64, 37); + h64 *= PRIME_MX1; + h64 = XXH_xorshift64(h64, 32); + return h64; +} + +/* + * This is a stronger avalanche, + * inspired by Pelle Evensen's rrmxmx + * preferable when input has not been previously mixed + */ +static XXH64_hash_t XXH3_rrmxmx(xxh_u64 h64, xxh_u64 len) +{ + /* this mix is inspired by Pelle Evensen's rrmxmx */ + h64 ^= XXH_rotl64(h64, 49) ^ XXH_rotl64(h64, 24); + h64 *= PRIME_MX2; + h64 ^= (h64 >> 35) + len ; + h64 *= PRIME_MX2; + return XXH_xorshift64(h64, 28); +} + + +/* ========================================== + * Short keys + * ========================================== + * One of the shortcomings of XXH32 and XXH64 was that their performance was + * sub-optimal on short lengths. It used an iterative algorithm which strongly + * favored lengths that were a multiple of 4 or 8. + * + * Instead of iterating over individual inputs, we use a set of single shot + * functions which piece together a range of lengths and operate in constant time. + * + * Additionally, the number of multiplies has been significantly reduced. This + * reduces latency, especially when emulating 64-bit multiplies on 32-bit. + * + * Depending on the platform, this may or may not be faster than XXH32, but it + * is almost guaranteed to be faster than XXH64. + */ + +/* + * At very short lengths, there isn't enough input to fully hide secrets, or use + * the entire secret. + * + * There is also only a limited amount of mixing we can do before significantly + * impacting performance. + * + * Therefore, we use different sections of the secret and always mix two secret + * samples with an XOR. This should have no effect on performance on the + * seedless or withSeed variants because everything _should_ be constant folded + * by modern compilers. + * + * The XOR mixing hides individual parts of the secret and increases entropy. + * + * This adds an extra layer of strength for custom secrets. + */ +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_1to3_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(1 <= len && len <= 3); + XXH_ASSERT(secret != NULL); + /* + * len = 1: combined = { input[0], 0x01, input[0], input[0] } + * len = 2: combined = { input[1], 0x02, input[0], input[1] } + * len = 3: combined = { input[2], 0x03, input[0], input[1] } + */ + { xxh_u8 const c1 = input[0]; + xxh_u8 const c2 = input[len >> 1]; + xxh_u8 const c3 = input[len - 1]; + xxh_u32 const combined = ((xxh_u32)c1 << 16) | ((xxh_u32)c2 << 24) + | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); + xxh_u64 const bitflip = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; + xxh_u64 const keyed = (xxh_u64)combined ^ bitflip; + return XXH64_avalanche(keyed); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_4to8_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(4 <= len && len <= 8); + seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; + { xxh_u32 const input1 = XXH_readLE32(input); + xxh_u32 const input2 = XXH_readLE32(input + len - 4); + xxh_u64 const bitflip = (XXH_readLE64(secret+8) ^ XXH_readLE64(secret+16)) - seed; + xxh_u64 const input64 = input2 + (((xxh_u64)input1) << 32); + xxh_u64 const keyed = input64 ^ bitflip; + return XXH3_rrmxmx(keyed, len); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_9to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(9 <= len && len <= 16); + { xxh_u64 const bitflip1 = (XXH_readLE64(secret+24) ^ XXH_readLE64(secret+32)) + seed; + xxh_u64 const bitflip2 = (XXH_readLE64(secret+40) ^ XXH_readLE64(secret+48)) - seed; + xxh_u64 const input_lo = XXH_readLE64(input) ^ bitflip1; + xxh_u64 const input_hi = XXH_readLE64(input + len - 8) ^ bitflip2; + xxh_u64 const acc = len + + XXH_swap64(input_lo) + input_hi + + XXH3_mul128_fold64(input_lo, input_hi); + return XXH3_avalanche(acc); + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_0to16_64b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(len <= 16); + { if (XXH_likely(len > 8)) return XXH3_len_9to16_64b(input, len, secret, seed); + if (XXH_likely(len >= 4)) return XXH3_len_4to8_64b(input, len, secret, seed); + if (len) return XXH3_len_1to3_64b(input, len, secret, seed); + return XXH64_avalanche(seed ^ (XXH_readLE64(secret+56) ^ XXH_readLE64(secret+64))); + } +} + +/* + * DISCLAIMER: There are known *seed-dependent* multicollisions here due to + * multiplication by zero, affecting hashes of lengths 17 to 240. + * + * However, they are very unlikely. + * + * Keep this in mind when using the unseeded XXH3_64bits() variant: As with all + * unseeded non-cryptographic hashes, it does not attempt to defend itself + * against specially crafted inputs, only random inputs. + * + * Compared to classic UMAC where a 1 in 2^31 chance of 4 consecutive bytes + * cancelling out the secret is taken an arbitrary number of times (addressed + * in XXH3_accumulate_512), this collision is very unlikely with random inputs + * and/or proper seeding: + * + * This only has a 1 in 2^63 chance of 8 consecutive bytes cancelling out, in a + * function that is only called up to 16 times per hash with up to 240 bytes of + * input. + * + * This is not too bad for a non-cryptographic hash function, especially with + * only 64 bit outputs. + * + * The 128-bit variant (which trades some speed for strength) is NOT affected + * by this, although it is always a good idea to use a proper seed if you care + * about strength. + */ +XXH_FORCE_INLINE xxh_u64 XXH3_mix16B(const xxh_u8* XXH_RESTRICT input, + const xxh_u8* XXH_RESTRICT secret, xxh_u64 seed64) +{ +#if defined(__GNUC__) && !defined(__clang__) /* GCC, not Clang */ \ + && defined(__i386__) && defined(__SSE2__) /* x86 + SSE2 */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable like XXH32 hack */ + /* + * UGLY HACK: + * GCC for x86 tends to autovectorize the 128-bit multiply, resulting in + * slower code. + * + * By forcing seed64 into a register, we disrupt the cost model and + * cause it to scalarize. See `XXH32_round()` + * + * FIXME: Clang's output is still _much_ faster -- On an AMD Ryzen 3600, + * XXH3_64bits @ len=240 runs at 4.6 GB/s with Clang 9, but 3.3 GB/s on + * GCC 9.2, despite both emitting scalar code. + * + * GCC generates much better scalar code than Clang for the rest of XXH3, + * which is why finding a more optimal codepath is an interest. + */ + XXH_COMPILER_GUARD(seed64); +#endif + { xxh_u64 const input_lo = XXH_readLE64(input); + xxh_u64 const input_hi = XXH_readLE64(input+8); + return XXH3_mul128_fold64( + input_lo ^ (XXH_readLE64(secret) + seed64), + input_hi ^ (XXH_readLE64(secret+8) - seed64) + ); + } +} + +/* For mid range keys, XXH3 uses a Mum-hash variant. */ +XXH_FORCE_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_17to128_64b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(16 < len && len <= 128); + + { xxh_u64 acc = len * XXH_PRIME64_1; +#if XXH_SIZE_OPT >= 1 + /* Smaller and cleaner, but slightly slower. */ + unsigned int i = (unsigned int)(len - 1) / 32; + do { + acc += XXH3_mix16B(input+16 * i, secret+32*i, seed); + acc += XXH3_mix16B(input+len-16*(i+1), secret+32*i+16, seed); + } while (i-- != 0); +#else + if (len > 32) { + if (len > 64) { + if (len > 96) { + acc += XXH3_mix16B(input+48, secret+96, seed); + acc += XXH3_mix16B(input+len-64, secret+112, seed); + } + acc += XXH3_mix16B(input+32, secret+64, seed); + acc += XXH3_mix16B(input+len-48, secret+80, seed); + } + acc += XXH3_mix16B(input+16, secret+32, seed); + acc += XXH3_mix16B(input+len-32, secret+48, seed); + } + acc += XXH3_mix16B(input+0, secret+0, seed); + acc += XXH3_mix16B(input+len-16, secret+16, seed); +#endif + return XXH3_avalanche(acc); + } +} + +XXH_NO_INLINE XXH_PUREF XXH64_hash_t +XXH3_len_129to240_64b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + + #define XXH3_MIDSIZE_STARTOFFSET 3 + #define XXH3_MIDSIZE_LASTOFFSET 17 + + { xxh_u64 acc = len * XXH_PRIME64_1; + xxh_u64 acc_end; + unsigned int const nbRounds = (unsigned int)len / 16; + unsigned int i; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + for (i=0; i<8; i++) { + acc += XXH3_mix16B(input+(16*i), secret+(16*i), seed); + } + /* last bytes */ + acc_end = XXH3_mix16B(input + len - 16, secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET, seed); + XXH_ASSERT(nbRounds >= 8); + acc = XXH3_avalanche(acc); +#if defined(__clang__) /* Clang */ \ + && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ + /* + * UGLY HACK: + * Clang for ARMv7-A tries to vectorize this loop, similar to GCC x86. + * In everywhere else, it uses scalar code. + * + * For 64->128-bit multiplies, even if the NEON was 100% optimal, it + * would still be slower than UMAAL (see XXH_mult64to128). + * + * Unfortunately, Clang doesn't handle the long multiplies properly and + * converts them to the nonexistent "vmulq_u64" intrinsic, which is then + * scalarized into an ugly mess of VMOV.32 instructions. + * + * This mess is difficult to avoid without turning autovectorization + * off completely, but they are usually relatively minor and/or not + * worth it to fix. + * + * This loop is the easiest to fix, as unlike XXH32, this pragma + * _actually works_ because it is a loop vectorization instead of an + * SLP vectorization. + */ + #pragma clang loop vectorize(disable) +#endif + for (i=8 ; i < nbRounds; i++) { + /* + * Prevents clang for unrolling the acc loop and interleaving with this one. + */ + XXH_COMPILER_GUARD(acc); + acc_end += XXH3_mix16B(input+(16*i), secret+(16*(i-8)) + XXH3_MIDSIZE_STARTOFFSET, seed); + } + return XXH3_avalanche(acc + acc_end); + } +} + + +/* ======= Long Keys ======= */ + +#define XXH_STRIPE_LEN 64 +#define XXH_SECRET_CONSUME_RATE 8 /* nb of secret bytes consumed at each accumulation */ +#define XXH_ACC_NB (XXH_STRIPE_LEN / sizeof(xxh_u64)) + +#ifdef XXH_OLD_NAMES +# define STRIPE_LEN XXH_STRIPE_LEN +# define ACC_NB XXH_ACC_NB +#endif + +#ifndef XXH_PREFETCH_DIST +# ifdef __clang__ +# define XXH_PREFETCH_DIST 320 +# else +# if (XXH_VECTOR == XXH_AVX512) +# define XXH_PREFETCH_DIST 512 +# else +# define XXH_PREFETCH_DIST 384 +# endif +# endif /* __clang__ */ +#endif /* XXH_PREFETCH_DIST */ + +/* + * These macros are to generate an XXH3_accumulate() function. + * The two arguments select the name suffix and target attribute. + * + * The name of this symbol is XXH3_accumulate_() and it calls + * XXH3_accumulate_512_(). + * + * It may be useful to hand implement this function if the compiler fails to + * optimize the inline function. + */ +#define XXH3_ACCUMULATE_TEMPLATE(name) \ +void \ +XXH3_accumulate_##name(xxh_u64* XXH_RESTRICT acc, \ + const xxh_u8* XXH_RESTRICT input, \ + const xxh_u8* XXH_RESTRICT secret, \ + size_t nbStripes) \ +{ \ + size_t n; \ + for (n = 0; n < nbStripes; n++ ) { \ + const xxh_u8* const in = input + n*XXH_STRIPE_LEN; \ + XXH_PREFETCH(in + XXH_PREFETCH_DIST); \ + XXH3_accumulate_512_##name( \ + acc, \ + in, \ + secret + n*XXH_SECRET_CONSUME_RATE); \ + } \ +} + + +XXH_FORCE_INLINE void XXH_writeLE64(void* dst, xxh_u64 v64) +{ + if (!XXH_CPU_LITTLE_ENDIAN) v64 = XXH_swap64(v64); + XXH_memcpy(dst, &v64, sizeof(v64)); +} + +/* Several intrinsic functions below are supposed to accept __int64 as argument, + * as documented in https://software.intel.com/sites/landingpage/IntrinsicsGuide/ . + * However, several environments do not define __int64 type, + * requiring a workaround. + */ +#if !defined (__VMS) \ + && (defined (__cplusplus) \ + || (defined (__STDC_VERSION__) && (__STDC_VERSION__ >= 199901L) /* C99 */) ) + typedef int64_t xxh_i64; +#else + /* the following type must have a width of 64-bit */ + typedef long long xxh_i64; +#endif + + +/* + * XXH3_accumulate_512 is the tightest loop for long inputs, and it is the most optimized. + * + * It is a hardened version of UMAC, based off of FARSH's implementation. + * + * This was chosen because it adapts quite well to 32-bit, 64-bit, and SIMD + * implementations, and it is ridiculously fast. + * + * We harden it by mixing the original input to the accumulators as well as the product. + * + * This means that in the (relatively likely) case of a multiply by zero, the + * original input is preserved. + * + * On 128-bit inputs, we swap 64-bit pairs when we add the input to improve + * cross-pollination, as otherwise the upper and lower halves would be + * essentially independent. + * + * This doesn't matter on 64-bit hashes since they all get merged together in + * the end, so we skip the extra step. + * + * Both XXH3_64bits and XXH3_128bits use this subroutine. + */ + +#if (XXH_VECTOR == XXH_AVX512) \ + || (defined(XXH_DISPATCH_AVX512) && XXH_DISPATCH_AVX512 != 0) + +#ifndef XXH_TARGET_AVX512 +# define XXH_TARGET_AVX512 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_accumulate_512_avx512(void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + __m512i* const xacc = (__m512i *) acc; + XXH_ASSERT((((size_t)acc) & 63) == 0); + XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); + + { + /* data_vec = input[0]; */ + __m512i const data_vec = _mm512_loadu_si512 (input); + /* key_vec = secret[0]; */ + __m512i const key_vec = _mm512_loadu_si512 (secret); + /* data_key = data_vec ^ key_vec; */ + __m512i const data_key = _mm512_xor_si512 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m512i const data_key_lo = _mm512_srli_epi64 (data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m512i const product = _mm512_mul_epu32 (data_key, data_key_lo); + /* xacc[0] += swap(data_vec); */ + __m512i const data_swap = _mm512_shuffle_epi32(data_vec, (_MM_PERM_ENUM)_MM_SHUFFLE(1, 0, 3, 2)); + __m512i const sum = _mm512_add_epi64(*xacc, data_swap); + /* xacc[0] += product; */ + *xacc = _mm512_add_epi64(product, sum); + } +} +XXH_FORCE_INLINE XXH_TARGET_AVX512 XXH3_ACCUMULATE_TEMPLATE(avx512) + +/* + * XXH3_scrambleAcc: Scrambles the accumulators to improve mixing. + * + * Multiplication isn't perfect, as explained by Google in HighwayHash: + * + * // Multiplication mixes/scrambles bytes 0-7 of the 64-bit result to + * // varying degrees. In descending order of goodness, bytes + * // 3 4 2 5 1 6 0 7 have quality 228 224 164 160 100 96 36 32. + * // As expected, the upper and lower bytes are much worse. + * + * Source: https://github.com/google/highwayhash/blob/0aaf66b/highwayhash/hh_avx2.h#L291 + * + * Since our algorithm uses a pseudorandom secret to add some variance into the + * mix, we don't need to (or want to) mix as often or as much as HighwayHash does. + * + * This isn't as tight as XXH3_accumulate, but still written in SIMD to avoid + * extraction. + * + * Both XXH3_64bits and XXH3_128bits use this subroutine. + */ + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_scrambleAcc_avx512(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 63) == 0); + XXH_STATIC_ASSERT(XXH_STRIPE_LEN == sizeof(__m512i)); + { __m512i* const xacc = (__m512i*) acc; + const __m512i prime32 = _mm512_set1_epi32((int)XXH_PRIME32_1); + + /* xacc[0] ^= (xacc[0] >> 47) */ + __m512i const acc_vec = *xacc; + __m512i const shifted = _mm512_srli_epi64 (acc_vec, 47); + /* xacc[0] ^= secret; */ + __m512i const key_vec = _mm512_loadu_si512 (secret); + __m512i const data_key = _mm512_ternarylogic_epi32(key_vec, acc_vec, shifted, 0x96 /* key_vec ^ acc_vec ^ shifted */); + + /* xacc[0] *= XXH_PRIME32_1; */ + __m512i const data_key_hi = _mm512_srli_epi64 (data_key, 32); + __m512i const prod_lo = _mm512_mul_epu32 (data_key, prime32); + __m512i const prod_hi = _mm512_mul_epu32 (data_key_hi, prime32); + *xacc = _mm512_add_epi64(prod_lo, _mm512_slli_epi64(prod_hi, 32)); + } +} + +XXH_FORCE_INLINE XXH_TARGET_AVX512 void +XXH3_initCustomSecret_avx512(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 63) == 0); + XXH_STATIC_ASSERT(XXH_SEC_ALIGN == 64); + XXH_ASSERT(((size_t)customSecret & 63) == 0); + (void)(&XXH_writeLE64); + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m512i); + __m512i const seed_pos = _mm512_set1_epi64((xxh_i64)seed64); + __m512i const seed = _mm512_mask_sub_epi64(seed_pos, 0xAA, _mm512_set1_epi8(0), seed_pos); + + const __m512i* const src = (const __m512i*) ((const void*) XXH3_kSecret); + __m512i* const dest = ( __m512i*) customSecret; + int i; + XXH_ASSERT(((size_t)src & 63) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dest & 63) == 0); + for (i=0; i < nbRounds; ++i) { + dest[i] = _mm512_add_epi64(_mm512_load_si512(src + i), seed); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_AVX2) \ + || (defined(XXH_DISPATCH_AVX2) && XXH_DISPATCH_AVX2 != 0) + +#ifndef XXH_TARGET_AVX2 +# define XXH_TARGET_AVX2 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void +XXH3_accumulate_512_avx2( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 31) == 0); + { __m256i* const xacc = (__m256i *) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xinput = (const __m256i *) input; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xsecret = (const __m256i *) secret; + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { + /* data_vec = xinput[i]; */ + __m256i const data_vec = _mm256_loadu_si256 (xinput+i); + /* key_vec = xsecret[i]; */ + __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); + /* data_key = data_vec ^ key_vec; */ + __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m256i const data_key_lo = _mm256_srli_epi64 (data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m256i const product = _mm256_mul_epu32 (data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m256i const data_swap = _mm256_shuffle_epi32(data_vec, _MM_SHUFFLE(1, 0, 3, 2)); + __m256i const sum = _mm256_add_epi64(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = _mm256_add_epi64(product, sum); + } } +} +XXH_FORCE_INLINE XXH_TARGET_AVX2 XXH3_ACCUMULATE_TEMPLATE(avx2) + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void +XXH3_scrambleAcc_avx2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 31) == 0); + { __m256i* const xacc = (__m256i*) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm256_loadu_si256 requires a const __m256i * pointer for some reason. */ + const __m256i* const xsecret = (const __m256i *) secret; + const __m256i prime32 = _mm256_set1_epi32((int)XXH_PRIME32_1); + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m256i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m256i const acc_vec = xacc[i]; + __m256i const shifted = _mm256_srli_epi64 (acc_vec, 47); + __m256i const data_vec = _mm256_xor_si256 (acc_vec, shifted); + /* xacc[i] ^= xsecret; */ + __m256i const key_vec = _mm256_loadu_si256 (xsecret+i); + __m256i const data_key = _mm256_xor_si256 (data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m256i const data_key_hi = _mm256_srli_epi64 (data_key, 32); + __m256i const prod_lo = _mm256_mul_epu32 (data_key, prime32); + __m256i const prod_hi = _mm256_mul_epu32 (data_key_hi, prime32); + xacc[i] = _mm256_add_epi64(prod_lo, _mm256_slli_epi64(prod_hi, 32)); + } + } +} + +XXH_FORCE_INLINE XXH_TARGET_AVX2 void XXH3_initCustomSecret_avx2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 31) == 0); + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE / sizeof(__m256i)) == 6); + XXH_STATIC_ASSERT(XXH_SEC_ALIGN <= 64); + (void)(&XXH_writeLE64); + XXH_PREFETCH(customSecret); + { __m256i const seed = _mm256_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64, (xxh_i64)(0U - seed64), (xxh_i64)seed64); + + const __m256i* const src = (const __m256i*) ((const void*) XXH3_kSecret); + __m256i* dest = ( __m256i*) customSecret; + +# if defined(__GNUC__) || defined(__clang__) + /* + * On GCC & Clang, marking 'dest' as modified will cause the compiler: + * - do not extract the secret from sse registers in the internal loop + * - use less common registers, and avoid pushing these reg into stack + */ + XXH_COMPILER_GUARD(dest); +# endif + XXH_ASSERT(((size_t)src & 31) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dest & 31) == 0); + + /* GCC -O2 need unroll loop manually */ + dest[0] = _mm256_add_epi64(_mm256_load_si256(src+0), seed); + dest[1] = _mm256_add_epi64(_mm256_load_si256(src+1), seed); + dest[2] = _mm256_add_epi64(_mm256_load_si256(src+2), seed); + dest[3] = _mm256_add_epi64(_mm256_load_si256(src+3), seed); + dest[4] = _mm256_add_epi64(_mm256_load_si256(src+4), seed); + dest[5] = _mm256_add_epi64(_mm256_load_si256(src+5), seed); + } +} + +#endif + +/* x86dispatch always generates SSE2 */ +#if (XXH_VECTOR == XXH_SSE2) || defined(XXH_X86DISPATCH) + +#ifndef XXH_TARGET_SSE2 +# define XXH_TARGET_SSE2 /* disable attribute target */ +#endif + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void +XXH3_accumulate_512_sse2( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + /* SSE2 is just a half-scale version of the AVX2 version. */ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { __m128i* const xacc = (__m128i *) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xinput = (const __m128i *) input; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xsecret = (const __m128i *) secret; + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { + /* data_vec = xinput[i]; */ + __m128i const data_vec = _mm_loadu_si128 (xinput+i); + /* key_vec = xsecret[i]; */ + __m128i const key_vec = _mm_loadu_si128 (xsecret+i); + /* data_key = data_vec ^ key_vec; */ + __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m128i const data_key_lo = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m128i const product = _mm_mul_epu32 (data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m128i const data_swap = _mm_shuffle_epi32(data_vec, _MM_SHUFFLE(1,0,3,2)); + __m128i const sum = _mm_add_epi64(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = _mm_add_epi64(product, sum); + } } +} +XXH_FORCE_INLINE XXH_TARGET_SSE2 XXH3_ACCUMULATE_TEMPLATE(sse2) + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void +XXH3_scrambleAcc_sse2(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { __m128i* const xacc = (__m128i*) acc; + /* Unaligned. This is mainly for pointer arithmetic, and because + * _mm_loadu_si128 requires a const __m128i * pointer for some reason. */ + const __m128i* const xsecret = (const __m128i *) secret; + const __m128i prime32 = _mm_set1_epi32((int)XXH_PRIME32_1); + + size_t i; + for (i=0; i < XXH_STRIPE_LEN/sizeof(__m128i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m128i const acc_vec = xacc[i]; + __m128i const shifted = _mm_srli_epi64 (acc_vec, 47); + __m128i const data_vec = _mm_xor_si128 (acc_vec, shifted); + /* xacc[i] ^= xsecret[i]; */ + __m128i const key_vec = _mm_loadu_si128 (xsecret+i); + __m128i const data_key = _mm_xor_si128 (data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m128i const data_key_hi = _mm_shuffle_epi32 (data_key, _MM_SHUFFLE(0, 3, 0, 1)); + __m128i const prod_lo = _mm_mul_epu32 (data_key, prime32); + __m128i const prod_hi = _mm_mul_epu32 (data_key_hi, prime32); + xacc[i] = _mm_add_epi64(prod_lo, _mm_slli_epi64(prod_hi, 32)); + } + } +} + +XXH_FORCE_INLINE XXH_TARGET_SSE2 void XXH3_initCustomSecret_sse2(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); + (void)(&XXH_writeLE64); + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / sizeof(__m128i); + +# if defined(_MSC_VER) && defined(_M_IX86) && _MSC_VER < 1900 + /* MSVC 32bit mode does not support _mm_set_epi64x before 2015 */ + XXH_ALIGN(16) const xxh_i64 seed64x2[2] = { (xxh_i64)seed64, (xxh_i64)(0U - seed64) }; + __m128i const seed = _mm_load_si128((__m128i const*)seed64x2); +# else + __m128i const seed = _mm_set_epi64x((xxh_i64)(0U - seed64), (xxh_i64)seed64); +# endif + int i; + + const void* const src16 = XXH3_kSecret; + __m128i* dst16 = (__m128i*) customSecret; +# if defined(__GNUC__) || defined(__clang__) + /* + * On GCC & Clang, marking 'dest' as modified will cause the compiler: + * - do not extract the secret from sse registers in the internal loop + * - use less common registers, and avoid pushing these reg into stack + */ + XXH_COMPILER_GUARD(dst16); +# endif + XXH_ASSERT(((size_t)src16 & 15) == 0); /* control alignment */ + XXH_ASSERT(((size_t)dst16 & 15) == 0); + + for (i=0; i < nbRounds; ++i) { + dst16[i] = _mm_add_epi64(_mm_load_si128((const __m128i *)src16+i), seed); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_NEON) + +/* forward declarations for the scalar routines */ +XXH_FORCE_INLINE void +XXH3_scalarRound(void* XXH_RESTRICT acc, void const* XXH_RESTRICT input, + void const* XXH_RESTRICT secret, size_t lane); + +XXH_FORCE_INLINE void +XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT secret, size_t lane); + +/*! + * @internal + * @brief The bulk processing loop for NEON and WASM SIMD128. + * + * The NEON code path is actually partially scalar when running on AArch64. This + * is to optimize the pipelining and can have up to 15% speedup depending on the + * CPU, and it also mitigates some GCC codegen issues. + * + * @see XXH3_NEON_LANES for configuring this and details about this optimization. + * + * NEON's 32-bit to 64-bit long multiply takes a half vector of 32-bit + * integers instead of the other platforms which mask full 64-bit vectors, + * so the setup is more complicated than just shifting right. + * + * Additionally, there is an optimization for 4 lanes at once noted below. + * + * Since, as stated, the most optimal amount of lanes for Cortexes is 6, + * there needs to be *three* versions of the accumulate operation used + * for the remaining 2 lanes. + * + * WASM's SIMD128 uses SIMDe's arm_neon.h polyfill because the intrinsics overlap + * nearly perfectly. + */ + +XXH_FORCE_INLINE void +XXH3_accumulate_512_neon( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + XXH_STATIC_ASSERT(XXH3_NEON_LANES > 0 && XXH3_NEON_LANES <= XXH_ACC_NB && XXH3_NEON_LANES % 2 == 0); + { /* GCC for darwin arm64 does not like aliasing here */ + xxh_aliasing_uint64x2_t* const xacc = (xxh_aliasing_uint64x2_t*) acc; + /* We don't use a uint32x4_t pointer because it causes bus errors on ARMv7. */ + uint8_t const* xinput = (const uint8_t *) input; + uint8_t const* xsecret = (const uint8_t *) secret; + + size_t i; +#ifdef __wasm_simd128__ + /* + * On WASM SIMD128, Clang emits direct address loads when XXH3_kSecret + * is constant propagated, which results in it converting it to this + * inside the loop: + * + * a = v128.load(XXH3_kSecret + 0 + $secret_offset, offset = 0) + * b = v128.load(XXH3_kSecret + 16 + $secret_offset, offset = 0) + * ... + * + * This requires a full 32-bit address immediate (and therefore a 6 byte + * instruction) as well as an add for each offset. + * + * Putting an asm guard prevents it from folding (at the cost of losing + * the alignment hint), and uses the free offset in `v128.load` instead + * of adding secret_offset each time which overall reduces code size by + * about a kilobyte and improves performance. + */ + XXH_COMPILER_GUARD(xsecret); +#endif + /* Scalar lanes use the normal scalarRound routine */ + for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { + XXH3_scalarRound(acc, input, secret, i); + } + i = 0; + /* 4 NEON lanes at a time. */ + for (; i+1 < XXH3_NEON_LANES / 2; i+=2) { + /* data_vec = xinput[i]; */ + uint64x2_t data_vec_1 = XXH_vld1q_u64(xinput + (i * 16)); + uint64x2_t data_vec_2 = XXH_vld1q_u64(xinput + ((i+1) * 16)); + /* key_vec = xsecret[i]; */ + uint64x2_t key_vec_1 = XXH_vld1q_u64(xsecret + (i * 16)); + uint64x2_t key_vec_2 = XXH_vld1q_u64(xsecret + ((i+1) * 16)); + /* data_swap = swap(data_vec) */ + uint64x2_t data_swap_1 = vextq_u64(data_vec_1, data_vec_1, 1); + uint64x2_t data_swap_2 = vextq_u64(data_vec_2, data_vec_2, 1); + /* data_key = data_vec ^ key_vec; */ + uint64x2_t data_key_1 = veorq_u64(data_vec_1, key_vec_1); + uint64x2_t data_key_2 = veorq_u64(data_vec_2, key_vec_2); + + /* + * If we reinterpret the 64x2 vectors as 32x4 vectors, we can use a + * de-interleave operation for 4 lanes in 1 step with `vuzpq_u32` to + * get one vector with the low 32 bits of each lane, and one vector + * with the high 32 bits of each lane. + * + * The intrinsic returns a double vector because the original ARMv7-a + * instruction modified both arguments in place. AArch64 and SIMD128 emit + * two instructions from this intrinsic. + * + * [ dk11L | dk11H | dk12L | dk12H ] -> [ dk11L | dk12L | dk21L | dk22L ] + * [ dk21L | dk21H | dk22L | dk22H ] -> [ dk11H | dk12H | dk21H | dk22H ] + */ + uint32x4x2_t unzipped = vuzpq_u32( + vreinterpretq_u32_u64(data_key_1), + vreinterpretq_u32_u64(data_key_2) + ); + /* data_key_lo = data_key & 0xFFFFFFFF */ + uint32x4_t data_key_lo = unzipped.val[0]; + /* data_key_hi = data_key >> 32 */ + uint32x4_t data_key_hi = unzipped.val[1]; + /* + * Then, we can split the vectors horizontally and multiply which, as for most + * widening intrinsics, have a variant that works on both high half vectors + * for free on AArch64. A similar instruction is available on SIMD128. + * + * sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi + */ + uint64x2_t sum_1 = XXH_vmlal_low_u32(data_swap_1, data_key_lo, data_key_hi); + uint64x2_t sum_2 = XXH_vmlal_high_u32(data_swap_2, data_key_lo, data_key_hi); + /* + * Clang reorders + * a += b * c; // umlal swap.2d, dkl.2s, dkh.2s + * c += a; // add acc.2d, acc.2d, swap.2d + * to + * c += a; // add acc.2d, acc.2d, swap.2d + * c += b * c; // umlal acc.2d, dkl.2s, dkh.2s + * + * While it would make sense in theory since the addition is faster, + * for reasons likely related to umlal being limited to certain NEON + * pipelines, this is worse. A compiler guard fixes this. + */ + XXH_COMPILER_GUARD_CLANG_NEON(sum_1); + XXH_COMPILER_GUARD_CLANG_NEON(sum_2); + /* xacc[i] = acc_vec + sum; */ + xacc[i] = vaddq_u64(xacc[i], sum_1); + xacc[i+1] = vaddq_u64(xacc[i+1], sum_2); + } + /* Operate on the remaining NEON lanes 2 at a time. */ + for (; i < XXH3_NEON_LANES / 2; i++) { + /* data_vec = xinput[i]; */ + uint64x2_t data_vec = XXH_vld1q_u64(xinput + (i * 16)); + /* key_vec = xsecret[i]; */ + uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); + /* acc_vec_2 = swap(data_vec) */ + uint64x2_t data_swap = vextq_u64(data_vec, data_vec, 1); + /* data_key = data_vec ^ key_vec; */ + uint64x2_t data_key = veorq_u64(data_vec, key_vec); + /* For two lanes, just use VMOVN and VSHRN. */ + /* data_key_lo = data_key & 0xFFFFFFFF; */ + uint32x2_t data_key_lo = vmovn_u64(data_key); + /* data_key_hi = data_key >> 32; */ + uint32x2_t data_key_hi = vshrn_n_u64(data_key, 32); + /* sum = data_swap + (u64x2) data_key_lo * (u64x2) data_key_hi; */ + uint64x2_t sum = vmlal_u32(data_swap, data_key_lo, data_key_hi); + /* Same Clang workaround as before */ + XXH_COMPILER_GUARD_CLANG_NEON(sum); + /* xacc[i] = acc_vec + sum; */ + xacc[i] = vaddq_u64 (xacc[i], sum); + } + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(neon) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_neon(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + + { xxh_aliasing_uint64x2_t* xacc = (xxh_aliasing_uint64x2_t*) acc; + uint8_t const* xsecret = (uint8_t const*) secret; + + size_t i; + /* WASM uses operator overloads and doesn't need these. */ +#ifndef __wasm_simd128__ + /* { prime32_1, prime32_1 } */ + uint32x2_t const kPrimeLo = vdup_n_u32(XXH_PRIME32_1); + /* { 0, prime32_1, 0, prime32_1 } */ + uint32x4_t const kPrimeHi = vreinterpretq_u32_u64(vdupq_n_u64((xxh_u64)XXH_PRIME32_1 << 32)); +#endif + + /* AArch64 uses both scalar and neon at the same time */ + for (i = XXH3_NEON_LANES; i < XXH_ACC_NB; i++) { + XXH3_scalarScrambleRound(acc, secret, i); + } + for (i=0; i < XXH3_NEON_LANES / 2; i++) { + /* xacc[i] ^= (xacc[i] >> 47); */ + uint64x2_t acc_vec = xacc[i]; + uint64x2_t shifted = vshrq_n_u64(acc_vec, 47); + uint64x2_t data_vec = veorq_u64(acc_vec, shifted); + + /* xacc[i] ^= xsecret[i]; */ + uint64x2_t key_vec = XXH_vld1q_u64(xsecret + (i * 16)); + uint64x2_t data_key = veorq_u64(data_vec, key_vec); + /* xacc[i] *= XXH_PRIME32_1 */ +#ifdef __wasm_simd128__ + /* SIMD128 has multiply by u64x2, use it instead of expanding and scalarizing */ + xacc[i] = data_key * XXH_PRIME32_1; +#else + /* + * Expanded version with portable NEON intrinsics + * + * lo(x) * lo(y) + (hi(x) * lo(y) << 32) + * + * prod_hi = hi(data_key) * lo(prime) << 32 + * + * Since we only need 32 bits of this multiply a trick can be used, reinterpreting the vector + * as a uint32x4_t and multiplying by { 0, prime, 0, prime } to cancel out the unwanted bits + * and avoid the shift. + */ + uint32x4_t prod_hi = vmulq_u32 (vreinterpretq_u32_u64(data_key), kPrimeHi); + /* Extract low bits for vmlal_u32 */ + uint32x2_t data_key_lo = vmovn_u64(data_key); + /* xacc[i] = prod_hi + lo(data_key) * XXH_PRIME32_1; */ + xacc[i] = vmlal_u32(vreinterpretq_u64_u32(prod_hi), data_key_lo, kPrimeLo); +#endif + } + } +} +#endif + +#if (XXH_VECTOR == XXH_VSX) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_vsx( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + /* presumed aligned */ + xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; + xxh_u8 const* const xinput = (xxh_u8 const*) input; /* no alignment restriction */ + xxh_u8 const* const xsecret = (xxh_u8 const*) secret; /* no alignment restriction */ + xxh_u64x2 const v32 = { 32, 32 }; + size_t i; + for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { + /* data_vec = xinput[i]; */ + xxh_u64x2 const data_vec = XXH_vec_loadu(xinput + 16*i); + /* key_vec = xsecret[i]; */ + xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); + xxh_u64x2 const data_key = data_vec ^ key_vec; + /* shuffled = (data_key << 32) | (data_key >> 32); */ + xxh_u32x4 const shuffled = (xxh_u32x4)vec_rl(data_key, v32); + /* product = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)shuffled & 0xFFFFFFFF); */ + xxh_u64x2 const product = XXH_vec_mulo((xxh_u32x4)data_key, shuffled); + /* acc_vec = xacc[i]; */ + xxh_u64x2 acc_vec = xacc[i]; + acc_vec += product; + + /* swap high and low halves */ +#ifdef __s390x__ + acc_vec += vec_permi(data_vec, data_vec, 2); +#else + acc_vec += vec_xxpermdi(data_vec, data_vec, 2); +#endif + xacc[i] = acc_vec; + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(vsx) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_vsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + + { xxh_aliasing_u64x2* const xacc = (xxh_aliasing_u64x2*) acc; + const xxh_u8* const xsecret = (const xxh_u8*) secret; + /* constants */ + xxh_u64x2 const v32 = { 32, 32 }; + xxh_u64x2 const v47 = { 47, 47 }; + xxh_u32x4 const prime = { XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1, XXH_PRIME32_1 }; + size_t i; + for (i = 0; i < XXH_STRIPE_LEN / sizeof(xxh_u64x2); i++) { + /* xacc[i] ^= (xacc[i] >> 47); */ + xxh_u64x2 const acc_vec = xacc[i]; + xxh_u64x2 const data_vec = acc_vec ^ (acc_vec >> v47); + + /* xacc[i] ^= xsecret[i]; */ + xxh_u64x2 const key_vec = XXH_vec_loadu(xsecret + 16*i); + xxh_u64x2 const data_key = data_vec ^ key_vec; + + /* xacc[i] *= XXH_PRIME32_1 */ + /* prod_lo = ((xxh_u64x2)data_key & 0xFFFFFFFF) * ((xxh_u64x2)prime & 0xFFFFFFFF); */ + xxh_u64x2 const prod_even = XXH_vec_mule((xxh_u32x4)data_key, prime); + /* prod_hi = ((xxh_u64x2)data_key >> 32) * ((xxh_u64x2)prime >> 32); */ + xxh_u64x2 const prod_odd = XXH_vec_mulo((xxh_u32x4)data_key, prime); + xacc[i] = prod_odd + (prod_even << v32); + } } +} + +#endif + +#if (XXH_VECTOR == XXH_SVE) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_sve( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + uint64_t *xacc = (uint64_t *)acc; + const uint64_t *xinput = (const uint64_t *)(const void *)input; + const uint64_t *xsecret = (const uint64_t *)(const void *)secret; + svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); + uint64_t element_count = svcntd(); + if (element_count >= 8) { + svbool_t mask = svptrue_pat_b64(SV_VL8); + svuint64_t vacc = svld1_u64(mask, xacc); + ACCRND(vacc, 0); + svst1_u64(mask, xacc, vacc); + } else if (element_count == 2) { /* sve128 */ + svbool_t mask = svptrue_pat_b64(SV_VL2); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 2); + svuint64_t acc2 = svld1_u64(mask, xacc + 4); + svuint64_t acc3 = svld1_u64(mask, xacc + 6); + ACCRND(acc0, 0); + ACCRND(acc1, 2); + ACCRND(acc2, 4); + ACCRND(acc3, 6); + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 2, acc1); + svst1_u64(mask, xacc + 4, acc2); + svst1_u64(mask, xacc + 6, acc3); + } else { + svbool_t mask = svptrue_pat_b64(SV_VL4); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 4); + ACCRND(acc0, 0); + ACCRND(acc1, 4); + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 4, acc1); + } +} + +XXH_FORCE_INLINE void +XXH3_accumulate_sve(xxh_u64* XXH_RESTRICT acc, + const xxh_u8* XXH_RESTRICT input, + const xxh_u8* XXH_RESTRICT secret, + size_t nbStripes) +{ + if (nbStripes != 0) { + uint64_t *xacc = (uint64_t *)acc; + const uint64_t *xinput = (const uint64_t *)(const void *)input; + const uint64_t *xsecret = (const uint64_t *)(const void *)secret; + svuint64_t kSwap = sveor_n_u64_z(svptrue_b64(), svindex_u64(0, 1), 1); + uint64_t element_count = svcntd(); + if (element_count >= 8) { + svbool_t mask = svptrue_pat_b64(SV_VL8); + svuint64_t vacc = svld1_u64(mask, xacc + 0); + do { + /* svprfd(svbool_t, void *, enum svfprop); */ + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(vacc, 0); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, vacc); + } else if (element_count == 2) { /* sve128 */ + svbool_t mask = svptrue_pat_b64(SV_VL2); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 2); + svuint64_t acc2 = svld1_u64(mask, xacc + 4); + svuint64_t acc3 = svld1_u64(mask, xacc + 6); + do { + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(acc0, 0); + ACCRND(acc1, 2); + ACCRND(acc2, 4); + ACCRND(acc3, 6); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 2, acc1); + svst1_u64(mask, xacc + 4, acc2); + svst1_u64(mask, xacc + 6, acc3); + } else { + svbool_t mask = svptrue_pat_b64(SV_VL4); + svuint64_t acc0 = svld1_u64(mask, xacc + 0); + svuint64_t acc1 = svld1_u64(mask, xacc + 4); + do { + svprfd(mask, xinput + 128, SV_PLDL1STRM); + ACCRND(acc0, 0); + ACCRND(acc1, 4); + xinput += 8; + xsecret += 1; + nbStripes--; + } while (nbStripes != 0); + + svst1_u64(mask, xacc + 0, acc0); + svst1_u64(mask, xacc + 4, acc1); + } + } +} + +#endif + +#if (XXH_VECTOR == XXH_LSX) +#define _LSX_SHUFFLE(z, y, x, w) (((z) << 6) | ((y) << 4) | ((x) << 2) | (w)) + +XXH_FORCE_INLINE void +XXH3_accumulate_512_lsx( void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { + __m128i* const xacc = (__m128i *) acc; + const __m128i* const xinput = (const __m128i *) input; + const __m128i* const xsecret = (const __m128i *) secret; + + for (size_t i = 0; i < XXH_STRIPE_LEN / sizeof(__m128i); i++) { + /* data_vec = xinput[i]; */ + __m128i const data_vec = __lsx_vld(xinput + i, 0); + /* key_vec = xsecret[i]; */ + __m128i const key_vec = __lsx_vld(xsecret + i, 0); + /* data_key = data_vec ^ key_vec; */ + __m128i const data_key = __lsx_vxor_v(data_vec, key_vec); + /* data_key_lo = data_key >> 32; */ + __m128i const data_key_lo = __lsx_vsrli_d(data_key, 32); + // __m128i const data_key_lo = __lsx_vsrli_d(data_key, 32); + /* product = (data_key & 0xffffffff) * (data_key_lo & 0xffffffff); */ + __m128i const product = __lsx_vmulwev_d_wu(data_key, data_key_lo); + /* xacc[i] += swap(data_vec); */ + __m128i const data_swap = __lsx_vshuf4i_w(data_vec, _LSX_SHUFFLE(1, 0, 3, 2)); + __m128i const sum = __lsx_vadd_d(xacc[i], data_swap); + /* xacc[i] += product; */ + xacc[i] = __lsx_vadd_d(product, sum); + } + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(lsx) + +XXH_FORCE_INLINE void +XXH3_scrambleAcc_lsx(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + XXH_ASSERT((((size_t)acc) & 15) == 0); + { + __m128i* const xacc = (__m128i*) acc; + const __m128i* const xsecret = (const __m128i *) secret; + const __m128i prime32 = __lsx_vreplgr2vr_w((int)XXH_PRIME32_1); + + for (size_t i = 0; i < XXH_STRIPE_LEN / sizeof(__m128i); i++) { + /* xacc[i] ^= (xacc[i] >> 47) */ + __m128i const acc_vec = xacc[i]; + __m128i const shifted = __lsx_vsrli_d(acc_vec, 47); + __m128i const data_vec = __lsx_vxor_v(acc_vec, shifted); + /* xacc[i] ^= xsecret[i]; */ + __m128i const key_vec = __lsx_vld(xsecret + i, 0); + __m128i const data_key = __lsx_vxor_v(data_vec, key_vec); + + /* xacc[i] *= XXH_PRIME32_1; */ + __m128i const data_key_hi = __lsx_vsrli_d(data_key, 32); + __m128i const prod_lo = __lsx_vmulwev_d_wu(data_key, prime32); + __m128i const prod_hi = __lsx_vmulwev_d_wu(data_key_hi, prime32); + xacc[i] = __lsx_vadd_d(prod_lo, __lsx_vslli_d(prod_hi, 32)); + } + } +} + +#endif + +/* scalar variants - universal */ + +#if defined(__aarch64__) && (defined(__GNUC__) || defined(__clang__)) +/* + * In XXH3_scalarRound(), GCC and Clang have a similar codegen issue, where they + * emit an excess mask and a full 64-bit multiply-add (MADD X-form). + * + * While this might not seem like much, as AArch64 is a 64-bit architecture, only + * big Cortex designs have a full 64-bit multiplier. + * + * On the little cores, the smaller 32-bit multiplier is used, and full 64-bit + * multiplies expand to 2-3 multiplies in microcode. This has a major penalty + * of up to 4 latency cycles and 2 stall cycles in the multiply pipeline. + * + * Thankfully, AArch64 still provides the 32-bit long multiply-add (UMADDL) which does + * not have this penalty and does the mask automatically. + */ +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) +{ + xxh_u64 ret; + /* note: %x = 64-bit register, %w = 32-bit register */ + __asm__("umaddl %x0, %w1, %w2, %x3" : "=r" (ret) : "r" (lhs), "r" (rhs), "r" (acc)); + return ret; +} +#else +XXH_FORCE_INLINE xxh_u64 +XXH_mult32to64_add64(xxh_u64 lhs, xxh_u64 rhs, xxh_u64 acc) +{ + return XXH_mult32to64((xxh_u32)lhs, (xxh_u32)rhs) + acc; +} +#endif + +/*! + * @internal + * @brief Scalar round for @ref XXH3_accumulate_512_scalar(). + * + * This is extracted to its own function because the NEON path uses a combination + * of NEON and scalar. + */ +XXH_FORCE_INLINE void +XXH3_scalarRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT input, + void const* XXH_RESTRICT secret, + size_t lane) +{ + xxh_u64* xacc = (xxh_u64*) acc; + xxh_u8 const* xinput = (xxh_u8 const*) input; + xxh_u8 const* xsecret = (xxh_u8 const*) secret; + XXH_ASSERT(lane < XXH_ACC_NB); + XXH_ASSERT(((size_t)acc & (XXH_ACC_ALIGN-1)) == 0); + { + xxh_u64 const data_val = XXH_readLE64(xinput + lane * 8); + xxh_u64 const data_key = data_val ^ XXH_readLE64(xsecret + lane * 8); + xacc[lane ^ 1] += data_val; /* swap adjacent lanes */ + xacc[lane] = XXH_mult32to64_add64(data_key /* & 0xFFFFFFFF */, data_key >> 32, xacc[lane]); + } +} + +/*! + * @internal + * @brief Processes a 64 byte block of data using the scalar path. + */ +XXH_FORCE_INLINE void +XXH3_accumulate_512_scalar(void* XXH_RESTRICT acc, + const void* XXH_RESTRICT input, + const void* XXH_RESTRICT secret) +{ + size_t i; + /* ARM GCC refuses to unroll this loop, resulting in a 24% slowdown on ARMv6. */ +#if defined(__GNUC__) && !defined(__clang__) \ + && (defined(__arm__) || defined(__thumb2__)) \ + && defined(__ARM_FEATURE_UNALIGNED) /* no unaligned access just wastes bytes */ \ + && XXH_SIZE_OPT <= 0 +# pragma GCC unroll 8 +#endif + for (i=0; i < XXH_ACC_NB; i++) { + XXH3_scalarRound(acc, input, secret, i); + } +} +XXH_FORCE_INLINE XXH3_ACCUMULATE_TEMPLATE(scalar) + +/*! + * @internal + * @brief Scalar scramble step for @ref XXH3_scrambleAcc_scalar(). + * + * This is extracted to its own function because the NEON path uses a combination + * of NEON and scalar. + */ +XXH_FORCE_INLINE void +XXH3_scalarScrambleRound(void* XXH_RESTRICT acc, + void const* XXH_RESTRICT secret, + size_t lane) +{ + xxh_u64* const xacc = (xxh_u64*) acc; /* presumed aligned */ + const xxh_u8* const xsecret = (const xxh_u8*) secret; /* no alignment restriction */ + XXH_ASSERT((((size_t)acc) & (XXH_ACC_ALIGN-1)) == 0); + XXH_ASSERT(lane < XXH_ACC_NB); + { + xxh_u64 const key64 = XXH_readLE64(xsecret + lane * 8); + xxh_u64 acc64 = xacc[lane]; + acc64 = XXH_xorshift64(acc64, 47); + acc64 ^= key64; + acc64 *= XXH_PRIME32_1; + xacc[lane] = acc64; + } +} + +/*! + * @internal + * @brief Scrambles the accumulators after a large chunk has been read + */ +XXH_FORCE_INLINE void +XXH3_scrambleAcc_scalar(void* XXH_RESTRICT acc, const void* XXH_RESTRICT secret) +{ + size_t i; + for (i=0; i < XXH_ACC_NB; i++) { + XXH3_scalarScrambleRound(acc, secret, i); + } +} + +XXH_FORCE_INLINE void +XXH3_initCustomSecret_scalar(void* XXH_RESTRICT customSecret, xxh_u64 seed64) +{ + /* + * We need a separate pointer for the hack below, + * which requires a non-const pointer. + * Any decent compiler will optimize this out otherwise. + */ + const xxh_u8* kSecretPtr = XXH3_kSecret; + XXH_STATIC_ASSERT((XXH_SECRET_DEFAULT_SIZE & 15) == 0); + +#if defined(__GNUC__) && defined(__aarch64__) + /* + * UGLY HACK: + * GCC and Clang generate a bunch of MOV/MOVK pairs for aarch64, and they are + * placed sequentially, in order, at the top of the unrolled loop. + * + * While MOVK is great for generating constants (2 cycles for a 64-bit + * constant compared to 4 cycles for LDR), it fights for bandwidth with + * the arithmetic instructions. + * + * I L S + * MOVK + * MOVK + * MOVK + * MOVK + * ADD + * SUB STR + * STR + * By forcing loads from memory (as the asm line causes the compiler to assume + * that XXH3_kSecretPtr has been changed), the pipelines are used more + * efficiently: + * I L S + * LDR + * ADD LDR + * SUB STR + * STR + * + * See XXH3_NEON_LANES for details on the pipsline. + * + * XXH3_64bits_withSeed, len == 256, Snapdragon 835 + * without hack: 2654.4 MB/s + * with hack: 3202.9 MB/s + */ + XXH_COMPILER_GUARD(kSecretPtr); +#endif + { int const nbRounds = XXH_SECRET_DEFAULT_SIZE / 16; + int i; + for (i=0; i < nbRounds; i++) { + /* + * The asm hack causes the compiler to assume that kSecretPtr aliases with + * customSecret, and on aarch64, this prevented LDP from merging two + * loads together for free. Putting the loads together before the stores + * properly generates LDP. + */ + xxh_u64 lo = XXH_readLE64(kSecretPtr + 16*i) + seed64; + xxh_u64 hi = XXH_readLE64(kSecretPtr + 16*i + 8) - seed64; + XXH_writeLE64((xxh_u8*)customSecret + 16*i, lo); + XXH_writeLE64((xxh_u8*)customSecret + 16*i + 8, hi); + } } +} + + +typedef void (*XXH3_f_accumulate)(xxh_u64* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, const xxh_u8* XXH_RESTRICT, size_t); +typedef void (*XXH3_f_scrambleAcc)(void* XXH_RESTRICT, const void*); +typedef void (*XXH3_f_initCustomSecret)(void* XXH_RESTRICT, xxh_u64); + + +#if (XXH_VECTOR == XXH_AVX512) + +#define XXH3_accumulate_512 XXH3_accumulate_512_avx512 +#define XXH3_accumulate XXH3_accumulate_avx512 +#define XXH3_scrambleAcc XXH3_scrambleAcc_avx512 +#define XXH3_initCustomSecret XXH3_initCustomSecret_avx512 + +#elif (XXH_VECTOR == XXH_AVX2) + +#define XXH3_accumulate_512 XXH3_accumulate_512_avx2 +#define XXH3_accumulate XXH3_accumulate_avx2 +#define XXH3_scrambleAcc XXH3_scrambleAcc_avx2 +#define XXH3_initCustomSecret XXH3_initCustomSecret_avx2 + +#elif (XXH_VECTOR == XXH_SSE2) + +#define XXH3_accumulate_512 XXH3_accumulate_512_sse2 +#define XXH3_accumulate XXH3_accumulate_sse2 +#define XXH3_scrambleAcc XXH3_scrambleAcc_sse2 +#define XXH3_initCustomSecret XXH3_initCustomSecret_sse2 + +#elif (XXH_VECTOR == XXH_NEON) + +#define XXH3_accumulate_512 XXH3_accumulate_512_neon +#define XXH3_accumulate XXH3_accumulate_neon +#define XXH3_scrambleAcc XXH3_scrambleAcc_neon +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_VSX) + +#define XXH3_accumulate_512 XXH3_accumulate_512_vsx +#define XXH3_accumulate XXH3_accumulate_vsx +#define XXH3_scrambleAcc XXH3_scrambleAcc_vsx +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_SVE) +#define XXH3_accumulate_512 XXH3_accumulate_512_sve +#define XXH3_accumulate XXH3_accumulate_sve +#define XXH3_scrambleAcc XXH3_scrambleAcc_scalar +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#elif (XXH_VECTOR == XXH_LSX) +#define XXH3_accumulate_512 XXH3_accumulate_512_lsx +#define XXH3_accumulate XXH3_accumulate_lsx +#define XXH3_scrambleAcc XXH3_scrambleAcc_lsx +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#else /* scalar */ + +#define XXH3_accumulate_512 XXH3_accumulate_512_scalar +#define XXH3_accumulate XXH3_accumulate_scalar +#define XXH3_scrambleAcc XXH3_scrambleAcc_scalar +#define XXH3_initCustomSecret XXH3_initCustomSecret_scalar + +#endif + +#if XXH_SIZE_OPT >= 1 /* don't do SIMD for initialization */ +# undef XXH3_initCustomSecret +# define XXH3_initCustomSecret XXH3_initCustomSecret_scalar +#endif + +XXH_FORCE_INLINE void +XXH3_hashLong_internal_loop(xxh_u64* XXH_RESTRICT acc, + const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + size_t const nbStripesPerBlock = (secretSize - XXH_STRIPE_LEN) / XXH_SECRET_CONSUME_RATE; + size_t const block_len = XXH_STRIPE_LEN * nbStripesPerBlock; + size_t const nb_blocks = (len - 1) / block_len; + + size_t n; + + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); + + for (n = 0; n < nb_blocks; n++) { + f_acc(acc, input + n*block_len, secret, nbStripesPerBlock); + f_scramble(acc, secret + secretSize - XXH_STRIPE_LEN); + } + + /* last partial block */ + XXH_ASSERT(len > XXH_STRIPE_LEN); + { size_t const nbStripes = ((len - 1) - (block_len * nb_blocks)) / XXH_STRIPE_LEN; + XXH_ASSERT(nbStripes <= (secretSize / XXH_SECRET_CONSUME_RATE)); + f_acc(acc, input + nb_blocks*block_len, secret, nbStripes); + + /* last stripe */ + { const xxh_u8* const p = input + len - XXH_STRIPE_LEN; +#define XXH_SECRET_LASTACC_START 7 /* not aligned on 8, last secret is different from acc & scrambler */ + XXH3_accumulate_512(acc, p, secret + secretSize - XXH_STRIPE_LEN - XXH_SECRET_LASTACC_START); + } } +} + +XXH_FORCE_INLINE xxh_u64 +XXH3_mix2Accs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret) +{ + return XXH3_mul128_fold64( + acc[0] ^ XXH_readLE64(secret), + acc[1] ^ XXH_readLE64(secret+8) ); +} + +static XXH_PUREF XXH64_hash_t +XXH3_mergeAccs(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 start) +{ + xxh_u64 result64 = start; + size_t i = 0; + + for (i = 0; i < 4; i++) { + result64 += XXH3_mix2Accs(acc+2*i, secret + 16*i); +#if defined(__clang__) /* Clang */ \ + && (defined(__arm__) || defined(__thumb__)) /* ARMv7 */ \ + && (defined(__ARM_NEON) || defined(__ARM_NEON__)) /* NEON */ \ + && !defined(XXH_ENABLE_AUTOVECTORIZE) /* Define to disable */ + /* + * UGLY HACK: + * Prevent autovectorization on Clang ARMv7-a. Exact same problem as + * the one in XXH3_len_129to240_64b. Speeds up shorter keys > 240b. + * XXH3_64bits, len == 256, Snapdragon 835: + * without hack: 2063.7 MB/s + * with hack: 2560.7 MB/s + */ + XXH_COMPILER_GUARD(result64); +#endif + } + + return XXH3_avalanche(result64); +} + +/* do not align on 8, so that the secret is different from the accumulator */ +#define XXH_SECRET_MERGEACCS_START 11 + +static XXH_PUREF XXH64_hash_t +XXH3_finalizeLong_64b(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, xxh_u64 len) +{ + return XXH3_mergeAccs(acc, secret + XXH_SECRET_MERGEACCS_START, len * XXH_PRIME64_1); +} + +#define XXH3_INIT_ACC { XXH_PRIME32_3, XXH_PRIME64_1, XXH_PRIME64_2, XXH_PRIME64_3, \ + XXH_PRIME64_4, XXH_PRIME32_2, XXH_PRIME64_5, XXH_PRIME32_1 } + +XXH_FORCE_INLINE XXH64_hash_t +XXH3_hashLong_64b_internal(const void* XXH_RESTRICT input, size_t len, + const void* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; + + XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, (const xxh_u8*)secret, secretSize, f_acc, f_scramble); + + /* converge into final hash */ + XXH_STATIC_ASSERT(sizeof(acc) == 64); + XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_finalizeLong_64b(acc, (const xxh_u8*)secret, (xxh_u64)len); +} + +/* + * It's important for performance to transmit secret's size (when it's static) + * so that the compiler can properly optimize the vectorized loop. + * This makes a big performance difference for "medium" keys (<1 KB) when using AVX instruction set. + * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE + * breaks -Og, this is XXH_NO_INLINE. + */ +XXH3_WITH_SECRET_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSecret(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; + return XXH3_hashLong_64b_internal(input, len, secret, secretLen, XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * It's preferable for performance that XXH3_hashLong is not inlined, + * as it results in a smaller function for small data, easier to the instruction cache. + * Note that inside this no_inline function, we do inline the internal loop, + * and provide a statically defined secret size to allow optimization of vector loop. + */ +XXH_NO_INLINE XXH_PUREF XXH64_hash_t +XXH3_hashLong_64b_default(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; (void)secret; (void)secretLen; + return XXH3_hashLong_64b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * XXH3_hashLong_64b_withSeed(): + * Generate a custom key based on alteration of default XXH3_kSecret with the seed, + * and then use this key for long mode hashing. + * + * This operation is decently fast but nonetheless costs a little bit of time. + * Try to avoid it whenever possible (typically when seed==0). + * + * It's important for performance that XXH3_hashLong is not inlined. Not sure + * why (uop cache maybe?), but the difference is large and easily measurable. + */ +XXH_FORCE_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSeed_internal(const void* input, size_t len, + XXH64_hash_t seed, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble, + XXH3_f_initCustomSecret f_initSec) +{ +#if XXH_SIZE_OPT <= 0 + if (seed == 0) + return XXH3_hashLong_64b_internal(input, len, + XXH3_kSecret, sizeof(XXH3_kSecret), + f_acc, f_scramble); +#endif + { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; + f_initSec(secret, seed); + return XXH3_hashLong_64b_internal(input, len, secret, sizeof(secret), + f_acc, f_scramble); + } +} + +/* + * It's important for performance that XXH3_hashLong is not inlined. + */ +XXH_NO_INLINE XXH64_hash_t +XXH3_hashLong_64b_withSeed(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed, const xxh_u8* XXH_RESTRICT secret, size_t secretLen) +{ + (void)secret; (void)secretLen; + return XXH3_hashLong_64b_withSeed_internal(input, len, seed, + XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); +} + + +typedef XXH64_hash_t (*XXH3_hashLong64_f)(const void* XXH_RESTRICT, size_t, + XXH64_hash_t, const xxh_u8* XXH_RESTRICT, size_t); + +XXH_FORCE_INLINE XXH64_hash_t +XXH3_64bits_internal(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, + XXH3_hashLong64_f f_hashLong) +{ + XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); + /* + * If an action is to be taken if `secretLen` condition is not respected, + * it should be done here. + * For now, it's a contract pre-condition. + * Adding a check and a branch here would cost performance at every hash. + * Also, note that function signature doesn't offer room to return an error. + */ + if (len <= 16) + return XXH3_len_0to16_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); + if (len <= 128) + return XXH3_len_17to128_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_len_129to240_64b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + return f_hashLong(input, len, seed64, (const xxh_u8*)secret, secretLen); +} + + +/* === Public entry point === */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t XXH3_64bits(XXH_NOESCAPE const void* input, size_t length) +{ + return XXH3_64bits_internal(input, length, 0, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_default); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSecret(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_64bits_internal(input, length, 0, secret, secretSize, XXH3_hashLong_64b_withSecret); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSeed(XXH_NOESCAPE const void* input, size_t length, XXH64_hash_t seed) +{ + return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), XXH3_hashLong_64b_withSeed); +} + +XXH_PUBLIC_API XXH64_hash_t +XXH3_64bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t length, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + if (length <= XXH3_MIDSIZE_MAX) + return XXH3_64bits_internal(input, length, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); + return XXH3_hashLong_64b_withSecret(input, length, seed, (const xxh_u8*)secret, secretSize); +} + + +/* === XXH3 streaming === */ +#ifndef XXH_NO_STREAM +/* + * Malloc's a pointer that is always aligned to @align. + * + * This must be freed with `XXH_alignedFree()`. + * + * malloc typically guarantees 16 byte alignment on 64-bit systems and 8 byte + * alignment on 32-bit. This isn't enough for the 32 byte aligned loads in AVX2 + * or on 32-bit, the 16 byte aligned loads in SSE2 and NEON. + * + * This underalignment previously caused a rather obvious crash which went + * completely unnoticed due to XXH3_createState() not actually being tested. + * Credit to RedSpah for noticing this bug. + * + * The alignment is done manually: Functions like posix_memalign or _mm_malloc + * are avoided: To maintain portability, we would have to write a fallback + * like this anyways, and besides, testing for the existence of library + * functions without relying on external build tools is impossible. + * + * The method is simple: Overallocate, manually align, and store the offset + * to the original behind the returned pointer. + * + * Align must be a power of 2 and 8 <= align <= 128. + */ +static XXH_MALLOCF void* XXH_alignedMalloc(size_t s, size_t align) +{ + XXH_ASSERT(align <= 128 && align >= 8); /* range check */ + XXH_ASSERT((align & (align-1)) == 0); /* power of 2 */ + XXH_ASSERT(s != 0 && s < (s + align)); /* empty/overflow */ + { /* Overallocate to make room for manual realignment and an offset byte */ + xxh_u8* base = (xxh_u8*)XXH_malloc(s + align); + if (base != NULL) { + /* + * Get the offset needed to align this pointer. + * + * Even if the returned pointer is aligned, there will always be + * at least one byte to store the offset to the original pointer. + */ + size_t offset = align - ((size_t)base & (align - 1)); /* base % align */ + /* Add the offset for the now-aligned pointer */ + xxh_u8* ptr = base + offset; + + XXH_ASSERT((size_t)ptr % align == 0); + + /* Store the offset immediately before the returned pointer. */ + ptr[-1] = (xxh_u8)offset; + return ptr; + } + return NULL; + } +} +/* + * Frees an aligned pointer allocated by XXH_alignedMalloc(). Don't pass + * normal malloc'd pointers, XXH_alignedMalloc has a specific data layout. + */ +static void XXH_alignedFree(void* p) +{ + if (p != NULL) { + xxh_u8* ptr = (xxh_u8*)p; + /* Get the offset byte we added in XXH_malloc. */ + xxh_u8 offset = ptr[-1]; + /* Free the original malloc'd pointer */ + xxh_u8* base = ptr - offset; + XXH_free(base); + } +} +/*! @ingroup XXH3_family */ +/*! + * @brief Allocate an @ref XXH3_state_t. + * + * @return An allocated pointer of @ref XXH3_state_t on success. + * @return `NULL` on failure. + * + * @note Must be freed with XXH3_freeState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH3_state_t* XXH3_createState(void) +{ + XXH3_state_t* const state = (XXH3_state_t*)XXH_alignedMalloc(sizeof(XXH3_state_t), 64); + if (state==NULL) return NULL; + XXH3_INITSTATE(state); + return state; +} + +/*! @ingroup XXH3_family */ +/*! + * @brief Frees an @ref XXH3_state_t. + * + * @param statePtr A pointer to an @ref XXH3_state_t allocated with @ref XXH3_createState(). + * + * @return @ref XXH_OK. + * + * @note Must be allocated with XXH3_createState(). + * + * @see @ref streaming_example "Streaming Example" + */ +XXH_PUBLIC_API XXH_errorcode XXH3_freeState(XXH3_state_t* statePtr) +{ + XXH_alignedFree(statePtr); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API void +XXH3_copyState(XXH_NOESCAPE XXH3_state_t* dst_state, XXH_NOESCAPE const XXH3_state_t* src_state) +{ + XXH_memcpy(dst_state, src_state, sizeof(*dst_state)); +} + +static void +XXH3_reset_internal(XXH3_state_t* statePtr, + XXH64_hash_t seed, + const void* secret, size_t secretSize) +{ + size_t const initStart = offsetof(XXH3_state_t, bufferedSize); + size_t const initLength = offsetof(XXH3_state_t, nbStripesPerBlock) - initStart; + XXH_ASSERT(offsetof(XXH3_state_t, nbStripesPerBlock) > initStart); + XXH_ASSERT(statePtr != NULL); + /* set members from bufferedSize to nbStripesPerBlock (excluded) to 0 */ + memset((char*)statePtr + initStart, 0, initLength); + statePtr->acc[0] = XXH_PRIME32_3; + statePtr->acc[1] = XXH_PRIME64_1; + statePtr->acc[2] = XXH_PRIME64_2; + statePtr->acc[3] = XXH_PRIME64_3; + statePtr->acc[4] = XXH_PRIME64_4; + statePtr->acc[5] = XXH_PRIME32_2; + statePtr->acc[6] = XXH_PRIME64_5; + statePtr->acc[7] = XXH_PRIME32_1; + statePtr->seed = seed; + statePtr->useSeed = (seed != 0); + statePtr->extSecret = (const unsigned char*)secret; + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); + statePtr->secretLimit = secretSize - XXH_STRIPE_LEN; + statePtr->nbStripesPerBlock = statePtr->secretLimit / XXH_SECRET_CONSUME_RATE; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) +{ + if (statePtr == NULL) return XXH_ERROR; + XXH3_reset_internal(statePtr, 0, XXH3_kSecret, XXH_SECRET_DEFAULT_SIZE); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + if (statePtr == NULL) return XXH_ERROR; + XXH3_reset_internal(statePtr, 0, secret, secretSize); + if (secret == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) +{ + if (statePtr == NULL) return XXH_ERROR; + if (seed==0) return XXH3_64bits_reset(statePtr); + if ((seed != statePtr->seed) || (statePtr->extSecret != NULL)) + XXH3_initCustomSecret(statePtr->customSecret, seed); + XXH3_reset_internal(statePtr, seed, NULL, XXH_SECRET_DEFAULT_SIZE); + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed64) +{ + if (statePtr == NULL) return XXH_ERROR; + if (secret == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; + XXH3_reset_internal(statePtr, seed64, secret, secretSize); + statePtr->useSeed = 1; /* always, even if seed64==0 */ + return XXH_OK; +} + +/*! + * @internal + * @brief Processes a large input for XXH3_update() and XXH3_digest_long(). + * + * Unlike XXH3_hashLong_internal_loop(), this can process data that overlaps a block. + * + * @param acc Pointer to the 8 accumulator lanes + * @param nbStripesSoFarPtr In/out pointer to the number of leftover stripes in the block* + * @param nbStripesPerBlock Number of stripes in a block + * @param input Input pointer + * @param nbStripes Number of stripes to process + * @param secret Secret pointer + * @param secretLimit Offset of the last block in @p secret + * @param f_acc Pointer to an XXH3_accumulate implementation + * @param f_scramble Pointer to an XXH3_scrambleAcc implementation + * @return Pointer past the end of @p input after processing + */ +XXH_FORCE_INLINE const xxh_u8 * +XXH3_consumeStripes(xxh_u64* XXH_RESTRICT acc, + size_t* XXH_RESTRICT nbStripesSoFarPtr, size_t nbStripesPerBlock, + const xxh_u8* XXH_RESTRICT input, size_t nbStripes, + const xxh_u8* XXH_RESTRICT secret, size_t secretLimit, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + const xxh_u8* initialSecret = secret + *nbStripesSoFarPtr * XXH_SECRET_CONSUME_RATE; + /* Process full blocks */ + if (nbStripes >= (nbStripesPerBlock - *nbStripesSoFarPtr)) { + /* Process the initial partial block... */ + size_t nbStripesThisIter = nbStripesPerBlock - *nbStripesSoFarPtr; + + do { + /* Accumulate and scramble */ + f_acc(acc, input, initialSecret, nbStripesThisIter); + f_scramble(acc, secret + secretLimit); + input += nbStripesThisIter * XXH_STRIPE_LEN; + nbStripes -= nbStripesThisIter; + /* Then continue the loop with the full block size */ + nbStripesThisIter = nbStripesPerBlock; + initialSecret = secret; + } while (nbStripes >= nbStripesPerBlock); + *nbStripesSoFarPtr = 0; + } + /* Process a partial block */ + if (nbStripes > 0) { + f_acc(acc, input, initialSecret, nbStripes); + input += nbStripes * XXH_STRIPE_LEN; + *nbStripesSoFarPtr += nbStripes; + } + /* Return end pointer */ + return input; +} + +#ifndef XXH3_STREAM_USE_STACK +# if XXH_SIZE_OPT <= 0 && !defined(__clang__) /* clang doesn't need additional stack space */ +# define XXH3_STREAM_USE_STACK 1 +# endif +#endif +/* + * Both XXH3_64bits_update and XXH3_128bits_update use this routine. + */ +XXH_FORCE_INLINE XXH_errorcode +XXH3_update(XXH3_state_t* XXH_RESTRICT const state, + const xxh_u8* XXH_RESTRICT input, size_t len, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + if (input==NULL) { + XXH_ASSERT(len == 0); + return XXH_OK; + } + + XXH_ASSERT(state != NULL); + { const xxh_u8* const bEnd = input + len; + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; +#if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 + /* For some reason, gcc and MSVC seem to suffer greatly + * when operating accumulators directly into state. + * Operating into stack space seems to enable proper optimization. + * clang, on the other hand, doesn't seem to need this trick */ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[8]; + XXH_memcpy(acc, state->acc, sizeof(acc)); +#else + xxh_u64* XXH_RESTRICT const acc = state->acc; +#endif + state->totalLen += len; + XXH_ASSERT(state->bufferedSize <= XXH3_INTERNALBUFFER_SIZE); + + /* small input : just fill in tmp buffer */ + if (len <= XXH3_INTERNALBUFFER_SIZE - state->bufferedSize) { + XXH_memcpy(state->buffer + state->bufferedSize, input, len); + state->bufferedSize += (XXH32_hash_t)len; + return XXH_OK; + } + + /* total input is now > XXH3_INTERNALBUFFER_SIZE */ + #define XXH3_INTERNALBUFFER_STRIPES (XXH3_INTERNALBUFFER_SIZE / XXH_STRIPE_LEN) + XXH_STATIC_ASSERT(XXH3_INTERNALBUFFER_SIZE % XXH_STRIPE_LEN == 0); /* clean multiple */ + + /* + * Internal buffer is partially filled (always, except at beginning) + * Complete it, then consume it. + */ + if (state->bufferedSize) { + size_t const loadSize = XXH3_INTERNALBUFFER_SIZE - state->bufferedSize; + XXH_memcpy(state->buffer + state->bufferedSize, input, loadSize); + input += loadSize; + XXH3_consumeStripes(acc, + &state->nbStripesSoFar, state->nbStripesPerBlock, + state->buffer, XXH3_INTERNALBUFFER_STRIPES, + secret, state->secretLimit, + f_acc, f_scramble); + state->bufferedSize = 0; + } + XXH_ASSERT(input < bEnd); + if (bEnd - input > XXH3_INTERNALBUFFER_SIZE) { + size_t nbStripes = (size_t)(bEnd - 1 - input) / XXH_STRIPE_LEN; + input = XXH3_consumeStripes(acc, + &state->nbStripesSoFar, state->nbStripesPerBlock, + input, nbStripes, + secret, state->secretLimit, + f_acc, f_scramble); + XXH_memcpy(state->buffer + sizeof(state->buffer) - XXH_STRIPE_LEN, input - XXH_STRIPE_LEN, XXH_STRIPE_LEN); + + } + /* Some remaining input (always) : buffer it */ + XXH_ASSERT(input < bEnd); + XXH_ASSERT(bEnd - input <= XXH3_INTERNALBUFFER_SIZE); + XXH_ASSERT(state->bufferedSize == 0); + XXH_memcpy(state->buffer, input, (size_t)(bEnd-input)); + state->bufferedSize = (XXH32_hash_t)(bEnd-input); +#if defined(XXH3_STREAM_USE_STACK) && XXH3_STREAM_USE_STACK >= 1 + /* save stack accumulators into state */ + XXH_memcpy(state->acc, acc, sizeof(acc)); +#endif + } + + return XXH_OK; +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_64bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_update(state, (const xxh_u8*)input, len, + XXH3_accumulate, XXH3_scrambleAcc); +} + + +XXH_FORCE_INLINE void +XXH3_digest_long (XXH64_hash_t* acc, + const XXH3_state_t* state, + const unsigned char* secret) +{ + xxh_u8 lastStripe[XXH_STRIPE_LEN]; + const xxh_u8* lastStripePtr; + + /* + * Digest on a local copy. This way, the state remains unaltered, and it can + * continue ingesting more input afterwards. + */ + XXH_memcpy(acc, state->acc, sizeof(state->acc)); + if (state->bufferedSize >= XXH_STRIPE_LEN) { + /* Consume remaining stripes then point to remaining data in buffer */ + size_t const nbStripes = (state->bufferedSize - 1) / XXH_STRIPE_LEN; + size_t nbStripesSoFar = state->nbStripesSoFar; + XXH3_consumeStripes(acc, + &nbStripesSoFar, state->nbStripesPerBlock, + state->buffer, nbStripes, + secret, state->secretLimit, + XXH3_accumulate, XXH3_scrambleAcc); + lastStripePtr = state->buffer + state->bufferedSize - XXH_STRIPE_LEN; + } else { /* bufferedSize < XXH_STRIPE_LEN */ + /* Copy to temp buffer */ + size_t const catchupSize = XXH_STRIPE_LEN - state->bufferedSize; + XXH_ASSERT(state->bufferedSize > 0); /* there is always some input buffered */ + XXH_memcpy(lastStripe, state->buffer + sizeof(state->buffer) - catchupSize, catchupSize); + XXH_memcpy(lastStripe + catchupSize, state->buffer, state->bufferedSize); + lastStripePtr = lastStripe; + } + /* Last stripe */ + XXH3_accumulate_512(acc, + lastStripePtr, + secret + state->secretLimit - XXH_SECRET_LASTACC_START); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH64_hash_t XXH3_64bits_digest (XXH_NOESCAPE const XXH3_state_t* state) +{ + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; + if (state->totalLen > XXH3_MIDSIZE_MAX) { + XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; + XXH3_digest_long(acc, state, secret); + return XXH3_finalizeLong_64b(acc, secret, (xxh_u64)state->totalLen); + } + /* totalLen <= XXH3_MIDSIZE_MAX: digesting a short input */ + if (state->useSeed) + return XXH3_64bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); + return XXH3_64bits_withSecret(state->buffer, (size_t)(state->totalLen), + secret, state->secretLimit + XXH_STRIPE_LEN); +} +#endif /* !XXH_NO_STREAM */ + + +/* ========================================== + * XXH3 128 bits (a.k.a XXH128) + * ========================================== + * XXH3's 128-bit variant has better mixing and strength than the 64-bit variant, + * even without counting the significantly larger output size. + * + * For example, extra steps are taken to avoid the seed-dependent collisions + * in 17-240 byte inputs (See XXH3_mix16B and XXH128_mix32B). + * + * This strength naturally comes at the cost of some speed, especially on short + * lengths. Note that longer hashes are about as fast as the 64-bit version + * due to it using only a slight modification of the 64-bit loop. + * + * XXH128 is also more oriented towards 64-bit machines. It is still extremely + * fast for a _128-bit_ hash on 32-bit (it usually clears XXH64). + */ + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_1to3_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + /* A doubled version of 1to3_64b with different constants. */ + XXH_ASSERT(input != NULL); + XXH_ASSERT(1 <= len && len <= 3); + XXH_ASSERT(secret != NULL); + /* + * len = 1: combinedl = { input[0], 0x01, input[0], input[0] } + * len = 2: combinedl = { input[1], 0x02, input[0], input[1] } + * len = 3: combinedl = { input[2], 0x03, input[0], input[1] } + */ + { xxh_u8 const c1 = input[0]; + xxh_u8 const c2 = input[len >> 1]; + xxh_u8 const c3 = input[len - 1]; + xxh_u32 const combinedl = ((xxh_u32)c1 <<16) | ((xxh_u32)c2 << 24) + | ((xxh_u32)c3 << 0) | ((xxh_u32)len << 8); + xxh_u32 const combinedh = XXH_rotl32(XXH_swap32(combinedl), 13); + xxh_u64 const bitflipl = (XXH_readLE32(secret) ^ XXH_readLE32(secret+4)) + seed; + xxh_u64 const bitfliph = (XXH_readLE32(secret+8) ^ XXH_readLE32(secret+12)) - seed; + xxh_u64 const keyed_lo = (xxh_u64)combinedl ^ bitflipl; + xxh_u64 const keyed_hi = (xxh_u64)combinedh ^ bitfliph; + XXH128_hash_t h128; + h128.low64 = XXH64_avalanche(keyed_lo); + h128.high64 = XXH64_avalanche(keyed_hi); + return h128; + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_4to8_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(4 <= len && len <= 8); + seed ^= (xxh_u64)XXH_swap32((xxh_u32)seed) << 32; + { xxh_u32 const input_lo = XXH_readLE32(input); + xxh_u32 const input_hi = XXH_readLE32(input + len - 4); + xxh_u64 const input_64 = input_lo + ((xxh_u64)input_hi << 32); + xxh_u64 const bitflip = (XXH_readLE64(secret+16) ^ XXH_readLE64(secret+24)) + seed; + xxh_u64 const keyed = input_64 ^ bitflip; + + /* Shift len to the left to ensure it is even, this avoids even multiplies. */ + XXH128_hash_t m128 = XXH_mult64to128(keyed, XXH_PRIME64_1 + (len << 2)); + + m128.high64 += (m128.low64 << 1); + m128.low64 ^= (m128.high64 >> 3); + + m128.low64 = XXH_xorshift64(m128.low64, 35); + m128.low64 *= PRIME_MX2; + m128.low64 = XXH_xorshift64(m128.low64, 28); + m128.high64 = XXH3_avalanche(m128.high64); + return m128; + } +} + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_9to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(input != NULL); + XXH_ASSERT(secret != NULL); + XXH_ASSERT(9 <= len && len <= 16); + { xxh_u64 const bitflipl = (XXH_readLE64(secret+32) ^ XXH_readLE64(secret+40)) - seed; + xxh_u64 const bitfliph = (XXH_readLE64(secret+48) ^ XXH_readLE64(secret+56)) + seed; + xxh_u64 const input_lo = XXH_readLE64(input); + xxh_u64 input_hi = XXH_readLE64(input + len - 8); + XXH128_hash_t m128 = XXH_mult64to128(input_lo ^ input_hi ^ bitflipl, XXH_PRIME64_1); + /* + * Put len in the middle of m128 to ensure that the length gets mixed to + * both the low and high bits in the 128x64 multiply below. + */ + m128.low64 += (xxh_u64)(len - 1) << 54; + input_hi ^= bitfliph; + /* + * Add the high 32 bits of input_hi to the high 32 bits of m128, then + * add the long product of the low 32 bits of input_hi and XXH_PRIME32_2 to + * the high 64 bits of m128. + * + * The best approach to this operation is different on 32-bit and 64-bit. + */ + if (sizeof(void *) < sizeof(xxh_u64)) { /* 32-bit */ + /* + * 32-bit optimized version, which is more readable. + * + * On 32-bit, it removes an ADC and delays a dependency between the two + * halves of m128.high64, but it generates an extra mask on 64-bit. + */ + m128.high64 += (input_hi & 0xFFFFFFFF00000000ULL) + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2); + } else { + /* + * 64-bit optimized (albeit more confusing) version. + * + * Uses some properties of addition and multiplication to remove the mask: + * + * Let: + * a = input_hi.lo = (input_hi & 0x00000000FFFFFFFF) + * b = input_hi.hi = (input_hi & 0xFFFFFFFF00000000) + * c = XXH_PRIME32_2 + * + * a + (b * c) + * Inverse Property: x + y - x == y + * a + (b * (1 + c - 1)) + * Distributive Property: x * (y + z) == (x * y) + (x * z) + * a + (b * 1) + (b * (c - 1)) + * Identity Property: x * 1 == x + * a + b + (b * (c - 1)) + * + * Substitute a, b, and c: + * input_hi.hi + input_hi.lo + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) + * + * Since input_hi.hi + input_hi.lo == input_hi, we get this: + * input_hi + ((xxh_u64)input_hi.lo * (XXH_PRIME32_2 - 1)) + */ + m128.high64 += input_hi + XXH_mult32to64((xxh_u32)input_hi, XXH_PRIME32_2 - 1); + } + /* m128 ^= XXH_swap64(m128 >> 64); */ + m128.low64 ^= XXH_swap64(m128.high64); + + { /* 128x64 multiply: h128 = m128 * XXH_PRIME64_2; */ + XXH128_hash_t h128 = XXH_mult64to128(m128.low64, XXH_PRIME64_2); + h128.high64 += m128.high64 * XXH_PRIME64_2; + + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = XXH3_avalanche(h128.high64); + return h128; + } } +} + +/* + * Assumption: `secret` size is >= XXH3_SECRET_SIZE_MIN + */ +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_0to16_128b(const xxh_u8* input, size_t len, const xxh_u8* secret, XXH64_hash_t seed) +{ + XXH_ASSERT(len <= 16); + { if (len > 8) return XXH3_len_9to16_128b(input, len, secret, seed); + if (len >= 4) return XXH3_len_4to8_128b(input, len, secret, seed); + if (len) return XXH3_len_1to3_128b(input, len, secret, seed); + { XXH128_hash_t h128; + xxh_u64 const bitflipl = XXH_readLE64(secret+64) ^ XXH_readLE64(secret+72); + xxh_u64 const bitfliph = XXH_readLE64(secret+80) ^ XXH_readLE64(secret+88); + h128.low64 = XXH64_avalanche(seed ^ bitflipl); + h128.high64 = XXH64_avalanche( seed ^ bitfliph); + return h128; + } } +} + +/* + * A bit slower than XXH3_mix16B, but handles multiply by zero better. + */ +XXH_FORCE_INLINE XXH128_hash_t +XXH128_mix32B(XXH128_hash_t acc, const xxh_u8* input_1, const xxh_u8* input_2, + const xxh_u8* secret, XXH64_hash_t seed) +{ + acc.low64 += XXH3_mix16B (input_1, secret+0, seed); + acc.low64 ^= XXH_readLE64(input_2) + XXH_readLE64(input_2 + 8); + acc.high64 += XXH3_mix16B (input_2, secret+16, seed); + acc.high64 ^= XXH_readLE64(input_1) + XXH_readLE64(input_1 + 8); + return acc; +} + + +XXH_FORCE_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_17to128_128b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(16 < len && len <= 128); + + { XXH128_hash_t acc; + acc.low64 = len * XXH_PRIME64_1; + acc.high64 = 0; + +#if XXH_SIZE_OPT >= 1 + { + /* Smaller, but slightly slower. */ + unsigned int i = (unsigned int)(len - 1) / 32; + do { + acc = XXH128_mix32B(acc, input+16*i, input+len-16*(i+1), secret+32*i, seed); + } while (i-- != 0); + } +#else + if (len > 32) { + if (len > 64) { + if (len > 96) { + acc = XXH128_mix32B(acc, input+48, input+len-64, secret+96, seed); + } + acc = XXH128_mix32B(acc, input+32, input+len-48, secret+64, seed); + } + acc = XXH128_mix32B(acc, input+16, input+len-32, secret+32, seed); + } + acc = XXH128_mix32B(acc, input, input+len-16, secret, seed); +#endif + { XXH128_hash_t h128; + h128.low64 = acc.low64 + acc.high64; + h128.high64 = (acc.low64 * XXH_PRIME64_1) + + (acc.high64 * XXH_PRIME64_4) + + ((len - seed) * XXH_PRIME64_2); + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); + return h128; + } + } +} + +XXH_NO_INLINE XXH_PUREF XXH128_hash_t +XXH3_len_129to240_128b(const xxh_u8* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH64_hash_t seed) +{ + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); (void)secretSize; + XXH_ASSERT(128 < len && len <= XXH3_MIDSIZE_MAX); + + { XXH128_hash_t acc; + unsigned i; + acc.low64 = len * XXH_PRIME64_1; + acc.high64 = 0; + /* + * We set as `i` as offset + 32. We do this so that unchanged + * `len` can be used as upper bound. This reaches a sweet spot + * where both x86 and aarch64 get simple agen and good codegen + * for the loop. + */ + for (i = 32; i < 160; i += 32) { + acc = XXH128_mix32B(acc, + input + i - 32, + input + i - 16, + secret + i - 32, + seed); + } + acc.low64 = XXH3_avalanche(acc.low64); + acc.high64 = XXH3_avalanche(acc.high64); + /* + * NB: `i <= len` will duplicate the last 32-bytes if + * len % 32 was zero. This is an unfortunate necessity to keep + * the hash result stable. + */ + for (i=160; i <= len; i += 32) { + acc = XXH128_mix32B(acc, + input + i - 32, + input + i - 16, + secret + XXH3_MIDSIZE_STARTOFFSET + i - 160, + seed); + } + /* last bytes */ + acc = XXH128_mix32B(acc, + input + len - 16, + input + len - 32, + secret + XXH3_SECRET_SIZE_MIN - XXH3_MIDSIZE_LASTOFFSET - 16, + (XXH64_hash_t)0 - seed); + + { XXH128_hash_t h128; + h128.low64 = acc.low64 + acc.high64; + h128.high64 = (acc.low64 * XXH_PRIME64_1) + + (acc.high64 * XXH_PRIME64_4) + + ((len - seed) * XXH_PRIME64_2); + h128.low64 = XXH3_avalanche(h128.low64); + h128.high64 = (XXH64_hash_t)0 - XXH3_avalanche(h128.high64); + return h128; + } + } +} + +static XXH_PUREF XXH128_hash_t +XXH3_finalizeLong_128b(const xxh_u64* XXH_RESTRICT acc, const xxh_u8* XXH_RESTRICT secret, size_t secretSize, xxh_u64 len) +{ + XXH128_hash_t h128; + h128.low64 = XXH3_finalizeLong_64b(acc, secret, len); + h128.high64 = XXH3_mergeAccs(acc, secret + secretSize + - XXH_STRIPE_LEN - XXH_SECRET_MERGEACCS_START, + ~(len * XXH_PRIME64_2)); + return h128; +} + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_hashLong_128b_internal(const void* XXH_RESTRICT input, size_t len, + const xxh_u8* XXH_RESTRICT secret, size_t secretSize, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble) +{ + XXH_ALIGN(XXH_ACC_ALIGN) xxh_u64 acc[XXH_ACC_NB] = XXH3_INIT_ACC; + + XXH3_hashLong_internal_loop(acc, (const xxh_u8*)input, len, secret, secretSize, f_acc, f_scramble); + + /* converge into final hash */ + XXH_STATIC_ASSERT(sizeof(acc) == 64); + XXH_ASSERT(secretSize >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_finalizeLong_128b(acc, secret, secretSize, (xxh_u64)len); +} + +/* + * It's important for performance that XXH3_hashLong() is not inlined. + */ +XXH_NO_INLINE XXH_PUREF XXH128_hash_t +XXH3_hashLong_128b_default(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; (void)secret; (void)secretLen; + return XXH3_hashLong_128b_internal(input, len, XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_accumulate, XXH3_scrambleAcc); +} + +/* + * It's important for performance to pass @p secretLen (when it's static) + * to the compiler, so that it can properly optimize the vectorized loop. + * + * When the secret size is unknown, or on GCC 12 where the mix of NO_INLINE and FORCE_INLINE + * breaks -Og, this is XXH_NO_INLINE. + */ +XXH3_WITH_SECRET_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSecret(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)seed64; + return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, secretLen, + XXH3_accumulate, XXH3_scrambleAcc); +} + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSeed_internal(const void* XXH_RESTRICT input, size_t len, + XXH64_hash_t seed64, + XXH3_f_accumulate f_acc, + XXH3_f_scrambleAcc f_scramble, + XXH3_f_initCustomSecret f_initSec) +{ + if (seed64 == 0) + return XXH3_hashLong_128b_internal(input, len, + XXH3_kSecret, sizeof(XXH3_kSecret), + f_acc, f_scramble); + { XXH_ALIGN(XXH_SEC_ALIGN) xxh_u8 secret[XXH_SECRET_DEFAULT_SIZE]; + f_initSec(secret, seed64); + return XXH3_hashLong_128b_internal(input, len, (const xxh_u8*)secret, sizeof(secret), + f_acc, f_scramble); + } +} + +/* + * It's important for performance that XXH3_hashLong is not inlined. + */ +XXH_NO_INLINE XXH128_hash_t +XXH3_hashLong_128b_withSeed(const void* input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen) +{ + (void)secret; (void)secretLen; + return XXH3_hashLong_128b_withSeed_internal(input, len, seed64, + XXH3_accumulate, XXH3_scrambleAcc, XXH3_initCustomSecret); +} + +typedef XXH128_hash_t (*XXH3_hashLong128_f)(const void* XXH_RESTRICT, size_t, + XXH64_hash_t, const void* XXH_RESTRICT, size_t); + +XXH_FORCE_INLINE XXH128_hash_t +XXH3_128bits_internal(const void* input, size_t len, + XXH64_hash_t seed64, const void* XXH_RESTRICT secret, size_t secretLen, + XXH3_hashLong128_f f_hl128) +{ + XXH_ASSERT(secretLen >= XXH3_SECRET_SIZE_MIN); + /* + * If an action is to be taken if `secret` conditions are not respected, + * it should be done here. + * For now, it's a contract pre-condition. + * Adding a check and a branch here would cost performance at every hash. + */ + if (len <= 16) + return XXH3_len_0to16_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, seed64); + if (len <= 128) + return XXH3_len_17to128_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_len_129to240_128b((const xxh_u8*)input, len, (const xxh_u8*)secret, secretLen, seed64); + return f_hl128(input, len, seed64, secret, secretLen); +} + + +/* === Public XXH128 API === */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t XXH3_128bits(XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_128bits_internal(input, len, 0, + XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_hashLong_128b_default); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSecret(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_128bits_internal(input, len, 0, + (const xxh_u8*)secret, secretSize, + XXH3_hashLong_128b_withSecret); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSeed(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ + return XXH3_128bits_internal(input, len, seed, + XXH3_kSecret, sizeof(XXH3_kSecret), + XXH3_hashLong_128b_withSeed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH3_128bits_withSecretandSeed(XXH_NOESCAPE const void* input, size_t len, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + if (len <= XXH3_MIDSIZE_MAX) + return XXH3_128bits_internal(input, len, seed, XXH3_kSecret, sizeof(XXH3_kSecret), NULL); + return XXH3_hashLong_128b_withSecret(input, len, seed, secret, secretSize); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH128(XXH_NOESCAPE const void* input, size_t len, XXH64_hash_t seed) +{ + return XXH3_128bits_withSeed(input, len, seed); +} + + +/* === XXH3 128-bit streaming === */ +#ifndef XXH_NO_STREAM +/* + * All initialization and update functions are identical to 64-bit streaming variant. + * The only difference is the finalization routine. + */ + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset(XXH_NOESCAPE XXH3_state_t* statePtr) +{ + return XXH3_64bits_reset(statePtr); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecret(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize) +{ + return XXH3_64bits_reset_withSecret(statePtr, secret, secretSize); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH64_hash_t seed) +{ + return XXH3_64bits_reset_withSeed(statePtr, seed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_reset_withSecretandSeed(XXH_NOESCAPE XXH3_state_t* statePtr, XXH_NOESCAPE const void* secret, size_t secretSize, XXH64_hash_t seed) +{ + return XXH3_64bits_reset_withSecretandSeed(statePtr, secret, secretSize, seed); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_128bits_update(XXH_NOESCAPE XXH3_state_t* state, XXH_NOESCAPE const void* input, size_t len) +{ + return XXH3_64bits_update(state, input, len); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t XXH3_128bits_digest (XXH_NOESCAPE const XXH3_state_t* state) +{ + const unsigned char* const secret = (state->extSecret == NULL) ? state->customSecret : state->extSecret; + if (state->totalLen > XXH3_MIDSIZE_MAX) { + XXH_ALIGN(XXH_ACC_ALIGN) XXH64_hash_t acc[XXH_ACC_NB]; + XXH3_digest_long(acc, state, secret); + XXH_ASSERT(state->secretLimit + XXH_STRIPE_LEN >= sizeof(acc) + XXH_SECRET_MERGEACCS_START); + return XXH3_finalizeLong_128b(acc, secret, state->secretLimit + XXH_STRIPE_LEN, (xxh_u64)state->totalLen); + } + /* len <= XXH3_MIDSIZE_MAX : short code */ + if (state->useSeed) + return XXH3_128bits_withSeed(state->buffer, (size_t)state->totalLen, state->seed); + return XXH3_128bits_withSecret(state->buffer, (size_t)(state->totalLen), + secret, state->secretLimit + XXH_STRIPE_LEN); +} +#endif /* !XXH_NO_STREAM */ +/* 128-bit utility functions */ + +#include /* memcmp, memcpy */ + +/* return : 1 is equal, 0 if different */ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API int XXH128_isEqual(XXH128_hash_t h1, XXH128_hash_t h2) +{ + /* note : XXH128_hash_t is compact, it has no padding byte */ + return !(memcmp(&h1, &h2, sizeof(h1))); +} + +/* This prototype is compatible with stdlib's qsort(). + * @return : >0 if *h128_1 > *h128_2 + * <0 if *h128_1 < *h128_2 + * =0 if *h128_1 == *h128_2 */ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API int XXH128_cmp(XXH_NOESCAPE const void* h128_1, XXH_NOESCAPE const void* h128_2) +{ + XXH128_hash_t const h1 = *(const XXH128_hash_t*)h128_1; + XXH128_hash_t const h2 = *(const XXH128_hash_t*)h128_2; + int const hcmp = (h1.high64 > h2.high64) - (h2.high64 > h1.high64); + /* note : bets that, in most cases, hash values are different */ + if (hcmp) return hcmp; + return (h1.low64 > h2.low64) - (h2.low64 > h1.low64); +} + + +/*====== Canonical representation ======*/ +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API void +XXH128_canonicalFromHash(XXH_NOESCAPE XXH128_canonical_t* dst, XXH128_hash_t hash) +{ + XXH_STATIC_ASSERT(sizeof(XXH128_canonical_t) == sizeof(XXH128_hash_t)); + if (XXH_CPU_LITTLE_ENDIAN) { + hash.high64 = XXH_swap64(hash.high64); + hash.low64 = XXH_swap64(hash.low64); + } + XXH_memcpy(dst, &hash.high64, sizeof(hash.high64)); + XXH_memcpy((char*)dst + sizeof(hash.high64), &hash.low64, sizeof(hash.low64)); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH128_hash_t +XXH128_hashFromCanonical(XXH_NOESCAPE const XXH128_canonical_t* src) +{ + XXH128_hash_t h; + h.high64 = XXH_readBE64(src); + h.low64 = XXH_readBE64(src->digest + 8); + return h; +} + + + +/* ========================================== + * Secret generators + * ========================================== + */ +#define XXH_MIN(x, y) (((x) > (y)) ? (y) : (x)) + +XXH_FORCE_INLINE void XXH3_combine16(void* dst, XXH128_hash_t h128) +{ + XXH_writeLE64( dst, XXH_readLE64(dst) ^ h128.low64 ); + XXH_writeLE64( (char*)dst+8, XXH_readLE64((char*)dst+8) ^ h128.high64 ); +} + +/*! @ingroup XXH3_family */ +XXH_PUBLIC_API XXH_errorcode +XXH3_generateSecret(XXH_NOESCAPE void* secretBuffer, size_t secretSize, XXH_NOESCAPE const void* customSeed, size_t customSeedSize) +{ +#if (XXH_DEBUGLEVEL >= 1) + XXH_ASSERT(secretBuffer != NULL); + XXH_ASSERT(secretSize >= XXH3_SECRET_SIZE_MIN); +#else + /* production mode, assert() are disabled */ + if (secretBuffer == NULL) return XXH_ERROR; + if (secretSize < XXH3_SECRET_SIZE_MIN) return XXH_ERROR; +#endif + + if (customSeedSize == 0) { + customSeed = XXH3_kSecret; + customSeedSize = XXH_SECRET_DEFAULT_SIZE; + } +#if (XXH_DEBUGLEVEL >= 1) + XXH_ASSERT(customSeed != NULL); +#else + if (customSeed == NULL) return XXH_ERROR; +#endif + + /* Fill secretBuffer with a copy of customSeed - repeat as needed */ + { size_t pos = 0; + while (pos < secretSize) { + size_t const toCopy = XXH_MIN((secretSize - pos), customSeedSize); + memcpy((char*)secretBuffer + pos, customSeed, toCopy); + pos += toCopy; + } } + + { size_t const nbSeg16 = secretSize / 16; + size_t n; + XXH128_canonical_t scrambler; + XXH128_canonicalFromHash(&scrambler, XXH128(customSeed, customSeedSize, 0)); + for (n=0; n +#include +#include + +#include "../runtime_reflection_containers/runtime_reflected_variable.h" + +const rsl::dynamic_string reflection_code_generator::generate_variable( + [[maybe_unused]] runtime_reflected_variable parsed_variable) +{ + /* + std::string pad = " "; + std::ostringstream out; + + out << "rythe::reflection_containers::runtime_reflected_variable(\n"; + + const char* name = parsed_variable.get_name().data(); + out << pad << "\"" << name << "\"\n"; + + const char* name_space = parsed_variable.get_namespace().data(); + out << pad << "\"" << name_space << "\"\n"; + + switch(parsed_variable.get_access_modifier()) + { + case reflection_properties::access_modifier::public_access: + out << pad << "reflection_properties::acess_modifier::public_access,\n"; + break; + case reflection_properties::access_modifier::protected_access: + out << pad << "reflection_properties::acess_modifier::protected_access,\n"; + break; + case reflection_properties::access_modifier::private_access: + out << pad << "reflection_properties::acess_modifier::private_access,\n"; + break; + default: + out << pad << "reflection_properties::acess_modifier::public_access,\n"; + break; + } + + out << pad << (parsed_variable.get_is_static() ? "true" : "false") << ",\n"; + out << pad << (parsed_variable.get_is_const() ? "true" : "false") << ",\n"; + out << pad << (parsed_variable.get_is_array() ? "true" : "false") << ",\n"; + out << pad << parsed_variable.get_array_size() << ",\n"; + + out << pad << parsed_variable.get_offset() << ",\n"; + out << pad << parsed_variable.get_size() << ",\n"; + out << pad << parsed_variable.get_alignment() << ",\n"; + + //const char* type_str = parsed_variable.get_type_id().get_type().data(); + //out << pad << "reflection_id(rsl::dynamic_string::from_string_length(" << type_str << ")),\n"; + + const auto& attributes = parsed_variable.get_attributes(); + out << pad << "rsl::dynamic_array{"; + if(!attributes.empty()) { out << "\n"; } + for (int i = 0; i < attributes.size(); ++i) { + const char* attribute = attributes[i].data(); + out << pad << " rsl::dynamic_string::from_buffer(" + << attribute << ", " << (attribute ? attributes.size() : 0) << ")"; + if (i + 1 < attributes) out << ","; + out << "\n"; + } + + if(!attributes.empty()) { out << pad; } + out << "}\n"; + + out << " )"; + + std::string string = out.str(); + return rsl::dynamic_string::from_buffer(string.c_str(), string.size()); + */ + return {}; +} + +reflection_code_generator::reflection_code_generator() {} + +reflection_code_generator::~reflection_code_generator() {} + +void reflection_code_generator::generate_reflected_file(const compile_reflected_file& compile_file) +{ + const rsl::string_view source_location = compile_file.get_source_location(); + const rsl::dynamic_string gen_path = get_gen_source_file(source_location); + std::ofstream file(gen_path.data()); + if(!file.is_open()) + { + std::cout << "Could not open file " << source_location.data() << " for writing.\n"; + return; + } + const uint64_t file_hash = compile_file.id.get_full_hash(); + + file << "#include \"../runtime_reflection_containers/runtime_reflected_class.h\"\n"; + file << "#include \"../runtime_reflection_containers/runtime_reflected_function.h\"\n"; + file << "#include \"../runtime_reflection_containers/runtime_reflected_variable.h\"\n"; + file << "#include \"../reflection_id/reflection_id.h\"\n"; + file << "#include \"../reflection_context/reflection_registration_registry.h\"\n"; + file << "#include \"../reflection_context/reflection_context.h\"\n"; + + file << "void register_reflection_file_" << file_hash << "()\n"; + file << "{\n"; + + // Top-level classes + for(const auto& cls : compile_file.get_class_container()) + { + generate_reflected_class(file, *cls.get(), ""); + } + + file << "};\n\n"; + + file << "struct reflection_file_registration_helper\n"; + file << "{\n"; + file << " reflection_file_registration_helper()\n"; + file << " {\n"; + file << + " reflection_registration_registry::instance().add(®ister_reflection_file_0);\n"; + file << " }\n"; + file << "};\n" << "static reflection_file_registration_helper registration_instance;"; +} + +rsl::dynamic_string reflection_code_generator::get_gen_source_file(rsl::string_view source_location) +{ + if(source_location.size() >= 256) + { + std::cout << "source_path is too long\n"; + return {}; + } + + char buffer[256]; + std::memcpy(buffer, source_location.data(), source_location.size()); + buffer[source_location.size()] = '\0'; + + char* last_dot = std::strrchr(buffer, '.'); + if(last_dot != nullptr) { *last_dot = '\0'; } + + strcat_s(buffer, "_generated.hpp"); + + return rsl::dynamic_string::from_buffer(buffer, std::strlen(buffer) + 1); +} + +/* +rsl::dynamic_string reflection_code_generator::generate_variable(rythe::reflection_containers::runtime_reflected_variable parsed_variable) { + + std::string pad = " "; + std::ostringstream out; + + out << "rythe::reflection_containers::runtime_reflected_variable(\n"; + + const char* name = parsed_variable.get_source_location().data(); + out << pad << "\"" << name << "\"\n"; + + const char* name_space = parsed_variable.get_namespace().data(); + out << pad << "\"" << name_space << "\"\n"; + + switch (parsed_variable.get_access_modifier()) { + case reflection_properties::acess_modifier::public_access: + out << pad << "reflection_properties::acess_modifier::public_access,\n"; break; + case reflection_properties::acess_modifier::protected_access: + out << pad << "reflection_properties::acess_modifier::protected_access,\n"; break; + case reflection_properties::acess_modifier::private_access: + out << pad << "reflection_properties::acess_modifier::private_access,\n"; break; + default: + out << pad << "reflection_properties::acess_modifier::public_access,\n"; break; + } + + out << pad << (parsed_variable.get_is_static() ? "true" : "false") << ",\n"; + out << pad << (parsed_variable.get_is_const() ? "true" : "false") << ",\n"; + out << pad << (parsed_variable.get_is_array() ? "true" : "false") << ",\n"; + out << pad << parsed_variable.get_array_size() << ",\n"; + + out << pad << parsed_variable.get_offset() << ",\n"; + out << pad << parsed_variable.get_size() << ",\n"; + out << pad << parsed_variable.get_alignment() << ",\n"; + + const char* type_str = parsed_variable.get_type_id().get_type().data(); + out << pad << "reflection_id(rsl::dynamic_string::from_buffer(" + << type_str << ", " << (type_str ? std::strlen(type_str) : 0) << ")),\n"; + + const auto& attributes = parsed_variable.get_attributes(); + out << pad << "rsl::dynamic_array{"; + if (!attributes.empty()) out << "\n"; + for (int i = 0; i < attributes.size(); ++i) { + const char* attribute = attributes[i].data(); + out << pad << " rsl::dynamic_string::from_buffer(" + << attribute << ", " << (attribute ? attributes.size() : 0) << ")"; + if (i + 1 < attributes) out << ","; + out << "\n"; + } + if (!attributes.empty()) out << pad; + out << "}\n"; + + out << " )"; + + std::string string = out.str(); + return rsl::dynamic_string::from_buffer(string.c_str(), string.size()); +} +*/ + +void reflection_code_generator::generate_reflected_variable( + std::ofstream& file, + const compile_reflected_variable& variable, + const std::string& parent_name) +{ + const std::string variable_name = "variable_" + std::to_string(variable.id.get_full_hash()); + + file << " runtime_reflected_variable " << variable_name << ";\n"; + file << " " << generate_reflection_id(variable.id, variable_name).data(); + file << " " << variable_name << ".offset = " << variable.offset << ";\n"; + file << " " << variable_name << ".type_spelling = " << + "rsl::dynamic_string::from_string_length(\"" << variable.type_spelling.data() << "\");\n"; + file << " " << parent_name << ".add_variable(std::move(" << variable_name << "));\n\n"; +} + +void reflection_code_generator::generate_reflected_class( + std::ofstream& file, + const compile_reflected_class& cls, + const std::string& parent_name) +{ + const std::string class_var = "class_" + std::to_string(cls.id.get_full_hash()); + + file << " runtime_reflected_class " << class_var << ";\n"; + file << " " << generate_reflection_id(cls.id, class_var).data(); + file << " " << class_var << ".type_spelling = rsl::dynamic_string::from_string_length(\"" << + cls.name.data() << "\");\n\n"; + + for(const auto& var : cls.get_variable_container()) + { + generate_reflected_variable(file, *var.get(), class_var); + } + + for(const auto& nested : cls.get_class_container()) + { + generate_reflected_class(file, *nested.get(), class_var); + } + + if(parent_name.empty()) + { + file << " " << "reflection_context::instance().add_class(std::move(" << class_var << + "));\n\n"; + } + else { file << " " << parent_name << ".add_class(std::move(" << class_var << "));\n\n"; } +} + +std::string reflection_code_generator::generate_reflection_id( + reflection_id id, + const std::string& owner_name) const +{ + std::string generate; + + generate += owner_name; + generate += ".id = reflection_id("; + + generate += std::to_string(id.get_name_hash()) += "ULL"; + generate += ", "; + + generate += std::to_string(id.get_structure_hash()) += "ULL"; + generate += ", "; + + generate += std::to_string(id.get_full_hash()) += "ULL"; + generate += ");\n"; + + return generate; +} + diff --git a/libraries/rythe/reflection-core/src/impl/code_gen/reflection_code_generator.h b/libraries/rythe/reflection-core/src/impl/code_gen/reflection_code_generator.h new file mode 100644 index 0000000..8c37b38 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/code_gen/reflection_code_generator.h @@ -0,0 +1,34 @@ +#pragma once +#include "../runtime_reflection_containers/runtime_reflected_variable.h" +#include +#include +#include "../compiletime_reflection_containers/compile_reflected_file.h" + +class reflection_code_generator +{ +public: + const rsl::dynamic_string generate_variable(runtime_reflected_variable parsed_variable); + + reflection_code_generator(); + ~reflection_code_generator(); + + void generate_reflected_variable( + std::ofstream& file, + const compile_reflected_variable& variable, + const std::string& parent_name); + + void generate_reflected_function( + std::ofstream& file, + const compile_reflected_function& variable, + const std::string& parent_name); + + void generate_reflected_class( + std::ofstream& file, + const compile_reflected_class& variable, + [[maybe_unused]] const std::string& parent_name); + + std::string generate_reflection_id(reflection_id id, const std::string& owner) const; + + void generate_reflected_file(const compile_reflected_file& file); + static rsl::dynamic_string get_gen_source_file(rsl::string_view source_location); +}; diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_class.cpp b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_class.cpp new file mode 100644 index 0000000..a9386a0 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_class.cpp @@ -0,0 +1,98 @@ +#include "compile_reflected_class.h" + +compile_reflected_class::compile_reflected_class(const CXCursor& cursor) + : compile_reflected_element(get_name_from_cursor(cursor)) +{ + this->compile_reflection_container::verify_typename(); + this->compile_reflection_container::verify_typename(); + this->compile_reflection_container::verify_typename(); +} + +compile_reflected_class::~compile_reflected_class() {} + +compile_reflected_class& compile_reflected_class::add_class(CXCursor& cursor) +{ + return this->compile_reflection_container::add_element(cursor); +} + +compile_reflected_variable& compile_reflected_class::add_variable(CXCursor& cursor, CXCursor& parent) +{ + return this->compile_reflection_container::add_element(cursor, parent); +} + +compile_reflected_function& compile_reflected_class::add_function(CXCursor& cursor, CXCursor& parent) +{ + return this->compile_reflection_container::add_element(cursor, parent); +} + +const std::vector>& +compile_reflected_class::get_class_container() const +{ + return this->compile_reflection_container::get_container(); +} + +const std::vector>& +compile_reflected_class::get_function_container() const +{ + return this->compile_reflection_container::get_container(); +} + +const std::vector>& +compile_reflected_class::get_variable_container() const +{ + return this->compile_reflection_container::get_container(); +} + +void compile_reflected_class::print(int indent) const +{ + compile_reflected_element::print(indent + 1); + + compile_reflection_container::print_container(indent + 1); + compile_reflection_container::print_container(indent + 1); + compile_reflection_container::print_container(indent + 1); +} + +rsl::id_type compile_reflected_class::compute_own_structure_hash() noexcept +{ + return SIZE_MAX; +} + +// This order should be always deterministic. +// Embedded classes -> functions -> variables. +// Important because same input should produce same hash. +rsl::id_type compile_reflected_class::compute_container_structure_hash() noexcept +{ + rsl::id_type hash = rsl::internal::hash::default_seed; + + this->compile_reflection_container::sort_container( + &compile_reflection_container::sort_by_name_comparator); + rsl::id_type class_container_hash = this->compile_reflection_container< + compile_reflected_class>::get_container_hash(); + + if(class_container_hash != SIZE_MAX) + { + hash = rsl::combine_hash(rsl::internal::hash::default_seed, hash, class_container_hash); + } + + this->compile_reflection_container::sort_container( + &compile_reflection_container::sort_by_offset_comparator); + rsl::id_type variable_container_hash = this->compile_reflection_container< + compile_reflected_variable>::get_container_hash(); + + if(variable_container_hash != SIZE_MAX) + { + hash = rsl::combine_hash(rsl::internal::hash::default_seed, hash, variable_container_hash); + } + + this->compile_reflection_container::sort_container( + &compile_reflection_container::sort_by_name_comparator); + rsl::id_type function_container_hash = this->compile_reflection_container< + compile_reflected_function>::get_container_hash(); + + if(function_container_hash != SIZE_MAX) + { + hash = rsl::combine_hash(rsl::internal::hash::default_seed, hash, function_container_hash); + } + + return hash; +} diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_class.h b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_class.h new file mode 100644 index 0000000..6b326e4 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_class.h @@ -0,0 +1,35 @@ +#pragma once +#include "compile_reflection_container.h" +#include "compile_reflected_variable.h" +#include "compile_reflected_function.h" + +#include + +struct compile_reflected_class + : public compile_reflected_element, + public compile_reflection_container, + public compile_reflection_container, + public compile_reflection_container +{ +public: + compile_reflected_class(const CXCursor& cursor); + compile_reflected_class(const compile_reflected_class&) = default; + compile_reflected_class(compile_reflected_class&&) noexcept = default; + ~compile_reflected_class(); + + compile_reflected_class& add_class(CXCursor& cursor); + compile_reflected_variable& add_variable(CXCursor& cursor, CXCursor& parent); + compile_reflected_function& add_function(CXCursor& cursor, CXCursor& parent); + + [[nodiscard]] const std::vector>& + get_class_container() const; + [[nodiscard]] const std::vector>& + get_function_container() const; + [[nodiscard]] const std::vector>& + get_variable_container() const; + + void print(int indent) const override; +private: + rsl::id_type compute_own_structure_hash() noexcept override; + rsl::id_type compute_container_structure_hash() noexcept override; +}; diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_element.cpp b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_element.cpp new file mode 100644 index 0000000..841dc64 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_element.cpp @@ -0,0 +1,60 @@ +#include "compile_reflected_element.h" + +#include +#include + +compile_reflected_element::compile_reflected_element(rsl::dynamic_string name) + : name(std::move(name)) + , id(reflection_id::null()) {} + +compile_reflected_element::compile_reflected_element(rsl::dynamic_string name, reflection_id id) + : name(std::move(name)) + , id(id) {} + +void compile_reflected_element::print(int indent) const +{ + for(auto i = 0; i < indent; i++) { std::cout << ' '; } + std::cout << "Name: " << this->name.data() << '\n'; + + id.print(indent); +} +void compile_reflected_element::set_full_hash() noexcept +{ + rsl::id_type name_hash = compute_name_hash(); + rsl::id_type structure_hash = compute_structure_hash(); + + id.set_name_hash(name_hash); + id.set_structure_hash(structure_hash); + id.set_full_hash(rsl::combine_hash(rsl::internal::hash::default_seed, name_hash, structure_hash)); +} + +rsl::id_type compile_reflected_element::compute_name_hash() const noexcept { return rsl::hash_string(name); } + +rsl::id_type compile_reflected_element::compute_structure_hash() noexcept +{ + rsl::id_type hash = rsl::internal::hash::default_seed; + + rsl::id_type own_hash = compute_own_structure_hash(); + if(own_hash != SIZE_MAX) { hash = rsl::combine_hash(rsl::internal::hash::default_seed, hash, own_hash); } + + rsl::id_type container_hash = compute_container_structure_hash(); + if(container_hash != SIZE_MAX) + { + hash = rsl::combine_hash(rsl::internal::hash::default_seed, hash, container_hash); + } + + return hash; +} + +rsl::id_type compile_reflected_element::compute_own_structure_hash() noexcept { return SIZE_MAX; } + +rsl::id_type compile_reflected_element::compute_container_structure_hash() noexcept { return SIZE_MAX; } + +rsl::dynamic_string compile_reflected_element::get_name_from_cursor(const CXCursor& cursor) +{ + CXString field_name = clang_getCursorSpelling(cursor); + const char* field_spelling = clang_getCString(field_name); + auto result = rsl::dynamic_string::from_string_length(field_spelling); + clang_disposeString(field_name); + return result; +} diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_element.h b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_element.h new file mode 100644 index 0000000..9bed3d8 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_element.h @@ -0,0 +1,30 @@ +#pragma once +#include "../reflection_id/reflection_id.h" +#include "clang-c/Index.h" +#include "rsl/containers" + +struct compile_reflected_element +{ +public: + rsl::dynamic_string name; + reflection_id id; + + compile_reflected_element(rsl::dynamic_string name); + compile_reflected_element(rsl::dynamic_string name, reflection_id id); + compile_reflected_element(const compile_reflected_element& other) = default; + compile_reflected_element(compile_reflected_element&& other) noexcept = default; + virtual ~compile_reflected_element() = default; + + virtual void print(int indent) const; + + void set_full_hash() noexcept; +private: +protected: + rsl::id_type compute_structure_hash() noexcept; + rsl::id_type compute_name_hash() const noexcept; + + virtual rsl::id_type compute_own_structure_hash() noexcept; + virtual rsl::id_type compute_container_structure_hash() noexcept; + + static rsl::dynamic_string get_name_from_cursor(const CXCursor& cursor); +}; diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_file.cpp b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_file.cpp new file mode 100644 index 0000000..79aa231 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_file.cpp @@ -0,0 +1,48 @@ +#include "compile_reflected_file.h" + +#include + +compile_reflected_file::compile_reflected_file(rsl::dynamic_string&& name, rsl::dynamic_string&& source_location) + : compile_reflected_element(std::move(name)) + , source_location(std::move(source_location)) +{ + this->compile_reflection_container::verify_typename(); + this->compile_reflection_container::verify_typename(); + this->compile_reflection_container::verify_typename(); +} + +compile_reflected_file::~compile_reflected_file() {} + +rsl::string_view compile_reflected_file::get_source_location() const { return source_location; } + +const std::vector>& +compile_reflected_file::get_class_container() const +{ + return this->compile_reflection_container::get_container(); +} + +const std::vector>& +compile_reflected_file::get_function_container() const +{ + return this->compile_reflection_container::get_container(); +} + +const std::vector>& +compile_reflected_file::get_variable_container() const +{ + return this->compile_reflection_container::get_container(); +} + + +void compile_reflected_file::print(int indent) const +{ + compile_reflected_element::print(indent + 1); + + for(auto i = 0; i < indent + 1; i++) { std::cout << ' '; } + std::cout << "Source location: " << this->source_location.data() << '\n'; + + this->compile_reflection_container::print_container(indent + 1); + this->compile_reflection_container::print_container(indent + 1); + this->compile_reflection_container::print_container(indent + 1); +} + diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_file.h b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_file.h new file mode 100644 index 0000000..adbfedc --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_file.h @@ -0,0 +1,34 @@ +#pragma once +#include "compile_reflected_class.h" +#include "compile_reflected_element.h" +#include "compile_reflected_variable.h" +#include "compile_reflected_function.h" +#include "clang-c/Index.h" + +struct compile_reflected_file + : public compile_reflected_element, + public compile_reflection_container, + public compile_reflection_container, + public compile_reflection_container +{ +public: + compile_reflected_file(rsl::dynamic_string&& name, rsl::dynamic_string&& source_location); + compile_reflected_file(const compile_reflected_file& other) = default; + compile_reflected_file(compile_reflected_file&& other) noexcept = default; + ~compile_reflected_file(); + + [[nodiscard]] const std::vector>& + get_class_container() const; + [[nodiscard]] const std::vector>& + get_function_container() const; + [[nodiscard]] const std::vector>& + get_variable_container() const; + + + void print(int indent) const override; + + rsl::string_view get_source_location() const; + +private: + const rsl::dynamic_string source_location; +}; diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_function.cpp b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_function.cpp new file mode 100644 index 0000000..a783865 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_function.cpp @@ -0,0 +1,26 @@ +#include "compile_reflected_function.h" + + +compile_reflected_function::compile_reflected_function(const CXCursor& cursor, const CXCursor& parent) + : compile_reflected_element(get_name_from_cursor(cursor)) +{ + CXType cursorType = clang_getCursorType(cursor); + CXString typeSpelling = clang_getTypeSpelling(cursorType); + const char* type_spelling_cstring = clang_getCString(typeSpelling); + + this->type_spelling = rsl::dynamic_string::from_string_length(type_spelling_cstring); + + clang_disposeString(typeSpelling); + + if(parent.kind) {} +} + +compile_reflected_function::~compile_reflected_function() {} + +void compile_reflected_function::print(int indent) const +{ + compile_reflected_element::print(indent + 1); + + for(auto i = 0; i < indent + 1; i++) { std::cout << ' '; } + std::cout << "Type: " << this->type_spelling.data() << '\n'; +} diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_function.h b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_function.h new file mode 100644 index 0000000..0c1a05a --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_function.h @@ -0,0 +1,19 @@ +#pragma once +#include "clang-c/Index.h" +#include +#include + +#include "compile_reflected_element.h" + +struct compile_reflected_function : public compile_reflected_element +{ +public: + rsl::dynamic_string type_spelling; + + compile_reflected_function(const CXCursor& cursor, const CXCursor& parent); + compile_reflected_function(const compile_reflected_function& other) = default; + compile_reflected_function(compile_reflected_function&& other) noexcept = default; + ~compile_reflected_function(); + + void print(int indent) const override; +}; diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_variable.cpp b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_variable.cpp new file mode 100644 index 0000000..8f1fb27 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_variable.cpp @@ -0,0 +1,47 @@ +#include "compile_reflected_variable.h" + +compile_reflected_variable::compile_reflected_variable(const CXCursor& cursor, const CXCursor& parent) + : compile_reflected_element(get_name_from_cursor(cursor)) +{ + CXType parent_type = clang_getCursorType(parent); + + CXType cursor_type = clang_getCursorType(cursor); + CXString type_spelling_string = clang_getTypeSpelling(cursor_type); + const char* type_spelling_var = clang_getCString(type_spelling_string); + + this->type_spelling = rsl::dynamic_string::from_string_length(type_spelling_var); + + long long offset_bits = clang_Type_getOffsetOf(parent_type, name.data()); + if(offset_bits < 0) { std::cout << "Offset bits are less than 0: " << offset_bits << '\n'; } + this->offset = static_cast(offset_bits) / 8; + + clang_disposeString(type_spelling_string); +} + +compile_reflected_variable::compile_reflected_variable( + rsl::dynamic_string name, + reflection_id type, + rsl::size_type offset) + : compile_reflected_element(name) + , offset(offset) { if(type.get_name_hash()) {} } + +compile_reflected_variable::~compile_reflected_variable() {} + +void compile_reflected_variable::print(int indent) const +{ + compile_reflected_element::print(indent + 1); + + for(auto i = 0; i < indent + 1; i++) { std::cout << ' '; } + std::cout << "Type: " << this->type_spelling.data() << '\n'; + + for(auto i = 0; i < indent + 1; i++) { std::cout << ' '; } + std::cout << "Offset in bytes: " << this->offset << '\n'; +} + +rsl::id_type compile_reflected_variable::compute_own_structure_hash() noexcept +{ + return rsl::combine_hash( + rsl::internal::hash::default_seed, + rsl::internal::hash::default_seed, + rsl::hash_string(type_spelling)); +} diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_variable.h b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_variable.h new file mode 100644 index 0000000..c8419f1 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflected_variable.h @@ -0,0 +1,23 @@ +#pragma once +#include +#include +#include "compile_reflected_element.h" +#include "../reflection_id/reflection_id.h" +#include "clang-c/Index.h" + +struct compile_reflected_variable : public compile_reflected_element +{ +public: + rsl::dynamic_string type_spelling; + rsl::size_type offset; + + compile_reflected_variable(const CXCursor& cursor, const CXCursor& parent); + compile_reflected_variable(rsl::dynamic_string name, reflection_id type, rsl::size_type offset); + ~compile_reflected_variable(); + compile_reflected_variable(const compile_reflected_variable& other) = default; + compile_reflected_variable(compile_reflected_variable&& other) noexcept = default; + + void print(int indent) const override; +private: + rsl::id_type compute_own_structure_hash() noexcept override; +}; diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflection_container.h b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflection_container.h new file mode 100644 index 0000000..439b1c8 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflection_container.h @@ -0,0 +1,36 @@ +#pragma once +#include +#include +#include +#include "clang-c/Index.h" +#include "compile_reflected_element.h" +#include + +template +class compile_reflection_container +{ +public: + T& add_element(const CXCursor& cursor); + T& add_element(const CXCursor& cursor, const CXCursor& parent); + + void print_container(int indent) const; + + // Need to me manually called in the owner class. + // Done so because using CRTP and class can have compile_reflection_container + void verify_typename() const; + + rsl::id_type get_container_hash() const noexcept; + + [[nodiscard]] const std::vector>& get_container() const; + +protected: + std::vector> container; + + template + void sort_container(Comparator&& comparator); + + static bool sort_by_name_comparator(const std::unique_ptr& a, const std::unique_ptr& b); + static bool sort_by_offset_comparator(const std::unique_ptr& a, const std::unique_ptr& b); +}; + +#include "compile_reflection_container.inl" diff --git a/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflection_container.inl b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflection_container.inl new file mode 100644 index 0000000..88ec743 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/compiletime_reflection_containers/compile_reflection_container.inl @@ -0,0 +1,82 @@ +#pragma once +template +T& compile_reflection_container::add_element(const CXCursor& cursor) +{ + container.emplace_back(std::make_unique(cursor)); + return *container.back(); +} + +template +T& compile_reflection_container::add_element(const CXCursor& cursor, const CXCursor& parent) +{ + container.emplace_back(std::make_unique(cursor, parent)); + return *container.back(); +} + +template +void compile_reflection_container::print_container(int indent) const +{ + for(const auto& element : container) + { + auto element_object = static_cast(element.get()); + if(!element_object) { std::cout << "Reflected container does not store reflected elements." << '\n'; } + element_object->print(indent); + } +} + +template +void compile_reflection_container::verify_typename() const +{ + static_assert(std::is_base_of_v, "T must inherit from compile_reflected_element"); +} + +template +const std::vector>& compile_reflection_container::get_container() const +{ + return container; +} + +template +rsl::id_type compile_reflection_container::get_container_hash() const noexcept +{ + if(container.size() == 0) { return SIZE_MAX; } + + rsl::id_type hash = rsl::internal::hash::default_seed; + for(const auto& element : container) + { + auto element_object = static_cast(element.get()); + hash = rsl::combine_hash(rsl::internal::hash::default_seed, hash, element_object->id.get_structure_hash()); + } + + return hash; +} + +template +template +void compile_reflection_container::sort_container(Comparator&& comparator) +{ + std::ranges::sort(container, std::forward(comparator)); +} + +template +bool compile_reflection_container::sort_by_name_comparator(const std::unique_ptr& a, const std::unique_ptr& b) +{ + rsl::dynamic_string a_string = a->name; + rsl::dynamic_string b_string = b->name; + const std::size_t min_length = a_string.size() < b_string.size() ? a_string.size() : b_string.size(); + + for(std::size_t i = 0; i < min_length; ++i) + { + if(a_string[i] < b_string[i]) { return true; } + if(a_string[i] > b_string[i]) { return false; } + } + return a_string.size() < b_string.size(); +} + +template +bool compile_reflection_container::sort_by_offset_comparator( + const std::unique_ptr& a, + const std::unique_ptr& b) +{ + return a->offset < b->offset; +} diff --git a/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_context.cpp b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_context.cpp new file mode 100644 index 0000000..48dd122 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_context.cpp @@ -0,0 +1,53 @@ +#include "reflection_context.h" +reflection_context& reflection_context::instance() +{ + static reflection_context ctx; + return ctx; +} + +runtime_reflected_class* reflection_context::find_class(reflection_id id) +{ + auto it = classes.find(id); + if(it != classes.end()) { return &it->second; } + return nullptr; +} + +void reflection_context::add_class(runtime_reflected_class&& cls) +{ + classes.emplace(cls.id, std::move(cls)); +} + +void reflection_context::clear() { classes.clear(); } + +void reflection_context::print_classes() const +{ + std::cout << "Registered reflection classes:\n"; + + for(const auto& [id, cls] : classes) + { + std::cout << "- reflection id: " << cls.id.get_full_hash() << ", type spelling: " << cls. + type_spelling + .data() + << "\n"; + + if(!cls.variables.empty()) + { + std::cout << " Variables:\n"; + for(const auto& var : cls.variables) + { + std::cout << " - " << var.type_spelling.data() << " (offset " << var.offset << + ", id: " << var.id.get_full_hash() << ")\n"; + } + } + + if(!cls.classes.empty()) + { + std::cout << " Nested Classes:\n"; + for(const auto& nested : cls.classes) + { + std::cout << " - Nested Class id: " << nested.id.get_full_hash() << + ", type spelling: " << nested.type_spelling.data() << "\n"; + } + } + } +} diff --git a/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_context.h b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_context.h new file mode 100644 index 0000000..37e4d49 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_context.h @@ -0,0 +1,29 @@ +#pragma once + +#include // for std::cout +#include +#include +#include +#include + +#include "../../runtime_reflection_containers.h" +class reflection_context +{ +public: + static reflection_context& instance(); + + reflection_context(const reflection_context&) = delete; + reflection_context& operator=(const reflection_context&) = delete; + + runtime_reflected_class* find_class(reflection_id id); + + void add_class(runtime_reflected_class&& cls); + + void clear(); + void print_classes() const; +private: + reflection_context() = default; + ~reflection_context() = default; + + std::unordered_map classes; +}; diff --git a/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_registration_registry.cpp b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_registration_registry.cpp new file mode 100644 index 0000000..ca40650 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_registration_registry.cpp @@ -0,0 +1,26 @@ +#include "reflection_registration_registry.h" + +#include + +reflection_registration_registry& reflection_registration_registry::instance() +{ + static reflection_registration_registry registry; + return registry; +} + +void reflection_registration_registry::add(void (*func)()) { functions.push_back(func); } + +void reflection_registration_registry::run_all() const +{ + std::cout << functions.size() << std::endl; + for(auto function : functions) + { + if(function) + { + std::cout << "lala"; + function(); + } + } +} + +void reflection_registration_registry::clear() { functions.clear(); } diff --git a/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_registration_registry.h b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_registration_registry.h new file mode 100644 index 0000000..9a9d539 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_context/reflection_registration_registry.h @@ -0,0 +1,23 @@ +#pragma once +#include +#include + +class reflection_registration_registry +{ +public: + static reflection_registration_registry& instance(); + + void add(void (*func)()); + + void run_all() const; + + void clear(); +private: + reflection_registration_registry() = default; + ~reflection_registration_registry() = default; + + reflection_registration_registry(const reflection_registration_registry&) = delete; + reflection_registration_registry& operator=(const reflection_registration_registry&) = delete; + + std::vector functions; +}; diff --git a/libraries/rythe/reflection-core/src/impl/reflection_id/reflection_id.cpp b/libraries/rythe/reflection-core/src/impl/reflection_id/reflection_id.cpp new file mode 100644 index 0000000..19e3ee8 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_id/reflection_id.cpp @@ -0,0 +1,63 @@ +#include "reflection_id.h" + +#include +#include // std::memcpy +#include +#include +#include + +reflection_id::reflection_id(rsl::id_type name_hash, rsl::id_type structure_hash, rsl::id_type full_hash) noexcept + : name_hash(name_hash) + , structure_hash(structure_hash) + , full_hash(full_hash) {} + +reflection_id reflection_id::null() noexcept { return reflection_id(); } +rsl::id_type reflection_id::get_name_hash() const { return name_hash; } + +rsl::id_type reflection_id::get_structure_hash() const { return structure_hash; } +rsl::id_type reflection_id::get_full_hash() const { return full_hash; } + +void reflection_id::set_name_hash(rsl::id_type hash) { name_hash = hash; } +void reflection_id::set_structure_hash(rsl::id_type hash) { structure_hash = hash; } +void reflection_id::set_full_hash(rsl::id_type hash) { full_hash = hash; } + +bool reflection_id::operator==(const reflection_id& other) const noexcept +{ + return name_hash == other.name_hash && structure_hash == other.structure_hash && full_hash == + other.full_hash; +} + +bool reflection_id::operator!=(const reflection_id& other) const noexcept { return !(*this == other); } + +void reflection_id::print(int indent) const +{ + for(auto i = 0; i < indent; i++) { std::cout << ' '; } + std::cout << "Full hash: " << this->get_full_hash() << '\n'; + + for(auto i = 0; i < indent + 1; i++) { std::cout << ' '; } + std::cout << "Name hash: " << this->get_name_hash() << '\n'; + + for(auto i = 0; i < indent + 1; i++) { std::cout << ' '; } + std::cout << "Structure hash: " << this->get_structure_hash() << '\n'; +} + +// TODO: returns dangling string view. FIX +rsl::string_view reflection_id::get_hash_value_hex_string(const rsl::id_type& hash) +{ + std::stringstream result; + result << "0x" << std::hex << hash; + return rsl::string_view::from_buffer(result.str().data(), result.str().size()); +} + +// TODO: returns dangling string view. FIX +rsl::string_view reflection_id::get_hash_value_decimal_string(const rsl::id_type& hash) +{ + std::stringstream result; + result << hash; + return rsl::string_view::from_buffer(result.str().data(), result.str().size()); +} + +reflection_id::reflection_id() noexcept + : name_hash(0) + , structure_hash(0) + , full_hash(0) {} diff --git a/libraries/rythe/reflection-core/src/impl/reflection_id/reflection_id.h b/libraries/rythe/reflection-core/src/impl/reflection_id/reflection_id.h new file mode 100644 index 0000000..21e8e8a --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_id/reflection_id.h @@ -0,0 +1,63 @@ +#pragma once + +#include +#include +#include +#include +#include + +#include "rsl/impl/containers/string.hpp" +#include "rsl/impl/math/basic/abs.hpp" +#include "rsl/impl/util/hash.hpp" + +struct reflection_id +{ +public: + reflection_id(rsl::id_type name_hash, rsl::id_type structure_hash, rsl::id_type full_hash) noexcept; + + reflection_id(const reflection_id&) = default; + reflection_id(reflection_id&&) = default; + ~reflection_id() = default; + + reflection_id& operator=(const reflection_id&) = default; + reflection_id& operator=(reflection_id&&) = default; + + static reflection_id null() noexcept; + + rsl::id_type get_name_hash() const; + rsl::id_type get_structure_hash() const; + rsl::id_type get_full_hash() const; + + void set_name_hash(rsl::id_type hash); + void set_structure_hash(rsl::id_type hash); + void set_full_hash(rsl::id_type hash); + + bool operator==(const reflection_id& other) const noexcept; + bool operator!=(const reflection_id& other) const noexcept; + + void print(int indent) const; + + static rsl::string_view get_hash_value_hex_string(const rsl::id_type& hash); + static rsl::string_view get_hash_value_decimal_string(const rsl::id_type& hash); + + reflection_id() noexcept; +private: + + rsl::id_type name_hash; + rsl::id_type structure_hash; + rsl::id_type full_hash; +}; + +// specialize std::hash +namespace std +{ + template<> + struct hash + { + std::size_t operator()(const reflection_id& id) const noexcept + { + // Combine the three 64-bit numbers into one hash + return id.get_full_hash(); + } + }; +} diff --git a/libraries/rythe/reflection-core/src/impl/reflection_parsers/ast_source_parser.cpp b/libraries/rythe/reflection-core/src/impl/reflection_parsers/ast_source_parser.cpp new file mode 100644 index 0000000..88eb5c2 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_parsers/ast_source_parser.cpp @@ -0,0 +1,177 @@ +#include "ast_source_parser.h" + +reflection_parsers::ast_source_parser::ast_source_parser() +{ + index = clang_createIndex(0, 0); + + code_generator = reflection_code_generator(); +} + +reflection_parsers::ast_source_parser::~ast_source_parser() {} + +void reflection_parsers::ast_source_parser::generate_reflection_files( + const std::unordered_set& folders) +{ + //std::cout << "reflection_parsers::ast_source_parser::generate_reflection_files" << std::endl; + index = clang_createIndex(0, 0); + + for(auto& folder : folders) + { + for(const auto& entry : std::filesystem::recursive_directory_iterator(folder)) + { + if(!entry.is_regular_file()) { continue; } + std::filesystem::path path = entry.path(); + std::string filename = path.filename().string(); + + if(filename.ends_with("_generated.h") || filename.ends_with("_generated.hpp")) + { + continue; + } + + if(path.extension() == ".hpp" || path.extension() == ".h") + { + ast_parse_file(path.string(), index); + } + } + } + //print_all_files(); + + for(auto iterator = all_files.begin(); iterator != all_files.end(); ++iterator) + { + code_generator.generate_reflected_file(iterator->second); + } +} + +void reflection_parsers::ast_source_parser::print_all_files() const +{ + for(const auto& element : all_files) { element.second.print(0); } +} + +CXChildVisitResult reflection_parsers::ast_source_parser::visitor_from_file( + CXCursor current_cursor, + CXCursor parent_cursor, + CXClientData client_data) +{ + if(parent_cursor.kind) {} + + auto* parent_file = static_cast(client_data); + CXCursorKind kind = clang_getCursorKind(current_cursor); + + /* + CXString kind_spelling = clang_getCursorKindSpelling(kind); + std::cout << clang_getCString(kind_spelling) << '\n'; + + clang_disposeString(kind_spelling); + */ + + switch(kind) + { + case CXCursor_ClassDecl: + { + compile_reflected_class& reflected_class = parent_file->compile_reflection_container< + compile_reflected_class>::add_element(current_cursor); + clang_visitChildren(current_cursor, visitor_from_class, &reflected_class); + reflected_class.set_full_hash(); + break; + } + case CXCursor_FieldDecl: + { + compile_reflected_variable& variable = parent_file->compile_reflection_container< + compile_reflected_variable>::add_element(current_cursor, parent_cursor); + variable.set_full_hash(); + break; + } + case CXCursor_FunctionDecl: + { + /*compile_reflected_function& function = parent_file->compile_reflection_container< + compile_reflected_function>::add_element(current_cursor, parent_cursor); + function.set_full_hash();*/ + break; + } + default: + break; + } + + + return CXChildVisit_Continue; +} + +CXChildVisitResult reflection_parsers::ast_source_parser::visitor_from_class( + CXCursor current_cursor, + CXCursor parent_cursor, + CXClientData client_data) +{ + if(parent_cursor.kind) {} + + auto* parent_class = static_cast(client_data); + CXCursorKind kind = clang_getCursorKind(current_cursor); + + switch(kind) + { + case CXCursor_ClassDecl: + { + compile_reflected_class& reflected_class = parent_class->add_class(current_cursor); + clang_visitChildren(current_cursor, visitor_from_class, &reflected_class); + reflected_class.set_full_hash(); + break; + } + case CXCursor_FieldDecl: + { + compile_reflected_variable& variable = parent_class->compile_reflection_container< + compile_reflected_variable>::add_element( + current_cursor, + parent_cursor); + variable.set_full_hash(); + break; + } + case CXCursor_FunctionDecl: + { + std::cout << "Indeed a function" << '\n'; + break; + } + default: + break; + } + return CXChildVisit_Continue; +} + +// Meant to pass temporary +void reflection_parsers::ast_source_parser::ast_parse_file(const std::string&& filePath, CXIndex cxIndex) +{ + std::cout << "Parsing file path: " << filePath << '\n'; + CXTranslationUnit unit = clang_parseTranslationUnit( + cxIndex, + filePath.data(), + nullptr, + 0, + nullptr, + 0, + CXTranslationUnit_None); + + if(unit == nullptr) + { + std::cerr << "Unable to parse translation unit. Quitting.\n"; + return; + } + CXCursor cursor = clang_getTranslationUnitCursor(unit); //Obtain a cursor at the root of the translation unit + + auto [iterator, inserted] = all_files.emplace( + filePath, + compile_reflected_file( + rsl::dynamic_string::from_string_length(extract_file_name(std::string_view(filePath)).data()), + rsl::dynamic_string::from_string_length(filePath.c_str()))); + + clang_visitChildren(cursor, visitor_from_file, &iterator->second); +} + +std::string_view reflection_parsers::ast_source_parser::extract_file_name(const std::string_view file_path) +{ + size_t last_slash_position = file_path.rfind('\\'); + std::string_view file_name; + if(last_slash_position != std::string::npos) + { + file_name = file_path.substr(last_slash_position + 1, file_path.size() - last_slash_position); + } + else { std::cerr << "File name is empty.\n"; } + return file_name; +} diff --git a/libraries/rythe/reflection-core/src/impl/reflection_parsers/ast_source_parser.h b/libraries/rythe/reflection-core/src/impl/reflection_parsers/ast_source_parser.h new file mode 100644 index 0000000..faf4132 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_parsers/ast_source_parser.h @@ -0,0 +1,54 @@ +#pragma once +#include +#include +#include +#include +#include "../code_gen/reflection_code_generator.h" +#include "../runtime_reflection_containers/runtime_reflected_variable.h" +#include "../compiletime_reflection_containers/compile_reflected_file.h" +#include "../compiletime_reflection_containers/compile_reflected_class.h" + +#include "rsl/containers" + +namespace reflection_parsers +{ + class ast_source_parser + { + public: + ast_source_parser(); + ~ast_source_parser(); + + void generate_reflection_files(const std::unordered_set& folders); + + void print_all_files() const; + private: + reflection_code_generator code_generator; + CXIndex index; + + std::map all_files; + + // If this function is a static function that is never needed outside ast_source_parser.cpp, then it would be best to keep + // it as a translation unit local function. You do this by moving this function into ast_source_parser.cpp and either marking + // it static, or by placing it into an anonymous namespace: + // https://en.cppreference.com/w/cpp/language/storage_duration.html#Internal_linkage + + static CXChildVisitResult visitor_from_file( + CXCursor current_cursor, + CXCursor parent_cursor, + CXClientData client_data); + static CXChildVisitResult visitor_from_class( + CXCursor current_cursor, + CXCursor parent_cursor, + CXClientData client_data); + + // You don't want to copy strings, either use a string_view if you need temporary read access or are going to copy, + // or use a string rvalue reference (rsl::dynamic_string&&) if you want to take ownership. + // Generally you'd want the ownership to not transfer and thus create the copy inside the function instead. + // However, you can use rsl::dynamic_string&& in order to force the calling code to create a temporary string object. + // The reason you'd do that is in order to make the caller aware of the allocation that will happen inside the function because + // you expect the function to be called in a place where allocations are not in the budget. + void ast_parse_file(const std::string&& filePath, CXIndex index); + + static std::string_view extract_file_name(const std::string_view filePath); + }; +} diff --git a/libraries/rythe/reflection-core/src/impl/reflection_properties/access_modifier.h b/libraries/rythe/reflection-core/src/impl/reflection_properties/access_modifier.h new file mode 100644 index 0000000..1ac23d7 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/reflection_properties/access_modifier.h @@ -0,0 +1,9 @@ +#pragma once + +namespace reflection_properties +{ + enum struct [[rythe_closed_enum]] access_modifier + { + private_access, protected_access, public_access + }; +} diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_enum.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_enum.cpp similarity index 100% rename from applications/reflection-experiments/src/reflection_containers/reflected_enum.cpp rename to libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_enum.cpp diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_enum.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_enum.h new file mode 100644 index 0000000..eb04481 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_enum.h @@ -0,0 +1,3 @@ +#pragma once + +struct reflected_enum {}; diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_global_variable.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_global_variable.cpp similarity index 100% rename from applications/reflection-experiments/src/reflection_containers/reflected_global_variable.cpp rename to libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_global_variable.cpp diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_global_variable.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_global_variable.h new file mode 100644 index 0000000..dc0d458 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_global_variable.h @@ -0,0 +1,4 @@ +#pragma once + +// is there a difference between a global variable and a variable in a struct? one is contained in a struct, and the other is contained in a namespace. +struct reflected_global_variable {}; diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_struct.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_struct.cpp similarity index 100% rename from applications/reflection-experiments/src/reflection_containers/reflected_struct.cpp rename to libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_struct.cpp diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_struct.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_struct.h similarity index 100% rename from applications/reflection-experiments/src/reflection_containers/reflected_struct.h rename to libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_struct.h diff --git a/applications/reflection-experiments/src/reflection_containers/reflected_template.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_template.cpp similarity index 100% rename from applications/reflection-experiments/src/reflection_containers/reflected_template.cpp rename to libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_template.cpp diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_template.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_template.h new file mode 100644 index 0000000..c9e76b4 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/reflected_template.h @@ -0,0 +1,3 @@ +#pragma once + +struct reflected_template {}; diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_class.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_class.cpp new file mode 100644 index 0000000..2a57fdc --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_class.cpp @@ -0,0 +1,39 @@ +#include "runtime_reflected_class.h" + +runtime_reflected_class::runtime_reflected_class() {} + +runtime_reflected_class::runtime_reflected_class( + rsl::dynamic_string&& name, + rsl::dynamic_string&& type_spelling, + reflection_id&& id) + : name(std::move(name)) + , type_spelling(std::move(type_spelling)) + , id(id) {} + +runtime_reflected_class::~runtime_reflected_class() {} + +[[nodiscard]] bool runtime_reflected_class::operator==(const runtime_reflected_class& other) const +{ + return this->id == other.id; +} + +[[nodiscard]] bool runtime_reflected_class::operator!=(const runtime_reflected_class& other) const +{ + return !this->operator==(other); +} + +void runtime_reflected_class::add_variable(runtime_reflected_variable&& variable) +{ + this->variables.push_back(std::move(variable)); +} + +void runtime_reflected_class::add_class(runtime_reflected_class&& cls) +{ + this->classes.push_back(std::move(cls)); +} + +void runtime_reflected_class::add_function(runtime_reflected_function&& function) +{ + this->functions.push_back(std::move(function)); +} + diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_class.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_class.h new file mode 100644 index 0000000..76093f6 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_class.h @@ -0,0 +1,33 @@ +#pragma once +#include "runtime_reflected_function.h" +#include "runtime_reflected_variable.h" +#include "../reflection_id/reflection_id.h" +#include "rsl/impl/containers/string.hpp" + +struct runtime_reflected_class +{ +public: + runtime_reflected_class(); + runtime_reflected_class(rsl::dynamic_string&& name, rsl::dynamic_string&& type_spelling, reflection_id&& id); + runtime_reflected_class(runtime_reflected_class&) = delete; + runtime_reflected_class(runtime_reflected_class&&) = default; + ~runtime_reflected_class(); + + bool operator==(const runtime_reflected_class& other) const; + bool operator!=(const runtime_reflected_class& other) const; + + runtime_reflected_class& operator=(const runtime_reflected_class&) = delete; + runtime_reflected_class& operator=(runtime_reflected_class&&) = delete; + + void add_variable(runtime_reflected_variable&& variable); + void add_class(runtime_reflected_class&& cls); + void add_function(runtime_reflected_function&& function); + + rsl::dynamic_string name; + rsl::dynamic_string type_spelling; + reflection_id id; + std::vector classes; + std::vector functions; + std::vector variables; +private: +}; diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_function.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_function.cpp new file mode 100644 index 0000000..159046c --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_function.cpp @@ -0,0 +1 @@ +#include "runtime_reflected_function.h" diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_function.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_function.h new file mode 100644 index 0000000..57f1586 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_function.h @@ -0,0 +1,59 @@ +#pragma once +#include +#include + +#include "../reflection_properties/access_modifier.h" +#include "../reflection_id/reflection_id.h" + +struct runtime_reflected_function + { + private: + rsl::hashed_string name; + rsl::dynamic_string current_namespace; + reflection_properties::access_modifier access_modifier; + + bool is_static; + bool is_const; + + void* memory_reference; + + reflection_id return_type_id; + rsl::dynamic_array parameters; + rsl::dynamic_array attributes; + public: + runtime_reflected_function( + rsl::hashed_string name, + rsl::dynamic_string current_namespace, + reflection_properties::access_modifier access_modifier, + bool is_static, + bool is_const, + void* memory_reference, + reflection_id return_type_id, + rsl::dynamic_array parameters, + rsl::dynamic_array attributes); + + const rsl::hashed_string& get_name() const; + const rsl::dynamic_string& get_namespace() const; + reflection_properties::access_modifier get_access_modifier() const; + + bool get_is_static() const; + bool get_is_const() const; + + const reflection_id& get_return_type() const; + const rsl::dynamic_array& get_parameters() const; + const rsl::dynamic_array& get_attributes() const; + + void set_name(const rsl::hashed_string& new_name); + void set_namespace(const rsl::dynamic_string& ns); + void set_access_modifier(reflection_properties::access_modifier modifier); + + void set_is_static(bool value); + void set_is_const(bool value); + + void set_memory_reference(void* mem); + void set_return_type(const reflection_id& id); + void set_parameters(const rsl::dynamic_array& params); + void set_attributes(const rsl::dynamic_array& attrs); + + void* get_value_ptr(void* obj); + }; diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_variable.cpp b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_variable.cpp new file mode 100644 index 0000000..548cf79 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_variable.cpp @@ -0,0 +1,111 @@ +#include "runtime_reflected_variable.h" + +runtime_reflected_variable::runtime_reflected_variable() {} + +runtime_reflected_variable::runtime_reflected_variable( + rsl::dynamic_string&& name, + reflection_id&& id, + rsl::size_type&& offset, + rsl::dynamic_string&& type_spelling) + : name(std::move(name)) + , id(id) + , offset(offset) + , type_spelling(std::move(type_spelling)) {} + + + +/* +rythe::reflection_containers::runtime_reflected_variable::runtime_reflected_variable( + rsl::hashed_string name, + rsl::dynamic_string current_namespace, + reflection_properties::access_modifier access_modifier, + bool is_static, + bool is_const, + bool is_array, + int array_size, + size_t offset, + size_t size, + uint16_t alignment, + reflection_id type_id, + rsl::dynamic_array attributes) + : name(name) + , current_namespace(current_namespace) + , access_modifier(access_modifier) + , is_static(is_static) + , is_const(is_const) + , is_array(is_array) + , array_size(array_size) + , offset(offset) + , size(size) + , alignment(alignment) + , type_id(type_id) + , attributes(attributes) {} + +void* rythe::reflection_containers::runtime_reflected_variable::get_value_ptr(void* obj) +{ + return reinterpret_cast(obj) + offset; +} + +const rsl::hashed_string& rythe::reflection_containers::runtime_reflected_variable::get_name() const { return this->name; } + +const rsl::dynamic_string& rythe::reflection_containers::runtime_reflected_variable::get_namespace() const +{ + return this->current_namespace; +} + +reflection_properties::access_modifier rythe::reflection_containers::runtime_reflected_variable::get_access_modifier() const +{ + return this->access_modifier; +} + +bool rythe::reflection_containers::runtime_reflected_variable::get_is_static() const { return this->is_static; } + +bool rythe::reflection_containers::runtime_reflected_variable::get_is_const() const { return this->is_const; } + +bool rythe::reflection_containers::runtime_reflected_variable::get_is_array() const { return this->is_array; } + +int rythe::reflection_containers::runtime_reflected_variable::get_array_size() const { return this->array_size; } + +size_t rythe::reflection_containers::runtime_reflected_variable::get_offset() const { return this->offset; } + +size_t rythe::reflection_containers::runtime_reflected_variable::get_size() const { return this->size; } + +uint16_t rythe::reflection_containers::runtime_reflected_variable::get_alignment() const { return this->alignment; } + +const reflection_id& rythe::reflection_containers::runtime_reflected_variable::get_type_id() const { return this->type_id; } + +const rsl::dynamic_array& rythe::reflection_containers::runtime_reflected_variable::get_attributes() const +{ + return this->attributes; +} + +void rythe::reflection_containers::runtime_reflected_variable::set_name(const rsl::hashed_string& name) { this->name = name; } + +void rythe::reflection_containers::runtime_reflected_variable::set_namespace(const rsl::dynamic_string& current_namespace) +{ + this->current_namespace = current_namespace; +} + +void rythe::reflection_containers::runtime_reflected_variable::set_access_modifier( + reflection_properties::access_modifier access_modifier) { this->access_modifier = access_modifier; } + +void rythe::reflection_containers::runtime_reflected_variable::set_is_static(bool is_static) { this->is_static = is_static; } + +void rythe::reflection_containers::runtime_reflected_variable::set_is_const(bool is_const) { this->is_const = is_const; } + +void rythe::reflection_containers::runtime_reflected_variable::set_is_array(bool is_array) { this->is_array = is_array; } + +void rythe::reflection_containers::runtime_reflected_variable::set_array_size(int array_size) { this->array_size = array_size; } + +void rythe::reflection_containers::runtime_reflected_variable::set_offset(size_t offset) { this->offset = offset; } + +void rythe::reflection_containers::runtime_reflected_variable::set_size(size_t size) { this->size = size; } + +void rythe::reflection_containers::runtime_reflected_variable::set_alignment(uint16_t alignment) +{ + this->alignment = alignment; +} + +void rythe::reflection_containers::runtime_reflected_variable::set_attributes( + const rsl::dynamic_array& attributes) { this->attributes = attributes; } +*/ diff --git a/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_variable.h b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_variable.h new file mode 100644 index 0000000..482d117 --- /dev/null +++ b/libraries/rythe/reflection-core/src/impl/runtime_reflection_containers/runtime_reflected_variable.h @@ -0,0 +1,126 @@ +#pragma once +#include +#include "../reflection_properties/access_modifier.h" +#include "../reflection_id/reflection_id.h" + + +struct runtime_reflected_variable +{ +public: + runtime_reflected_variable(); + runtime_reflected_variable( + rsl::dynamic_string&& name, + reflection_id&& id, + rsl::size_type&& offset, + rsl::dynamic_string&& type_spelling); + + rsl::dynamic_string name; + reflection_id id; + rsl::size_type offset; + rsl::dynamic_string type_spelling; +}; + + + + + +/* +namespace rythe::reflection_containers +{ + struct runtime_reflected_variable + { + rsl::hashed_string name; + rsl::dynamic_string current_namespace; + reflection_properties::access_modifier access_modifier; + + bool is_static; + bool is_const; + + bool is_array; + int array_size; + + size_t offset; + size_t size; + uint16_t alignment; + + reflection_id type_id; + + rsl::dynamic_array attributes; + + runtime_reflected_variable( + rsl::hashed_string name, + rsl::dynamic_string current_namespace, + reflection_properties::access_modifier access_modifier, + bool is_static, + bool is_const, + bool is_array, + int array_size, + size_t offset, + size_t size, + uint16_t alignment, + reflection_id type_id, + rsl::dynamic_array attributes); + + + template + runtime_reflected_variable( + rsl::hashed_string name, + rsl::dynamic_string current_namespace, + reflection_properties::access_modifier access_modifier, + bool is_static, + bool is_const, + bool is_array, + int array_size, + size_t offset, + reflection_id type_id, + rsl::dynamic_array attributes) + : name(name) + , current_namespace(current_namespace) + , access_modifier(access_modifier) + , is_static(is_static) + , is_const(is_const) + , is_array(is_array) + , array_size(array_size) + , offset(offset) + , size(sizeof(T)) + , alignment(alignof(T)) + , type_id(type_id) + , attributes(attributes) {} + + void* get_value_ptr(void* obj); + + const rsl::hashed_string& get_name() const; + const rsl::dynamic_string& get_namespace() const; + reflection_properties::access_modifier get_access_modifier() const; + + bool get_is_static() const; + bool get_is_const() const; + + bool get_is_array() const; + int get_array_size() const; + + size_t get_offset() const; + size_t get_size() const; + uint16_t get_alignment() const; + + const reflection_id& get_type_id() const; + const rsl::dynamic_array& get_attributes() const; + + void set_name(const rsl::hashed_string& n); + void set_namespace(const rsl::dynamic_string& ns); + void set_access_modifier(reflection_properties::access_modifier mod); + + void set_is_static(bool v); + void set_is_const(bool v); + + void set_is_array(bool v); + void set_array_size(int sz); + + void set_offset(size_t o); + void set_size(size_t s); + void set_alignment(uint16_t a); + + void set_attributes(const rsl::dynamic_array& attrs); + }; +} +*/ diff --git a/libraries/rythe/reflection-core/src/runtime_reflection_containers.h b/libraries/rythe/reflection-core/src/runtime_reflection_containers.h new file mode 100644 index 0000000..68c8a73 --- /dev/null +++ b/libraries/rythe/reflection-core/src/runtime_reflection_containers.h @@ -0,0 +1,5 @@ +#pragma once +#include "impl/runtime_reflection_containers/runtime_reflected_class.h" +#include "impl/runtime_reflection_containers/runtime_reflected_function.h" +#include "impl/runtime_reflection_containers/runtime_reflected_variable.h" +#include "impl/reflection_id/reflection_id.h" diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/BuildSystem.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/BuildSystem.h new file mode 100644 index 0000000..57e16af --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/BuildSystem.h @@ -0,0 +1,153 @@ +/*==-- clang-c/BuildSystem.h - Utilities for use by build systems -*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides various utilities for use by build systems. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_BUILDSYSTEM_H +#define LLVM_CLANG_C_BUILDSYSTEM_H + +#include "clang-c/CXErrorCode.h" +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * \defgroup BUILD_SYSTEM Build system utilities + * @{ + */ + +/** + * Return the timestamp for use with Clang's + * \c -fbuild-session-timestamp= option. + */ +CINDEX_LINKAGE unsigned long long clang_getBuildSessionTimestamp(void); + +/** + * Object encapsulating information about overlaying virtual + * file/directories over the real file system. + */ +typedef struct CXVirtualFileOverlayImpl *CXVirtualFileOverlay; + +/** + * Create a \c CXVirtualFileOverlay object. + * Must be disposed with \c clang_VirtualFileOverlay_dispose(). + * + * \param options is reserved, always pass 0. + */ +CINDEX_LINKAGE CXVirtualFileOverlay +clang_VirtualFileOverlay_create(unsigned options); + +/** + * Map an absolute virtual file path to an absolute real one. + * The virtual path must be canonicalized (not contain "."/".."). + * \returns 0 for success, non-zero to indicate an error. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_VirtualFileOverlay_addFileMapping(CXVirtualFileOverlay, + const char *virtualPath, + const char *realPath); + +/** + * Set the case sensitivity for the \c CXVirtualFileOverlay object. + * The \c CXVirtualFileOverlay object is case-sensitive by default, this + * option can be used to override the default. + * \returns 0 for success, non-zero to indicate an error. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_VirtualFileOverlay_setCaseSensitivity(CXVirtualFileOverlay, + int caseSensitive); + +/** + * Write out the \c CXVirtualFileOverlay object to a char buffer. + * + * \param options is reserved, always pass 0. + * \param out_buffer_ptr pointer to receive the buffer pointer, which should be + * disposed using \c clang_free(). + * \param out_buffer_size pointer to receive the buffer size. + * \returns 0 for success, non-zero to indicate an error. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_VirtualFileOverlay_writeToBuffer(CXVirtualFileOverlay, unsigned options, + char **out_buffer_ptr, + unsigned *out_buffer_size); + +/** + * free memory allocated by libclang, such as the buffer returned by + * \c CXVirtualFileOverlay() or \c clang_ModuleMapDescriptor_writeToBuffer(). + * + * \param buffer memory pointer to free. + */ +CINDEX_LINKAGE void clang_free(void *buffer); + +/** + * Dispose a \c CXVirtualFileOverlay object. + */ +CINDEX_LINKAGE void clang_VirtualFileOverlay_dispose(CXVirtualFileOverlay); + +/** + * Object encapsulating information about a module.modulemap file. + */ +typedef struct CXModuleMapDescriptorImpl *CXModuleMapDescriptor; + +/** + * Create a \c CXModuleMapDescriptor object. + * Must be disposed with \c clang_ModuleMapDescriptor_dispose(). + * + * \param options is reserved, always pass 0. + */ +CINDEX_LINKAGE CXModuleMapDescriptor +clang_ModuleMapDescriptor_create(unsigned options); + +/** + * Sets the framework module name that the module.modulemap describes. + * \returns 0 for success, non-zero to indicate an error. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_ModuleMapDescriptor_setFrameworkModuleName(CXModuleMapDescriptor, + const char *name); + +/** + * Sets the umbrella header name that the module.modulemap describes. + * \returns 0 for success, non-zero to indicate an error. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_ModuleMapDescriptor_setUmbrellaHeader(CXModuleMapDescriptor, + const char *name); + +/** + * Write out the \c CXModuleMapDescriptor object to a char buffer. + * + * \param options is reserved, always pass 0. + * \param out_buffer_ptr pointer to receive the buffer pointer, which should be + * disposed using \c clang_free(). + * \param out_buffer_size pointer to receive the buffer size. + * \returns 0 for success, non-zero to indicate an error. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_ModuleMapDescriptor_writeToBuffer(CXModuleMapDescriptor, unsigned options, + char **out_buffer_ptr, + unsigned *out_buffer_size); + +/** + * Dispose a \c CXModuleMapDescriptor object. + */ +CINDEX_LINKAGE void clang_ModuleMapDescriptor_dispose(CXModuleMapDescriptor); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif /* CLANG_C_BUILD_SYSTEM_H */ + diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXCompilationDatabase.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXCompilationDatabase.h new file mode 100644 index 0000000..2b336e5 --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXCompilationDatabase.h @@ -0,0 +1,174 @@ +/*===-- clang-c/CXCompilationDatabase.h - Compilation database ---*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides a public interface to use CompilationDatabase without *| +|* the full Clang C++ API. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_CXCOMPILATIONDATABASE_H +#define LLVM_CLANG_C_CXCOMPILATIONDATABASE_H + +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** \defgroup COMPILATIONDB CompilationDatabase functions + * \ingroup CINDEX + * + * @{ + */ + +/** + * A compilation database holds all information used to compile files in a + * project. For each file in the database, it can be queried for the working + * directory or the command line used for the compiler invocation. + * + * Must be freed by \c clang_CompilationDatabase_dispose + */ +typedef void * CXCompilationDatabase; + +/** + * Contains the results of a search in the compilation database + * + * When searching for the compile command for a file, the compilation db can + * return several commands, as the file may have been compiled with + * different options in different places of the project. This choice of compile + * commands is wrapped in this opaque data structure. It must be freed by + * \c clang_CompileCommands_dispose. + */ +typedef void * CXCompileCommands; + +/** + * Represents the command line invocation to compile a specific file. + */ +typedef void * CXCompileCommand; + +/** + * Error codes for Compilation Database + */ +typedef enum { + /* + * No error occurred + */ + CXCompilationDatabase_NoError = 0, + + /* + * Database can not be loaded + */ + CXCompilationDatabase_CanNotLoadDatabase = 1 + +} CXCompilationDatabase_Error; + +/** + * Creates a compilation database from the database found in directory + * buildDir. For example, CMake can output a compile_commands.json which can + * be used to build the database. + * + * It must be freed by \c clang_CompilationDatabase_dispose. + */ +CINDEX_LINKAGE CXCompilationDatabase +clang_CompilationDatabase_fromDirectory(const char *BuildDir, + CXCompilationDatabase_Error *ErrorCode); + +/** + * Free the given compilation database + */ +CINDEX_LINKAGE void +clang_CompilationDatabase_dispose(CXCompilationDatabase); + +/** + * Find the compile commands used for a file. The compile commands + * must be freed by \c clang_CompileCommands_dispose. + */ +CINDEX_LINKAGE CXCompileCommands +clang_CompilationDatabase_getCompileCommands(CXCompilationDatabase, + const char *CompleteFileName); + +/** + * Get all the compile commands in the given compilation database. + */ +CINDEX_LINKAGE CXCompileCommands +clang_CompilationDatabase_getAllCompileCommands(CXCompilationDatabase); + +/** + * Free the given CompileCommands + */ +CINDEX_LINKAGE void clang_CompileCommands_dispose(CXCompileCommands); + +/** + * Get the number of CompileCommand we have for a file + */ +CINDEX_LINKAGE unsigned +clang_CompileCommands_getSize(CXCompileCommands); + +/** + * Get the I'th CompileCommand for a file + * + * Note : 0 <= i < clang_CompileCommands_getSize(CXCompileCommands) + */ +CINDEX_LINKAGE CXCompileCommand +clang_CompileCommands_getCommand(CXCompileCommands, unsigned I); + +/** + * Get the working directory where the CompileCommand was executed from + */ +CINDEX_LINKAGE CXString +clang_CompileCommand_getDirectory(CXCompileCommand); + +/** + * Get the filename associated with the CompileCommand. + */ +CINDEX_LINKAGE CXString +clang_CompileCommand_getFilename(CXCompileCommand); + +/** + * Get the number of arguments in the compiler invocation. + * + */ +CINDEX_LINKAGE unsigned +clang_CompileCommand_getNumArgs(CXCompileCommand); + +/** + * Get the I'th argument value in the compiler invocations + * + * Invariant : + * - argument 0 is the compiler executable + */ +CINDEX_LINKAGE CXString +clang_CompileCommand_getArg(CXCompileCommand, unsigned I); + +/** + * Get the number of source mappings for the compiler invocation. + */ +CINDEX_LINKAGE unsigned +clang_CompileCommand_getNumMappedSources(CXCompileCommand); + +/** + * Get the I'th mapped source path for the compiler invocation. + */ +CINDEX_LINKAGE CXString +clang_CompileCommand_getMappedSourcePath(CXCompileCommand, unsigned I); + +/** + * Get the I'th mapped source content for the compiler invocation. + */ +CINDEX_LINKAGE CXString +clang_CompileCommand_getMappedSourceContent(CXCompileCommand, unsigned I); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif + diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXDiagnostic.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXDiagnostic.h new file mode 100644 index 0000000..911d001 --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXDiagnostic.h @@ -0,0 +1,379 @@ +/*===-- clang-c/CXDiagnostic.h - C Index Diagnostics --------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides the interface to C Index diagnostics. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_CXDIAGNOSTIC_H +#define LLVM_CLANG_C_CXDIAGNOSTIC_H + +#include "clang-c/CXSourceLocation.h" +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * \defgroup CINDEX_DIAG Diagnostic reporting + * + * @{ + */ + +/** + * Describes the severity of a particular diagnostic. + */ +enum CXDiagnosticSeverity { + /** + * A diagnostic that has been suppressed, e.g., by a command-line + * option. + */ + CXDiagnostic_Ignored = 0, + + /** + * This diagnostic is a note that should be attached to the + * previous (non-note) diagnostic. + */ + CXDiagnostic_Note = 1, + + /** + * This diagnostic indicates suspicious code that may not be + * wrong. + */ + CXDiagnostic_Warning = 2, + + /** + * This diagnostic indicates that the code is ill-formed. + */ + CXDiagnostic_Error = 3, + + /** + * This diagnostic indicates that the code is ill-formed such + * that future parser recovery is unlikely to produce useful + * results. + */ + CXDiagnostic_Fatal = 4 +}; + +/** + * A single diagnostic, containing the diagnostic's severity, + * location, text, source ranges, and fix-it hints. + */ +typedef void *CXDiagnostic; + +/** + * A group of CXDiagnostics. + */ +typedef void *CXDiagnosticSet; + +/** + * Determine the number of diagnostics in a CXDiagnosticSet. + */ +CINDEX_LINKAGE unsigned clang_getNumDiagnosticsInSet(CXDiagnosticSet Diags); + +/** + * Retrieve a diagnostic associated with the given CXDiagnosticSet. + * + * \param Diags the CXDiagnosticSet to query. + * \param Index the zero-based diagnostic number to retrieve. + * + * \returns the requested diagnostic. This diagnostic must be freed + * via a call to \c clang_disposeDiagnostic(). + */ +CINDEX_LINKAGE CXDiagnostic clang_getDiagnosticInSet(CXDiagnosticSet Diags, + unsigned Index); + +/** + * Describes the kind of error that occurred (if any) in a call to + * \c clang_loadDiagnostics. + */ +enum CXLoadDiag_Error { + /** + * Indicates that no error occurred. + */ + CXLoadDiag_None = 0, + + /** + * Indicates that an unknown error occurred while attempting to + * deserialize diagnostics. + */ + CXLoadDiag_Unknown = 1, + + /** + * Indicates that the file containing the serialized diagnostics + * could not be opened. + */ + CXLoadDiag_CannotLoad = 2, + + /** + * Indicates that the serialized diagnostics file is invalid or + * corrupt. + */ + CXLoadDiag_InvalidFile = 3 +}; + +/** + * Deserialize a set of diagnostics from a Clang diagnostics bitcode + * file. + * + * \param file The name of the file to deserialize. + * \param error A pointer to a enum value recording if there was a problem + * deserializing the diagnostics. + * \param errorString A pointer to a CXString for recording the error string + * if the file was not successfully loaded. + * + * \returns A loaded CXDiagnosticSet if successful, and NULL otherwise. These + * diagnostics should be released using clang_disposeDiagnosticSet(). + */ +CINDEX_LINKAGE CXDiagnosticSet clang_loadDiagnostics( + const char *file, enum CXLoadDiag_Error *error, CXString *errorString); + +/** + * Release a CXDiagnosticSet and all of its contained diagnostics. + */ +CINDEX_LINKAGE void clang_disposeDiagnosticSet(CXDiagnosticSet Diags); + +/** + * Retrieve the child diagnostics of a CXDiagnostic. + * + * This CXDiagnosticSet does not need to be released by + * clang_disposeDiagnosticSet. + */ +CINDEX_LINKAGE CXDiagnosticSet clang_getChildDiagnostics(CXDiagnostic D); + +/** + * Destroy a diagnostic. + */ +CINDEX_LINKAGE void clang_disposeDiagnostic(CXDiagnostic Diagnostic); + +/** + * Options to control the display of diagnostics. + * + * The values in this enum are meant to be combined to customize the + * behavior of \c clang_formatDiagnostic(). + */ +enum CXDiagnosticDisplayOptions { + /** + * Display the source-location information where the + * diagnostic was located. + * + * When set, diagnostics will be prefixed by the file, line, and + * (optionally) column to which the diagnostic refers. For example, + * + * \code + * test.c:28: warning: extra tokens at end of #endif directive + * \endcode + * + * This option corresponds to the clang flag \c -fshow-source-location. + */ + CXDiagnostic_DisplaySourceLocation = 0x01, + + /** + * If displaying the source-location information of the + * diagnostic, also include the column number. + * + * This option corresponds to the clang flag \c -fshow-column. + */ + CXDiagnostic_DisplayColumn = 0x02, + + /** + * If displaying the source-location information of the + * diagnostic, also include information about source ranges in a + * machine-parsable format. + * + * This option corresponds to the clang flag + * \c -fdiagnostics-print-source-range-info. + */ + CXDiagnostic_DisplaySourceRanges = 0x04, + + /** + * Display the option name associated with this diagnostic, if any. + * + * The option name displayed (e.g., -Wconversion) will be placed in brackets + * after the diagnostic text. This option corresponds to the clang flag + * \c -fdiagnostics-show-option. + */ + CXDiagnostic_DisplayOption = 0x08, + + /** + * Display the category number associated with this diagnostic, if any. + * + * The category number is displayed within brackets after the diagnostic text. + * This option corresponds to the clang flag + * \c -fdiagnostics-show-category=id. + */ + CXDiagnostic_DisplayCategoryId = 0x10, + + /** + * Display the category name associated with this diagnostic, if any. + * + * The category name is displayed within brackets after the diagnostic text. + * This option corresponds to the clang flag + * \c -fdiagnostics-show-category=name. + */ + CXDiagnostic_DisplayCategoryName = 0x20 +}; + +/** + * Format the given diagnostic in a manner that is suitable for display. + * + * This routine will format the given diagnostic to a string, rendering + * the diagnostic according to the various options given. The + * \c clang_defaultDiagnosticDisplayOptions() function returns the set of + * options that most closely mimics the behavior of the clang compiler. + * + * \param Diagnostic The diagnostic to print. + * + * \param Options A set of options that control the diagnostic display, + * created by combining \c CXDiagnosticDisplayOptions values. + * + * \returns A new string containing for formatted diagnostic. + */ +CINDEX_LINKAGE CXString clang_formatDiagnostic(CXDiagnostic Diagnostic, + unsigned Options); + +/** + * Retrieve the set of display options most similar to the + * default behavior of the clang compiler. + * + * \returns A set of display options suitable for use with \c + * clang_formatDiagnostic(). + */ +CINDEX_LINKAGE unsigned clang_defaultDiagnosticDisplayOptions(void); + +/** + * Determine the severity of the given diagnostic. + */ +CINDEX_LINKAGE enum CXDiagnosticSeverity + clang_getDiagnosticSeverity(CXDiagnostic); + +/** + * Retrieve the source location of the given diagnostic. + * + * This location is where Clang would print the caret ('^') when + * displaying the diagnostic on the command line. + */ +CINDEX_LINKAGE CXSourceLocation clang_getDiagnosticLocation(CXDiagnostic); + +/** + * Retrieve the text of the given diagnostic. + */ +CINDEX_LINKAGE CXString clang_getDiagnosticSpelling(CXDiagnostic); + +/** + * Retrieve the name of the command-line option that enabled this + * diagnostic. + * + * \param Diag The diagnostic to be queried. + * + * \param Disable If non-NULL, will be set to the option that disables this + * diagnostic (if any). + * + * \returns A string that contains the command-line option used to enable this + * warning, such as "-Wconversion" or "-pedantic". + */ +CINDEX_LINKAGE CXString clang_getDiagnosticOption(CXDiagnostic Diag, + CXString *Disable); + +/** + * Retrieve the category number for this diagnostic. + * + * Diagnostics can be categorized into groups along with other, related + * diagnostics (e.g., diagnostics under the same warning flag). This routine + * retrieves the category number for the given diagnostic. + * + * \returns The number of the category that contains this diagnostic, or zero + * if this diagnostic is uncategorized. + */ +CINDEX_LINKAGE unsigned clang_getDiagnosticCategory(CXDiagnostic); + +/** + * Retrieve the name of a particular diagnostic category. This + * is now deprecated. Use clang_getDiagnosticCategoryText() + * instead. + * + * \param Category A diagnostic category number, as returned by + * \c clang_getDiagnosticCategory(). + * + * \returns The name of the given diagnostic category. + */ +CINDEX_DEPRECATED CINDEX_LINKAGE CXString +clang_getDiagnosticCategoryName(unsigned Category); + +/** + * Retrieve the diagnostic category text for a given diagnostic. + * + * \returns The text of the given diagnostic category. + */ +CINDEX_LINKAGE CXString clang_getDiagnosticCategoryText(CXDiagnostic); + +/** + * Determine the number of source ranges associated with the given + * diagnostic. + */ +CINDEX_LINKAGE unsigned clang_getDiagnosticNumRanges(CXDiagnostic); + +/** + * Retrieve a source range associated with the diagnostic. + * + * A diagnostic's source ranges highlight important elements in the source + * code. On the command line, Clang displays source ranges by + * underlining them with '~' characters. + * + * \param Diagnostic the diagnostic whose range is being extracted. + * + * \param Range the zero-based index specifying which range to + * + * \returns the requested source range. + */ +CINDEX_LINKAGE CXSourceRange clang_getDiagnosticRange(CXDiagnostic Diagnostic, + unsigned Range); + +/** + * Determine the number of fix-it hints associated with the + * given diagnostic. + */ +CINDEX_LINKAGE unsigned clang_getDiagnosticNumFixIts(CXDiagnostic Diagnostic); + +/** + * Retrieve the replacement information for a given fix-it. + * + * Fix-its are described in terms of a source range whose contents + * should be replaced by a string. This approach generalizes over + * three kinds of operations: removal of source code (the range covers + * the code to be removed and the replacement string is empty), + * replacement of source code (the range covers the code to be + * replaced and the replacement string provides the new code), and + * insertion (both the start and end of the range point at the + * insertion location, and the replacement string provides the text to + * insert). + * + * \param Diagnostic The diagnostic whose fix-its are being queried. + * + * \param FixIt The zero-based index of the fix-it. + * + * \param ReplacementRange The source range whose contents will be + * replaced with the returned replacement string. Note that source + * ranges are half-open ranges [a, b), so the source code should be + * replaced from a and up to (but not including) b. + * + * \returns A string containing text that should be replace the source + * code indicated by the \c ReplacementRange. + */ +CINDEX_LINKAGE CXString clang_getDiagnosticFixIt( + CXDiagnostic Diagnostic, unsigned FixIt, CXSourceRange *ReplacementRange); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXErrorCode.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXErrorCode.h new file mode 100644 index 0000000..b3a0b9d --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXErrorCode.h @@ -0,0 +1,62 @@ +/*===-- clang-c/CXErrorCode.h - C Index Error Codes --------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides the CXErrorCode enumerators. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_CXERRORCODE_H +#define LLVM_CLANG_C_CXERRORCODE_H + +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * Error codes returned by libclang routines. + * + * Zero (\c CXError_Success) is the only error code indicating success. Other + * error codes, including not yet assigned non-zero values, indicate errors. + */ +enum CXErrorCode { + /** + * No error. + */ + CXError_Success = 0, + + /** + * A generic error code, no further details are available. + * + * Errors of this kind can get their own specific error codes in future + * libclang versions. + */ + CXError_Failure = 1, + + /** + * libclang crashed while performing the requested operation. + */ + CXError_Crashed = 2, + + /** + * The function detected that the arguments violate the function + * contract. + */ + CXError_InvalidArguments = 3, + + /** + * An AST deserialization error has occurred. + */ + CXError_ASTReadError = 4 +}; + +LLVM_CLANG_C_EXTERN_C_END + +#endif + diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXFile.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXFile.h new file mode 100644 index 0000000..c48f58c --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXFile.h @@ -0,0 +1,83 @@ +/*===-- clang-c/CXFile.h - C Index File ---------------------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides the interface to C Index files. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_CXFILE_H +#define LLVM_CLANG_C_CXFILE_H + +#include + +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * \defgroup CINDEX_FILES File manipulation routines + * + * @{ + */ + +/** + * A particular source file that is part of a translation unit. + */ +typedef void *CXFile; + +/** + * Retrieve the complete file and path name of the given file. + */ +CINDEX_LINKAGE CXString clang_getFileName(CXFile SFile); + +/** + * Retrieve the last modification time of the given file. + */ +CINDEX_LINKAGE time_t clang_getFileTime(CXFile SFile); + +/** + * Uniquely identifies a CXFile, that refers to the same underlying file, + * across an indexing session. + */ +typedef struct { + unsigned long long data[3]; +} CXFileUniqueID; + +/** + * Retrieve the unique ID for the given \c file. + * + * \param file the file to get the ID for. + * \param outID stores the returned CXFileUniqueID. + * \returns If there was a failure getting the unique ID, returns non-zero, + * otherwise returns 0. + */ +CINDEX_LINKAGE int clang_getFileUniqueID(CXFile file, CXFileUniqueID *outID); + +/** + * Returns non-zero if the \c file1 and \c file2 point to the same file, + * or they are both NULL. + */ +CINDEX_LINKAGE int clang_File_isEqual(CXFile file1, CXFile file2); + +/** + * Returns the real path name of \c file. + * + * An empty string may be returned. Use \c clang_getFileName() in that case. + */ +CINDEX_LINKAGE CXString clang_File_tryGetRealPathName(CXFile file); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXSourceLocation.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXSourceLocation.h new file mode 100644 index 0000000..4218021 --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXSourceLocation.h @@ -0,0 +1,296 @@ +/*===-- clang-c/CXSourceLocation.h - C Index Source Location ------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides the interface to C Index source locations. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_CXSOURCE_LOCATION_H +#define LLVM_CLANG_C_CXSOURCE_LOCATION_H + +#include "clang-c/CXFile.h" +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * \defgroup CINDEX_LOCATIONS Physical source locations + * + * Clang represents physical source locations in its abstract syntax tree in + * great detail, with file, line, and column information for the majority of + * the tokens parsed in the source code. These data types and functions are + * used to represent source location information, either for a particular + * point in the program or for a range of points in the program, and extract + * specific location information from those data types. + * + * @{ + */ + +/** + * Identifies a specific source location within a translation + * unit. + * + * Use clang_getExpansionLocation() or clang_getSpellingLocation() + * to map a source location to a particular file, line, and column. + */ +typedef struct { + const void *ptr_data[2]; + unsigned int_data; +} CXSourceLocation; + +/** + * Identifies a half-open character range in the source code. + * + * Use clang_getRangeStart() and clang_getRangeEnd() to retrieve the + * starting and end locations from a source range, respectively. + */ +typedef struct { + const void *ptr_data[2]; + unsigned begin_int_data; + unsigned end_int_data; +} CXSourceRange; + +/** + * Retrieve a NULL (invalid) source location. + */ +CINDEX_LINKAGE CXSourceLocation clang_getNullLocation(void); + +/** + * Determine whether two source locations, which must refer into + * the same translation unit, refer to exactly the same point in the source + * code. + * + * \returns non-zero if the source locations refer to the same location, zero + * if they refer to different locations. + */ +CINDEX_LINKAGE unsigned clang_equalLocations(CXSourceLocation loc1, + CXSourceLocation loc2); + +/** + * Determine for two source locations if the first comes + * strictly before the second one in the source code. + * + * \returns non-zero if the first source location comes + * strictly before the second one, zero otherwise. + */ +CINDEX_LINKAGE unsigned clang_isBeforeInTranslationUnit(CXSourceLocation loc1, + CXSourceLocation loc2); + +/** + * Returns non-zero if the given source location is in a system header. + */ +CINDEX_LINKAGE int clang_Location_isInSystemHeader(CXSourceLocation location); + +/** + * Returns non-zero if the given source location is in the main file of + * the corresponding translation unit. + */ +CINDEX_LINKAGE int clang_Location_isFromMainFile(CXSourceLocation location); + +/** + * Retrieve a NULL (invalid) source range. + */ +CINDEX_LINKAGE CXSourceRange clang_getNullRange(void); + +/** + * Retrieve a source range given the beginning and ending source + * locations. + */ +CINDEX_LINKAGE CXSourceRange clang_getRange(CXSourceLocation begin, + CXSourceLocation end); + +/** + * Determine whether two ranges are equivalent. + * + * \returns non-zero if the ranges are the same, zero if they differ. + */ +CINDEX_LINKAGE unsigned clang_equalRanges(CXSourceRange range1, + CXSourceRange range2); + +/** + * Returns non-zero if \p range is null. + */ +CINDEX_LINKAGE int clang_Range_isNull(CXSourceRange range); + +/** + * Retrieve the file, line, column, and offset represented by + * the given source location. + * + * If the location refers into a macro expansion, retrieves the + * location of the macro expansion. + * + * \param location the location within a source file that will be decomposed + * into its parts. + * + * \param file [out] if non-NULL, will be set to the file to which the given + * source location points. + * + * \param line [out] if non-NULL, will be set to the line to which the given + * source location points. + * + * \param column [out] if non-NULL, will be set to the column to which the given + * source location points. + * + * \param offset [out] if non-NULL, will be set to the offset into the + * buffer to which the given source location points. + */ +CINDEX_LINKAGE void clang_getExpansionLocation(CXSourceLocation location, + CXFile *file, unsigned *line, + unsigned *column, + unsigned *offset); + +/** + * Retrieve the file, line and column represented by the given source + * location, as specified in a # line directive. + * + * Example: given the following source code in a file somefile.c + * + * \code + * #123 "dummy.c" 1 + * + * static int func(void) + * { + * return 0; + * } + * \endcode + * + * the location information returned by this function would be + * + * File: dummy.c Line: 124 Column: 12 + * + * whereas clang_getExpansionLocation would have returned + * + * File: somefile.c Line: 3 Column: 12 + * + * \param location the location within a source file that will be decomposed + * into its parts. + * + * \param filename [out] if non-NULL, will be set to the filename of the + * source location. Note that filenames returned will be for "virtual" files, + * which don't necessarily exist on the machine running clang - e.g. when + * parsing preprocessed output obtained from a different environment. If + * a non-NULL value is passed in, remember to dispose of the returned value + * using \c clang_disposeString() once you've finished with it. For an invalid + * source location, an empty string is returned. + * + * \param line [out] if non-NULL, will be set to the line number of the + * source location. For an invalid source location, zero is returned. + * + * \param column [out] if non-NULL, will be set to the column number of the + * source location. For an invalid source location, zero is returned. + */ +CINDEX_LINKAGE void clang_getPresumedLocation(CXSourceLocation location, + CXString *filename, + unsigned *line, unsigned *column); + +/** + * Legacy API to retrieve the file, line, column, and offset represented + * by the given source location. + * + * This interface has been replaced by the newer interface + * #clang_getExpansionLocation(). See that interface's documentation for + * details. + */ +CINDEX_LINKAGE void clang_getInstantiationLocation(CXSourceLocation location, + CXFile *file, unsigned *line, + unsigned *column, + unsigned *offset); + +/** + * Retrieve the file, line, column, and offset represented by + * the given source location. + * + * If the location refers into a macro instantiation, return where the + * location was originally spelled in the source file. + * + * \param location the location within a source file that will be decomposed + * into its parts. + * + * \param file [out] if non-NULL, will be set to the file to which the given + * source location points. + * + * \param line [out] if non-NULL, will be set to the line to which the given + * source location points. + * + * \param column [out] if non-NULL, will be set to the column to which the given + * source location points. + * + * \param offset [out] if non-NULL, will be set to the offset into the + * buffer to which the given source location points. + */ +CINDEX_LINKAGE void clang_getSpellingLocation(CXSourceLocation location, + CXFile *file, unsigned *line, + unsigned *column, + unsigned *offset); + +/** + * Retrieve the file, line, column, and offset represented by + * the given source location. + * + * If the location refers into a macro expansion, return where the macro was + * expanded or where the macro argument was written, if the location points at + * a macro argument. + * + * \param location the location within a source file that will be decomposed + * into its parts. + * + * \param file [out] if non-NULL, will be set to the file to which the given + * source location points. + * + * \param line [out] if non-NULL, will be set to the line to which the given + * source location points. + * + * \param column [out] if non-NULL, will be set to the column to which the given + * source location points. + * + * \param offset [out] if non-NULL, will be set to the offset into the + * buffer to which the given source location points. + */ +CINDEX_LINKAGE void clang_getFileLocation(CXSourceLocation location, + CXFile *file, unsigned *line, + unsigned *column, unsigned *offset); + +/** + * Retrieve a source location representing the first character within a + * source range. + */ +CINDEX_LINKAGE CXSourceLocation clang_getRangeStart(CXSourceRange range); + +/** + * Retrieve a source location representing the last character within a + * source range. + */ +CINDEX_LINKAGE CXSourceLocation clang_getRangeEnd(CXSourceRange range); + +/** + * Identifies an array of ranges. + */ +typedef struct { + /** The number of ranges in the \c ranges array. */ + unsigned count; + /** + * An array of \c CXSourceRanges. + */ + CXSourceRange *ranges; +} CXSourceRangeList; + +/** + * Destroy the given \c CXSourceRangeList. + */ +CINDEX_LINKAGE void clang_disposeSourceRangeList(CXSourceRangeList *ranges); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXString.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXString.h new file mode 100644 index 0000000..63dce4d --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/CXString.h @@ -0,0 +1,73 @@ +/*===-- clang-c/CXString.h - C Index strings --------------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides the interface to C Index strings. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_CXSTRING_H +#define LLVM_CLANG_C_CXSTRING_H + +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * \defgroup CINDEX_STRING String manipulation routines + * \ingroup CINDEX + * + * @{ + */ + +/** + * A character string. + * + * The \c CXString type is used to return strings from the interface when + * the ownership of that string might differ from one call to the next. + * Use \c clang_getCString() to retrieve the string data and, once finished + * with the string data, call \c clang_disposeString() to free the string. + */ +typedef struct { + const void *data; + unsigned private_flags; +} CXString; + +typedef struct { + CXString *Strings; + unsigned Count; +} CXStringSet; + +/** + * Retrieve the character data associated with the given string. + * + * The returned data is a reference and not owned by the user. This data + * is only valid while the `CXString` is valid. This function is similar + * to `std::string::c_str()`. + */ +CINDEX_LINKAGE const char *clang_getCString(CXString string); + +/** + * Free the given string. + */ +CINDEX_LINKAGE void clang_disposeString(CXString string); + +/** + * Free the given string set. + */ +CINDEX_LINKAGE void clang_disposeStringSet(CXStringSet *set); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif + diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Documentation.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Documentation.h new file mode 100644 index 0000000..e04c50a --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Documentation.h @@ -0,0 +1,619 @@ +/*==-- clang-c/Documentation.h - Utilities for comment processing -*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides a supplementary interface for inspecting *| +|* documentation comments. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_DOCUMENTATION_H +#define LLVM_CLANG_C_DOCUMENTATION_H + +#include "clang-c/CXErrorCode.h" +#include "clang-c/ExternC.h" +#include "clang-c/Index.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * \defgroup CINDEX_COMMENT Comment introspection + * + * The routines in this group provide access to information in documentation + * comments. These facilities are distinct from the core and may be subject to + * their own schedule of stability and deprecation. + * + * @{ + */ + +/** + * A parsed comment. + */ +typedef struct { + const void *ASTNode; + CXTranslationUnit TranslationUnit; +} CXComment; + +/** + * Given a cursor that represents a documentable entity (e.g., + * declaration), return the associated parsed comment as a + * \c CXComment_FullComment AST node. + */ +CINDEX_LINKAGE CXComment clang_Cursor_getParsedComment(CXCursor C); + +/** + * Describes the type of the comment AST node (\c CXComment). A comment + * node can be considered block content (e. g., paragraph), inline content + * (plain text) or neither (the root AST node). + */ +enum CXCommentKind { + /** + * Null comment. No AST node is constructed at the requested location + * because there is no text or a syntax error. + */ + CXComment_Null = 0, + + /** + * Plain text. Inline content. + */ + CXComment_Text = 1, + + /** + * A command with word-like arguments that is considered inline content. + * + * For example: \\c command. + */ + CXComment_InlineCommand = 2, + + /** + * HTML start tag with attributes (name-value pairs). Considered + * inline content. + * + * For example: + * \verbatim + *

+ * \endverbatim + */ + CXComment_HTMLStartTag = 3, + + /** + * HTML end tag. Considered inline content. + * + * For example: + * \verbatim + * + * \endverbatim + */ + CXComment_HTMLEndTag = 4, + + /** + * A paragraph, contains inline comment. The paragraph itself is + * block content. + */ + CXComment_Paragraph = 5, + + /** + * A command that has zero or more word-like arguments (number of + * word-like arguments depends on command name) and a paragraph as an + * argument. Block command is block content. + * + * Paragraph argument is also a child of the block command. + * + * For example: \has 0 word-like arguments and a paragraph argument. + * + * AST nodes of special kinds that parser knows about (e. g., \\param + * command) have their own node kinds. + */ + CXComment_BlockCommand = 6, + + /** + * A \\param or \\arg command that describes the function parameter + * (name, passing direction, description). + * + * For example: \\param [in] ParamName description. + */ + CXComment_ParamCommand = 7, + + /** + * A \\tparam command that describes a template parameter (name and + * description). + * + * For example: \\tparam T description. + */ + CXComment_TParamCommand = 8, + + /** + * A verbatim block command (e. g., preformatted code). Verbatim + * block has an opening and a closing command and contains multiple lines of + * text (\c CXComment_VerbatimBlockLine child nodes). + * + * For example: + * \\verbatim + * aaa + * \\endverbatim + */ + CXComment_VerbatimBlockCommand = 9, + + /** + * A line of text that is contained within a + * CXComment_VerbatimBlockCommand node. + */ + CXComment_VerbatimBlockLine = 10, + + /** + * A verbatim line command. Verbatim line has an opening command, + * a single line of text (up to the newline after the opening command) and + * has no closing command. + */ + CXComment_VerbatimLine = 11, + + /** + * A full comment attached to a declaration, contains block content. + */ + CXComment_FullComment = 12 +}; + +/** + * The most appropriate rendering mode for an inline command, chosen on + * command semantics in Doxygen. + */ +enum CXCommentInlineCommandRenderKind { + /** + * Command argument should be rendered in a normal font. + */ + CXCommentInlineCommandRenderKind_Normal, + + /** + * Command argument should be rendered in a bold font. + */ + CXCommentInlineCommandRenderKind_Bold, + + /** + * Command argument should be rendered in a monospaced font. + */ + CXCommentInlineCommandRenderKind_Monospaced, + + /** + * Command argument should be rendered emphasized (typically italic + * font). + */ + CXCommentInlineCommandRenderKind_Emphasized, + + /** + * Command argument should not be rendered (since it only defines an anchor). + */ + CXCommentInlineCommandRenderKind_Anchor +}; + +/** + * Describes parameter passing direction for \\param or \\arg command. + */ +enum CXCommentParamPassDirection { + /** + * The parameter is an input parameter. + */ + CXCommentParamPassDirection_In, + + /** + * The parameter is an output parameter. + */ + CXCommentParamPassDirection_Out, + + /** + * The parameter is an input and output parameter. + */ + CXCommentParamPassDirection_InOut +}; + +/** + * \param Comment AST node of any kind. + * + * \returns the type of the AST node. + */ +CINDEX_LINKAGE enum CXCommentKind clang_Comment_getKind(CXComment Comment); + +/** + * \param Comment AST node of any kind. + * + * \returns number of children of the AST node. + */ +CINDEX_LINKAGE unsigned clang_Comment_getNumChildren(CXComment Comment); + +/** + * \param Comment AST node of any kind. + * + * \param ChildIdx child index (zero-based). + * + * \returns the specified child of the AST node. + */ +CINDEX_LINKAGE +CXComment clang_Comment_getChild(CXComment Comment, unsigned ChildIdx); + +/** + * A \c CXComment_Paragraph node is considered whitespace if it contains + * only \c CXComment_Text nodes that are empty or whitespace. + * + * Other AST nodes (except \c CXComment_Paragraph and \c CXComment_Text) are + * never considered whitespace. + * + * \returns non-zero if \c Comment is whitespace. + */ +CINDEX_LINKAGE unsigned clang_Comment_isWhitespace(CXComment Comment); + +/** + * \returns non-zero if \c Comment is inline content and has a newline + * immediately following it in the comment text. Newlines between paragraphs + * do not count. + */ +CINDEX_LINKAGE +unsigned clang_InlineContentComment_hasTrailingNewline(CXComment Comment); + +/** + * \param Comment a \c CXComment_Text AST node. + * + * \returns text contained in the AST node. + */ +CINDEX_LINKAGE CXString clang_TextComment_getText(CXComment Comment); + +/** + * \param Comment a \c CXComment_InlineCommand AST node. + * + * \returns name of the inline command. + */ +CINDEX_LINKAGE +CXString clang_InlineCommandComment_getCommandName(CXComment Comment); + +/** + * \param Comment a \c CXComment_InlineCommand AST node. + * + * \returns the most appropriate rendering mode, chosen on command + * semantics in Doxygen. + */ +CINDEX_LINKAGE enum CXCommentInlineCommandRenderKind +clang_InlineCommandComment_getRenderKind(CXComment Comment); + +/** + * \param Comment a \c CXComment_InlineCommand AST node. + * + * \returns number of command arguments. + */ +CINDEX_LINKAGE +unsigned clang_InlineCommandComment_getNumArgs(CXComment Comment); + +/** + * \param Comment a \c CXComment_InlineCommand AST node. + * + * \param ArgIdx argument index (zero-based). + * + * \returns text of the specified argument. + */ +CINDEX_LINKAGE +CXString clang_InlineCommandComment_getArgText(CXComment Comment, + unsigned ArgIdx); + +/** + * \param Comment a \c CXComment_HTMLStartTag or \c CXComment_HTMLEndTag AST + * node. + * + * \returns HTML tag name. + */ +CINDEX_LINKAGE CXString clang_HTMLTagComment_getTagName(CXComment Comment); + +/** + * \param Comment a \c CXComment_HTMLStartTag AST node. + * + * \returns non-zero if tag is self-closing (for example, <br />). + */ +CINDEX_LINKAGE +unsigned clang_HTMLStartTagComment_isSelfClosing(CXComment Comment); + +/** + * \param Comment a \c CXComment_HTMLStartTag AST node. + * + * \returns number of attributes (name-value pairs) attached to the start tag. + */ +CINDEX_LINKAGE unsigned clang_HTMLStartTag_getNumAttrs(CXComment Comment); + +/** + * \param Comment a \c CXComment_HTMLStartTag AST node. + * + * \param AttrIdx attribute index (zero-based). + * + * \returns name of the specified attribute. + */ +CINDEX_LINKAGE +CXString clang_HTMLStartTag_getAttrName(CXComment Comment, unsigned AttrIdx); + +/** + * \param Comment a \c CXComment_HTMLStartTag AST node. + * + * \param AttrIdx attribute index (zero-based). + * + * \returns value of the specified attribute. + */ +CINDEX_LINKAGE +CXString clang_HTMLStartTag_getAttrValue(CXComment Comment, unsigned AttrIdx); + +/** + * \param Comment a \c CXComment_BlockCommand AST node. + * + * \returns name of the block command. + */ +CINDEX_LINKAGE +CXString clang_BlockCommandComment_getCommandName(CXComment Comment); + +/** + * \param Comment a \c CXComment_BlockCommand AST node. + * + * \returns number of word-like arguments. + */ +CINDEX_LINKAGE +unsigned clang_BlockCommandComment_getNumArgs(CXComment Comment); + +/** + * \param Comment a \c CXComment_BlockCommand AST node. + * + * \param ArgIdx argument index (zero-based). + * + * \returns text of the specified word-like argument. + */ +CINDEX_LINKAGE +CXString clang_BlockCommandComment_getArgText(CXComment Comment, + unsigned ArgIdx); + +/** + * \param Comment a \c CXComment_BlockCommand or + * \c CXComment_VerbatimBlockCommand AST node. + * + * \returns paragraph argument of the block command. + */ +CINDEX_LINKAGE +CXComment clang_BlockCommandComment_getParagraph(CXComment Comment); + +/** + * \param Comment a \c CXComment_ParamCommand AST node. + * + * \returns parameter name. + */ +CINDEX_LINKAGE +CXString clang_ParamCommandComment_getParamName(CXComment Comment); + +/** + * \param Comment a \c CXComment_ParamCommand AST node. + * + * \returns non-zero if the parameter that this AST node represents was found + * in the function prototype and \c clang_ParamCommandComment_getParamIndex + * function will return a meaningful value. + */ +CINDEX_LINKAGE +unsigned clang_ParamCommandComment_isParamIndexValid(CXComment Comment); + +/** + * \param Comment a \c CXComment_ParamCommand AST node. + * + * \returns zero-based parameter index in function prototype. + */ +CINDEX_LINKAGE +unsigned clang_ParamCommandComment_getParamIndex(CXComment Comment); + +/** + * \param Comment a \c CXComment_ParamCommand AST node. + * + * \returns non-zero if parameter passing direction was specified explicitly in + * the comment. + */ +CINDEX_LINKAGE +unsigned clang_ParamCommandComment_isDirectionExplicit(CXComment Comment); + +/** + * \param Comment a \c CXComment_ParamCommand AST node. + * + * \returns parameter passing direction. + */ +CINDEX_LINKAGE +enum CXCommentParamPassDirection clang_ParamCommandComment_getDirection( + CXComment Comment); + +/** + * \param Comment a \c CXComment_TParamCommand AST node. + * + * \returns template parameter name. + */ +CINDEX_LINKAGE +CXString clang_TParamCommandComment_getParamName(CXComment Comment); + +/** + * \param Comment a \c CXComment_TParamCommand AST node. + * + * \returns non-zero if the parameter that this AST node represents was found + * in the template parameter list and + * \c clang_TParamCommandComment_getDepth and + * \c clang_TParamCommandComment_getIndex functions will return a meaningful + * value. + */ +CINDEX_LINKAGE +unsigned clang_TParamCommandComment_isParamPositionValid(CXComment Comment); + +/** + * \param Comment a \c CXComment_TParamCommand AST node. + * + * \returns zero-based nesting depth of this parameter in the template parameter list. + * + * For example, + * \verbatim + * template class TT> + * void test(TT aaa); + * \endverbatim + * for C and TT nesting depth is 0, + * for T nesting depth is 1. + */ +CINDEX_LINKAGE +unsigned clang_TParamCommandComment_getDepth(CXComment Comment); + +/** + * \param Comment a \c CXComment_TParamCommand AST node. + * + * \returns zero-based parameter index in the template parameter list at a + * given nesting depth. + * + * For example, + * \verbatim + * template class TT> + * void test(TT aaa); + * \endverbatim + * for C and TT nesting depth is 0, so we can ask for index at depth 0: + * at depth 0 C's index is 0, TT's index is 1. + * + * For T nesting depth is 1, so we can ask for index at depth 0 and 1: + * at depth 0 T's index is 1 (same as TT's), + * at depth 1 T's index is 0. + */ +CINDEX_LINKAGE +unsigned clang_TParamCommandComment_getIndex(CXComment Comment, unsigned Depth); + +/** + * \param Comment a \c CXComment_VerbatimBlockLine AST node. + * + * \returns text contained in the AST node. + */ +CINDEX_LINKAGE +CXString clang_VerbatimBlockLineComment_getText(CXComment Comment); + +/** + * \param Comment a \c CXComment_VerbatimLine AST node. + * + * \returns text contained in the AST node. + */ +CINDEX_LINKAGE CXString clang_VerbatimLineComment_getText(CXComment Comment); + +/** + * Convert an HTML tag AST node to string. + * + * \param Comment a \c CXComment_HTMLStartTag or \c CXComment_HTMLEndTag AST + * node. + * + * \returns string containing an HTML tag. + */ +CINDEX_LINKAGE CXString clang_HTMLTagComment_getAsString(CXComment Comment); + +/** + * Convert a given full parsed comment to an HTML fragment. + * + * Specific details of HTML layout are subject to change. Don't try to parse + * this HTML back into an AST, use other APIs instead. + * + * Currently the following CSS classes are used: + * \li "para-brief" for \paragraph and equivalent commands; + * \li "para-returns" for \\returns paragraph and equivalent commands; + * \li "word-returns" for the "Returns" word in \\returns paragraph. + * + * Function argument documentation is rendered as a \ list with arguments + * sorted in function prototype order. CSS classes used: + * \li "param-name-index-NUMBER" for parameter name (\); + * \li "param-descr-index-NUMBER" for parameter description (\); + * \li "param-name-index-invalid" and "param-descr-index-invalid" are used if + * parameter index is invalid. + * + * Template parameter documentation is rendered as a \ list with + * parameters sorted in template parameter list order. CSS classes used: + * \li "tparam-name-index-NUMBER" for parameter name (\); + * \li "tparam-descr-index-NUMBER" for parameter description (\); + * \li "tparam-name-index-other" and "tparam-descr-index-other" are used for + * names inside template template parameters; + * \li "tparam-name-index-invalid" and "tparam-descr-index-invalid" are used if + * parameter position is invalid. + * + * \param Comment a \c CXComment_FullComment AST node. + * + * \returns string containing an HTML fragment. + */ +CINDEX_LINKAGE CXString clang_FullComment_getAsHTML(CXComment Comment); + +/** + * Convert a given full parsed comment to an XML document. + * + * A Relax NG schema for the XML can be found in comment-xml-schema.rng file + * inside clang source tree. + * + * \param Comment a \c CXComment_FullComment AST node. + * + * \returns string containing an XML document. + */ +CINDEX_LINKAGE CXString clang_FullComment_getAsXML(CXComment Comment); + +/** + * CXAPISet is an opaque type that represents a data structure containing all + * the API information for a given translation unit. This can be used for a + * single symbol symbol graph for a given symbol. + */ +typedef struct CXAPISetImpl *CXAPISet; + +/** + * Traverses the translation unit to create a \c CXAPISet. + * + * \param tu is the \c CXTranslationUnit to build the \c CXAPISet for. + * + * \param out_api is a pointer to the output of this function. It is needs to be + * disposed of by calling clang_disposeAPISet. + * + * \returns Error code indicating success or failure of the APISet creation. + */ +CINDEX_LINKAGE enum CXErrorCode clang_createAPISet(CXTranslationUnit tu, + CXAPISet *out_api); + +/** + * Dispose of an APISet. + * + * The provided \c CXAPISet can not be used after this function is called. + */ +CINDEX_LINKAGE void clang_disposeAPISet(CXAPISet api); + +/** + * Generate a single symbol symbol graph for the given USR. Returns a null + * string if the associated symbol can not be found in the provided \c CXAPISet. + * + * The output contains the symbol graph as well as some additional information + * about related symbols. + * + * \param usr is a string containing the USR of the symbol to generate the + * symbol graph for. + * + * \param api the \c CXAPISet to look for the symbol in. + * + * \returns a string containing the serialized symbol graph representation for + * the symbol being queried or a null string if it can not be found in the + * APISet. + */ +CINDEX_LINKAGE CXString clang_getSymbolGraphForUSR(const char *usr, + CXAPISet api); + +/** + * Generate a single symbol symbol graph for the declaration at the given + * cursor. Returns a null string if the AST node for the cursor isn't a + * declaration. + * + * The output contains the symbol graph as well as some additional information + * about related symbols. + * + * \param cursor the declaration for which to generate the single symbol symbol + * graph. + * + * \returns a string containing the serialized symbol graph representation for + * the symbol being queried or a null string if it can not be found in the + * APISet. + */ +CINDEX_LINKAGE CXString clang_getSymbolGraphForCursor(CXCursor cursor); + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif /* CLANG_C_DOCUMENTATION_H */ + diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/ExternC.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/ExternC.h new file mode 100644 index 0000000..384f24d --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/ExternC.h @@ -0,0 +1,39 @@ +/*===- clang-c/ExternC.h - Wrapper for 'extern "C"' ---------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This file defines an 'extern "C"' wrapper. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_EXTERN_C_H +#define LLVM_CLANG_C_EXTERN_C_H + +#ifdef __clang__ +#define LLVM_CLANG_C_STRICT_PROTOTYPES_BEGIN \ + _Pragma("clang diagnostic push") \ + _Pragma("clang diagnostic error \"-Wstrict-prototypes\"") +#define LLVM_CLANG_C_STRICT_PROTOTYPES_END _Pragma("clang diagnostic pop") +#else +#define LLVM_CLANG_C_STRICT_PROTOTYPES_BEGIN +#define LLVM_CLANG_C_STRICT_PROTOTYPES_END +#endif + +#ifdef __cplusplus +#define LLVM_CLANG_C_EXTERN_C_BEGIN \ + extern "C" { \ + LLVM_CLANG_C_STRICT_PROTOTYPES_BEGIN +#define LLVM_CLANG_C_EXTERN_C_END \ + LLVM_CLANG_C_STRICT_PROTOTYPES_END \ + } +#else +#define LLVM_CLANG_C_EXTERN_C_BEGIN LLVM_CLANG_C_STRICT_PROTOTYPES_BEGIN +#define LLVM_CLANG_C_EXTERN_C_END LLVM_CLANG_C_STRICT_PROTOTYPES_END +#endif + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/FatalErrorHandler.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/FatalErrorHandler.h new file mode 100644 index 0000000..22f34fa --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/FatalErrorHandler.h @@ -0,0 +1,32 @@ +/*===-- clang-c/FatalErrorHandler.h - Fatal Error Handling --------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_FATAL_ERROR_HANDLER_H +#define LLVM_CLANG_C_FATAL_ERROR_HANDLER_H + +#include "clang-c/ExternC.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** + * Installs error handler that prints error message to stderr and calls abort(). + * Replaces currently installed error handler (if any). + */ +void clang_install_aborting_llvm_fatal_error_handler(void); + +/** + * Removes currently installed error handler (if any). + * If no error handler is intalled, the default strategy is to print error + * message to stderr and call exit(1). + */ +void clang_uninstall_llvm_fatal_error_handler(void); + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Index.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Index.h new file mode 100644 index 0000000..aac5d1f --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Index.h @@ -0,0 +1,6831 @@ +/*===-- clang-c/Index.h - Indexing Public C Interface -------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides a public interface to a Clang library for extracting *| +|* high-level symbol information from source files without exposing the full *| +|* Clang C++ API. *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_INDEX_H +#define LLVM_CLANG_C_INDEX_H + +#include "clang-c/BuildSystem.h" +#include "clang-c/CXDiagnostic.h" +#include "clang-c/CXErrorCode.h" +#include "clang-c/CXFile.h" +#include "clang-c/CXSourceLocation.h" +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Platform.h" + +/** + * The version constants for the libclang API. + * CINDEX_VERSION_MINOR should increase when there are API additions. + * CINDEX_VERSION_MAJOR is intended for "major" source/ABI breaking changes. + * + * The policy about the libclang API was always to keep it source and ABI + * compatible, thus CINDEX_VERSION_MAJOR is expected to remain stable. + */ +#define CINDEX_VERSION_MAJOR 0 +#define CINDEX_VERSION_MINOR 64 + +#define CINDEX_VERSION_ENCODE(major, minor) (((major)*10000) + ((minor)*1)) + +#define CINDEX_VERSION \ + CINDEX_VERSION_ENCODE(CINDEX_VERSION_MAJOR, CINDEX_VERSION_MINOR) + +#define CINDEX_VERSION_STRINGIZE_(major, minor) #major "." #minor +#define CINDEX_VERSION_STRINGIZE(major, minor) \ + CINDEX_VERSION_STRINGIZE_(major, minor) + +#define CINDEX_VERSION_STRING \ + CINDEX_VERSION_STRINGIZE(CINDEX_VERSION_MAJOR, CINDEX_VERSION_MINOR) + +#ifndef __has_feature +#define __has_feature(feature) 0 +#endif + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/** \defgroup CINDEX libclang: C Interface to Clang + * + * The C Interface to Clang provides a relatively small API that exposes + * facilities for parsing source code into an abstract syntax tree (AST), + * loading already-parsed ASTs, traversing the AST, associating + * physical source locations with elements within the AST, and other + * facilities that support Clang-based development tools. + * + * This C interface to Clang will never provide all of the information + * representation stored in Clang's C++ AST, nor should it: the intent is to + * maintain an API that is relatively stable from one release to the next, + * providing only the basic functionality needed to support development tools. + * + * To avoid namespace pollution, data types are prefixed with "CX" and + * functions are prefixed with "clang_". + * + * @{ + */ + +/** + * An "index" that consists of a set of translation units that would + * typically be linked together into an executable or library. + */ +typedef void *CXIndex; + +/** + * An opaque type representing target information for a given translation + * unit. + */ +typedef struct CXTargetInfoImpl *CXTargetInfo; + +/** + * A single translation unit, which resides in an index. + */ +typedef struct CXTranslationUnitImpl *CXTranslationUnit; + +/** + * Opaque pointer representing client data that will be passed through + * to various callbacks and visitors. + */ +typedef void *CXClientData; + +/** + * Provides the contents of a file that has not yet been saved to disk. + * + * Each CXUnsavedFile instance provides the name of a file on the + * system along with the current contents of that file that have not + * yet been saved to disk. + */ +struct CXUnsavedFile { + /** + * The file whose contents have not yet been saved. + * + * This file must already exist in the file system. + */ + const char *Filename; + + /** + * A buffer containing the unsaved contents of this file. + */ + const char *Contents; + + /** + * The length of the unsaved contents of this buffer. + */ + unsigned long Length; +}; + +/** + * Describes the availability of a particular entity, which indicates + * whether the use of this entity will result in a warning or error due to + * it being deprecated or unavailable. + */ +enum CXAvailabilityKind { + /** + * The entity is available. + */ + CXAvailability_Available, + /** + * The entity is available, but has been deprecated (and its use is + * not recommended). + */ + CXAvailability_Deprecated, + /** + * The entity is not available; any use of it will be an error. + */ + CXAvailability_NotAvailable, + /** + * The entity is available, but not accessible; any use of it will be + * an error. + */ + CXAvailability_NotAccessible +}; + +/** + * Describes a version number of the form major.minor.subminor. + */ +typedef struct CXVersion { + /** + * The major version number, e.g., the '10' in '10.7.3'. A negative + * value indicates that there is no version number at all. + */ + int Major; + /** + * The minor version number, e.g., the '7' in '10.7.3'. This value + * will be negative if no minor version number was provided, e.g., for + * version '10'. + */ + int Minor; + /** + * The subminor version number, e.g., the '3' in '10.7.3'. This value + * will be negative if no minor or subminor version number was provided, + * e.g., in version '10' or '10.7'. + */ + int Subminor; +} CXVersion; + +/** + * Describes the exception specification of a cursor. + * + * A negative value indicates that the cursor is not a function declaration. + */ +enum CXCursor_ExceptionSpecificationKind { + /** + * The cursor has no exception specification. + */ + CXCursor_ExceptionSpecificationKind_None, + + /** + * The cursor has exception specification throw() + */ + CXCursor_ExceptionSpecificationKind_DynamicNone, + + /** + * The cursor has exception specification throw(T1, T2) + */ + CXCursor_ExceptionSpecificationKind_Dynamic, + + /** + * The cursor has exception specification throw(...). + */ + CXCursor_ExceptionSpecificationKind_MSAny, + + /** + * The cursor has exception specification basic noexcept. + */ + CXCursor_ExceptionSpecificationKind_BasicNoexcept, + + /** + * The cursor has exception specification computed noexcept. + */ + CXCursor_ExceptionSpecificationKind_ComputedNoexcept, + + /** + * The exception specification has not yet been evaluated. + */ + CXCursor_ExceptionSpecificationKind_Unevaluated, + + /** + * The exception specification has not yet been instantiated. + */ + CXCursor_ExceptionSpecificationKind_Uninstantiated, + + /** + * The exception specification has not been parsed yet. + */ + CXCursor_ExceptionSpecificationKind_Unparsed, + + /** + * The cursor has a __declspec(nothrow) exception specification. + */ + CXCursor_ExceptionSpecificationKind_NoThrow +}; + +/** + * Provides a shared context for creating translation units. + * + * It provides two options: + * + * - excludeDeclarationsFromPCH: When non-zero, allows enumeration of "local" + * declarations (when loading any new translation units). A "local" declaration + * is one that belongs in the translation unit itself and not in a precompiled + * header that was used by the translation unit. If zero, all declarations + * will be enumerated. + * + * Here is an example: + * + * \code + * // excludeDeclsFromPCH = 1, displayDiagnostics=1 + * Idx = clang_createIndex(1, 1); + * + * // IndexTest.pch was produced with the following command: + * // "clang -x c IndexTest.h -emit-ast -o IndexTest.pch" + * TU = clang_createTranslationUnit(Idx, "IndexTest.pch"); + * + * // This will load all the symbols from 'IndexTest.pch' + * clang_visitChildren(clang_getTranslationUnitCursor(TU), + * TranslationUnitVisitor, 0); + * clang_disposeTranslationUnit(TU); + * + * // This will load all the symbols from 'IndexTest.c', excluding symbols + * // from 'IndexTest.pch'. + * char *args[] = { "-Xclang", "-include-pch=IndexTest.pch" }; + * TU = clang_createTranslationUnitFromSourceFile(Idx, "IndexTest.c", 2, args, + * 0, 0); + * clang_visitChildren(clang_getTranslationUnitCursor(TU), + * TranslationUnitVisitor, 0); + * clang_disposeTranslationUnit(TU); + * \endcode + * + * This process of creating the 'pch', loading it separately, and using it (via + * -include-pch) allows 'excludeDeclsFromPCH' to remove redundant callbacks + * (which gives the indexer the same performance benefit as the compiler). + */ +CINDEX_LINKAGE CXIndex clang_createIndex(int excludeDeclarationsFromPCH, + int displayDiagnostics); + +/** + * Destroy the given index. + * + * The index must not be destroyed until all of the translation units created + * within that index have been destroyed. + */ +CINDEX_LINKAGE void clang_disposeIndex(CXIndex index); + +typedef enum { + /** + * Use the default value of an option that may depend on the process + * environment. + */ + CXChoice_Default = 0, + /** + * Enable the option. + */ + CXChoice_Enabled = 1, + /** + * Disable the option. + */ + CXChoice_Disabled = 2 +} CXChoice; + +typedef enum { + /** + * Used to indicate that no special CXIndex options are needed. + */ + CXGlobalOpt_None = 0x0, + + /** + * Used to indicate that threads that libclang creates for indexing + * purposes should use background priority. + * + * Affects #clang_indexSourceFile, #clang_indexTranslationUnit, + * #clang_parseTranslationUnit, #clang_saveTranslationUnit. + */ + CXGlobalOpt_ThreadBackgroundPriorityForIndexing = 0x1, + + /** + * Used to indicate that threads that libclang creates for editing + * purposes should use background priority. + * + * Affects #clang_reparseTranslationUnit, #clang_codeCompleteAt, + * #clang_annotateTokens + */ + CXGlobalOpt_ThreadBackgroundPriorityForEditing = 0x2, + + /** + * Used to indicate that all threads that libclang creates should use + * background priority. + */ + CXGlobalOpt_ThreadBackgroundPriorityForAll = + CXGlobalOpt_ThreadBackgroundPriorityForIndexing | + CXGlobalOpt_ThreadBackgroundPriorityForEditing + +} CXGlobalOptFlags; + +/** + * Index initialization options. + * + * 0 is the default value of each member of this struct except for Size. + * Initialize the struct in one of the following three ways to avoid adapting + * code each time a new member is added to it: + * \code + * CXIndexOptions Opts; + * memset(&Opts, 0, sizeof(Opts)); + * Opts.Size = sizeof(CXIndexOptions); + * \endcode + * or explicitly initialize the first data member and zero-initialize the rest: + * \code + * CXIndexOptions Opts = { sizeof(CXIndexOptions) }; + * \endcode + * or to prevent the -Wmissing-field-initializers warning for the above version: + * \code + * CXIndexOptions Opts{}; + * Opts.Size = sizeof(CXIndexOptions); + * \endcode + */ +typedef struct CXIndexOptions { + /** + * The size of struct CXIndexOptions used for option versioning. + * + * Always initialize this member to sizeof(CXIndexOptions), or assign + * sizeof(CXIndexOptions) to it right after creating a CXIndexOptions object. + */ + unsigned Size; + /** + * A CXChoice enumerator that specifies the indexing priority policy. + * \sa CXGlobalOpt_ThreadBackgroundPriorityForIndexing + */ + unsigned char ThreadBackgroundPriorityForIndexing; + /** + * A CXChoice enumerator that specifies the editing priority policy. + * \sa CXGlobalOpt_ThreadBackgroundPriorityForEditing + */ + unsigned char ThreadBackgroundPriorityForEditing; + /** + * \see clang_createIndex() + */ + unsigned ExcludeDeclarationsFromPCH : 1; + /** + * \see clang_createIndex() + */ + unsigned DisplayDiagnostics : 1; + /** + * Store PCH in memory. If zero, PCH are stored in temporary files. + */ + unsigned StorePreamblesInMemory : 1; + unsigned /*Reserved*/ : 13; + + /** + * The path to a directory, in which to store temporary PCH files. If null or + * empty, the default system temporary directory is used. These PCH files are + * deleted on clean exit but stay on disk if the program crashes or is killed. + * + * This option is ignored if \a StorePreamblesInMemory is non-zero. + * + * Libclang does not create the directory at the specified path in the file + * system. Therefore it must exist, or storing PCH files will fail. + */ + const char *PreambleStoragePath; + /** + * Specifies a path which will contain log files for certain libclang + * invocations. A null value implies that libclang invocations are not logged. + */ + const char *InvocationEmissionPath; +} CXIndexOptions; + +/** + * Provides a shared context for creating translation units. + * + * Call this function instead of clang_createIndex() if you need to configure + * the additional options in CXIndexOptions. + * + * \returns The created index or null in case of error, such as an unsupported + * value of options->Size. + * + * For example: + * \code + * CXIndex createIndex(const char *ApplicationTemporaryPath) { + * const int ExcludeDeclarationsFromPCH = 1; + * const int DisplayDiagnostics = 1; + * CXIndex Idx; + * #if CINDEX_VERSION_MINOR >= 64 + * CXIndexOptions Opts; + * memset(&Opts, 0, sizeof(Opts)); + * Opts.Size = sizeof(CXIndexOptions); + * Opts.ThreadBackgroundPriorityForIndexing = 1; + * Opts.ExcludeDeclarationsFromPCH = ExcludeDeclarationsFromPCH; + * Opts.DisplayDiagnostics = DisplayDiagnostics; + * Opts.PreambleStoragePath = ApplicationTemporaryPath; + * Idx = clang_createIndexWithOptions(&Opts); + * if (Idx) + * return Idx; + * fprintf(stderr, + * "clang_createIndexWithOptions() failed. " + * "CINDEX_VERSION_MINOR = %d, sizeof(CXIndexOptions) = %u\n", + * CINDEX_VERSION_MINOR, Opts.Size); + * #else + * (void)ApplicationTemporaryPath; + * #endif + * Idx = clang_createIndex(ExcludeDeclarationsFromPCH, DisplayDiagnostics); + * clang_CXIndex_setGlobalOptions( + * Idx, clang_CXIndex_getGlobalOptions(Idx) | + * CXGlobalOpt_ThreadBackgroundPriorityForIndexing); + * return Idx; + * } + * \endcode + * + * \sa clang_createIndex() + */ +CINDEX_LINKAGE CXIndex +clang_createIndexWithOptions(const CXIndexOptions *options); + +/** + * Sets general options associated with a CXIndex. + * + * This function is DEPRECATED. Set + * CXIndexOptions::ThreadBackgroundPriorityForIndexing and/or + * CXIndexOptions::ThreadBackgroundPriorityForEditing and call + * clang_createIndexWithOptions() instead. + * + * For example: + * \code + * CXIndex idx = ...; + * clang_CXIndex_setGlobalOptions(idx, + * clang_CXIndex_getGlobalOptions(idx) | + * CXGlobalOpt_ThreadBackgroundPriorityForIndexing); + * \endcode + * + * \param options A bitmask of options, a bitwise OR of CXGlobalOpt_XXX flags. + */ +CINDEX_LINKAGE void clang_CXIndex_setGlobalOptions(CXIndex, unsigned options); + +/** + * Gets the general options associated with a CXIndex. + * + * This function allows to obtain the final option values used by libclang after + * specifying the option policies via CXChoice enumerators. + * + * \returns A bitmask of options, a bitwise OR of CXGlobalOpt_XXX flags that + * are associated with the given CXIndex object. + */ +CINDEX_LINKAGE unsigned clang_CXIndex_getGlobalOptions(CXIndex); + +/** + * Sets the invocation emission path option in a CXIndex. + * + * This function is DEPRECATED. Set CXIndexOptions::InvocationEmissionPath and + * call clang_createIndexWithOptions() instead. + * + * The invocation emission path specifies a path which will contain log + * files for certain libclang invocations. A null value (default) implies that + * libclang invocations are not logged.. + */ +CINDEX_LINKAGE void +clang_CXIndex_setInvocationEmissionPathOption(CXIndex, const char *Path); + +/** + * Determine whether the given header is guarded against + * multiple inclusions, either with the conventional + * \#ifndef/\#define/\#endif macro guards or with \#pragma once. + */ +CINDEX_LINKAGE unsigned clang_isFileMultipleIncludeGuarded(CXTranslationUnit tu, + CXFile file); + +/** + * Retrieve a file handle within the given translation unit. + * + * \param tu the translation unit + * + * \param file_name the name of the file. + * + * \returns the file handle for the named file in the translation unit \p tu, + * or a NULL file handle if the file was not a part of this translation unit. + */ +CINDEX_LINKAGE CXFile clang_getFile(CXTranslationUnit tu, + const char *file_name); + +/** + * Retrieve the buffer associated with the given file. + * + * \param tu the translation unit + * + * \param file the file for which to retrieve the buffer. + * + * \param size [out] if non-NULL, will be set to the size of the buffer. + * + * \returns a pointer to the buffer in memory that holds the contents of + * \p file, or a NULL pointer when the file is not loaded. + */ +CINDEX_LINKAGE const char *clang_getFileContents(CXTranslationUnit tu, + CXFile file, size_t *size); + +/** + * Retrieves the source location associated with a given file/line/column + * in a particular translation unit. + */ +CINDEX_LINKAGE CXSourceLocation clang_getLocation(CXTranslationUnit tu, + CXFile file, unsigned line, + unsigned column); +/** + * Retrieves the source location associated with a given character offset + * in a particular translation unit. + */ +CINDEX_LINKAGE CXSourceLocation clang_getLocationForOffset(CXTranslationUnit tu, + CXFile file, + unsigned offset); + +/** + * Retrieve all ranges that were skipped by the preprocessor. + * + * The preprocessor will skip lines when they are surrounded by an + * if/ifdef/ifndef directive whose condition does not evaluate to true. + */ +CINDEX_LINKAGE CXSourceRangeList *clang_getSkippedRanges(CXTranslationUnit tu, + CXFile file); + +/** + * Retrieve all ranges from all files that were skipped by the + * preprocessor. + * + * The preprocessor will skip lines when they are surrounded by an + * if/ifdef/ifndef directive whose condition does not evaluate to true. + */ +CINDEX_LINKAGE CXSourceRangeList * +clang_getAllSkippedRanges(CXTranslationUnit tu); + +/** + * Determine the number of diagnostics produced for the given + * translation unit. + */ +CINDEX_LINKAGE unsigned clang_getNumDiagnostics(CXTranslationUnit Unit); + +/** + * Retrieve a diagnostic associated with the given translation unit. + * + * \param Unit the translation unit to query. + * \param Index the zero-based diagnostic number to retrieve. + * + * \returns the requested diagnostic. This diagnostic must be freed + * via a call to \c clang_disposeDiagnostic(). + */ +CINDEX_LINKAGE CXDiagnostic clang_getDiagnostic(CXTranslationUnit Unit, + unsigned Index); + +/** + * Retrieve the complete set of diagnostics associated with a + * translation unit. + * + * \param Unit the translation unit to query. + */ +CINDEX_LINKAGE CXDiagnosticSet +clang_getDiagnosticSetFromTU(CXTranslationUnit Unit); + +/** + * \defgroup CINDEX_TRANSLATION_UNIT Translation unit manipulation + * + * The routines in this group provide the ability to create and destroy + * translation units from files, either by parsing the contents of the files or + * by reading in a serialized representation of a translation unit. + * + * @{ + */ + +/** + * Get the original translation unit source file name. + */ +CINDEX_LINKAGE CXString +clang_getTranslationUnitSpelling(CXTranslationUnit CTUnit); + +/** + * Return the CXTranslationUnit for a given source file and the provided + * command line arguments one would pass to the compiler. + * + * Note: The 'source_filename' argument is optional. If the caller provides a + * NULL pointer, the name of the source file is expected to reside in the + * specified command line arguments. + * + * Note: When encountered in 'clang_command_line_args', the following options + * are ignored: + * + * '-c' + * '-emit-ast' + * '-fsyntax-only' + * '-o \' (both '-o' and '\' are ignored) + * + * \param CIdx The index object with which the translation unit will be + * associated. + * + * \param source_filename The name of the source file to load, or NULL if the + * source file is included in \p clang_command_line_args. + * + * \param num_clang_command_line_args The number of command-line arguments in + * \p clang_command_line_args. + * + * \param clang_command_line_args The command-line arguments that would be + * passed to the \c clang executable if it were being invoked out-of-process. + * These command-line options will be parsed and will affect how the translation + * unit is parsed. Note that the following options are ignored: '-c', + * '-emit-ast', '-fsyntax-only' (which is the default), and '-o \'. + * + * \param num_unsaved_files the number of unsaved file entries in \p + * unsaved_files. + * + * \param unsaved_files the files that have not yet been saved to disk + * but may be required for code completion, including the contents of + * those files. The contents and name of these files (as specified by + * CXUnsavedFile) are copied when necessary, so the client only needs to + * guarantee their validity until the call to this function returns. + */ +CINDEX_LINKAGE CXTranslationUnit clang_createTranslationUnitFromSourceFile( + CXIndex CIdx, const char *source_filename, int num_clang_command_line_args, + const char *const *clang_command_line_args, unsigned num_unsaved_files, + struct CXUnsavedFile *unsaved_files); + +/** + * Same as \c clang_createTranslationUnit2, but returns + * the \c CXTranslationUnit instead of an error code. In case of an error this + * routine returns a \c NULL \c CXTranslationUnit, without further detailed + * error codes. + */ +CINDEX_LINKAGE CXTranslationUnit +clang_createTranslationUnit(CXIndex CIdx, const char *ast_filename); + +/** + * Create a translation unit from an AST file (\c -emit-ast). + * + * \param[out] out_TU A non-NULL pointer to store the created + * \c CXTranslationUnit. + * + * \returns Zero on success, otherwise returns an error code. + */ +CINDEX_LINKAGE enum CXErrorCode +clang_createTranslationUnit2(CXIndex CIdx, const char *ast_filename, + CXTranslationUnit *out_TU); + +/** + * Flags that control the creation of translation units. + * + * The enumerators in this enumeration type are meant to be bitwise + * ORed together to specify which options should be used when + * constructing the translation unit. + */ +enum CXTranslationUnit_Flags { + /** + * Used to indicate that no special translation-unit options are + * needed. + */ + CXTranslationUnit_None = 0x0, + + /** + * Used to indicate that the parser should construct a "detailed" + * preprocessing record, including all macro definitions and instantiations. + * + * Constructing a detailed preprocessing record requires more memory + * and time to parse, since the information contained in the record + * is usually not retained. However, it can be useful for + * applications that require more detailed information about the + * behavior of the preprocessor. + */ + CXTranslationUnit_DetailedPreprocessingRecord = 0x01, + + /** + * Used to indicate that the translation unit is incomplete. + * + * When a translation unit is considered "incomplete", semantic + * analysis that is typically performed at the end of the + * translation unit will be suppressed. For example, this suppresses + * the completion of tentative declarations in C and of + * instantiation of implicitly-instantiation function templates in + * C++. This option is typically used when parsing a header with the + * intent of producing a precompiled header. + */ + CXTranslationUnit_Incomplete = 0x02, + + /** + * Used to indicate that the translation unit should be built with an + * implicit precompiled header for the preamble. + * + * An implicit precompiled header is used as an optimization when a + * particular translation unit is likely to be reparsed many times + * when the sources aren't changing that often. In this case, an + * implicit precompiled header will be built containing all of the + * initial includes at the top of the main file (what we refer to as + * the "preamble" of the file). In subsequent parses, if the + * preamble or the files in it have not changed, \c + * clang_reparseTranslationUnit() will re-use the implicit + * precompiled header to improve parsing performance. + */ + CXTranslationUnit_PrecompiledPreamble = 0x04, + + /** + * Used to indicate that the translation unit should cache some + * code-completion results with each reparse of the source file. + * + * Caching of code-completion results is a performance optimization that + * introduces some overhead to reparsing but improves the performance of + * code-completion operations. + */ + CXTranslationUnit_CacheCompletionResults = 0x08, + + /** + * Used to indicate that the translation unit will be serialized with + * \c clang_saveTranslationUnit. + * + * This option is typically used when parsing a header with the intent of + * producing a precompiled header. + */ + CXTranslationUnit_ForSerialization = 0x10, + + /** + * DEPRECATED: Enabled chained precompiled preambles in C++. + * + * Note: this is a *temporary* option that is available only while + * we are testing C++ precompiled preamble support. It is deprecated. + */ + CXTranslationUnit_CXXChainedPCH = 0x20, + + /** + * Used to indicate that function/method bodies should be skipped while + * parsing. + * + * This option can be used to search for declarations/definitions while + * ignoring the usages. + */ + CXTranslationUnit_SkipFunctionBodies = 0x40, + + /** + * Used to indicate that brief documentation comments should be + * included into the set of code completions returned from this translation + * unit. + */ + CXTranslationUnit_IncludeBriefCommentsInCodeCompletion = 0x80, + + /** + * Used to indicate that the precompiled preamble should be created on + * the first parse. Otherwise it will be created on the first reparse. This + * trades runtime on the first parse (serializing the preamble takes time) for + * reduced runtime on the second parse (can now reuse the preamble). + */ + CXTranslationUnit_CreatePreambleOnFirstParse = 0x100, + + /** + * Do not stop processing when fatal errors are encountered. + * + * When fatal errors are encountered while parsing a translation unit, + * semantic analysis is typically stopped early when compiling code. A common + * source for fatal errors are unresolvable include files. For the + * purposes of an IDE, this is undesirable behavior and as much information + * as possible should be reported. Use this flag to enable this behavior. + */ + CXTranslationUnit_KeepGoing = 0x200, + + /** + * Sets the preprocessor in a mode for parsing a single file only. + */ + CXTranslationUnit_SingleFileParse = 0x400, + + /** + * Used in combination with CXTranslationUnit_SkipFunctionBodies to + * constrain the skipping of function bodies to the preamble. + * + * The function bodies of the main file are not skipped. + */ + CXTranslationUnit_LimitSkipFunctionBodiesToPreamble = 0x800, + + /** + * Used to indicate that attributed types should be included in CXType. + */ + CXTranslationUnit_IncludeAttributedTypes = 0x1000, + + /** + * Used to indicate that implicit attributes should be visited. + */ + CXTranslationUnit_VisitImplicitAttributes = 0x2000, + + /** + * Used to indicate that non-errors from included files should be ignored. + * + * If set, clang_getDiagnosticSetFromTU() will not report e.g. warnings from + * included files anymore. This speeds up clang_getDiagnosticSetFromTU() for + * the case where these warnings are not of interest, as for an IDE for + * example, which typically shows only the diagnostics in the main file. + */ + CXTranslationUnit_IgnoreNonErrorsFromIncludedFiles = 0x4000, + + /** + * Tells the preprocessor not to skip excluded conditional blocks. + */ + CXTranslationUnit_RetainExcludedConditionalBlocks = 0x8000 +}; + +/** + * Returns the set of flags that is suitable for parsing a translation + * unit that is being edited. + * + * The set of flags returned provide options for \c clang_parseTranslationUnit() + * to indicate that the translation unit is likely to be reparsed many times, + * either explicitly (via \c clang_reparseTranslationUnit()) or implicitly + * (e.g., by code completion (\c clang_codeCompletionAt())). The returned flag + * set contains an unspecified set of optimizations (e.g., the precompiled + * preamble) geared toward improving the performance of these routines. The + * set of optimizations enabled may change from one version to the next. + */ +CINDEX_LINKAGE unsigned clang_defaultEditingTranslationUnitOptions(void); + +/** + * Same as \c clang_parseTranslationUnit2, but returns + * the \c CXTranslationUnit instead of an error code. In case of an error this + * routine returns a \c NULL \c CXTranslationUnit, without further detailed + * error codes. + */ +CINDEX_LINKAGE CXTranslationUnit clang_parseTranslationUnit( + CXIndex CIdx, const char *source_filename, + const char *const *command_line_args, int num_command_line_args, + struct CXUnsavedFile *unsaved_files, unsigned num_unsaved_files, + unsigned options); + +/** + * Parse the given source file and the translation unit corresponding + * to that file. + * + * This routine is the main entry point for the Clang C API, providing the + * ability to parse a source file into a translation unit that can then be + * queried by other functions in the API. This routine accepts a set of + * command-line arguments so that the compilation can be configured in the same + * way that the compiler is configured on the command line. + * + * \param CIdx The index object with which the translation unit will be + * associated. + * + * \param source_filename The name of the source file to load, or NULL if the + * source file is included in \c command_line_args. + * + * \param command_line_args The command-line arguments that would be + * passed to the \c clang executable if it were being invoked out-of-process. + * These command-line options will be parsed and will affect how the translation + * unit is parsed. Note that the following options are ignored: '-c', + * '-emit-ast', '-fsyntax-only' (which is the default), and '-o \'. + * + * \param num_command_line_args The number of command-line arguments in + * \c command_line_args. + * + * \param unsaved_files the files that have not yet been saved to disk + * but may be required for parsing, including the contents of + * those files. The contents and name of these files (as specified by + * CXUnsavedFile) are copied when necessary, so the client only needs to + * guarantee their validity until the call to this function returns. + * + * \param num_unsaved_files the number of unsaved file entries in \p + * unsaved_files. + * + * \param options A bitmask of options that affects how the translation unit + * is managed but not its compilation. This should be a bitwise OR of the + * CXTranslationUnit_XXX flags. + * + * \param[out] out_TU A non-NULL pointer to store the created + * \c CXTranslationUnit, describing the parsed code and containing any + * diagnostics produced by the compiler. + * + * \returns Zero on success, otherwise returns an error code. + */ +CINDEX_LINKAGE enum CXErrorCode clang_parseTranslationUnit2( + CXIndex CIdx, const char *source_filename, + const char *const *command_line_args, int num_command_line_args, + struct CXUnsavedFile *unsaved_files, unsigned num_unsaved_files, + unsigned options, CXTranslationUnit *out_TU); + +/** + * Same as clang_parseTranslationUnit2 but requires a full command line + * for \c command_line_args including argv[0]. This is useful if the standard + * library paths are relative to the binary. + */ +CINDEX_LINKAGE enum CXErrorCode clang_parseTranslationUnit2FullArgv( + CXIndex CIdx, const char *source_filename, + const char *const *command_line_args, int num_command_line_args, + struct CXUnsavedFile *unsaved_files, unsigned num_unsaved_files, + unsigned options, CXTranslationUnit *out_TU); + +/** + * Flags that control how translation units are saved. + * + * The enumerators in this enumeration type are meant to be bitwise + * ORed together to specify which options should be used when + * saving the translation unit. + */ +enum CXSaveTranslationUnit_Flags { + /** + * Used to indicate that no special saving options are needed. + */ + CXSaveTranslationUnit_None = 0x0 +}; + +/** + * Returns the set of flags that is suitable for saving a translation + * unit. + * + * The set of flags returned provide options for + * \c clang_saveTranslationUnit() by default. The returned flag + * set contains an unspecified set of options that save translation units with + * the most commonly-requested data. + */ +CINDEX_LINKAGE unsigned clang_defaultSaveOptions(CXTranslationUnit TU); + +/** + * Describes the kind of error that occurred (if any) in a call to + * \c clang_saveTranslationUnit(). + */ +enum CXSaveError { + /** + * Indicates that no error occurred while saving a translation unit. + */ + CXSaveError_None = 0, + + /** + * Indicates that an unknown error occurred while attempting to save + * the file. + * + * This error typically indicates that file I/O failed when attempting to + * write the file. + */ + CXSaveError_Unknown = 1, + + /** + * Indicates that errors during translation prevented this attempt + * to save the translation unit. + * + * Errors that prevent the translation unit from being saved can be + * extracted using \c clang_getNumDiagnostics() and \c clang_getDiagnostic(). + */ + CXSaveError_TranslationErrors = 2, + + /** + * Indicates that the translation unit to be saved was somehow + * invalid (e.g., NULL). + */ + CXSaveError_InvalidTU = 3 +}; + +/** + * Saves a translation unit into a serialized representation of + * that translation unit on disk. + * + * Any translation unit that was parsed without error can be saved + * into a file. The translation unit can then be deserialized into a + * new \c CXTranslationUnit with \c clang_createTranslationUnit() or, + * if it is an incomplete translation unit that corresponds to a + * header, used as a precompiled header when parsing other translation + * units. + * + * \param TU The translation unit to save. + * + * \param FileName The file to which the translation unit will be saved. + * + * \param options A bitmask of options that affects how the translation unit + * is saved. This should be a bitwise OR of the + * CXSaveTranslationUnit_XXX flags. + * + * \returns A value that will match one of the enumerators of the CXSaveError + * enumeration. Zero (CXSaveError_None) indicates that the translation unit was + * saved successfully, while a non-zero value indicates that a problem occurred. + */ +CINDEX_LINKAGE int clang_saveTranslationUnit(CXTranslationUnit TU, + const char *FileName, + unsigned options); + +/** + * Suspend a translation unit in order to free memory associated with it. + * + * A suspended translation unit uses significantly less memory but on the other + * side does not support any other calls than \c clang_reparseTranslationUnit + * to resume it or \c clang_disposeTranslationUnit to dispose it completely. + */ +CINDEX_LINKAGE unsigned clang_suspendTranslationUnit(CXTranslationUnit); + +/** + * Destroy the specified CXTranslationUnit object. + */ +CINDEX_LINKAGE void clang_disposeTranslationUnit(CXTranslationUnit); + +/** + * Flags that control the reparsing of translation units. + * + * The enumerators in this enumeration type are meant to be bitwise + * ORed together to specify which options should be used when + * reparsing the translation unit. + */ +enum CXReparse_Flags { + /** + * Used to indicate that no special reparsing options are needed. + */ + CXReparse_None = 0x0 +}; + +/** + * Returns the set of flags that is suitable for reparsing a translation + * unit. + * + * The set of flags returned provide options for + * \c clang_reparseTranslationUnit() by default. The returned flag + * set contains an unspecified set of optimizations geared toward common uses + * of reparsing. The set of optimizations enabled may change from one version + * to the next. + */ +CINDEX_LINKAGE unsigned clang_defaultReparseOptions(CXTranslationUnit TU); + +/** + * Reparse the source files that produced this translation unit. + * + * This routine can be used to re-parse the source files that originally + * created the given translation unit, for example because those source files + * have changed (either on disk or as passed via \p unsaved_files). The + * source code will be reparsed with the same command-line options as it + * was originally parsed. + * + * Reparsing a translation unit invalidates all cursors and source locations + * that refer into that translation unit. This makes reparsing a translation + * unit semantically equivalent to destroying the translation unit and then + * creating a new translation unit with the same command-line arguments. + * However, it may be more efficient to reparse a translation + * unit using this routine. + * + * \param TU The translation unit whose contents will be re-parsed. The + * translation unit must originally have been built with + * \c clang_createTranslationUnitFromSourceFile(). + * + * \param num_unsaved_files The number of unsaved file entries in \p + * unsaved_files. + * + * \param unsaved_files The files that have not yet been saved to disk + * but may be required for parsing, including the contents of + * those files. The contents and name of these files (as specified by + * CXUnsavedFile) are copied when necessary, so the client only needs to + * guarantee their validity until the call to this function returns. + * + * \param options A bitset of options composed of the flags in CXReparse_Flags. + * The function \c clang_defaultReparseOptions() produces a default set of + * options recommended for most uses, based on the translation unit. + * + * \returns 0 if the sources could be reparsed. A non-zero error code will be + * returned if reparsing was impossible, such that the translation unit is + * invalid. In such cases, the only valid call for \c TU is + * \c clang_disposeTranslationUnit(TU). The error codes returned by this + * routine are described by the \c CXErrorCode enum. + */ +CINDEX_LINKAGE int +clang_reparseTranslationUnit(CXTranslationUnit TU, unsigned num_unsaved_files, + struct CXUnsavedFile *unsaved_files, + unsigned options); + +/** + * Categorizes how memory is being used by a translation unit. + */ +enum CXTUResourceUsageKind { + CXTUResourceUsage_AST = 1, + CXTUResourceUsage_Identifiers = 2, + CXTUResourceUsage_Selectors = 3, + CXTUResourceUsage_GlobalCompletionResults = 4, + CXTUResourceUsage_SourceManagerContentCache = 5, + CXTUResourceUsage_AST_SideTables = 6, + CXTUResourceUsage_SourceManager_Membuffer_Malloc = 7, + CXTUResourceUsage_SourceManager_Membuffer_MMap = 8, + CXTUResourceUsage_ExternalASTSource_Membuffer_Malloc = 9, + CXTUResourceUsage_ExternalASTSource_Membuffer_MMap = 10, + CXTUResourceUsage_Preprocessor = 11, + CXTUResourceUsage_PreprocessingRecord = 12, + CXTUResourceUsage_SourceManager_DataStructures = 13, + CXTUResourceUsage_Preprocessor_HeaderSearch = 14, + CXTUResourceUsage_MEMORY_IN_BYTES_BEGIN = CXTUResourceUsage_AST, + CXTUResourceUsage_MEMORY_IN_BYTES_END = + CXTUResourceUsage_Preprocessor_HeaderSearch, + + CXTUResourceUsage_First = CXTUResourceUsage_AST, + CXTUResourceUsage_Last = CXTUResourceUsage_Preprocessor_HeaderSearch +}; + +/** + * Returns the human-readable null-terminated C string that represents + * the name of the memory category. This string should never be freed. + */ +CINDEX_LINKAGE +const char *clang_getTUResourceUsageName(enum CXTUResourceUsageKind kind); + +typedef struct CXTUResourceUsageEntry { + /* The memory usage category. */ + enum CXTUResourceUsageKind kind; + /* Amount of resources used. + The units will depend on the resource kind. */ + unsigned long amount; +} CXTUResourceUsageEntry; + +/** + * The memory usage of a CXTranslationUnit, broken into categories. + */ +typedef struct CXTUResourceUsage { + /* Private data member, used for queries. */ + void *data; + + /* The number of entries in the 'entries' array. */ + unsigned numEntries; + + /* An array of key-value pairs, representing the breakdown of memory + usage. */ + CXTUResourceUsageEntry *entries; + +} CXTUResourceUsage; + +/** + * Return the memory usage of a translation unit. This object + * should be released with clang_disposeCXTUResourceUsage(). + */ +CINDEX_LINKAGE CXTUResourceUsage +clang_getCXTUResourceUsage(CXTranslationUnit TU); + +CINDEX_LINKAGE void clang_disposeCXTUResourceUsage(CXTUResourceUsage usage); + +/** + * Get target information for this translation unit. + * + * The CXTargetInfo object cannot outlive the CXTranslationUnit object. + */ +CINDEX_LINKAGE CXTargetInfo +clang_getTranslationUnitTargetInfo(CXTranslationUnit CTUnit); + +/** + * Destroy the CXTargetInfo object. + */ +CINDEX_LINKAGE void clang_TargetInfo_dispose(CXTargetInfo Info); + +/** + * Get the normalized target triple as a string. + * + * Returns the empty string in case of any error. + */ +CINDEX_LINKAGE CXString clang_TargetInfo_getTriple(CXTargetInfo Info); + +/** + * Get the pointer width of the target in bits. + * + * Returns -1 in case of error. + */ +CINDEX_LINKAGE int clang_TargetInfo_getPointerWidth(CXTargetInfo Info); + +/** + * @} + */ + +/** + * Describes the kind of entity that a cursor refers to. + */ +enum CXCursorKind { + /* Declarations */ + /** + * A declaration whose specific kind is not exposed via this + * interface. + * + * Unexposed declarations have the same operations as any other kind + * of declaration; one can extract their location information, + * spelling, find their definitions, etc. However, the specific kind + * of the declaration is not reported. + */ + CXCursor_UnexposedDecl = 1, + /** A C or C++ struct. */ + CXCursor_StructDecl = 2, + /** A C or C++ union. */ + CXCursor_UnionDecl = 3, + /** A C++ class. */ + CXCursor_ClassDecl = 4, + /** An enumeration. */ + CXCursor_EnumDecl = 5, + /** + * A field (in C) or non-static data member (in C++) in a + * struct, union, or C++ class. + */ + CXCursor_FieldDecl = 6, + /** An enumerator constant. */ + CXCursor_EnumConstantDecl = 7, + /** A function. */ + CXCursor_FunctionDecl = 8, + /** A variable. */ + CXCursor_VarDecl = 9, + /** A function or method parameter. */ + CXCursor_ParmDecl = 10, + /** An Objective-C \@interface. */ + CXCursor_ObjCInterfaceDecl = 11, + /** An Objective-C \@interface for a category. */ + CXCursor_ObjCCategoryDecl = 12, + /** An Objective-C \@protocol declaration. */ + CXCursor_ObjCProtocolDecl = 13, + /** An Objective-C \@property declaration. */ + CXCursor_ObjCPropertyDecl = 14, + /** An Objective-C instance variable. */ + CXCursor_ObjCIvarDecl = 15, + /** An Objective-C instance method. */ + CXCursor_ObjCInstanceMethodDecl = 16, + /** An Objective-C class method. */ + CXCursor_ObjCClassMethodDecl = 17, + /** An Objective-C \@implementation. */ + CXCursor_ObjCImplementationDecl = 18, + /** An Objective-C \@implementation for a category. */ + CXCursor_ObjCCategoryImplDecl = 19, + /** A typedef. */ + CXCursor_TypedefDecl = 20, + /** A C++ class method. */ + CXCursor_CXXMethod = 21, + /** A C++ namespace. */ + CXCursor_Namespace = 22, + /** A linkage specification, e.g. 'extern "C"'. */ + CXCursor_LinkageSpec = 23, + /** A C++ constructor. */ + CXCursor_Constructor = 24, + /** A C++ destructor. */ + CXCursor_Destructor = 25, + /** A C++ conversion function. */ + CXCursor_ConversionFunction = 26, + /** A C++ template type parameter. */ + CXCursor_TemplateTypeParameter = 27, + /** A C++ non-type template parameter. */ + CXCursor_NonTypeTemplateParameter = 28, + /** A C++ template template parameter. */ + CXCursor_TemplateTemplateParameter = 29, + /** A C++ function template. */ + CXCursor_FunctionTemplate = 30, + /** A C++ class template. */ + CXCursor_ClassTemplate = 31, + /** A C++ class template partial specialization. */ + CXCursor_ClassTemplatePartialSpecialization = 32, + /** A C++ namespace alias declaration. */ + CXCursor_NamespaceAlias = 33, + /** A C++ using directive. */ + CXCursor_UsingDirective = 34, + /** A C++ using declaration. */ + CXCursor_UsingDeclaration = 35, + /** A C++ alias declaration */ + CXCursor_TypeAliasDecl = 36, + /** An Objective-C \@synthesize definition. */ + CXCursor_ObjCSynthesizeDecl = 37, + /** An Objective-C \@dynamic definition. */ + CXCursor_ObjCDynamicDecl = 38, + /** An access specifier. */ + CXCursor_CXXAccessSpecifier = 39, + + CXCursor_FirstDecl = CXCursor_UnexposedDecl, + CXCursor_LastDecl = CXCursor_CXXAccessSpecifier, + + /* References */ + CXCursor_FirstRef = 40, /* Decl references */ + CXCursor_ObjCSuperClassRef = 40, + CXCursor_ObjCProtocolRef = 41, + CXCursor_ObjCClassRef = 42, + /** + * A reference to a type declaration. + * + * A type reference occurs anywhere where a type is named but not + * declared. For example, given: + * + * \code + * typedef unsigned size_type; + * size_type size; + * \endcode + * + * The typedef is a declaration of size_type (CXCursor_TypedefDecl), + * while the type of the variable "size" is referenced. The cursor + * referenced by the type of size is the typedef for size_type. + */ + CXCursor_TypeRef = 43, + CXCursor_CXXBaseSpecifier = 44, + /** + * A reference to a class template, function template, template + * template parameter, or class template partial specialization. + */ + CXCursor_TemplateRef = 45, + /** + * A reference to a namespace or namespace alias. + */ + CXCursor_NamespaceRef = 46, + /** + * A reference to a member of a struct, union, or class that occurs in + * some non-expression context, e.g., a designated initializer. + */ + CXCursor_MemberRef = 47, + /** + * A reference to a labeled statement. + * + * This cursor kind is used to describe the jump to "start_over" in the + * goto statement in the following example: + * + * \code + * start_over: + * ++counter; + * + * goto start_over; + * \endcode + * + * A label reference cursor refers to a label statement. + */ + CXCursor_LabelRef = 48, + + /** + * A reference to a set of overloaded functions or function templates + * that has not yet been resolved to a specific function or function template. + * + * An overloaded declaration reference cursor occurs in C++ templates where + * a dependent name refers to a function. For example: + * + * \code + * template void swap(T&, T&); + * + * struct X { ... }; + * void swap(X&, X&); + * + * template + * void reverse(T* first, T* last) { + * while (first < last - 1) { + * swap(*first, *--last); + * ++first; + * } + * } + * + * struct Y { }; + * void swap(Y&, Y&); + * \endcode + * + * Here, the identifier "swap" is associated with an overloaded declaration + * reference. In the template definition, "swap" refers to either of the two + * "swap" functions declared above, so both results will be available. At + * instantiation time, "swap" may also refer to other functions found via + * argument-dependent lookup (e.g., the "swap" function at the end of the + * example). + * + * The functions \c clang_getNumOverloadedDecls() and + * \c clang_getOverloadedDecl() can be used to retrieve the definitions + * referenced by this cursor. + */ + CXCursor_OverloadedDeclRef = 49, + + /** + * A reference to a variable that occurs in some non-expression + * context, e.g., a C++ lambda capture list. + */ + CXCursor_VariableRef = 50, + + CXCursor_LastRef = CXCursor_VariableRef, + + /* Error conditions */ + CXCursor_FirstInvalid = 70, + CXCursor_InvalidFile = 70, + CXCursor_NoDeclFound = 71, + CXCursor_NotImplemented = 72, + CXCursor_InvalidCode = 73, + CXCursor_LastInvalid = CXCursor_InvalidCode, + + /* Expressions */ + CXCursor_FirstExpr = 100, + + /** + * An expression whose specific kind is not exposed via this + * interface. + * + * Unexposed expressions have the same operations as any other kind + * of expression; one can extract their location information, + * spelling, children, etc. However, the specific kind of the + * expression is not reported. + */ + CXCursor_UnexposedExpr = 100, + + /** + * An expression that refers to some value declaration, such + * as a function, variable, or enumerator. + */ + CXCursor_DeclRefExpr = 101, + + /** + * An expression that refers to a member of a struct, union, + * class, Objective-C class, etc. + */ + CXCursor_MemberRefExpr = 102, + + /** An expression that calls a function. */ + CXCursor_CallExpr = 103, + + /** An expression that sends a message to an Objective-C + object or class. */ + CXCursor_ObjCMessageExpr = 104, + + /** An expression that represents a block literal. */ + CXCursor_BlockExpr = 105, + + /** An integer literal. + */ + CXCursor_IntegerLiteral = 106, + + /** A floating point number literal. + */ + CXCursor_FloatingLiteral = 107, + + /** An imaginary number literal. + */ + CXCursor_ImaginaryLiteral = 108, + + /** A string literal. + */ + CXCursor_StringLiteral = 109, + + /** A character literal. + */ + CXCursor_CharacterLiteral = 110, + + /** A parenthesized expression, e.g. "(1)". + * + * This AST node is only formed if full location information is requested. + */ + CXCursor_ParenExpr = 111, + + /** This represents the unary-expression's (except sizeof and + * alignof). + */ + CXCursor_UnaryOperator = 112, + + /** [C99 6.5.2.1] Array Subscripting. + */ + CXCursor_ArraySubscriptExpr = 113, + + /** A builtin binary operation expression such as "x + y" or + * "x <= y". + */ + CXCursor_BinaryOperator = 114, + + /** Compound assignment such as "+=". + */ + CXCursor_CompoundAssignOperator = 115, + + /** The ?: ternary operator. + */ + CXCursor_ConditionalOperator = 116, + + /** An explicit cast in C (C99 6.5.4) or a C-style cast in C++ + * (C++ [expr.cast]), which uses the syntax (Type)expr. + * + * For example: (int)f. + */ + CXCursor_CStyleCastExpr = 117, + + /** [C99 6.5.2.5] + */ + CXCursor_CompoundLiteralExpr = 118, + + /** Describes an C or C++ initializer list. + */ + CXCursor_InitListExpr = 119, + + /** The GNU address of label extension, representing &&label. + */ + CXCursor_AddrLabelExpr = 120, + + /** This is the GNU Statement Expression extension: ({int X=4; X;}) + */ + CXCursor_StmtExpr = 121, + + /** Represents a C11 generic selection. + */ + CXCursor_GenericSelectionExpr = 122, + + /** Implements the GNU __null extension, which is a name for a null + * pointer constant that has integral type (e.g., int or long) and is the same + * size and alignment as a pointer. + * + * The __null extension is typically only used by system headers, which define + * NULL as __null in C++ rather than using 0 (which is an integer that may not + * match the size of a pointer). + */ + CXCursor_GNUNullExpr = 123, + + /** C++'s static_cast<> expression. + */ + CXCursor_CXXStaticCastExpr = 124, + + /** C++'s dynamic_cast<> expression. + */ + CXCursor_CXXDynamicCastExpr = 125, + + /** C++'s reinterpret_cast<> expression. + */ + CXCursor_CXXReinterpretCastExpr = 126, + + /** C++'s const_cast<> expression. + */ + CXCursor_CXXConstCastExpr = 127, + + /** Represents an explicit C++ type conversion that uses "functional" + * notion (C++ [expr.type.conv]). + * + * Example: + * \code + * x = int(0.5); + * \endcode + */ + CXCursor_CXXFunctionalCastExpr = 128, + + /** A C++ typeid expression (C++ [expr.typeid]). + */ + CXCursor_CXXTypeidExpr = 129, + + /** [C++ 2.13.5] C++ Boolean Literal. + */ + CXCursor_CXXBoolLiteralExpr = 130, + + /** [C++0x 2.14.7] C++ Pointer Literal. + */ + CXCursor_CXXNullPtrLiteralExpr = 131, + + /** Represents the "this" expression in C++ + */ + CXCursor_CXXThisExpr = 132, + + /** [C++ 15] C++ Throw Expression. + * + * This handles 'throw' and 'throw' assignment-expression. When + * assignment-expression isn't present, Op will be null. + */ + CXCursor_CXXThrowExpr = 133, + + /** A new expression for memory allocation and constructor calls, e.g: + * "new CXXNewExpr(foo)". + */ + CXCursor_CXXNewExpr = 134, + + /** A delete expression for memory deallocation and destructor calls, + * e.g. "delete[] pArray". + */ + CXCursor_CXXDeleteExpr = 135, + + /** A unary expression. (noexcept, sizeof, or other traits) + */ + CXCursor_UnaryExpr = 136, + + /** An Objective-C string literal i.e. @"foo". + */ + CXCursor_ObjCStringLiteral = 137, + + /** An Objective-C \@encode expression. + */ + CXCursor_ObjCEncodeExpr = 138, + + /** An Objective-C \@selector expression. + */ + CXCursor_ObjCSelectorExpr = 139, + + /** An Objective-C \@protocol expression. + */ + CXCursor_ObjCProtocolExpr = 140, + + /** An Objective-C "bridged" cast expression, which casts between + * Objective-C pointers and C pointers, transferring ownership in the process. + * + * \code + * NSString *str = (__bridge_transfer NSString *)CFCreateString(); + * \endcode + */ + CXCursor_ObjCBridgedCastExpr = 141, + + /** Represents a C++0x pack expansion that produces a sequence of + * expressions. + * + * A pack expansion expression contains a pattern (which itself is an + * expression) followed by an ellipsis. For example: + * + * \code + * template + * void forward(F f, Types &&...args) { + * f(static_cast(args)...); + * } + * \endcode + */ + CXCursor_PackExpansionExpr = 142, + + /** Represents an expression that computes the length of a parameter + * pack. + * + * \code + * template + * struct count { + * static const unsigned value = sizeof...(Types); + * }; + * \endcode + */ + CXCursor_SizeOfPackExpr = 143, + + /* Represents a C++ lambda expression that produces a local function + * object. + * + * \code + * void abssort(float *x, unsigned N) { + * std::sort(x, x + N, + * [](float a, float b) { + * return std::abs(a) < std::abs(b); + * }); + * } + * \endcode + */ + CXCursor_LambdaExpr = 144, + + /** Objective-c Boolean Literal. + */ + CXCursor_ObjCBoolLiteralExpr = 145, + + /** Represents the "self" expression in an Objective-C method. + */ + CXCursor_ObjCSelfExpr = 146, + + /** OpenMP 5.0 [2.1.5, Array Section]. + * OpenACC 3.3 [2.7.1, Data Specification for Data Clauses (Sub Arrays)] + */ + CXCursor_ArraySectionExpr = 147, + + /** Represents an @available(...) check. + */ + CXCursor_ObjCAvailabilityCheckExpr = 148, + + /** + * Fixed point literal + */ + CXCursor_FixedPointLiteral = 149, + + /** OpenMP 5.0 [2.1.4, Array Shaping]. + */ + CXCursor_OMPArrayShapingExpr = 150, + + /** + * OpenMP 5.0 [2.1.6 Iterators] + */ + CXCursor_OMPIteratorExpr = 151, + + /** OpenCL's addrspace_cast<> expression. + */ + CXCursor_CXXAddrspaceCastExpr = 152, + + /** + * Expression that references a C++20 concept. + */ + CXCursor_ConceptSpecializationExpr = 153, + + /** + * Expression that references a C++20 requires expression. + */ + CXCursor_RequiresExpr = 154, + + /** + * Expression that references a C++20 parenthesized list aggregate + * initializer. + */ + CXCursor_CXXParenListInitExpr = 155, + + /** + * Represents a C++26 pack indexing expression. + */ + CXCursor_PackIndexingExpr = 156, + + CXCursor_LastExpr = CXCursor_PackIndexingExpr, + + /* Statements */ + CXCursor_FirstStmt = 200, + /** + * A statement whose specific kind is not exposed via this + * interface. + * + * Unexposed statements have the same operations as any other kind of + * statement; one can extract their location information, spelling, + * children, etc. However, the specific kind of the statement is not + * reported. + */ + CXCursor_UnexposedStmt = 200, + + /** A labelled statement in a function. + * + * This cursor kind is used to describe the "start_over:" label statement in + * the following example: + * + * \code + * start_over: + * ++counter; + * \endcode + * + */ + CXCursor_LabelStmt = 201, + + /** A group of statements like { stmt stmt }. + * + * This cursor kind is used to describe compound statements, e.g. function + * bodies. + */ + CXCursor_CompoundStmt = 202, + + /** A case statement. + */ + CXCursor_CaseStmt = 203, + + /** A default statement. + */ + CXCursor_DefaultStmt = 204, + + /** An if statement + */ + CXCursor_IfStmt = 205, + + /** A switch statement. + */ + CXCursor_SwitchStmt = 206, + + /** A while statement. + */ + CXCursor_WhileStmt = 207, + + /** A do statement. + */ + CXCursor_DoStmt = 208, + + /** A for statement. + */ + CXCursor_ForStmt = 209, + + /** A goto statement. + */ + CXCursor_GotoStmt = 210, + + /** An indirect goto statement. + */ + CXCursor_IndirectGotoStmt = 211, + + /** A continue statement. + */ + CXCursor_ContinueStmt = 212, + + /** A break statement. + */ + CXCursor_BreakStmt = 213, + + /** A return statement. + */ + CXCursor_ReturnStmt = 214, + + /** A GCC inline assembly statement extension. + */ + CXCursor_GCCAsmStmt = 215, + CXCursor_AsmStmt = CXCursor_GCCAsmStmt, + + /** Objective-C's overall \@try-\@catch-\@finally statement. + */ + CXCursor_ObjCAtTryStmt = 216, + + /** Objective-C's \@catch statement. + */ + CXCursor_ObjCAtCatchStmt = 217, + + /** Objective-C's \@finally statement. + */ + CXCursor_ObjCAtFinallyStmt = 218, + + /** Objective-C's \@throw statement. + */ + CXCursor_ObjCAtThrowStmt = 219, + + /** Objective-C's \@synchronized statement. + */ + CXCursor_ObjCAtSynchronizedStmt = 220, + + /** Objective-C's autorelease pool statement. + */ + CXCursor_ObjCAutoreleasePoolStmt = 221, + + /** Objective-C's collection statement. + */ + CXCursor_ObjCForCollectionStmt = 222, + + /** C++'s catch statement. + */ + CXCursor_CXXCatchStmt = 223, + + /** C++'s try statement. + */ + CXCursor_CXXTryStmt = 224, + + /** C++'s for (* : *) statement. + */ + CXCursor_CXXForRangeStmt = 225, + + /** Windows Structured Exception Handling's try statement. + */ + CXCursor_SEHTryStmt = 226, + + /** Windows Structured Exception Handling's except statement. + */ + CXCursor_SEHExceptStmt = 227, + + /** Windows Structured Exception Handling's finally statement. + */ + CXCursor_SEHFinallyStmt = 228, + + /** A MS inline assembly statement extension. + */ + CXCursor_MSAsmStmt = 229, + + /** The null statement ";": C99 6.8.3p3. + * + * This cursor kind is used to describe the null statement. + */ + CXCursor_NullStmt = 230, + + /** Adaptor class for mixing declarations with statements and + * expressions. + */ + CXCursor_DeclStmt = 231, + + /** OpenMP parallel directive. + */ + CXCursor_OMPParallelDirective = 232, + + /** OpenMP SIMD directive. + */ + CXCursor_OMPSimdDirective = 233, + + /** OpenMP for directive. + */ + CXCursor_OMPForDirective = 234, + + /** OpenMP sections directive. + */ + CXCursor_OMPSectionsDirective = 235, + + /** OpenMP section directive. + */ + CXCursor_OMPSectionDirective = 236, + + /** OpenMP single directive. + */ + CXCursor_OMPSingleDirective = 237, + + /** OpenMP parallel for directive. + */ + CXCursor_OMPParallelForDirective = 238, + + /** OpenMP parallel sections directive. + */ + CXCursor_OMPParallelSectionsDirective = 239, + + /** OpenMP task directive. + */ + CXCursor_OMPTaskDirective = 240, + + /** OpenMP master directive. + */ + CXCursor_OMPMasterDirective = 241, + + /** OpenMP critical directive. + */ + CXCursor_OMPCriticalDirective = 242, + + /** OpenMP taskyield directive. + */ + CXCursor_OMPTaskyieldDirective = 243, + + /** OpenMP barrier directive. + */ + CXCursor_OMPBarrierDirective = 244, + + /** OpenMP taskwait directive. + */ + CXCursor_OMPTaskwaitDirective = 245, + + /** OpenMP flush directive. + */ + CXCursor_OMPFlushDirective = 246, + + /** Windows Structured Exception Handling's leave statement. + */ + CXCursor_SEHLeaveStmt = 247, + + /** OpenMP ordered directive. + */ + CXCursor_OMPOrderedDirective = 248, + + /** OpenMP atomic directive. + */ + CXCursor_OMPAtomicDirective = 249, + + /** OpenMP for SIMD directive. + */ + CXCursor_OMPForSimdDirective = 250, + + /** OpenMP parallel for SIMD directive. + */ + CXCursor_OMPParallelForSimdDirective = 251, + + /** OpenMP target directive. + */ + CXCursor_OMPTargetDirective = 252, + + /** OpenMP teams directive. + */ + CXCursor_OMPTeamsDirective = 253, + + /** OpenMP taskgroup directive. + */ + CXCursor_OMPTaskgroupDirective = 254, + + /** OpenMP cancellation point directive. + */ + CXCursor_OMPCancellationPointDirective = 255, + + /** OpenMP cancel directive. + */ + CXCursor_OMPCancelDirective = 256, + + /** OpenMP target data directive. + */ + CXCursor_OMPTargetDataDirective = 257, + + /** OpenMP taskloop directive. + */ + CXCursor_OMPTaskLoopDirective = 258, + + /** OpenMP taskloop simd directive. + */ + CXCursor_OMPTaskLoopSimdDirective = 259, + + /** OpenMP distribute directive. + */ + CXCursor_OMPDistributeDirective = 260, + + /** OpenMP target enter data directive. + */ + CXCursor_OMPTargetEnterDataDirective = 261, + + /** OpenMP target exit data directive. + */ + CXCursor_OMPTargetExitDataDirective = 262, + + /** OpenMP target parallel directive. + */ + CXCursor_OMPTargetParallelDirective = 263, + + /** OpenMP target parallel for directive. + */ + CXCursor_OMPTargetParallelForDirective = 264, + + /** OpenMP target update directive. + */ + CXCursor_OMPTargetUpdateDirective = 265, + + /** OpenMP distribute parallel for directive. + */ + CXCursor_OMPDistributeParallelForDirective = 266, + + /** OpenMP distribute parallel for simd directive. + */ + CXCursor_OMPDistributeParallelForSimdDirective = 267, + + /** OpenMP distribute simd directive. + */ + CXCursor_OMPDistributeSimdDirective = 268, + + /** OpenMP target parallel for simd directive. + */ + CXCursor_OMPTargetParallelForSimdDirective = 269, + + /** OpenMP target simd directive. + */ + CXCursor_OMPTargetSimdDirective = 270, + + /** OpenMP teams distribute directive. + */ + CXCursor_OMPTeamsDistributeDirective = 271, + + /** OpenMP teams distribute simd directive. + */ + CXCursor_OMPTeamsDistributeSimdDirective = 272, + + /** OpenMP teams distribute parallel for simd directive. + */ + CXCursor_OMPTeamsDistributeParallelForSimdDirective = 273, + + /** OpenMP teams distribute parallel for directive. + */ + CXCursor_OMPTeamsDistributeParallelForDirective = 274, + + /** OpenMP target teams directive. + */ + CXCursor_OMPTargetTeamsDirective = 275, + + /** OpenMP target teams distribute directive. + */ + CXCursor_OMPTargetTeamsDistributeDirective = 276, + + /** OpenMP target teams distribute parallel for directive. + */ + CXCursor_OMPTargetTeamsDistributeParallelForDirective = 277, + + /** OpenMP target teams distribute parallel for simd directive. + */ + CXCursor_OMPTargetTeamsDistributeParallelForSimdDirective = 278, + + /** OpenMP target teams distribute simd directive. + */ + CXCursor_OMPTargetTeamsDistributeSimdDirective = 279, + + /** C++2a std::bit_cast expression. + */ + CXCursor_BuiltinBitCastExpr = 280, + + /** OpenMP master taskloop directive. + */ + CXCursor_OMPMasterTaskLoopDirective = 281, + + /** OpenMP parallel master taskloop directive. + */ + CXCursor_OMPParallelMasterTaskLoopDirective = 282, + + /** OpenMP master taskloop simd directive. + */ + CXCursor_OMPMasterTaskLoopSimdDirective = 283, + + /** OpenMP parallel master taskloop simd directive. + */ + CXCursor_OMPParallelMasterTaskLoopSimdDirective = 284, + + /** OpenMP parallel master directive. + */ + CXCursor_OMPParallelMasterDirective = 285, + + /** OpenMP depobj directive. + */ + CXCursor_OMPDepobjDirective = 286, + + /** OpenMP scan directive. + */ + CXCursor_OMPScanDirective = 287, + + /** OpenMP tile directive. + */ + CXCursor_OMPTileDirective = 288, + + /** OpenMP canonical loop. + */ + CXCursor_OMPCanonicalLoop = 289, + + /** OpenMP interop directive. + */ + CXCursor_OMPInteropDirective = 290, + + /** OpenMP dispatch directive. + */ + CXCursor_OMPDispatchDirective = 291, + + /** OpenMP masked directive. + */ + CXCursor_OMPMaskedDirective = 292, + + /** OpenMP unroll directive. + */ + CXCursor_OMPUnrollDirective = 293, + + /** OpenMP metadirective directive. + */ + CXCursor_OMPMetaDirective = 294, + + /** OpenMP loop directive. + */ + CXCursor_OMPGenericLoopDirective = 295, + + /** OpenMP teams loop directive. + */ + CXCursor_OMPTeamsGenericLoopDirective = 296, + + /** OpenMP target teams loop directive. + */ + CXCursor_OMPTargetTeamsGenericLoopDirective = 297, + + /** OpenMP parallel loop directive. + */ + CXCursor_OMPParallelGenericLoopDirective = 298, + + /** OpenMP target parallel loop directive. + */ + CXCursor_OMPTargetParallelGenericLoopDirective = 299, + + /** OpenMP parallel masked directive. + */ + CXCursor_OMPParallelMaskedDirective = 300, + + /** OpenMP masked taskloop directive. + */ + CXCursor_OMPMaskedTaskLoopDirective = 301, + + /** OpenMP masked taskloop simd directive. + */ + CXCursor_OMPMaskedTaskLoopSimdDirective = 302, + + /** OpenMP parallel masked taskloop directive. + */ + CXCursor_OMPParallelMaskedTaskLoopDirective = 303, + + /** OpenMP parallel masked taskloop simd directive. + */ + CXCursor_OMPParallelMaskedTaskLoopSimdDirective = 304, + + /** OpenMP error directive. + */ + CXCursor_OMPErrorDirective = 305, + + /** OpenMP scope directive. + */ + CXCursor_OMPScopeDirective = 306, + + /** OpenMP reverse directive. + */ + CXCursor_OMPReverseDirective = 307, + + /** OpenMP interchange directive. + */ + CXCursor_OMPInterchangeDirective = 308, + + /** OpenMP assume directive. + */ + CXCursor_OMPAssumeDirective = 309, + + /** OpenACC Compute Construct. + */ + CXCursor_OpenACCComputeConstruct = 320, + + /** OpenACC Loop Construct. + */ + CXCursor_OpenACCLoopConstruct = 321, + + /** OpenACC Combined Constructs. + */ + CXCursor_OpenACCCombinedConstruct = 322, + + /** OpenACC data Construct. + */ + CXCursor_OpenACCDataConstruct = 323, + + /** OpenACC enter data Construct. + */ + CXCursor_OpenACCEnterDataConstruct = 324, + + /** OpenACC exit data Construct. + */ + CXCursor_OpenACCExitDataConstruct = 325, + + /** OpenACC host_data Construct. + */ + CXCursor_OpenACCHostDataConstruct = 326, + + /** OpenACC wait Construct. + */ + CXCursor_OpenACCWaitConstruct = 327, + + /** OpenACC init Construct. + */ + CXCursor_OpenACCInitConstruct = 328, + + /** OpenACC shutdown Construct. + */ + CXCursor_OpenACCShutdownConstruct = 329, + + /** OpenACC set Construct. + */ + CXCursor_OpenACCSetConstruct = 330, + + /** OpenACC update Construct. + */ + CXCursor_OpenACCUpdateConstruct = 331, + + CXCursor_LastStmt = CXCursor_OpenACCUpdateConstruct, + + /** + * Cursor that represents the translation unit itself. + * + * The translation unit cursor exists primarily to act as the root + * cursor for traversing the contents of a translation unit. + */ + CXCursor_TranslationUnit = 350, + + /* Attributes */ + CXCursor_FirstAttr = 400, + /** + * An attribute whose specific kind is not exposed via this + * interface. + */ + CXCursor_UnexposedAttr = 400, + + CXCursor_IBActionAttr = 401, + CXCursor_IBOutletAttr = 402, + CXCursor_IBOutletCollectionAttr = 403, + CXCursor_CXXFinalAttr = 404, + CXCursor_CXXOverrideAttr = 405, + CXCursor_AnnotateAttr = 406, + CXCursor_AsmLabelAttr = 407, + CXCursor_PackedAttr = 408, + CXCursor_PureAttr = 409, + CXCursor_ConstAttr = 410, + CXCursor_NoDuplicateAttr = 411, + CXCursor_CUDAConstantAttr = 412, + CXCursor_CUDADeviceAttr = 413, + CXCursor_CUDAGlobalAttr = 414, + CXCursor_CUDAHostAttr = 415, + CXCursor_CUDASharedAttr = 416, + CXCursor_VisibilityAttr = 417, + CXCursor_DLLExport = 418, + CXCursor_DLLImport = 419, + CXCursor_NSReturnsRetained = 420, + CXCursor_NSReturnsNotRetained = 421, + CXCursor_NSReturnsAutoreleased = 422, + CXCursor_NSConsumesSelf = 423, + CXCursor_NSConsumed = 424, + CXCursor_ObjCException = 425, + CXCursor_ObjCNSObject = 426, + CXCursor_ObjCIndependentClass = 427, + CXCursor_ObjCPreciseLifetime = 428, + CXCursor_ObjCReturnsInnerPointer = 429, + CXCursor_ObjCRequiresSuper = 430, + CXCursor_ObjCRootClass = 431, + CXCursor_ObjCSubclassingRestricted = 432, + CXCursor_ObjCExplicitProtocolImpl = 433, + CXCursor_ObjCDesignatedInitializer = 434, + CXCursor_ObjCRuntimeVisible = 435, + CXCursor_ObjCBoxable = 436, + CXCursor_FlagEnum = 437, + CXCursor_ConvergentAttr = 438, + CXCursor_WarnUnusedAttr = 439, + CXCursor_WarnUnusedResultAttr = 440, + CXCursor_AlignedAttr = 441, + CXCursor_LastAttr = CXCursor_AlignedAttr, + + /* Preprocessing */ + CXCursor_PreprocessingDirective = 500, + CXCursor_MacroDefinition = 501, + CXCursor_MacroExpansion = 502, + CXCursor_MacroInstantiation = CXCursor_MacroExpansion, + CXCursor_InclusionDirective = 503, + CXCursor_FirstPreprocessing = CXCursor_PreprocessingDirective, + CXCursor_LastPreprocessing = CXCursor_InclusionDirective, + + /* Extra Declarations */ + /** + * A module import declaration. + */ + CXCursor_ModuleImportDecl = 600, + CXCursor_TypeAliasTemplateDecl = 601, + /** + * A static_assert or _Static_assert node + */ + CXCursor_StaticAssert = 602, + /** + * a friend declaration. + */ + CXCursor_FriendDecl = 603, + /** + * a concept declaration. + */ + CXCursor_ConceptDecl = 604, + + CXCursor_FirstExtraDecl = CXCursor_ModuleImportDecl, + CXCursor_LastExtraDecl = CXCursor_ConceptDecl, + + /** + * A code completion overload candidate. + */ + CXCursor_OverloadCandidate = 700 +}; + +/** + * A cursor representing some element in the abstract syntax tree for + * a translation unit. + * + * The cursor abstraction unifies the different kinds of entities in a + * program--declaration, statements, expressions, references to declarations, + * etc.--under a single "cursor" abstraction with a common set of operations. + * Common operation for a cursor include: getting the physical location in + * a source file where the cursor points, getting the name associated with a + * cursor, and retrieving cursors for any child nodes of a particular cursor. + * + * Cursors can be produced in two specific ways. + * clang_getTranslationUnitCursor() produces a cursor for a translation unit, + * from which one can use clang_visitChildren() to explore the rest of the + * translation unit. clang_getCursor() maps from a physical source location + * to the entity that resides at that location, allowing one to map from the + * source code into the AST. + */ +typedef struct { + enum CXCursorKind kind; + int xdata; + const void *data[3]; +} CXCursor; + +/** + * \defgroup CINDEX_CURSOR_MANIP Cursor manipulations + * + * @{ + */ + +/** + * Retrieve the NULL cursor, which represents no entity. + */ +CINDEX_LINKAGE CXCursor clang_getNullCursor(void); + +/** + * Retrieve the cursor that represents the given translation unit. + * + * The translation unit cursor can be used to start traversing the + * various declarations within the given translation unit. + */ +CINDEX_LINKAGE CXCursor clang_getTranslationUnitCursor(CXTranslationUnit); + +/** + * Determine whether two cursors are equivalent. + */ +CINDEX_LINKAGE unsigned clang_equalCursors(CXCursor, CXCursor); + +/** + * Returns non-zero if \p cursor is null. + */ +CINDEX_LINKAGE int clang_Cursor_isNull(CXCursor cursor); + +/** + * Compute a hash value for the given cursor. + */ +CINDEX_LINKAGE unsigned clang_hashCursor(CXCursor); + +/** + * Retrieve the kind of the given cursor. + */ +CINDEX_LINKAGE enum CXCursorKind clang_getCursorKind(CXCursor); + +/** + * Determine whether the given cursor kind represents a declaration. + */ +CINDEX_LINKAGE unsigned clang_isDeclaration(enum CXCursorKind); + +/** + * Determine whether the given declaration is invalid. + * + * A declaration is invalid if it could not be parsed successfully. + * + * \returns non-zero if the cursor represents a declaration and it is + * invalid, otherwise NULL. + */ +CINDEX_LINKAGE unsigned clang_isInvalidDeclaration(CXCursor); + +/** + * Determine whether the given cursor kind represents a simple + * reference. + * + * Note that other kinds of cursors (such as expressions) can also refer to + * other cursors. Use clang_getCursorReferenced() to determine whether a + * particular cursor refers to another entity. + */ +CINDEX_LINKAGE unsigned clang_isReference(enum CXCursorKind); + +/** + * Determine whether the given cursor kind represents an expression. + */ +CINDEX_LINKAGE unsigned clang_isExpression(enum CXCursorKind); + +/** + * Determine whether the given cursor kind represents a statement. + */ +CINDEX_LINKAGE unsigned clang_isStatement(enum CXCursorKind); + +/** + * Determine whether the given cursor kind represents an attribute. + */ +CINDEX_LINKAGE unsigned clang_isAttribute(enum CXCursorKind); + +/** + * Determine whether the given cursor has any attributes. + */ +CINDEX_LINKAGE unsigned clang_Cursor_hasAttrs(CXCursor C); + +/** + * Determine whether the given cursor kind represents an invalid + * cursor. + */ +CINDEX_LINKAGE unsigned clang_isInvalid(enum CXCursorKind); + +/** + * Determine whether the given cursor kind represents a translation + * unit. + */ +CINDEX_LINKAGE unsigned clang_isTranslationUnit(enum CXCursorKind); + +/*** + * Determine whether the given cursor represents a preprocessing + * element, such as a preprocessor directive or macro instantiation. + */ +CINDEX_LINKAGE unsigned clang_isPreprocessing(enum CXCursorKind); + +/*** + * Determine whether the given cursor represents a currently + * unexposed piece of the AST (e.g., CXCursor_UnexposedStmt). + */ +CINDEX_LINKAGE unsigned clang_isUnexposed(enum CXCursorKind); + +/** + * Describe the linkage of the entity referred to by a cursor. + */ +enum CXLinkageKind { + /** This value indicates that no linkage information is available + * for a provided CXCursor. */ + CXLinkage_Invalid, + /** + * This is the linkage for variables, parameters, and so on that + * have automatic storage. This covers normal (non-extern) local variables. + */ + CXLinkage_NoLinkage, + /** This is the linkage for static variables and static functions. */ + CXLinkage_Internal, + /** This is the linkage for entities with external linkage that live + * in C++ anonymous namespaces.*/ + CXLinkage_UniqueExternal, + /** This is the linkage for entities with true, external linkage. */ + CXLinkage_External +}; + +/** + * Determine the linkage of the entity referred to by a given cursor. + */ +CINDEX_LINKAGE enum CXLinkageKind clang_getCursorLinkage(CXCursor cursor); + +enum CXVisibilityKind { + /** This value indicates that no visibility information is available + * for a provided CXCursor. */ + CXVisibility_Invalid, + + /** Symbol not seen by the linker. */ + CXVisibility_Hidden, + /** Symbol seen by the linker but resolves to a symbol inside this object. */ + CXVisibility_Protected, + /** Symbol seen by the linker and acts like a normal symbol. */ + CXVisibility_Default +}; + +/** + * Describe the visibility of the entity referred to by a cursor. + * + * This returns the default visibility if not explicitly specified by + * a visibility attribute. The default visibility may be changed by + * commandline arguments. + * + * \param cursor The cursor to query. + * + * \returns The visibility of the cursor. + */ +CINDEX_LINKAGE enum CXVisibilityKind clang_getCursorVisibility(CXCursor cursor); + +/** + * Determine the availability of the entity that this cursor refers to, + * taking the current target platform into account. + * + * \param cursor The cursor to query. + * + * \returns The availability of the cursor. + */ +CINDEX_LINKAGE enum CXAvailabilityKind +clang_getCursorAvailability(CXCursor cursor); + +/** + * Describes the availability of a given entity on a particular platform, e.g., + * a particular class might only be available on Mac OS 10.7 or newer. + */ +typedef struct CXPlatformAvailability { + /** + * A string that describes the platform for which this structure + * provides availability information. + * + * Possible values are "ios" or "macos". + */ + CXString Platform; + /** + * The version number in which this entity was introduced. + */ + CXVersion Introduced; + /** + * The version number in which this entity was deprecated (but is + * still available). + */ + CXVersion Deprecated; + /** + * The version number in which this entity was obsoleted, and therefore + * is no longer available. + */ + CXVersion Obsoleted; + /** + * Whether the entity is unconditionally unavailable on this platform. + */ + int Unavailable; + /** + * An optional message to provide to a user of this API, e.g., to + * suggest replacement APIs. + */ + CXString Message; +} CXPlatformAvailability; + +/** + * Determine the availability of the entity that this cursor refers to + * on any platforms for which availability information is known. + * + * \param cursor The cursor to query. + * + * \param always_deprecated If non-NULL, will be set to indicate whether the + * entity is deprecated on all platforms. + * + * \param deprecated_message If non-NULL, will be set to the message text + * provided along with the unconditional deprecation of this entity. The client + * is responsible for deallocating this string. + * + * \param always_unavailable If non-NULL, will be set to indicate whether the + * entity is unavailable on all platforms. + * + * \param unavailable_message If non-NULL, will be set to the message text + * provided along with the unconditional unavailability of this entity. The + * client is responsible for deallocating this string. + * + * \param availability If non-NULL, an array of CXPlatformAvailability instances + * that will be populated with platform availability information, up to either + * the number of platforms for which availability information is available (as + * returned by this function) or \c availability_size, whichever is smaller. + * + * \param availability_size The number of elements available in the + * \c availability array. + * + * \returns The number of platforms (N) for which availability information is + * available (which is unrelated to \c availability_size). + * + * Note that the client is responsible for calling + * \c clang_disposeCXPlatformAvailability to free each of the + * platform-availability structures returned. There are + * \c min(N, availability_size) such structures. + */ +CINDEX_LINKAGE int clang_getCursorPlatformAvailability( + CXCursor cursor, int *always_deprecated, CXString *deprecated_message, + int *always_unavailable, CXString *unavailable_message, + CXPlatformAvailability *availability, int availability_size); + +/** + * Free the memory associated with a \c CXPlatformAvailability structure. + */ +CINDEX_LINKAGE void +clang_disposeCXPlatformAvailability(CXPlatformAvailability *availability); + +/** + * If cursor refers to a variable declaration and it has initializer returns + * cursor referring to the initializer otherwise return null cursor. + */ +CINDEX_LINKAGE CXCursor clang_Cursor_getVarDeclInitializer(CXCursor cursor); + +/** + * If cursor refers to a variable declaration that has global storage returns 1. + * If cursor refers to a variable declaration that doesn't have global storage + * returns 0. Otherwise returns -1. + */ +CINDEX_LINKAGE int clang_Cursor_hasVarDeclGlobalStorage(CXCursor cursor); + +/** + * If cursor refers to a variable declaration that has external storage + * returns 1. If cursor refers to a variable declaration that doesn't have + * external storage returns 0. Otherwise returns -1. + */ +CINDEX_LINKAGE int clang_Cursor_hasVarDeclExternalStorage(CXCursor cursor); + +/** + * Describe the "language" of the entity referred to by a cursor. + */ +enum CXLanguageKind { + CXLanguage_Invalid = 0, + CXLanguage_C, + CXLanguage_ObjC, + CXLanguage_CPlusPlus +}; + +/** + * Determine the "language" of the entity referred to by a given cursor. + */ +CINDEX_LINKAGE enum CXLanguageKind clang_getCursorLanguage(CXCursor cursor); + +/** + * Describe the "thread-local storage (TLS) kind" of the declaration + * referred to by a cursor. + */ +enum CXTLSKind { CXTLS_None = 0, CXTLS_Dynamic, CXTLS_Static }; + +/** + * Determine the "thread-local storage (TLS) kind" of the declaration + * referred to by a cursor. + */ +CINDEX_LINKAGE enum CXTLSKind clang_getCursorTLSKind(CXCursor cursor); + +/** + * Returns the translation unit that a cursor originated from. + */ +CINDEX_LINKAGE CXTranslationUnit clang_Cursor_getTranslationUnit(CXCursor); + +/** + * A fast container representing a set of CXCursors. + */ +typedef struct CXCursorSetImpl *CXCursorSet; + +/** + * Creates an empty CXCursorSet. + */ +CINDEX_LINKAGE CXCursorSet clang_createCXCursorSet(void); + +/** + * Disposes a CXCursorSet and releases its associated memory. + */ +CINDEX_LINKAGE void clang_disposeCXCursorSet(CXCursorSet cset); + +/** + * Queries a CXCursorSet to see if it contains a specific CXCursor. + * + * \returns non-zero if the set contains the specified cursor. + */ +CINDEX_LINKAGE unsigned clang_CXCursorSet_contains(CXCursorSet cset, + CXCursor cursor); + +/** + * Inserts a CXCursor into a CXCursorSet. + * + * \returns zero if the CXCursor was already in the set, and non-zero otherwise. + */ +CINDEX_LINKAGE unsigned clang_CXCursorSet_insert(CXCursorSet cset, + CXCursor cursor); + +/** + * Determine the semantic parent of the given cursor. + * + * The semantic parent of a cursor is the cursor that semantically contains + * the given \p cursor. For many declarations, the lexical and semantic parents + * are equivalent (the lexical parent is returned by + * \c clang_getCursorLexicalParent()). They diverge when declarations or + * definitions are provided out-of-line. For example: + * + * \code + * class C { + * void f(); + * }; + * + * void C::f() { } + * \endcode + * + * In the out-of-line definition of \c C::f, the semantic parent is + * the class \c C, of which this function is a member. The lexical parent is + * the place where the declaration actually occurs in the source code; in this + * case, the definition occurs in the translation unit. In general, the + * lexical parent for a given entity can change without affecting the semantics + * of the program, and the lexical parent of different declarations of the + * same entity may be different. Changing the semantic parent of a declaration, + * on the other hand, can have a major impact on semantics, and redeclarations + * of a particular entity should all have the same semantic context. + * + * In the example above, both declarations of \c C::f have \c C as their + * semantic context, while the lexical context of the first \c C::f is \c C + * and the lexical context of the second \c C::f is the translation unit. + * + * For global declarations, the semantic parent is the translation unit. + */ +CINDEX_LINKAGE CXCursor clang_getCursorSemanticParent(CXCursor cursor); + +/** + * Determine the lexical parent of the given cursor. + * + * The lexical parent of a cursor is the cursor in which the given \p cursor + * was actually written. For many declarations, the lexical and semantic parents + * are equivalent (the semantic parent is returned by + * \c clang_getCursorSemanticParent()). They diverge when declarations or + * definitions are provided out-of-line. For example: + * + * \code + * class C { + * void f(); + * }; + * + * void C::f() { } + * \endcode + * + * In the out-of-line definition of \c C::f, the semantic parent is + * the class \c C, of which this function is a member. The lexical parent is + * the place where the declaration actually occurs in the source code; in this + * case, the definition occurs in the translation unit. In general, the + * lexical parent for a given entity can change without affecting the semantics + * of the program, and the lexical parent of different declarations of the + * same entity may be different. Changing the semantic parent of a declaration, + * on the other hand, can have a major impact on semantics, and redeclarations + * of a particular entity should all have the same semantic context. + * + * In the example above, both declarations of \c C::f have \c C as their + * semantic context, while the lexical context of the first \c C::f is \c C + * and the lexical context of the second \c C::f is the translation unit. + * + * For declarations written in the global scope, the lexical parent is + * the translation unit. + */ +CINDEX_LINKAGE CXCursor clang_getCursorLexicalParent(CXCursor cursor); + +/** + * Determine the set of methods that are overridden by the given + * method. + * + * In both Objective-C and C++, a method (aka virtual member function, + * in C++) can override a virtual method in a base class. For + * Objective-C, a method is said to override any method in the class's + * base class, its protocols, or its categories' protocols, that has the same + * selector and is of the same kind (class or instance). + * If no such method exists, the search continues to the class's superclass, + * its protocols, and its categories, and so on. A method from an Objective-C + * implementation is considered to override the same methods as its + * corresponding method in the interface. + * + * For C++, a virtual member function overrides any virtual member + * function with the same signature that occurs in its base + * classes. With multiple inheritance, a virtual member function can + * override several virtual member functions coming from different + * base classes. + * + * In all cases, this function determines the immediate overridden + * method, rather than all of the overridden methods. For example, if + * a method is originally declared in a class A, then overridden in B + * (which in inherits from A) and also in C (which inherited from B), + * then the only overridden method returned from this function when + * invoked on C's method will be B's method. The client may then + * invoke this function again, given the previously-found overridden + * methods, to map out the complete method-override set. + * + * \param cursor A cursor representing an Objective-C or C++ + * method. This routine will compute the set of methods that this + * method overrides. + * + * \param overridden A pointer whose pointee will be replaced with a + * pointer to an array of cursors, representing the set of overridden + * methods. If there are no overridden methods, the pointee will be + * set to NULL. The pointee must be freed via a call to + * \c clang_disposeOverriddenCursors(). + * + * \param num_overridden A pointer to the number of overridden + * functions, will be set to the number of overridden functions in the + * array pointed to by \p overridden. + */ +CINDEX_LINKAGE void clang_getOverriddenCursors(CXCursor cursor, + CXCursor **overridden, + unsigned *num_overridden); + +/** + * Free the set of overridden cursors returned by \c + * clang_getOverriddenCursors(). + */ +CINDEX_LINKAGE void clang_disposeOverriddenCursors(CXCursor *overridden); + +/** + * Retrieve the file that is included by the given inclusion directive + * cursor. + */ +CINDEX_LINKAGE CXFile clang_getIncludedFile(CXCursor cursor); + +/** + * @} + */ + +/** + * \defgroup CINDEX_CURSOR_SOURCE Mapping between cursors and source code + * + * Cursors represent a location within the Abstract Syntax Tree (AST). These + * routines help map between cursors and the physical locations where the + * described entities occur in the source code. The mapping is provided in + * both directions, so one can map from source code to the AST and back. + * + * @{ + */ + +/** + * Map a source location to the cursor that describes the entity at that + * location in the source code. + * + * clang_getCursor() maps an arbitrary source location within a translation + * unit down to the most specific cursor that describes the entity at that + * location. For example, given an expression \c x + y, invoking + * clang_getCursor() with a source location pointing to "x" will return the + * cursor for "x"; similarly for "y". If the cursor points anywhere between + * "x" or "y" (e.g., on the + or the whitespace around it), clang_getCursor() + * will return a cursor referring to the "+" expression. + * + * \returns a cursor representing the entity at the given source location, or + * a NULL cursor if no such entity can be found. + */ +CINDEX_LINKAGE CXCursor clang_getCursor(CXTranslationUnit, CXSourceLocation); + +/** + * Retrieve the physical location of the source constructor referenced + * by the given cursor. + * + * The location of a declaration is typically the location of the name of that + * declaration, where the name of that declaration would occur if it is + * unnamed, or some keyword that introduces that particular declaration. + * The location of a reference is where that reference occurs within the + * source code. + */ +CINDEX_LINKAGE CXSourceLocation clang_getCursorLocation(CXCursor); + +/** + * Retrieve the physical extent of the source construct referenced by + * the given cursor. + * + * The extent of a cursor starts with the file/line/column pointing at the + * first character within the source construct that the cursor refers to and + * ends with the last character within that source construct. For a + * declaration, the extent covers the declaration itself. For a reference, + * the extent covers the location of the reference (e.g., where the referenced + * entity was actually used). + */ +CINDEX_LINKAGE CXSourceRange clang_getCursorExtent(CXCursor); + +/** + * @} + */ + +/** + * \defgroup CINDEX_TYPES Type information for CXCursors + * + * @{ + */ + +/** + * Describes the kind of type + */ +enum CXTypeKind { + /** + * Represents an invalid type (e.g., where no type is available). + */ + CXType_Invalid = 0, + + /** + * A type whose specific kind is not exposed via this + * interface. + */ + CXType_Unexposed = 1, + + /* Builtin types */ + CXType_Void = 2, + CXType_Bool = 3, + CXType_Char_U = 4, + CXType_UChar = 5, + CXType_Char16 = 6, + CXType_Char32 = 7, + CXType_UShort = 8, + CXType_UInt = 9, + CXType_ULong = 10, + CXType_ULongLong = 11, + CXType_UInt128 = 12, + CXType_Char_S = 13, + CXType_SChar = 14, + CXType_WChar = 15, + CXType_Short = 16, + CXType_Int = 17, + CXType_Long = 18, + CXType_LongLong = 19, + CXType_Int128 = 20, + CXType_Float = 21, + CXType_Double = 22, + CXType_LongDouble = 23, + CXType_NullPtr = 24, + CXType_Overload = 25, + CXType_Dependent = 26, + CXType_ObjCId = 27, + CXType_ObjCClass = 28, + CXType_ObjCSel = 29, + CXType_Float128 = 30, + CXType_Half = 31, + CXType_Float16 = 32, + CXType_ShortAccum = 33, + CXType_Accum = 34, + CXType_LongAccum = 35, + CXType_UShortAccum = 36, + CXType_UAccum = 37, + CXType_ULongAccum = 38, + CXType_BFloat16 = 39, + CXType_Ibm128 = 40, + CXType_FirstBuiltin = CXType_Void, + CXType_LastBuiltin = CXType_Ibm128, + + CXType_Complex = 100, + CXType_Pointer = 101, + CXType_BlockPointer = 102, + CXType_LValueReference = 103, + CXType_RValueReference = 104, + CXType_Record = 105, + CXType_Enum = 106, + CXType_Typedef = 107, + CXType_ObjCInterface = 108, + CXType_ObjCObjectPointer = 109, + CXType_FunctionNoProto = 110, + CXType_FunctionProto = 111, + CXType_ConstantArray = 112, + CXType_Vector = 113, + CXType_IncompleteArray = 114, + CXType_VariableArray = 115, + CXType_DependentSizedArray = 116, + CXType_MemberPointer = 117, + CXType_Auto = 118, + + /** + * Represents a type that was referred to using an elaborated type keyword. + * + * E.g., struct S, or via a qualified name, e.g., N::M::type, or both. + */ + CXType_Elaborated = 119, + + /* OpenCL PipeType. */ + CXType_Pipe = 120, + + /* OpenCL builtin types. */ + CXType_OCLImage1dRO = 121, + CXType_OCLImage1dArrayRO = 122, + CXType_OCLImage1dBufferRO = 123, + CXType_OCLImage2dRO = 124, + CXType_OCLImage2dArrayRO = 125, + CXType_OCLImage2dDepthRO = 126, + CXType_OCLImage2dArrayDepthRO = 127, + CXType_OCLImage2dMSAARO = 128, + CXType_OCLImage2dArrayMSAARO = 129, + CXType_OCLImage2dMSAADepthRO = 130, + CXType_OCLImage2dArrayMSAADepthRO = 131, + CXType_OCLImage3dRO = 132, + CXType_OCLImage1dWO = 133, + CXType_OCLImage1dArrayWO = 134, + CXType_OCLImage1dBufferWO = 135, + CXType_OCLImage2dWO = 136, + CXType_OCLImage2dArrayWO = 137, + CXType_OCLImage2dDepthWO = 138, + CXType_OCLImage2dArrayDepthWO = 139, + CXType_OCLImage2dMSAAWO = 140, + CXType_OCLImage2dArrayMSAAWO = 141, + CXType_OCLImage2dMSAADepthWO = 142, + CXType_OCLImage2dArrayMSAADepthWO = 143, + CXType_OCLImage3dWO = 144, + CXType_OCLImage1dRW = 145, + CXType_OCLImage1dArrayRW = 146, + CXType_OCLImage1dBufferRW = 147, + CXType_OCLImage2dRW = 148, + CXType_OCLImage2dArrayRW = 149, + CXType_OCLImage2dDepthRW = 150, + CXType_OCLImage2dArrayDepthRW = 151, + CXType_OCLImage2dMSAARW = 152, + CXType_OCLImage2dArrayMSAARW = 153, + CXType_OCLImage2dMSAADepthRW = 154, + CXType_OCLImage2dArrayMSAADepthRW = 155, + CXType_OCLImage3dRW = 156, + CXType_OCLSampler = 157, + CXType_OCLEvent = 158, + CXType_OCLQueue = 159, + CXType_OCLReserveID = 160, + + CXType_ObjCObject = 161, + CXType_ObjCTypeParam = 162, + CXType_Attributed = 163, + + CXType_OCLIntelSubgroupAVCMcePayload = 164, + CXType_OCLIntelSubgroupAVCImePayload = 165, + CXType_OCLIntelSubgroupAVCRefPayload = 166, + CXType_OCLIntelSubgroupAVCSicPayload = 167, + CXType_OCLIntelSubgroupAVCMceResult = 168, + CXType_OCLIntelSubgroupAVCImeResult = 169, + CXType_OCLIntelSubgroupAVCRefResult = 170, + CXType_OCLIntelSubgroupAVCSicResult = 171, + CXType_OCLIntelSubgroupAVCImeResultSingleReferenceStreamout = 172, + CXType_OCLIntelSubgroupAVCImeResultDualReferenceStreamout = 173, + CXType_OCLIntelSubgroupAVCImeSingleReferenceStreamin = 174, + CXType_OCLIntelSubgroupAVCImeDualReferenceStreamin = 175, + + /* Old aliases for AVC OpenCL extension types. */ + CXType_OCLIntelSubgroupAVCImeResultSingleRefStreamout = 172, + CXType_OCLIntelSubgroupAVCImeResultDualRefStreamout = 173, + CXType_OCLIntelSubgroupAVCImeSingleRefStreamin = 174, + CXType_OCLIntelSubgroupAVCImeDualRefStreamin = 175, + + CXType_ExtVector = 176, + CXType_Atomic = 177, + CXType_BTFTagAttributed = 178, + + /* HLSL Types */ + CXType_HLSLResource = 179, + CXType_HLSLAttributedResource = 180 +}; + +/** + * Describes the calling convention of a function type + */ +enum CXCallingConv { + CXCallingConv_Default = 0, + CXCallingConv_C = 1, + CXCallingConv_X86StdCall = 2, + CXCallingConv_X86FastCall = 3, + CXCallingConv_X86ThisCall = 4, + CXCallingConv_X86Pascal = 5, + CXCallingConv_AAPCS = 6, + CXCallingConv_AAPCS_VFP = 7, + CXCallingConv_X86RegCall = 8, + CXCallingConv_IntelOclBicc = 9, + CXCallingConv_Win64 = 10, + /* Alias for compatibility with older versions of API. */ + CXCallingConv_X86_64Win64 = CXCallingConv_Win64, + CXCallingConv_X86_64SysV = 11, + CXCallingConv_X86VectorCall = 12, + CXCallingConv_Swift = 13, + CXCallingConv_PreserveMost = 14, + CXCallingConv_PreserveAll = 15, + CXCallingConv_AArch64VectorCall = 16, + CXCallingConv_SwiftAsync = 17, + CXCallingConv_AArch64SVEPCS = 18, + CXCallingConv_M68kRTD = 19, + CXCallingConv_PreserveNone = 20, + CXCallingConv_RISCVVectorCall = 21, + + CXCallingConv_Invalid = 100, + CXCallingConv_Unexposed = 200 +}; + +/** + * The type of an element in the abstract syntax tree. + * + */ +typedef struct { + enum CXTypeKind kind; + void *data[2]; +} CXType; + +/** + * Retrieve the type of a CXCursor (if any). + */ +CINDEX_LINKAGE CXType clang_getCursorType(CXCursor C); + +/** + * Pretty-print the underlying type using the rules of the + * language of the translation unit from which it came. + * + * If the type is invalid, an empty string is returned. + */ +CINDEX_LINKAGE CXString clang_getTypeSpelling(CXType CT); + +/** + * Retrieve the underlying type of a typedef declaration. + * + * If the cursor does not reference a typedef declaration, an invalid type is + * returned. + */ +CINDEX_LINKAGE CXType clang_getTypedefDeclUnderlyingType(CXCursor C); + +/** + * Retrieve the integer type of an enum declaration. + * + * If the cursor does not reference an enum declaration, an invalid type is + * returned. + */ +CINDEX_LINKAGE CXType clang_getEnumDeclIntegerType(CXCursor C); + +/** + * Retrieve the integer value of an enum constant declaration as a signed + * long long. + * + * If the cursor does not reference an enum constant declaration, LLONG_MIN is + * returned. Since this is also potentially a valid constant value, the kind of + * the cursor must be verified before calling this function. + */ +CINDEX_LINKAGE long long clang_getEnumConstantDeclValue(CXCursor C); + +/** + * Retrieve the integer value of an enum constant declaration as an unsigned + * long long. + * + * If the cursor does not reference an enum constant declaration, ULLONG_MAX is + * returned. Since this is also potentially a valid constant value, the kind of + * the cursor must be verified before calling this function. + */ +CINDEX_LINKAGE unsigned long long +clang_getEnumConstantDeclUnsignedValue(CXCursor C); + +/** + * Returns non-zero if the cursor specifies a Record member that is a bit-field. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isBitField(CXCursor C); + +/** + * Retrieve the bit width of a bit-field declaration as an integer. + * + * If the cursor does not reference a bit-field, or if the bit-field's width + * expression cannot be evaluated, -1 is returned. + * + * For example: + * \code + * if (clang_Cursor_isBitField(Cursor)) { + * int Width = clang_getFieldDeclBitWidth(Cursor); + * if (Width != -1) { + * // The bit-field width is not value-dependent. + * } + * } + * \endcode + */ +CINDEX_LINKAGE int clang_getFieldDeclBitWidth(CXCursor C); + +/** + * Retrieve the number of non-variadic arguments associated with a given + * cursor. + * + * The number of arguments can be determined for calls as well as for + * declarations of functions or methods. For other cursors -1 is returned. + */ +CINDEX_LINKAGE int clang_Cursor_getNumArguments(CXCursor C); + +/** + * Retrieve the argument cursor of a function or method. + * + * The argument cursor can be determined for calls as well as for declarations + * of functions or methods. For other cursors and for invalid indices, an + * invalid cursor is returned. + */ +CINDEX_LINKAGE CXCursor clang_Cursor_getArgument(CXCursor C, unsigned i); + +/** + * Describes the kind of a template argument. + * + * See the definition of llvm::clang::TemplateArgument::ArgKind for full + * element descriptions. + */ +enum CXTemplateArgumentKind { + CXTemplateArgumentKind_Null, + CXTemplateArgumentKind_Type, + CXTemplateArgumentKind_Declaration, + CXTemplateArgumentKind_NullPtr, + CXTemplateArgumentKind_Integral, + CXTemplateArgumentKind_Template, + CXTemplateArgumentKind_TemplateExpansion, + CXTemplateArgumentKind_Expression, + CXTemplateArgumentKind_Pack, + /* Indicates an error case, preventing the kind from being deduced. */ + CXTemplateArgumentKind_Invalid +}; + +/** + * Returns the number of template args of a function, struct, or class decl + * representing a template specialization. + * + * If the argument cursor cannot be converted into a template function + * declaration, -1 is returned. + * + * For example, for the following declaration and specialization: + * template + * void foo() { ... } + * + * template <> + * void foo(); + * + * The value 3 would be returned from this call. + */ +CINDEX_LINKAGE int clang_Cursor_getNumTemplateArguments(CXCursor C); + +/** + * Retrieve the kind of the I'th template argument of the CXCursor C. + * + * If the argument CXCursor does not represent a FunctionDecl, StructDecl, or + * ClassTemplatePartialSpecialization, an invalid template argument kind is + * returned. + * + * For example, for the following declaration and specialization: + * template + * void foo() { ... } + * + * template <> + * void foo(); + * + * For I = 0, 1, and 2, Type, Integral, and Integral will be returned, + * respectively. + */ +CINDEX_LINKAGE enum CXTemplateArgumentKind +clang_Cursor_getTemplateArgumentKind(CXCursor C, unsigned I); + +/** + * Retrieve a CXType representing the type of a TemplateArgument of a + * function decl representing a template specialization. + * + * If the argument CXCursor does not represent a FunctionDecl, StructDecl, + * ClassDecl or ClassTemplatePartialSpecialization whose I'th template argument + * has a kind of CXTemplateArgKind_Integral, an invalid type is returned. + * + * For example, for the following declaration and specialization: + * template + * void foo() { ... } + * + * template <> + * void foo(); + * + * If called with I = 0, "float", will be returned. + * Invalid types will be returned for I == 1 or 2. + */ +CINDEX_LINKAGE CXType clang_Cursor_getTemplateArgumentType(CXCursor C, + unsigned I); + +/** + * Retrieve the value of an Integral TemplateArgument (of a function + * decl representing a template specialization) as a signed long long. + * + * It is undefined to call this function on a CXCursor that does not represent a + * FunctionDecl, StructDecl, ClassDecl or ClassTemplatePartialSpecialization + * whose I'th template argument is not an integral value. + * + * For example, for the following declaration and specialization: + * template + * void foo() { ... } + * + * template <> + * void foo(); + * + * If called with I = 1 or 2, -7 or true will be returned, respectively. + * For I == 0, this function's behavior is undefined. + */ +CINDEX_LINKAGE long long clang_Cursor_getTemplateArgumentValue(CXCursor C, + unsigned I); + +/** + * Retrieve the value of an Integral TemplateArgument (of a function + * decl representing a template specialization) as an unsigned long long. + * + * It is undefined to call this function on a CXCursor that does not represent a + * FunctionDecl, StructDecl, ClassDecl or ClassTemplatePartialSpecialization or + * whose I'th template argument is not an integral value. + * + * For example, for the following declaration and specialization: + * template + * void foo() { ... } + * + * template <> + * void foo(); + * + * If called with I = 1 or 2, 2147483649 or true will be returned, respectively. + * For I == 0, this function's behavior is undefined. + */ +CINDEX_LINKAGE unsigned long long +clang_Cursor_getTemplateArgumentUnsignedValue(CXCursor C, unsigned I); + +/** + * Determine whether two CXTypes represent the same type. + * + * \returns non-zero if the CXTypes represent the same type and + * zero otherwise. + */ +CINDEX_LINKAGE unsigned clang_equalTypes(CXType A, CXType B); + +/** + * Return the canonical type for a CXType. + * + * Clang's type system explicitly models typedefs and all the ways + * a specific type can be represented. The canonical type is the underlying + * type with all the "sugar" removed. For example, if 'T' is a typedef + * for 'int', the canonical type for 'T' would be 'int'. + */ +CINDEX_LINKAGE CXType clang_getCanonicalType(CXType T); + +/** + * Determine whether a CXType has the "const" qualifier set, + * without looking through typedefs that may have added "const" at a + * different level. + */ +CINDEX_LINKAGE unsigned clang_isConstQualifiedType(CXType T); + +/** + * Determine whether a CXCursor that is a macro, is + * function like. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isMacroFunctionLike(CXCursor C); + +/** + * Determine whether a CXCursor that is a macro, is a + * builtin one. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isMacroBuiltin(CXCursor C); + +/** + * Determine whether a CXCursor that is a function declaration, is an + * inline declaration. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isFunctionInlined(CXCursor C); + +/** + * Determine whether a CXType has the "volatile" qualifier set, + * without looking through typedefs that may have added "volatile" at + * a different level. + */ +CINDEX_LINKAGE unsigned clang_isVolatileQualifiedType(CXType T); + +/** + * Determine whether a CXType has the "restrict" qualifier set, + * without looking through typedefs that may have added "restrict" at a + * different level. + */ +CINDEX_LINKAGE unsigned clang_isRestrictQualifiedType(CXType T); + +/** + * Returns the address space of the given type. + */ +CINDEX_LINKAGE unsigned clang_getAddressSpace(CXType T); + +/** + * Returns the typedef name of the given type. + */ +CINDEX_LINKAGE CXString clang_getTypedefName(CXType CT); + +/** + * For pointer types, returns the type of the pointee. + */ +CINDEX_LINKAGE CXType clang_getPointeeType(CXType T); + +/** + * Retrieve the unqualified variant of the given type, removing as + * little sugar as possible. + * + * For example, given the following series of typedefs: + * + * \code + * typedef int Integer; + * typedef const Integer CInteger; + * typedef CInteger DifferenceType; + * \endcode + * + * Executing \c clang_getUnqualifiedType() on a \c CXType that + * represents \c DifferenceType, will desugar to a type representing + * \c Integer, that has no qualifiers. + * + * And, executing \c clang_getUnqualifiedType() on the type of the + * first argument of the following function declaration: + * + * \code + * void foo(const int); + * \endcode + * + * Will return a type representing \c int, removing the \c const + * qualifier. + * + * Sugar over array types is not desugared. + * + * A type can be checked for qualifiers with \c + * clang_isConstQualifiedType(), \c clang_isVolatileQualifiedType() + * and \c clang_isRestrictQualifiedType(). + * + * A type that resulted from a call to \c clang_getUnqualifiedType + * will return \c false for all of the above calls. + */ +CINDEX_LINKAGE CXType clang_getUnqualifiedType(CXType CT); + +/** + * For reference types (e.g., "const int&"), returns the type that the + * reference refers to (e.g "const int"). + * + * Otherwise, returns the type itself. + * + * A type that has kind \c CXType_LValueReference or + * \c CXType_RValueReference is a reference type. + */ +CINDEX_LINKAGE CXType clang_getNonReferenceType(CXType CT); + +/** + * Return the cursor for the declaration of the given type. + */ +CINDEX_LINKAGE CXCursor clang_getTypeDeclaration(CXType T); + +/** + * Returns the Objective-C type encoding for the specified declaration. + */ +CINDEX_LINKAGE CXString clang_getDeclObjCTypeEncoding(CXCursor C); + +/** + * Returns the Objective-C type encoding for the specified CXType. + */ +CINDEX_LINKAGE CXString clang_Type_getObjCEncoding(CXType type); + +/** + * Retrieve the spelling of a given CXTypeKind. + */ +CINDEX_LINKAGE CXString clang_getTypeKindSpelling(enum CXTypeKind K); + +/** + * Retrieve the calling convention associated with a function type. + * + * If a non-function type is passed in, CXCallingConv_Invalid is returned. + */ +CINDEX_LINKAGE enum CXCallingConv clang_getFunctionTypeCallingConv(CXType T); + +/** + * Retrieve the return type associated with a function type. + * + * If a non-function type is passed in, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_getResultType(CXType T); + +/** + * Retrieve the exception specification type associated with a function type. + * This is a value of type CXCursor_ExceptionSpecificationKind. + * + * If a non-function type is passed in, an error code of -1 is returned. + */ +CINDEX_LINKAGE int clang_getExceptionSpecificationType(CXType T); + +/** + * Retrieve the number of non-variadic parameters associated with a + * function type. + * + * If a non-function type is passed in, -1 is returned. + */ +CINDEX_LINKAGE int clang_getNumArgTypes(CXType T); + +/** + * Retrieve the type of a parameter of a function type. + * + * If a non-function type is passed in or the function does not have enough + * parameters, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_getArgType(CXType T, unsigned i); + +/** + * Retrieves the base type of the ObjCObjectType. + * + * If the type is not an ObjC object, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_Type_getObjCObjectBaseType(CXType T); + +/** + * Retrieve the number of protocol references associated with an ObjC object/id. + * + * If the type is not an ObjC object, 0 is returned. + */ +CINDEX_LINKAGE unsigned clang_Type_getNumObjCProtocolRefs(CXType T); + +/** + * Retrieve the decl for a protocol reference for an ObjC object/id. + * + * If the type is not an ObjC object or there are not enough protocol + * references, an invalid cursor is returned. + */ +CINDEX_LINKAGE CXCursor clang_Type_getObjCProtocolDecl(CXType T, unsigned i); + +/** + * Retrieve the number of type arguments associated with an ObjC object. + * + * If the type is not an ObjC object, 0 is returned. + */ +CINDEX_LINKAGE unsigned clang_Type_getNumObjCTypeArgs(CXType T); + +/** + * Retrieve a type argument associated with an ObjC object. + * + * If the type is not an ObjC or the index is not valid, + * an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_Type_getObjCTypeArg(CXType T, unsigned i); + +/** + * Return 1 if the CXType is a variadic function type, and 0 otherwise. + */ +CINDEX_LINKAGE unsigned clang_isFunctionTypeVariadic(CXType T); + +/** + * Retrieve the return type associated with a given cursor. + * + * This only returns a valid type if the cursor refers to a function or method. + */ +CINDEX_LINKAGE CXType clang_getCursorResultType(CXCursor C); + +/** + * Retrieve the exception specification type associated with a given cursor. + * This is a value of type CXCursor_ExceptionSpecificationKind. + * + * This only returns a valid result if the cursor refers to a function or + * method. + */ +CINDEX_LINKAGE int clang_getCursorExceptionSpecificationType(CXCursor C); + +/** + * Return 1 if the CXType is a POD (plain old data) type, and 0 + * otherwise. + */ +CINDEX_LINKAGE unsigned clang_isPODType(CXType T); + +/** + * Return the element type of an array, complex, or vector type. + * + * If a type is passed in that is not an array, complex, or vector type, + * an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_getElementType(CXType T); + +/** + * Return the number of elements of an array or vector type. + * + * If a type is passed in that is not an array or vector type, + * -1 is returned. + */ +CINDEX_LINKAGE long long clang_getNumElements(CXType T); + +/** + * Return the element type of an array type. + * + * If a non-array type is passed in, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_getArrayElementType(CXType T); + +/** + * Return the array size of a constant array. + * + * If a non-array type is passed in, -1 is returned. + */ +CINDEX_LINKAGE long long clang_getArraySize(CXType T); + +/** + * Retrieve the type named by the qualified-id. + * + * If a non-elaborated type is passed in, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_Type_getNamedType(CXType T); + +/** + * Determine if a typedef is 'transparent' tag. + * + * A typedef is considered 'transparent' if it shares a name and spelling + * location with its underlying tag type, as is the case with the NS_ENUM macro. + * + * \returns non-zero if transparent and zero otherwise. + */ +CINDEX_LINKAGE unsigned clang_Type_isTransparentTagTypedef(CXType T); + +enum CXTypeNullabilityKind { + /** + * Values of this type can never be null. + */ + CXTypeNullability_NonNull = 0, + /** + * Values of this type can be null. + */ + CXTypeNullability_Nullable = 1, + /** + * Whether values of this type can be null is (explicitly) + * unspecified. This captures a (fairly rare) case where we + * can't conclude anything about the nullability of the type even + * though it has been considered. + */ + CXTypeNullability_Unspecified = 2, + /** + * Nullability is not applicable to this type. + */ + CXTypeNullability_Invalid = 3, + + /** + * Generally behaves like Nullable, except when used in a block parameter that + * was imported into a swift async method. There, swift will assume that the + * parameter can get null even if no error occurred. _Nullable parameters are + * assumed to only get null on error. + */ + CXTypeNullability_NullableResult = 4 +}; + +/** + * Retrieve the nullability kind of a pointer type. + */ +CINDEX_LINKAGE enum CXTypeNullabilityKind clang_Type_getNullability(CXType T); + +/** + * List the possible error codes for \c clang_Type_getSizeOf, + * \c clang_Type_getAlignOf, \c clang_Type_getOffsetOf, + * \c clang_Cursor_getOffsetOf, and \c clang_getOffsetOfBase. + * + * A value of this enumeration type can be returned if the target type is not + * a valid argument to sizeof, alignof or offsetof. + */ +enum CXTypeLayoutError { + /** + * Type is of kind CXType_Invalid. + */ + CXTypeLayoutError_Invalid = -1, + /** + * The type is an incomplete Type. + */ + CXTypeLayoutError_Incomplete = -2, + /** + * The type is a dependent Type. + */ + CXTypeLayoutError_Dependent = -3, + /** + * The type is not a constant size type. + */ + CXTypeLayoutError_NotConstantSize = -4, + /** + * The Field name is not valid for this record. + */ + CXTypeLayoutError_InvalidFieldName = -5, + /** + * The type is undeduced. + */ + CXTypeLayoutError_Undeduced = -6 +}; + +/** + * Return the alignment of a type in bytes as per C++[expr.alignof] + * standard. + * + * If the type declaration is invalid, CXTypeLayoutError_Invalid is returned. + * If the type declaration is an incomplete type, CXTypeLayoutError_Incomplete + * is returned. + * If the type declaration is a dependent type, CXTypeLayoutError_Dependent is + * returned. + * If the type declaration is not a constant size type, + * CXTypeLayoutError_NotConstantSize is returned. + */ +CINDEX_LINKAGE long long clang_Type_getAlignOf(CXType T); + +/** + * Return the class type of an member pointer type. + * + * If a non-member-pointer type is passed in, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_Type_getClassType(CXType T); + +/** + * Return the size of a type in bytes as per C++[expr.sizeof] standard. + * + * If the type declaration is invalid, CXTypeLayoutError_Invalid is returned. + * If the type declaration is an incomplete type, CXTypeLayoutError_Incomplete + * is returned. + * If the type declaration is a dependent type, CXTypeLayoutError_Dependent is + * returned. + */ +CINDEX_LINKAGE long long clang_Type_getSizeOf(CXType T); + +/** + * Return the offset of a field named S in a record of type T in bits + * as it would be returned by __offsetof__ as per C++11[18.2p4] + * + * If the cursor is not a record field declaration, CXTypeLayoutError_Invalid + * is returned. + * If the field's type declaration is an incomplete type, + * CXTypeLayoutError_Incomplete is returned. + * If the field's type declaration is a dependent type, + * CXTypeLayoutError_Dependent is returned. + * If the field's name S is not found, + * CXTypeLayoutError_InvalidFieldName is returned. + */ +CINDEX_LINKAGE long long clang_Type_getOffsetOf(CXType T, const char *S); + +/** + * Return the type that was modified by this attributed type. + * + * If the type is not an attributed type, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_Type_getModifiedType(CXType T); + +/** + * Gets the type contained by this atomic type. + * + * If a non-atomic type is passed in, an invalid type is returned. + */ +CINDEX_LINKAGE CXType clang_Type_getValueType(CXType CT); + +/** + * Return the offset of the field represented by the Cursor. + * + * If the cursor is not a field declaration, -1 is returned. + * If the cursor semantic parent is not a record field declaration, + * CXTypeLayoutError_Invalid is returned. + * If the field's type declaration is an incomplete type, + * CXTypeLayoutError_Incomplete is returned. + * If the field's type declaration is a dependent type, + * CXTypeLayoutError_Dependent is returned. + * If the field's name S is not found, + * CXTypeLayoutError_InvalidFieldName is returned. + */ +CINDEX_LINKAGE long long clang_Cursor_getOffsetOfField(CXCursor C); + +/** + * Determine whether the given cursor represents an anonymous + * tag or namespace + */ +CINDEX_LINKAGE unsigned clang_Cursor_isAnonymous(CXCursor C); + +/** + * Determine whether the given cursor represents an anonymous record + * declaration. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isAnonymousRecordDecl(CXCursor C); + +/** + * Determine whether the given cursor represents an inline namespace + * declaration. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isInlineNamespace(CXCursor C); + +enum CXRefQualifierKind { + /** No ref-qualifier was provided. */ + CXRefQualifier_None = 0, + /** An lvalue ref-qualifier was provided (\c &). */ + CXRefQualifier_LValue, + /** An rvalue ref-qualifier was provided (\c &&). */ + CXRefQualifier_RValue +}; + +/** + * Returns the number of template arguments for given template + * specialization, or -1 if type \c T is not a template specialization. + */ +CINDEX_LINKAGE int clang_Type_getNumTemplateArguments(CXType T); + +/** + * Returns the type template argument of a template class specialization + * at given index. + * + * This function only returns template type arguments and does not handle + * template template arguments or variadic packs. + */ +CINDEX_LINKAGE CXType clang_Type_getTemplateArgumentAsType(CXType T, + unsigned i); + +/** + * Retrieve the ref-qualifier kind of a function or method. + * + * The ref-qualifier is returned for C++ functions or methods. For other types + * or non-C++ declarations, CXRefQualifier_None is returned. + */ +CINDEX_LINKAGE enum CXRefQualifierKind clang_Type_getCXXRefQualifier(CXType T); + +/** + * Returns 1 if the base class specified by the cursor with kind + * CX_CXXBaseSpecifier is virtual. + */ +CINDEX_LINKAGE unsigned clang_isVirtualBase(CXCursor); + +/** + * Returns the offset in bits of a CX_CXXBaseSpecifier relative to the parent + * class. + * + * Returns a small negative number if the offset cannot be computed. See + * CXTypeLayoutError for error codes. + */ +CINDEX_LINKAGE long long clang_getOffsetOfBase(CXCursor Parent, CXCursor Base); + +/** + * Represents the C++ access control level to a base class for a + * cursor with kind CX_CXXBaseSpecifier. + */ +enum CX_CXXAccessSpecifier { + CX_CXXInvalidAccessSpecifier, + CX_CXXPublic, + CX_CXXProtected, + CX_CXXPrivate +}; + +/** + * Returns the access control level for the referenced object. + * + * If the cursor refers to a C++ declaration, its access control level within + * its parent scope is returned. Otherwise, if the cursor refers to a base + * specifier or access specifier, the specifier itself is returned. + */ +CINDEX_LINKAGE enum CX_CXXAccessSpecifier clang_getCXXAccessSpecifier(CXCursor); + +/** + * Represents the storage classes as declared in the source. CX_SC_Invalid + * was added for the case that the passed cursor in not a declaration. + */ +enum CX_StorageClass { + CX_SC_Invalid, + CX_SC_None, + CX_SC_Extern, + CX_SC_Static, + CX_SC_PrivateExtern, + CX_SC_OpenCLWorkGroupLocal, + CX_SC_Auto, + CX_SC_Register +}; + +/** + * Represents a specific kind of binary operator which can appear at a cursor. + */ +enum CX_BinaryOperatorKind { + CX_BO_Invalid = 0, + CX_BO_PtrMemD = 1, + CX_BO_PtrMemI = 2, + CX_BO_Mul = 3, + CX_BO_Div = 4, + CX_BO_Rem = 5, + CX_BO_Add = 6, + CX_BO_Sub = 7, + CX_BO_Shl = 8, + CX_BO_Shr = 9, + CX_BO_Cmp = 10, + CX_BO_LT = 11, + CX_BO_GT = 12, + CX_BO_LE = 13, + CX_BO_GE = 14, + CX_BO_EQ = 15, + CX_BO_NE = 16, + CX_BO_And = 17, + CX_BO_Xor = 18, + CX_BO_Or = 19, + CX_BO_LAnd = 20, + CX_BO_LOr = 21, + CX_BO_Assign = 22, + CX_BO_MulAssign = 23, + CX_BO_DivAssign = 24, + CX_BO_RemAssign = 25, + CX_BO_AddAssign = 26, + CX_BO_SubAssign = 27, + CX_BO_ShlAssign = 28, + CX_BO_ShrAssign = 29, + CX_BO_AndAssign = 30, + CX_BO_XorAssign = 31, + CX_BO_OrAssign = 32, + CX_BO_Comma = 33, + CX_BO_LAST = CX_BO_Comma +}; + +/** + * \brief Returns the operator code for the binary operator. + */ +CINDEX_LINKAGE enum CX_BinaryOperatorKind +clang_Cursor_getBinaryOpcode(CXCursor C); + +/** + * \brief Returns a string containing the spelling of the binary operator. + */ +CINDEX_LINKAGE CXString +clang_Cursor_getBinaryOpcodeStr(enum CX_BinaryOperatorKind Op); + +/** + * Returns the storage class for a function or variable declaration. + * + * If the passed in Cursor is not a function or variable declaration, + * CX_SC_Invalid is returned else the storage class. + */ +CINDEX_LINKAGE enum CX_StorageClass clang_Cursor_getStorageClass(CXCursor); + +/** + * Determine the number of overloaded declarations referenced by a + * \c CXCursor_OverloadedDeclRef cursor. + * + * \param cursor The cursor whose overloaded declarations are being queried. + * + * \returns The number of overloaded declarations referenced by \c cursor. If it + * is not a \c CXCursor_OverloadedDeclRef cursor, returns 0. + */ +CINDEX_LINKAGE unsigned clang_getNumOverloadedDecls(CXCursor cursor); + +/** + * Retrieve a cursor for one of the overloaded declarations referenced + * by a \c CXCursor_OverloadedDeclRef cursor. + * + * \param cursor The cursor whose overloaded declarations are being queried. + * + * \param index The zero-based index into the set of overloaded declarations in + * the cursor. + * + * \returns A cursor representing the declaration referenced by the given + * \c cursor at the specified \c index. If the cursor does not have an + * associated set of overloaded declarations, or if the index is out of bounds, + * returns \c clang_getNullCursor(); + */ +CINDEX_LINKAGE CXCursor clang_getOverloadedDecl(CXCursor cursor, + unsigned index); + +/** + * @} + */ + +/** + * \defgroup CINDEX_ATTRIBUTES Information for attributes + * + * @{ + */ + +/** + * For cursors representing an iboutletcollection attribute, + * this function returns the collection element type. + * + */ +CINDEX_LINKAGE CXType clang_getIBOutletCollectionType(CXCursor); + +/** + * @} + */ + +/** + * \defgroup CINDEX_CURSOR_TRAVERSAL Traversing the AST with cursors + * + * These routines provide the ability to traverse the abstract syntax tree + * using cursors. + * + * @{ + */ + +/** + * Describes how the traversal of the children of a particular + * cursor should proceed after visiting a particular child cursor. + * + * A value of this enumeration type should be returned by each + * \c CXCursorVisitor to indicate how clang_visitChildren() proceed. + */ +enum CXChildVisitResult { + /** + * Terminates the cursor traversal. + */ + CXChildVisit_Break, + /** + * Continues the cursor traversal with the next sibling of + * the cursor just visited, without visiting its children. + */ + CXChildVisit_Continue, + /** + * Recursively traverse the children of this cursor, using + * the same visitor and client data. + */ + CXChildVisit_Recurse +}; + +/** + * Visitor invoked for each cursor found by a traversal. + * + * This visitor function will be invoked for each cursor found by + * clang_visitCursorChildren(). Its first argument is the cursor being + * visited, its second argument is the parent visitor for that cursor, + * and its third argument is the client data provided to + * clang_visitCursorChildren(). + * + * The visitor should return one of the \c CXChildVisitResult values + * to direct clang_visitCursorChildren(). + */ +typedef enum CXChildVisitResult (*CXCursorVisitor)(CXCursor cursor, + CXCursor parent, + CXClientData client_data); + +/** + * Visit the children of a particular cursor. + * + * This function visits all the direct children of the given cursor, + * invoking the given \p visitor function with the cursors of each + * visited child. The traversal may be recursive, if the visitor returns + * \c CXChildVisit_Recurse. The traversal may also be ended prematurely, if + * the visitor returns \c CXChildVisit_Break. + * + * \param parent the cursor whose child may be visited. All kinds of + * cursors can be visited, including invalid cursors (which, by + * definition, have no children). + * + * \param visitor the visitor function that will be invoked for each + * child of \p parent. + * + * \param client_data pointer data supplied by the client, which will + * be passed to the visitor each time it is invoked. + * + * \returns a non-zero value if the traversal was terminated + * prematurely by the visitor returning \c CXChildVisit_Break. + */ +CINDEX_LINKAGE unsigned clang_visitChildren(CXCursor parent, + CXCursorVisitor visitor, + CXClientData client_data); +/** + * Visitor invoked for each cursor found by a traversal. + * + * This visitor block will be invoked for each cursor found by + * clang_visitChildrenWithBlock(). Its first argument is the cursor being + * visited, its second argument is the parent visitor for that cursor. + * + * The visitor should return one of the \c CXChildVisitResult values + * to direct clang_visitChildrenWithBlock(). + */ +#if __has_feature(blocks) +typedef enum CXChildVisitResult (^CXCursorVisitorBlock)(CXCursor cursor, + CXCursor parent); +#else +typedef struct _CXChildVisitResult *CXCursorVisitorBlock; +#endif + +/** + * Visits the children of a cursor using the specified block. Behaves + * identically to clang_visitChildren() in all other respects. + */ +CINDEX_LINKAGE unsigned +clang_visitChildrenWithBlock(CXCursor parent, CXCursorVisitorBlock block); + +/** + * @} + */ + +/** + * \defgroup CINDEX_CURSOR_XREF Cross-referencing in the AST + * + * These routines provide the ability to determine references within and + * across translation units, by providing the names of the entities referenced + * by cursors, follow reference cursors to the declarations they reference, + * and associate declarations with their definitions. + * + * @{ + */ + +/** + * Retrieve a Unified Symbol Resolution (USR) for the entity referenced + * by the given cursor. + * + * A Unified Symbol Resolution (USR) is a string that identifies a particular + * entity (function, class, variable, etc.) within a program. USRs can be + * compared across translation units to determine, e.g., when references in + * one translation refer to an entity defined in another translation unit. + */ +CINDEX_LINKAGE CXString clang_getCursorUSR(CXCursor); + +/** + * Construct a USR for a specified Objective-C class. + */ +CINDEX_LINKAGE CXString clang_constructUSR_ObjCClass(const char *class_name); + +/** + * Construct a USR for a specified Objective-C category. + */ +CINDEX_LINKAGE CXString clang_constructUSR_ObjCCategory( + const char *class_name, const char *category_name); + +/** + * Construct a USR for a specified Objective-C protocol. + */ +CINDEX_LINKAGE CXString +clang_constructUSR_ObjCProtocol(const char *protocol_name); + +/** + * Construct a USR for a specified Objective-C instance variable and + * the USR for its containing class. + */ +CINDEX_LINKAGE CXString clang_constructUSR_ObjCIvar(const char *name, + CXString classUSR); + +/** + * Construct a USR for a specified Objective-C method and + * the USR for its containing class. + */ +CINDEX_LINKAGE CXString clang_constructUSR_ObjCMethod(const char *name, + unsigned isInstanceMethod, + CXString classUSR); + +/** + * Construct a USR for a specified Objective-C property and the USR + * for its containing class. + */ +CINDEX_LINKAGE CXString clang_constructUSR_ObjCProperty(const char *property, + CXString classUSR); + +/** + * Retrieve a name for the entity referenced by this cursor. + */ +CINDEX_LINKAGE CXString clang_getCursorSpelling(CXCursor); + +/** + * Retrieve a range for a piece that forms the cursors spelling name. + * Most of the times there is only one range for the complete spelling but for + * Objective-C methods and Objective-C message expressions, there are multiple + * pieces for each selector identifier. + * + * \param pieceIndex the index of the spelling name piece. If this is greater + * than the actual number of pieces, it will return a NULL (invalid) range. + * + * \param options Reserved. + */ +CINDEX_LINKAGE CXSourceRange clang_Cursor_getSpellingNameRange( + CXCursor, unsigned pieceIndex, unsigned options); + +/** + * Opaque pointer representing a policy that controls pretty printing + * for \c clang_getCursorPrettyPrinted. + */ +typedef void *CXPrintingPolicy; + +/** + * Properties for the printing policy. + * + * See \c clang::PrintingPolicy for more information. + */ +enum CXPrintingPolicyProperty { + CXPrintingPolicy_Indentation, + CXPrintingPolicy_SuppressSpecifiers, + CXPrintingPolicy_SuppressTagKeyword, + CXPrintingPolicy_IncludeTagDefinition, + CXPrintingPolicy_SuppressScope, + CXPrintingPolicy_SuppressUnwrittenScope, + CXPrintingPolicy_SuppressInitializers, + CXPrintingPolicy_ConstantArraySizeAsWritten, + CXPrintingPolicy_AnonymousTagLocations, + CXPrintingPolicy_SuppressStrongLifetime, + CXPrintingPolicy_SuppressLifetimeQualifiers, + CXPrintingPolicy_SuppressTemplateArgsInCXXConstructors, + CXPrintingPolicy_Bool, + CXPrintingPolicy_Restrict, + CXPrintingPolicy_Alignof, + CXPrintingPolicy_UnderscoreAlignof, + CXPrintingPolicy_UseVoidForZeroParams, + CXPrintingPolicy_TerseOutput, + CXPrintingPolicy_PolishForDeclaration, + CXPrintingPolicy_Half, + CXPrintingPolicy_MSWChar, + CXPrintingPolicy_IncludeNewlines, + CXPrintingPolicy_MSVCFormatting, + CXPrintingPolicy_ConstantsAsWritten, + CXPrintingPolicy_SuppressImplicitBase, + CXPrintingPolicy_FullyQualifiedName, + + CXPrintingPolicy_LastProperty = CXPrintingPolicy_FullyQualifiedName +}; + +/** + * Get a property value for the given printing policy. + */ +CINDEX_LINKAGE unsigned +clang_PrintingPolicy_getProperty(CXPrintingPolicy Policy, + enum CXPrintingPolicyProperty Property); + +/** + * Set a property value for the given printing policy. + */ +CINDEX_LINKAGE void +clang_PrintingPolicy_setProperty(CXPrintingPolicy Policy, + enum CXPrintingPolicyProperty Property, + unsigned Value); + +/** + * Retrieve the default policy for the cursor. + * + * The policy should be released after use with \c + * clang_PrintingPolicy_dispose. + */ +CINDEX_LINKAGE CXPrintingPolicy clang_getCursorPrintingPolicy(CXCursor); + +/** + * Release a printing policy. + */ +CINDEX_LINKAGE void clang_PrintingPolicy_dispose(CXPrintingPolicy Policy); + +/** + * Pretty print declarations. + * + * \param Cursor The cursor representing a declaration. + * + * \param Policy The policy to control the entities being printed. If + * NULL, a default policy is used. + * + * \returns The pretty printed declaration or the empty string for + * other cursors. + */ +CINDEX_LINKAGE CXString clang_getCursorPrettyPrinted(CXCursor Cursor, + CXPrintingPolicy Policy); + +/** + * Pretty-print the underlying type using a custom printing policy. + * + * If the type is invalid, an empty string is returned. + */ +CINDEX_LINKAGE CXString clang_getTypePrettyPrinted(CXType CT, + CXPrintingPolicy cxPolicy); + +/** + * Retrieve the display name for the entity referenced by this cursor. + * + * The display name contains extra information that helps identify the cursor, + * such as the parameters of a function or template or the arguments of a + * class template specialization. + */ +CINDEX_LINKAGE CXString clang_getCursorDisplayName(CXCursor); + +/** For a cursor that is a reference, retrieve a cursor representing the + * entity that it references. + * + * Reference cursors refer to other entities in the AST. For example, an + * Objective-C superclass reference cursor refers to an Objective-C class. + * This function produces the cursor for the Objective-C class from the + * cursor for the superclass reference. If the input cursor is a declaration or + * definition, it returns that declaration or definition unchanged. + * Otherwise, returns the NULL cursor. + */ +CINDEX_LINKAGE CXCursor clang_getCursorReferenced(CXCursor); + +/** + * For a cursor that is either a reference to or a declaration + * of some entity, retrieve a cursor that describes the definition of + * that entity. + * + * Some entities can be declared multiple times within a translation + * unit, but only one of those declarations can also be a + * definition. For example, given: + * + * \code + * int f(int, int); + * int g(int x, int y) { return f(x, y); } + * int f(int a, int b) { return a + b; } + * int f(int, int); + * \endcode + * + * there are three declarations of the function "f", but only the + * second one is a definition. The clang_getCursorDefinition() + * function will take any cursor pointing to a declaration of "f" + * (the first or fourth lines of the example) or a cursor referenced + * that uses "f" (the call to "f' inside "g") and will return a + * declaration cursor pointing to the definition (the second "f" + * declaration). + * + * If given a cursor for which there is no corresponding definition, + * e.g., because there is no definition of that entity within this + * translation unit, returns a NULL cursor. + */ +CINDEX_LINKAGE CXCursor clang_getCursorDefinition(CXCursor); + +/** + * Determine whether the declaration pointed to by this cursor + * is also a definition of that entity. + */ +CINDEX_LINKAGE unsigned clang_isCursorDefinition(CXCursor); + +/** + * Retrieve the canonical cursor corresponding to the given cursor. + * + * In the C family of languages, many kinds of entities can be declared several + * times within a single translation unit. For example, a structure type can + * be forward-declared (possibly multiple times) and later defined: + * + * \code + * struct X; + * struct X; + * struct X { + * int member; + * }; + * \endcode + * + * The declarations and the definition of \c X are represented by three + * different cursors, all of which are declarations of the same underlying + * entity. One of these cursor is considered the "canonical" cursor, which + * is effectively the representative for the underlying entity. One can + * determine if two cursors are declarations of the same underlying entity by + * comparing their canonical cursors. + * + * \returns The canonical cursor for the entity referred to by the given cursor. + */ +CINDEX_LINKAGE CXCursor clang_getCanonicalCursor(CXCursor); + +/** + * If the cursor points to a selector identifier in an Objective-C + * method or message expression, this returns the selector index. + * + * After getting a cursor with #clang_getCursor, this can be called to + * determine if the location points to a selector identifier. + * + * \returns The selector index if the cursor is an Objective-C method or message + * expression and the cursor is pointing to a selector identifier, or -1 + * otherwise. + */ +CINDEX_LINKAGE int clang_Cursor_getObjCSelectorIndex(CXCursor); + +/** + * Given a cursor pointing to a C++ method call or an Objective-C + * message, returns non-zero if the method/message is "dynamic", meaning: + * + * For a C++ method: the call is virtual. + * For an Objective-C message: the receiver is an object instance, not 'super' + * or a specific class. + * + * If the method/message is "static" or the cursor does not point to a + * method/message, it will return zero. + */ +CINDEX_LINKAGE int clang_Cursor_isDynamicCall(CXCursor C); + +/** + * Given a cursor pointing to an Objective-C message or property + * reference, or C++ method call, returns the CXType of the receiver. + */ +CINDEX_LINKAGE CXType clang_Cursor_getReceiverType(CXCursor C); + +/** + * Property attributes for a \c CXCursor_ObjCPropertyDecl. + */ +typedef enum { + CXObjCPropertyAttr_noattr = 0x00, + CXObjCPropertyAttr_readonly = 0x01, + CXObjCPropertyAttr_getter = 0x02, + CXObjCPropertyAttr_assign = 0x04, + CXObjCPropertyAttr_readwrite = 0x08, + CXObjCPropertyAttr_retain = 0x10, + CXObjCPropertyAttr_copy = 0x20, + CXObjCPropertyAttr_nonatomic = 0x40, + CXObjCPropertyAttr_setter = 0x80, + CXObjCPropertyAttr_atomic = 0x100, + CXObjCPropertyAttr_weak = 0x200, + CXObjCPropertyAttr_strong = 0x400, + CXObjCPropertyAttr_unsafe_unretained = 0x800, + CXObjCPropertyAttr_class = 0x1000 +} CXObjCPropertyAttrKind; + +/** + * Given a cursor that represents a property declaration, return the + * associated property attributes. The bits are formed from + * \c CXObjCPropertyAttrKind. + * + * \param reserved Reserved for future use, pass 0. + */ +CINDEX_LINKAGE unsigned +clang_Cursor_getObjCPropertyAttributes(CXCursor C, unsigned reserved); + +/** + * Given a cursor that represents a property declaration, return the + * name of the method that implements the getter. + */ +CINDEX_LINKAGE CXString clang_Cursor_getObjCPropertyGetterName(CXCursor C); + +/** + * Given a cursor that represents a property declaration, return the + * name of the method that implements the setter, if any. + */ +CINDEX_LINKAGE CXString clang_Cursor_getObjCPropertySetterName(CXCursor C); + +/** + * 'Qualifiers' written next to the return and parameter types in + * Objective-C method declarations. + */ +typedef enum { + CXObjCDeclQualifier_None = 0x0, + CXObjCDeclQualifier_In = 0x1, + CXObjCDeclQualifier_Inout = 0x2, + CXObjCDeclQualifier_Out = 0x4, + CXObjCDeclQualifier_Bycopy = 0x8, + CXObjCDeclQualifier_Byref = 0x10, + CXObjCDeclQualifier_Oneway = 0x20 +} CXObjCDeclQualifierKind; + +/** + * Given a cursor that represents an Objective-C method or parameter + * declaration, return the associated Objective-C qualifiers for the return + * type or the parameter respectively. The bits are formed from + * CXObjCDeclQualifierKind. + */ +CINDEX_LINKAGE unsigned clang_Cursor_getObjCDeclQualifiers(CXCursor C); + +/** + * Given a cursor that represents an Objective-C method or property + * declaration, return non-zero if the declaration was affected by "\@optional". + * Returns zero if the cursor is not such a declaration or it is "\@required". + */ +CINDEX_LINKAGE unsigned clang_Cursor_isObjCOptional(CXCursor C); + +/** + * Returns non-zero if the given cursor is a variadic function or method. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isVariadic(CXCursor C); + +/** + * Returns non-zero if the given cursor points to a symbol marked with + * external_source_symbol attribute. + * + * \param language If non-NULL, and the attribute is present, will be set to + * the 'language' string from the attribute. + * + * \param definedIn If non-NULL, and the attribute is present, will be set to + * the 'definedIn' string from the attribute. + * + * \param isGenerated If non-NULL, and the attribute is present, will be set to + * non-zero if the 'generated_declaration' is set in the attribute. + */ +CINDEX_LINKAGE unsigned clang_Cursor_isExternalSymbol(CXCursor C, + CXString *language, + CXString *definedIn, + unsigned *isGenerated); + +/** + * Given a cursor that represents a declaration, return the associated + * comment's source range. The range may include multiple consecutive comments + * with whitespace in between. + */ +CINDEX_LINKAGE CXSourceRange clang_Cursor_getCommentRange(CXCursor C); + +/** + * Given a cursor that represents a declaration, return the associated + * comment text, including comment markers. + */ +CINDEX_LINKAGE CXString clang_Cursor_getRawCommentText(CXCursor C); + +/** + * Given a cursor that represents a documentable entity (e.g., + * declaration), return the associated \paragraph; otherwise return the + * first paragraph. + */ +CINDEX_LINKAGE CXString clang_Cursor_getBriefCommentText(CXCursor C); + +/** + * @} + */ + +/** \defgroup CINDEX_MANGLE Name Mangling API Functions + * + * @{ + */ + +/** + * Retrieve the CXString representing the mangled name of the cursor. + */ +CINDEX_LINKAGE CXString clang_Cursor_getMangling(CXCursor); + +/** + * Retrieve the CXStrings representing the mangled symbols of the C++ + * constructor or destructor at the cursor. + */ +CINDEX_LINKAGE CXStringSet *clang_Cursor_getCXXManglings(CXCursor); + +/** + * Retrieve the CXStrings representing the mangled symbols of the ObjC + * class interface or implementation at the cursor. + */ +CINDEX_LINKAGE CXStringSet *clang_Cursor_getObjCManglings(CXCursor); + +/** + * @} + */ + +/** + * \defgroup CINDEX_MODULE Module introspection + * + * The functions in this group provide access to information about modules. + * + * @{ + */ + +typedef void *CXModule; + +/** + * Given a CXCursor_ModuleImportDecl cursor, return the associated module. + */ +CINDEX_LINKAGE CXModule clang_Cursor_getModule(CXCursor C); + +/** + * Given a CXFile header file, return the module that contains it, if one + * exists. + */ +CINDEX_LINKAGE CXModule clang_getModuleForFile(CXTranslationUnit, CXFile); + +/** + * \param Module a module object. + * + * \returns the module file where the provided module object came from. + */ +CINDEX_LINKAGE CXFile clang_Module_getASTFile(CXModule Module); + +/** + * \param Module a module object. + * + * \returns the parent of a sub-module or NULL if the given module is top-level, + * e.g. for 'std.vector' it will return the 'std' module. + */ +CINDEX_LINKAGE CXModule clang_Module_getParent(CXModule Module); + +/** + * \param Module a module object. + * + * \returns the name of the module, e.g. for the 'std.vector' sub-module it + * will return "vector". + */ +CINDEX_LINKAGE CXString clang_Module_getName(CXModule Module); + +/** + * \param Module a module object. + * + * \returns the full name of the module, e.g. "std.vector". + */ +CINDEX_LINKAGE CXString clang_Module_getFullName(CXModule Module); + +/** + * \param Module a module object. + * + * \returns non-zero if the module is a system one. + */ +CINDEX_LINKAGE int clang_Module_isSystem(CXModule Module); + +/** + * \param Module a module object. + * + * \returns the number of top level headers associated with this module. + */ +CINDEX_LINKAGE unsigned clang_Module_getNumTopLevelHeaders(CXTranslationUnit, + CXModule Module); + +/** + * \param Module a module object. + * + * \param Index top level header index (zero-based). + * + * \returns the specified top level header associated with the module. + */ +CINDEX_LINKAGE +CXFile clang_Module_getTopLevelHeader(CXTranslationUnit, CXModule Module, + unsigned Index); + +/** + * @} + */ + +/** + * \defgroup CINDEX_CPP C++ AST introspection + * + * The routines in this group provide access information in the ASTs specific + * to C++ language features. + * + * @{ + */ + +/** + * Determine if a C++ constructor is a converting constructor. + */ +CINDEX_LINKAGE unsigned +clang_CXXConstructor_isConvertingConstructor(CXCursor C); + +/** + * Determine if a C++ constructor is a copy constructor. + */ +CINDEX_LINKAGE unsigned clang_CXXConstructor_isCopyConstructor(CXCursor C); + +/** + * Determine if a C++ constructor is the default constructor. + */ +CINDEX_LINKAGE unsigned clang_CXXConstructor_isDefaultConstructor(CXCursor C); + +/** + * Determine if a C++ constructor is a move constructor. + */ +CINDEX_LINKAGE unsigned clang_CXXConstructor_isMoveConstructor(CXCursor C); + +/** + * Determine if a C++ field is declared 'mutable'. + */ +CINDEX_LINKAGE unsigned clang_CXXField_isMutable(CXCursor C); + +/** + * Determine if a C++ method is declared '= default'. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isDefaulted(CXCursor C); + +/** + * Determine if a C++ method is declared '= delete'. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isDeleted(CXCursor C); + +/** + * Determine if a C++ member function or member function template is + * pure virtual. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isPureVirtual(CXCursor C); + +/** + * Determine if a C++ member function or member function template is + * declared 'static'. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isStatic(CXCursor C); + +/** + * Determine if a C++ member function or member function template is + * explicitly declared 'virtual' or if it overrides a virtual method from + * one of the base classes. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isVirtual(CXCursor C); + +/** + * Determine if a C++ member function is a copy-assignment operator, + * returning 1 if such is the case and 0 otherwise. + * + * > A copy-assignment operator `X::operator=` is a non-static, + * > non-template member function of _class_ `X` with exactly one + * > parameter of type `X`, `X&`, `const X&`, `volatile X&` or `const + * > volatile X&`. + * + * That is, for example, the `operator=` in: + * + * class Foo { + * bool operator=(const volatile Foo&); + * }; + * + * Is a copy-assignment operator, while the `operator=` in: + * + * class Bar { + * bool operator=(const int&); + * }; + * + * Is not. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isCopyAssignmentOperator(CXCursor C); + +/** + * Determine if a C++ member function is a move-assignment operator, + * returning 1 if such is the case and 0 otherwise. + * + * > A move-assignment operator `X::operator=` is a non-static, + * > non-template member function of _class_ `X` with exactly one + * > parameter of type `X&&`, `const X&&`, `volatile X&&` or `const + * > volatile X&&`. + * + * That is, for example, the `operator=` in: + * + * class Foo { + * bool operator=(const volatile Foo&&); + * }; + * + * Is a move-assignment operator, while the `operator=` in: + * + * class Bar { + * bool operator=(const int&&); + * }; + * + * Is not. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isMoveAssignmentOperator(CXCursor C); + +/** + * Determines if a C++ constructor or conversion function was declared + * explicit, returning 1 if such is the case and 0 otherwise. + * + * Constructors or conversion functions are declared explicit through + * the use of the explicit specifier. + * + * For example, the following constructor and conversion function are + * not explicit as they lack the explicit specifier: + * + * class Foo { + * Foo(); + * operator int(); + * }; + * + * While the following constructor and conversion function are + * explicit as they are declared with the explicit specifier. + * + * class Foo { + * explicit Foo(); + * explicit operator int(); + * }; + * + * This function will return 0 when given a cursor pointing to one of + * the former declarations and it will return 1 for a cursor pointing + * to the latter declarations. + * + * The explicit specifier allows the user to specify a + * conditional compile-time expression whose value decides + * whether the marked element is explicit or not. + * + * For example: + * + * constexpr bool foo(int i) { return i % 2 == 0; } + * + * class Foo { + * explicit(foo(1)) Foo(); + * explicit(foo(2)) operator int(); + * } + * + * This function will return 0 for the constructor and 1 for + * the conversion function. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isExplicit(CXCursor C); + +/** + * Determine if a C++ record is abstract, i.e. whether a class or struct + * has a pure virtual member function. + */ +CINDEX_LINKAGE unsigned clang_CXXRecord_isAbstract(CXCursor C); + +/** + * Determine if an enum declaration refers to a scoped enum. + */ +CINDEX_LINKAGE unsigned clang_EnumDecl_isScoped(CXCursor C); + +/** + * Determine if a C++ member function or member function template is + * declared 'const'. + */ +CINDEX_LINKAGE unsigned clang_CXXMethod_isConst(CXCursor C); + +/** + * Given a cursor that represents a template, determine + * the cursor kind of the specializations would be generated by instantiating + * the template. + * + * This routine can be used to determine what flavor of function template, + * class template, or class template partial specialization is stored in the + * cursor. For example, it can describe whether a class template cursor is + * declared with "struct", "class" or "union". + * + * \param C The cursor to query. This cursor should represent a template + * declaration. + * + * \returns The cursor kind of the specializations that would be generated + * by instantiating the template \p C. If \p C is not a template, returns + * \c CXCursor_NoDeclFound. + */ +CINDEX_LINKAGE enum CXCursorKind clang_getTemplateCursorKind(CXCursor C); + +/** + * Given a cursor that may represent a specialization or instantiation + * of a template, retrieve the cursor that represents the template that it + * specializes or from which it was instantiated. + * + * This routine determines the template involved both for explicit + * specializations of templates and for implicit instantiations of the template, + * both of which are referred to as "specializations". For a class template + * specialization (e.g., \c std::vector), this routine will return + * either the primary template (\c std::vector) or, if the specialization was + * instantiated from a class template partial specialization, the class template + * partial specialization. For a class template partial specialization and a + * function template specialization (including instantiations), this + * this routine will return the specialized template. + * + * For members of a class template (e.g., member functions, member classes, or + * static data members), returns the specialized or instantiated member. + * Although not strictly "templates" in the C++ language, members of class + * templates have the same notions of specializations and instantiations that + * templates do, so this routine treats them similarly. + * + * \param C A cursor that may be a specialization of a template or a member + * of a template. + * + * \returns If the given cursor is a specialization or instantiation of a + * template or a member thereof, the template or member that it specializes or + * from which it was instantiated. Otherwise, returns a NULL cursor. + */ +CINDEX_LINKAGE CXCursor clang_getSpecializedCursorTemplate(CXCursor C); + +/** + * Given a cursor that references something else, return the source range + * covering that reference. + * + * \param C A cursor pointing to a member reference, a declaration reference, or + * an operator call. + * \param NameFlags A bitset with three independent flags: + * CXNameRange_WantQualifier, CXNameRange_WantTemplateArgs, and + * CXNameRange_WantSinglePiece. + * \param PieceIndex For contiguous names or when passing the flag + * CXNameRange_WantSinglePiece, only one piece with index 0 is + * available. When the CXNameRange_WantSinglePiece flag is not passed for a + * non-contiguous names, this index can be used to retrieve the individual + * pieces of the name. See also CXNameRange_WantSinglePiece. + * + * \returns The piece of the name pointed to by the given cursor. If there is no + * name, or if the PieceIndex is out-of-range, a null-cursor will be returned. + */ +CINDEX_LINKAGE CXSourceRange clang_getCursorReferenceNameRange( + CXCursor C, unsigned NameFlags, unsigned PieceIndex); + +enum CXNameRefFlags { + /** + * Include the nested-name-specifier, e.g. Foo:: in x.Foo::y, in the + * range. + */ + CXNameRange_WantQualifier = 0x1, + + /** + * Include the explicit template arguments, e.g. \ in x.f, + * in the range. + */ + CXNameRange_WantTemplateArgs = 0x2, + + /** + * If the name is non-contiguous, return the full spanning range. + * + * Non-contiguous names occur in Objective-C when a selector with two or more + * parameters is used, or in C++ when using an operator: + * \code + * [object doSomething:here withValue:there]; // Objective-C + * return some_vector[1]; // C++ + * \endcode + */ + CXNameRange_WantSinglePiece = 0x4 +}; + +/** + * @} + */ + +/** + * \defgroup CINDEX_LEX Token extraction and manipulation + * + * The routines in this group provide access to the tokens within a + * translation unit, along with a semantic mapping of those tokens to + * their corresponding cursors. + * + * @{ + */ + +/** + * Describes a kind of token. + */ +typedef enum CXTokenKind { + /** + * A token that contains some kind of punctuation. + */ + CXToken_Punctuation, + + /** + * A language keyword. + */ + CXToken_Keyword, + + /** + * An identifier (that is not a keyword). + */ + CXToken_Identifier, + + /** + * A numeric, string, or character literal. + */ + CXToken_Literal, + + /** + * A comment. + */ + CXToken_Comment +} CXTokenKind; + +/** + * Describes a single preprocessing token. + */ +typedef struct { + unsigned int_data[4]; + void *ptr_data; +} CXToken; + +/** + * Get the raw lexical token starting with the given location. + * + * \param TU the translation unit whose text is being tokenized. + * + * \param Location the source location with which the token starts. + * + * \returns The token starting with the given location or NULL if no such token + * exist. The returned pointer must be freed with clang_disposeTokens before the + * translation unit is destroyed. + */ +CINDEX_LINKAGE CXToken *clang_getToken(CXTranslationUnit TU, + CXSourceLocation Location); + +/** + * Determine the kind of the given token. + */ +CINDEX_LINKAGE CXTokenKind clang_getTokenKind(CXToken); + +/** + * Determine the spelling of the given token. + * + * The spelling of a token is the textual representation of that token, e.g., + * the text of an identifier or keyword. + */ +CINDEX_LINKAGE CXString clang_getTokenSpelling(CXTranslationUnit, CXToken); + +/** + * Retrieve the source location of the given token. + */ +CINDEX_LINKAGE CXSourceLocation clang_getTokenLocation(CXTranslationUnit, + CXToken); + +/** + * Retrieve a source range that covers the given token. + */ +CINDEX_LINKAGE CXSourceRange clang_getTokenExtent(CXTranslationUnit, CXToken); + +/** + * Tokenize the source code described by the given range into raw + * lexical tokens. + * + * \param TU the translation unit whose text is being tokenized. + * + * \param Range the source range in which text should be tokenized. All of the + * tokens produced by tokenization will fall within this source range, + * + * \param Tokens this pointer will be set to point to the array of tokens + * that occur within the given source range. The returned pointer must be + * freed with clang_disposeTokens() before the translation unit is destroyed. + * + * \param NumTokens will be set to the number of tokens in the \c *Tokens + * array. + * + */ +CINDEX_LINKAGE void clang_tokenize(CXTranslationUnit TU, CXSourceRange Range, + CXToken **Tokens, unsigned *NumTokens); + +/** + * Annotate the given set of tokens by providing cursors for each token + * that can be mapped to a specific entity within the abstract syntax tree. + * + * This token-annotation routine is equivalent to invoking + * clang_getCursor() for the source locations of each of the + * tokens. The cursors provided are filtered, so that only those + * cursors that have a direct correspondence to the token are + * accepted. For example, given a function call \c f(x), + * clang_getCursor() would provide the following cursors: + * + * * when the cursor is over the 'f', a DeclRefExpr cursor referring to 'f'. + * * when the cursor is over the '(' or the ')', a CallExpr referring to 'f'. + * * when the cursor is over the 'x', a DeclRefExpr cursor referring to 'x'. + * + * Only the first and last of these cursors will occur within the + * annotate, since the tokens "f" and "x' directly refer to a function + * and a variable, respectively, but the parentheses are just a small + * part of the full syntax of the function call expression, which is + * not provided as an annotation. + * + * \param TU the translation unit that owns the given tokens. + * + * \param Tokens the set of tokens to annotate. + * + * \param NumTokens the number of tokens in \p Tokens. + * + * \param Cursors an array of \p NumTokens cursors, whose contents will be + * replaced with the cursors corresponding to each token. + */ +CINDEX_LINKAGE void clang_annotateTokens(CXTranslationUnit TU, CXToken *Tokens, + unsigned NumTokens, CXCursor *Cursors); + +/** + * Free the given set of tokens. + */ +CINDEX_LINKAGE void clang_disposeTokens(CXTranslationUnit TU, CXToken *Tokens, + unsigned NumTokens); + +/** + * @} + */ + +/** + * \defgroup CINDEX_DEBUG Debugging facilities + * + * These routines are used for testing and debugging, only, and should not + * be relied upon. + * + * @{ + */ + +/* for debug/testing */ +CINDEX_LINKAGE CXString clang_getCursorKindSpelling(enum CXCursorKind Kind); +CINDEX_LINKAGE void clang_getDefinitionSpellingAndExtent( + CXCursor, const char **startBuf, const char **endBuf, unsigned *startLine, + unsigned *startColumn, unsigned *endLine, unsigned *endColumn); +CINDEX_LINKAGE void clang_enableStackTraces(void); +CINDEX_LINKAGE void clang_executeOnThread(void (*fn)(void *), void *user_data, + unsigned stack_size); + +/** + * @} + */ + +/** + * \defgroup CINDEX_CODE_COMPLET Code completion + * + * Code completion involves taking an (incomplete) source file, along with + * knowledge of where the user is actively editing that file, and suggesting + * syntactically- and semantically-valid constructs that the user might want to + * use at that particular point in the source code. These data structures and + * routines provide support for code completion. + * + * @{ + */ + +/** + * A semantic string that describes a code-completion result. + * + * A semantic string that describes the formatting of a code-completion + * result as a single "template" of text that should be inserted into the + * source buffer when a particular code-completion result is selected. + * Each semantic string is made up of some number of "chunks", each of which + * contains some text along with a description of what that text means, e.g., + * the name of the entity being referenced, whether the text chunk is part of + * the template, or whether it is a "placeholder" that the user should replace + * with actual code,of a specific kind. See \c CXCompletionChunkKind for a + * description of the different kinds of chunks. + */ +typedef void *CXCompletionString; + +/** + * A single result of code completion. + */ +typedef struct { + /** + * The kind of entity that this completion refers to. + * + * The cursor kind will be a macro, keyword, or a declaration (one of the + * *Decl cursor kinds), describing the entity that the completion is + * referring to. + * + * \todo In the future, we would like to provide a full cursor, to allow + * the client to extract additional information from declaration. + */ + enum CXCursorKind CursorKind; + + /** + * The code-completion string that describes how to insert this + * code-completion result into the editing buffer. + */ + CXCompletionString CompletionString; +} CXCompletionResult; + +/** + * Describes a single piece of text within a code-completion string. + * + * Each "chunk" within a code-completion string (\c CXCompletionString) is + * either a piece of text with a specific "kind" that describes how that text + * should be interpreted by the client or is another completion string. + */ +enum CXCompletionChunkKind { + /** + * A code-completion string that describes "optional" text that + * could be a part of the template (but is not required). + * + * The Optional chunk is the only kind of chunk that has a code-completion + * string for its representation, which is accessible via + * \c clang_getCompletionChunkCompletionString(). The code-completion string + * describes an additional part of the template that is completely optional. + * For example, optional chunks can be used to describe the placeholders for + * arguments that match up with defaulted function parameters, e.g. given: + * + * \code + * void f(int x, float y = 3.14, double z = 2.71828); + * \endcode + * + * The code-completion string for this function would contain: + * - a TypedText chunk for "f". + * - a LeftParen chunk for "(". + * - a Placeholder chunk for "int x" + * - an Optional chunk containing the remaining defaulted arguments, e.g., + * - a Comma chunk for "," + * - a Placeholder chunk for "float y" + * - an Optional chunk containing the last defaulted argument: + * - a Comma chunk for "," + * - a Placeholder chunk for "double z" + * - a RightParen chunk for ")" + * + * There are many ways to handle Optional chunks. Two simple approaches are: + * - Completely ignore optional chunks, in which case the template for the + * function "f" would only include the first parameter ("int x"). + * - Fully expand all optional chunks, in which case the template for the + * function "f" would have all of the parameters. + */ + CXCompletionChunk_Optional, + /** + * Text that a user would be expected to type to get this + * code-completion result. + * + * There will be exactly one "typed text" chunk in a semantic string, which + * will typically provide the spelling of a keyword or the name of a + * declaration that could be used at the current code point. Clients are + * expected to filter the code-completion results based on the text in this + * chunk. + */ + CXCompletionChunk_TypedText, + /** + * Text that should be inserted as part of a code-completion result. + * + * A "text" chunk represents text that is part of the template to be + * inserted into user code should this particular code-completion result + * be selected. + */ + CXCompletionChunk_Text, + /** + * Placeholder text that should be replaced by the user. + * + * A "placeholder" chunk marks a place where the user should insert text + * into the code-completion template. For example, placeholders might mark + * the function parameters for a function declaration, to indicate that the + * user should provide arguments for each of those parameters. The actual + * text in a placeholder is a suggestion for the text to display before + * the user replaces the placeholder with real code. + */ + CXCompletionChunk_Placeholder, + /** + * Informative text that should be displayed but never inserted as + * part of the template. + * + * An "informative" chunk contains annotations that can be displayed to + * help the user decide whether a particular code-completion result is the + * right option, but which is not part of the actual template to be inserted + * by code completion. + */ + CXCompletionChunk_Informative, + /** + * Text that describes the current parameter when code-completion is + * referring to function call, message send, or template specialization. + * + * A "current parameter" chunk occurs when code-completion is providing + * information about a parameter corresponding to the argument at the + * code-completion point. For example, given a function + * + * \code + * int add(int x, int y); + * \endcode + * + * and the source code \c add(, where the code-completion point is after the + * "(", the code-completion string will contain a "current parameter" chunk + * for "int x", indicating that the current argument will initialize that + * parameter. After typing further, to \c add(17, (where the code-completion + * point is after the ","), the code-completion string will contain a + * "current parameter" chunk to "int y". + */ + CXCompletionChunk_CurrentParameter, + /** + * A left parenthesis ('('), used to initiate a function call or + * signal the beginning of a function parameter list. + */ + CXCompletionChunk_LeftParen, + /** + * A right parenthesis (')'), used to finish a function call or + * signal the end of a function parameter list. + */ + CXCompletionChunk_RightParen, + /** + * A left bracket ('['). + */ + CXCompletionChunk_LeftBracket, + /** + * A right bracket (']'). + */ + CXCompletionChunk_RightBracket, + /** + * A left brace ('{'). + */ + CXCompletionChunk_LeftBrace, + /** + * A right brace ('}'). + */ + CXCompletionChunk_RightBrace, + /** + * A left angle bracket ('<'). + */ + CXCompletionChunk_LeftAngle, + /** + * A right angle bracket ('>'). + */ + CXCompletionChunk_RightAngle, + /** + * A comma separator (','). + */ + CXCompletionChunk_Comma, + /** + * Text that specifies the result type of a given result. + * + * This special kind of informative chunk is not meant to be inserted into + * the text buffer. Rather, it is meant to illustrate the type that an + * expression using the given completion string would have. + */ + CXCompletionChunk_ResultType, + /** + * A colon (':'). + */ + CXCompletionChunk_Colon, + /** + * A semicolon (';'). + */ + CXCompletionChunk_SemiColon, + /** + * An '=' sign. + */ + CXCompletionChunk_Equal, + /** + * Horizontal space (' '). + */ + CXCompletionChunk_HorizontalSpace, + /** + * Vertical space ('\\n'), after which it is generally a good idea to + * perform indentation. + */ + CXCompletionChunk_VerticalSpace +}; + +/** + * Determine the kind of a particular chunk within a completion string. + * + * \param completion_string the completion string to query. + * + * \param chunk_number the 0-based index of the chunk in the completion string. + * + * \returns the kind of the chunk at the index \c chunk_number. + */ +CINDEX_LINKAGE enum CXCompletionChunkKind +clang_getCompletionChunkKind(CXCompletionString completion_string, + unsigned chunk_number); + +/** + * Retrieve the text associated with a particular chunk within a + * completion string. + * + * \param completion_string the completion string to query. + * + * \param chunk_number the 0-based index of the chunk in the completion string. + * + * \returns the text associated with the chunk at index \c chunk_number. + */ +CINDEX_LINKAGE CXString clang_getCompletionChunkText( + CXCompletionString completion_string, unsigned chunk_number); + +/** + * Retrieve the completion string associated with a particular chunk + * within a completion string. + * + * \param completion_string the completion string to query. + * + * \param chunk_number the 0-based index of the chunk in the completion string. + * + * \returns the completion string associated with the chunk at index + * \c chunk_number. + */ +CINDEX_LINKAGE CXCompletionString clang_getCompletionChunkCompletionString( + CXCompletionString completion_string, unsigned chunk_number); + +/** + * Retrieve the number of chunks in the given code-completion string. + */ +CINDEX_LINKAGE unsigned +clang_getNumCompletionChunks(CXCompletionString completion_string); + +/** + * Determine the priority of this code completion. + * + * The priority of a code completion indicates how likely it is that this + * particular completion is the completion that the user will select. The + * priority is selected by various internal heuristics. + * + * \param completion_string The completion string to query. + * + * \returns The priority of this completion string. Smaller values indicate + * higher-priority (more likely) completions. + */ +CINDEX_LINKAGE unsigned +clang_getCompletionPriority(CXCompletionString completion_string); + +/** + * Determine the availability of the entity that this code-completion + * string refers to. + * + * \param completion_string The completion string to query. + * + * \returns The availability of the completion string. + */ +CINDEX_LINKAGE enum CXAvailabilityKind +clang_getCompletionAvailability(CXCompletionString completion_string); + +/** + * Retrieve the number of annotations associated with the given + * completion string. + * + * \param completion_string the completion string to query. + * + * \returns the number of annotations associated with the given completion + * string. + */ +CINDEX_LINKAGE unsigned +clang_getCompletionNumAnnotations(CXCompletionString completion_string); + +/** + * Retrieve the annotation associated with the given completion string. + * + * \param completion_string the completion string to query. + * + * \param annotation_number the 0-based index of the annotation of the + * completion string. + * + * \returns annotation string associated with the completion at index + * \c annotation_number, or a NULL string if that annotation is not available. + */ +CINDEX_LINKAGE CXString clang_getCompletionAnnotation( + CXCompletionString completion_string, unsigned annotation_number); + +/** + * Retrieve the parent context of the given completion string. + * + * The parent context of a completion string is the semantic parent of + * the declaration (if any) that the code completion represents. For example, + * a code completion for an Objective-C method would have the method's class + * or protocol as its context. + * + * \param completion_string The code completion string whose parent is + * being queried. + * + * \param kind DEPRECATED: always set to CXCursor_NotImplemented if non-NULL. + * + * \returns The name of the completion parent, e.g., "NSObject" if + * the completion string represents a method in the NSObject class. + */ +CINDEX_LINKAGE CXString clang_getCompletionParent( + CXCompletionString completion_string, enum CXCursorKind *kind); + +/** + * Retrieve the brief documentation comment attached to the declaration + * that corresponds to the given completion string. + */ +CINDEX_LINKAGE CXString +clang_getCompletionBriefComment(CXCompletionString completion_string); + +/** + * Retrieve a completion string for an arbitrary declaration or macro + * definition cursor. + * + * \param cursor The cursor to query. + * + * \returns A non-context-sensitive completion string for declaration and macro + * definition cursors, or NULL for other kinds of cursors. + */ +CINDEX_LINKAGE CXCompletionString +clang_getCursorCompletionString(CXCursor cursor); + +/** + * Contains the results of code-completion. + * + * This data structure contains the results of code completion, as + * produced by \c clang_codeCompleteAt(). Its contents must be freed by + * \c clang_disposeCodeCompleteResults. + */ +typedef struct { + /** + * The code-completion results. + */ + CXCompletionResult *Results; + + /** + * The number of code-completion results stored in the + * \c Results array. + */ + unsigned NumResults; +} CXCodeCompleteResults; + +/** + * Retrieve the number of fix-its for the given completion index. + * + * Calling this makes sense only if CXCodeComplete_IncludeCompletionsWithFixIts + * option was set. + * + * \param results The structure keeping all completion results + * + * \param completion_index The index of the completion + * + * \return The number of fix-its which must be applied before the completion at + * completion_index can be applied + */ +CINDEX_LINKAGE unsigned +clang_getCompletionNumFixIts(CXCodeCompleteResults *results, + unsigned completion_index); + +/** + * Fix-its that *must* be applied before inserting the text for the + * corresponding completion. + * + * By default, clang_codeCompleteAt() only returns completions with empty + * fix-its. Extra completions with non-empty fix-its should be explicitly + * requested by setting CXCodeComplete_IncludeCompletionsWithFixIts. + * + * For the clients to be able to compute position of the cursor after applying + * fix-its, the following conditions are guaranteed to hold for + * replacement_range of the stored fix-its: + * - Ranges in the fix-its are guaranteed to never contain the completion + * point (or identifier under completion point, if any) inside them, except + * at the start or at the end of the range. + * - If a fix-it range starts or ends with completion point (or starts or + * ends after the identifier under completion point), it will contain at + * least one character. It allows to unambiguously recompute completion + * point after applying the fix-it. + * + * The intuition is that provided fix-its change code around the identifier we + * complete, but are not allowed to touch the identifier itself or the + * completion point. One example of completions with corrections are the ones + * replacing '.' with '->' and vice versa: + * + * std::unique_ptr> vec_ptr; + * In 'vec_ptr.^', one of the completions is 'push_back', it requires + * replacing '.' with '->'. + * In 'vec_ptr->^', one of the completions is 'release', it requires + * replacing '->' with '.'. + * + * \param results The structure keeping all completion results + * + * \param completion_index The index of the completion + * + * \param fixit_index The index of the fix-it for the completion at + * completion_index + * + * \param replacement_range The fix-it range that must be replaced before the + * completion at completion_index can be applied + * + * \returns The fix-it string that must replace the code at replacement_range + * before the completion at completion_index can be applied + */ +CINDEX_LINKAGE CXString clang_getCompletionFixIt( + CXCodeCompleteResults *results, unsigned completion_index, + unsigned fixit_index, CXSourceRange *replacement_range); + +/** + * Flags that can be passed to \c clang_codeCompleteAt() to + * modify its behavior. + * + * The enumerators in this enumeration can be bitwise-OR'd together to + * provide multiple options to \c clang_codeCompleteAt(). + */ +enum CXCodeComplete_Flags { + /** + * Whether to include macros within the set of code + * completions returned. + */ + CXCodeComplete_IncludeMacros = 0x01, + + /** + * Whether to include code patterns for language constructs + * within the set of code completions, e.g., for loops. + */ + CXCodeComplete_IncludeCodePatterns = 0x02, + + /** + * Whether to include brief documentation within the set of code + * completions returned. + */ + CXCodeComplete_IncludeBriefComments = 0x04, + + /** + * Whether to speed up completion by omitting top- or namespace-level entities + * defined in the preamble. There's no guarantee any particular entity is + * omitted. This may be useful if the headers are indexed externally. + */ + CXCodeComplete_SkipPreamble = 0x08, + + /** + * Whether to include completions with small + * fix-its, e.g. change '.' to '->' on member access, etc. + */ + CXCodeComplete_IncludeCompletionsWithFixIts = 0x10 +}; + +/** + * Bits that represent the context under which completion is occurring. + * + * The enumerators in this enumeration may be bitwise-OR'd together if multiple + * contexts are occurring simultaneously. + */ +enum CXCompletionContext { + /** + * The context for completions is unexposed, as only Clang results + * should be included. (This is equivalent to having no context bits set.) + */ + CXCompletionContext_Unexposed = 0, + + /** + * Completions for any possible type should be included in the results. + */ + CXCompletionContext_AnyType = 1 << 0, + + /** + * Completions for any possible value (variables, function calls, etc.) + * should be included in the results. + */ + CXCompletionContext_AnyValue = 1 << 1, + /** + * Completions for values that resolve to an Objective-C object should + * be included in the results. + */ + CXCompletionContext_ObjCObjectValue = 1 << 2, + /** + * Completions for values that resolve to an Objective-C selector + * should be included in the results. + */ + CXCompletionContext_ObjCSelectorValue = 1 << 3, + /** + * Completions for values that resolve to a C++ class type should be + * included in the results. + */ + CXCompletionContext_CXXClassTypeValue = 1 << 4, + + /** + * Completions for fields of the member being accessed using the dot + * operator should be included in the results. + */ + CXCompletionContext_DotMemberAccess = 1 << 5, + /** + * Completions for fields of the member being accessed using the arrow + * operator should be included in the results. + */ + CXCompletionContext_ArrowMemberAccess = 1 << 6, + /** + * Completions for properties of the Objective-C object being accessed + * using the dot operator should be included in the results. + */ + CXCompletionContext_ObjCPropertyAccess = 1 << 7, + + /** + * Completions for enum tags should be included in the results. + */ + CXCompletionContext_EnumTag = 1 << 8, + /** + * Completions for union tags should be included in the results. + */ + CXCompletionContext_UnionTag = 1 << 9, + /** + * Completions for struct tags should be included in the results. + */ + CXCompletionContext_StructTag = 1 << 10, + + /** + * Completions for C++ class names should be included in the results. + */ + CXCompletionContext_ClassTag = 1 << 11, + /** + * Completions for C++ namespaces and namespace aliases should be + * included in the results. + */ + CXCompletionContext_Namespace = 1 << 12, + /** + * Completions for C++ nested name specifiers should be included in + * the results. + */ + CXCompletionContext_NestedNameSpecifier = 1 << 13, + + /** + * Completions for Objective-C interfaces (classes) should be included + * in the results. + */ + CXCompletionContext_ObjCInterface = 1 << 14, + /** + * Completions for Objective-C protocols should be included in + * the results. + */ + CXCompletionContext_ObjCProtocol = 1 << 15, + /** + * Completions for Objective-C categories should be included in + * the results. + */ + CXCompletionContext_ObjCCategory = 1 << 16, + /** + * Completions for Objective-C instance messages should be included + * in the results. + */ + CXCompletionContext_ObjCInstanceMessage = 1 << 17, + /** + * Completions for Objective-C class messages should be included in + * the results. + */ + CXCompletionContext_ObjCClassMessage = 1 << 18, + /** + * Completions for Objective-C selector names should be included in + * the results. + */ + CXCompletionContext_ObjCSelectorName = 1 << 19, + + /** + * Completions for preprocessor macro names should be included in + * the results. + */ + CXCompletionContext_MacroName = 1 << 20, + + /** + * Natural language completions should be included in the results. + */ + CXCompletionContext_NaturalLanguage = 1 << 21, + + /** + * #include file completions should be included in the results. + */ + CXCompletionContext_IncludedFile = 1 << 22, + + /** + * The current context is unknown, so set all contexts. + */ + CXCompletionContext_Unknown = ((1 << 23) - 1) +}; + +/** + * Returns a default set of code-completion options that can be + * passed to\c clang_codeCompleteAt(). + */ +CINDEX_LINKAGE unsigned clang_defaultCodeCompleteOptions(void); + +/** + * Perform code completion at a given location in a translation unit. + * + * This function performs code completion at a particular file, line, and + * column within source code, providing results that suggest potential + * code snippets based on the context of the completion. The basic model + * for code completion is that Clang will parse a complete source file, + * performing syntax checking up to the location where code-completion has + * been requested. At that point, a special code-completion token is passed + * to the parser, which recognizes this token and determines, based on the + * current location in the C/Objective-C/C++ grammar and the state of + * semantic analysis, what completions to provide. These completions are + * returned via a new \c CXCodeCompleteResults structure. + * + * Code completion itself is meant to be triggered by the client when the + * user types punctuation characters or whitespace, at which point the + * code-completion location will coincide with the cursor. For example, if \c p + * is a pointer, code-completion might be triggered after the "-" and then + * after the ">" in \c p->. When the code-completion location is after the ">", + * the completion results will provide, e.g., the members of the struct that + * "p" points to. The client is responsible for placing the cursor at the + * beginning of the token currently being typed, then filtering the results + * based on the contents of the token. For example, when code-completing for + * the expression \c p->get, the client should provide the location just after + * the ">" (e.g., pointing at the "g") to this code-completion hook. Then, the + * client can filter the results based on the current token text ("get"), only + * showing those results that start with "get". The intent of this interface + * is to separate the relatively high-latency acquisition of code-completion + * results from the filtering of results on a per-character basis, which must + * have a lower latency. + * + * \param TU The translation unit in which code-completion should + * occur. The source files for this translation unit need not be + * completely up-to-date (and the contents of those source files may + * be overridden via \p unsaved_files). Cursors referring into the + * translation unit may be invalidated by this invocation. + * + * \param complete_filename The name of the source file where code + * completion should be performed. This filename may be any file + * included in the translation unit. + * + * \param complete_line The line at which code-completion should occur. + * + * \param complete_column The column at which code-completion should occur. + * Note that the column should point just after the syntactic construct that + * initiated code completion, and not in the middle of a lexical token. + * + * \param unsaved_files the Files that have not yet been saved to disk + * but may be required for parsing or code completion, including the + * contents of those files. The contents and name of these files (as + * specified by CXUnsavedFile) are copied when necessary, so the + * client only needs to guarantee their validity until the call to + * this function returns. + * + * \param num_unsaved_files The number of unsaved file entries in \p + * unsaved_files. + * + * \param options Extra options that control the behavior of code + * completion, expressed as a bitwise OR of the enumerators of the + * CXCodeComplete_Flags enumeration. The + * \c clang_defaultCodeCompleteOptions() function returns a default set + * of code-completion options. + * + * \returns If successful, a new \c CXCodeCompleteResults structure + * containing code-completion results, which should eventually be + * freed with \c clang_disposeCodeCompleteResults(). If code + * completion fails, returns NULL. + */ +CINDEX_LINKAGE +CXCodeCompleteResults * +clang_codeCompleteAt(CXTranslationUnit TU, const char *complete_filename, + unsigned complete_line, unsigned complete_column, + struct CXUnsavedFile *unsaved_files, + unsigned num_unsaved_files, unsigned options); + +/** + * Sort the code-completion results in case-insensitive alphabetical + * order. + * + * \param Results The set of results to sort. + * \param NumResults The number of results in \p Results. + */ +CINDEX_LINKAGE +void clang_sortCodeCompletionResults(CXCompletionResult *Results, + unsigned NumResults); + +/** + * Free the given set of code-completion results. + */ +CINDEX_LINKAGE +void clang_disposeCodeCompleteResults(CXCodeCompleteResults *Results); + +/** + * Determine the number of diagnostics produced prior to the + * location where code completion was performed. + */ +CINDEX_LINKAGE +unsigned clang_codeCompleteGetNumDiagnostics(CXCodeCompleteResults *Results); + +/** + * Retrieve a diagnostic associated with the given code completion. + * + * \param Results the code completion results to query. + * \param Index the zero-based diagnostic number to retrieve. + * + * \returns the requested diagnostic. This diagnostic must be freed + * via a call to \c clang_disposeDiagnostic(). + */ +CINDEX_LINKAGE +CXDiagnostic clang_codeCompleteGetDiagnostic(CXCodeCompleteResults *Results, + unsigned Index); + +/** + * Determines what completions are appropriate for the context + * the given code completion. + * + * \param Results the code completion results to query + * + * \returns the kinds of completions that are appropriate for use + * along with the given code completion results. + */ +CINDEX_LINKAGE +unsigned long long +clang_codeCompleteGetContexts(CXCodeCompleteResults *Results); + +/** + * Returns the cursor kind for the container for the current code + * completion context. The container is only guaranteed to be set for + * contexts where a container exists (i.e. member accesses or Objective-C + * message sends); if there is not a container, this function will return + * CXCursor_InvalidCode. + * + * \param Results the code completion results to query + * + * \param IsIncomplete on return, this value will be false if Clang has complete + * information about the container. If Clang does not have complete + * information, this value will be true. + * + * \returns the container kind, or CXCursor_InvalidCode if there is not a + * container + */ +CINDEX_LINKAGE +enum CXCursorKind +clang_codeCompleteGetContainerKind(CXCodeCompleteResults *Results, + unsigned *IsIncomplete); + +/** + * Returns the USR for the container for the current code completion + * context. If there is not a container for the current context, this + * function will return the empty string. + * + * \param Results the code completion results to query + * + * \returns the USR for the container + */ +CINDEX_LINKAGE +CXString clang_codeCompleteGetContainerUSR(CXCodeCompleteResults *Results); + +/** + * Returns the currently-entered selector for an Objective-C message + * send, formatted like "initWithFoo:bar:". Only guaranteed to return a + * non-empty string for CXCompletionContext_ObjCInstanceMessage and + * CXCompletionContext_ObjCClassMessage. + * + * \param Results the code completion results to query + * + * \returns the selector (or partial selector) that has been entered thus far + * for an Objective-C message send. + */ +CINDEX_LINKAGE +CXString clang_codeCompleteGetObjCSelector(CXCodeCompleteResults *Results); + +/** + * @} + */ + +/** + * \defgroup CINDEX_MISC Miscellaneous utility functions + * + * @{ + */ + +/** + * Return a version string, suitable for showing to a user, but not + * intended to be parsed (the format is not guaranteed to be stable). + */ +CINDEX_LINKAGE CXString clang_getClangVersion(void); + +/** + * Enable/disable crash recovery. + * + * \param isEnabled Flag to indicate if crash recovery is enabled. A non-zero + * value enables crash recovery, while 0 disables it. + */ +CINDEX_LINKAGE void clang_toggleCrashRecovery(unsigned isEnabled); + +/** + * Visitor invoked for each file in a translation unit + * (used with clang_getInclusions()). + * + * This visitor function will be invoked by clang_getInclusions() for each + * file included (either at the top-level or by \#include directives) within + * a translation unit. The first argument is the file being included, and + * the second and third arguments provide the inclusion stack. The + * array is sorted in order of immediate inclusion. For example, + * the first element refers to the location that included 'included_file'. + */ +typedef void (*CXInclusionVisitor)(CXFile included_file, + CXSourceLocation *inclusion_stack, + unsigned include_len, + CXClientData client_data); + +/** + * Visit the set of preprocessor inclusions in a translation unit. + * The visitor function is called with the provided data for every included + * file. This does not include headers included by the PCH file (unless one + * is inspecting the inclusions in the PCH file itself). + */ +CINDEX_LINKAGE void clang_getInclusions(CXTranslationUnit tu, + CXInclusionVisitor visitor, + CXClientData client_data); + +typedef enum { + CXEval_Int = 1, + CXEval_Float = 2, + CXEval_ObjCStrLiteral = 3, + CXEval_StrLiteral = 4, + CXEval_CFStr = 5, + CXEval_Other = 6, + + CXEval_UnExposed = 0 + +} CXEvalResultKind; + +/** + * Evaluation result of a cursor + */ +typedef void *CXEvalResult; + +/** + * If cursor is a statement declaration tries to evaluate the + * statement and if its variable, tries to evaluate its initializer, + * into its corresponding type. + * If it's an expression, tries to evaluate the expression. + */ +CINDEX_LINKAGE CXEvalResult clang_Cursor_Evaluate(CXCursor C); + +/** + * Returns the kind of the evaluated result. + */ +CINDEX_LINKAGE CXEvalResultKind clang_EvalResult_getKind(CXEvalResult E); + +/** + * Returns the evaluation result as integer if the + * kind is Int. + */ +CINDEX_LINKAGE int clang_EvalResult_getAsInt(CXEvalResult E); + +/** + * Returns the evaluation result as a long long integer if the + * kind is Int. This prevents overflows that may happen if the result is + * returned with clang_EvalResult_getAsInt. + */ +CINDEX_LINKAGE long long clang_EvalResult_getAsLongLong(CXEvalResult E); + +/** + * Returns a non-zero value if the kind is Int and the evaluation + * result resulted in an unsigned integer. + */ +CINDEX_LINKAGE unsigned clang_EvalResult_isUnsignedInt(CXEvalResult E); + +/** + * Returns the evaluation result as an unsigned integer if + * the kind is Int and clang_EvalResult_isUnsignedInt is non-zero. + */ +CINDEX_LINKAGE unsigned long long +clang_EvalResult_getAsUnsigned(CXEvalResult E); + +/** + * Returns the evaluation result as double if the + * kind is double. + */ +CINDEX_LINKAGE double clang_EvalResult_getAsDouble(CXEvalResult E); + +/** + * Returns the evaluation result as a constant string if the + * kind is other than Int or float. User must not free this pointer, + * instead call clang_EvalResult_dispose on the CXEvalResult returned + * by clang_Cursor_Evaluate. + */ +CINDEX_LINKAGE const char *clang_EvalResult_getAsStr(CXEvalResult E); + +/** + * Disposes the created Eval memory. + */ +CINDEX_LINKAGE void clang_EvalResult_dispose(CXEvalResult E); +/** + * @} + */ + +/** \defgroup CINDEX_REMAPPING Remapping functions + * + * @{ + */ + +/** + * A remapping of original source files and their translated files. + */ +typedef void *CXRemapping; + +/** + * Retrieve a remapping. + * + * \param path the path that contains metadata about remappings. + * + * \returns the requested remapping. This remapping must be freed + * via a call to \c clang_remap_dispose(). Can return NULL if an error occurred. + */ +CINDEX_LINKAGE CXRemapping clang_getRemappings(const char *path); + +/** + * Retrieve a remapping. + * + * \param filePaths pointer to an array of file paths containing remapping info. + * + * \param numFiles number of file paths. + * + * \returns the requested remapping. This remapping must be freed + * via a call to \c clang_remap_dispose(). Can return NULL if an error occurred. + */ +CINDEX_LINKAGE +CXRemapping clang_getRemappingsFromFileList(const char **filePaths, + unsigned numFiles); + +/** + * Determine the number of remappings. + */ +CINDEX_LINKAGE unsigned clang_remap_getNumFiles(CXRemapping); + +/** + * Get the original and the associated filename from the remapping. + * + * \param original If non-NULL, will be set to the original filename. + * + * \param transformed If non-NULL, will be set to the filename that the original + * is associated with. + */ +CINDEX_LINKAGE void clang_remap_getFilenames(CXRemapping, unsigned index, + CXString *original, + CXString *transformed); + +/** + * Dispose the remapping. + */ +CINDEX_LINKAGE void clang_remap_dispose(CXRemapping); + +/** + * @} + */ + +/** \defgroup CINDEX_HIGH Higher level API functions + * + * @{ + */ + +enum CXVisitorResult { CXVisit_Break, CXVisit_Continue }; + +typedef struct CXCursorAndRangeVisitor { + void *context; + enum CXVisitorResult (*visit)(void *context, CXCursor, CXSourceRange); +} CXCursorAndRangeVisitor; + +typedef enum { + /** + * Function returned successfully. + */ + CXResult_Success = 0, + /** + * One of the parameters was invalid for the function. + */ + CXResult_Invalid = 1, + /** + * The function was terminated by a callback (e.g. it returned + * CXVisit_Break) + */ + CXResult_VisitBreak = 2 + +} CXResult; + +/** + * Find references of a declaration in a specific file. + * + * \param cursor pointing to a declaration or a reference of one. + * + * \param file to search for references. + * + * \param visitor callback that will receive pairs of CXCursor/CXSourceRange for + * each reference found. + * The CXSourceRange will point inside the file; if the reference is inside + * a macro (and not a macro argument) the CXSourceRange will be invalid. + * + * \returns one of the CXResult enumerators. + */ +CINDEX_LINKAGE CXResult clang_findReferencesInFile( + CXCursor cursor, CXFile file, CXCursorAndRangeVisitor visitor); + +/** + * Find #import/#include directives in a specific file. + * + * \param TU translation unit containing the file to query. + * + * \param file to search for #import/#include directives. + * + * \param visitor callback that will receive pairs of CXCursor/CXSourceRange for + * each directive found. + * + * \returns one of the CXResult enumerators. + */ +CINDEX_LINKAGE CXResult clang_findIncludesInFile( + CXTranslationUnit TU, CXFile file, CXCursorAndRangeVisitor visitor); + +#if __has_feature(blocks) +typedef enum CXVisitorResult (^CXCursorAndRangeVisitorBlock)(CXCursor, + CXSourceRange); +#else +typedef struct _CXCursorAndRangeVisitorBlock *CXCursorAndRangeVisitorBlock; +#endif + +CINDEX_LINKAGE +CXResult clang_findReferencesInFileWithBlock(CXCursor, CXFile, + CXCursorAndRangeVisitorBlock); + +CINDEX_LINKAGE +CXResult clang_findIncludesInFileWithBlock(CXTranslationUnit, CXFile, + CXCursorAndRangeVisitorBlock); + +/** + * The client's data object that is associated with a CXFile. + */ +typedef void *CXIdxClientFile; + +/** + * The client's data object that is associated with a semantic entity. + */ +typedef void *CXIdxClientEntity; + +/** + * The client's data object that is associated with a semantic container + * of entities. + */ +typedef void *CXIdxClientContainer; + +/** + * The client's data object that is associated with an AST file (PCH + * or module). + */ +typedef void *CXIdxClientASTFile; + +/** + * Source location passed to index callbacks. + */ +typedef struct { + void *ptr_data[2]; + unsigned int_data; +} CXIdxLoc; + +/** + * Data for ppIncludedFile callback. + */ +typedef struct { + /** + * Location of '#' in the \#include/\#import directive. + */ + CXIdxLoc hashLoc; + /** + * Filename as written in the \#include/\#import directive. + */ + const char *filename; + /** + * The actual file that the \#include/\#import directive resolved to. + */ + CXFile file; + int isImport; + int isAngled; + /** + * Non-zero if the directive was automatically turned into a module + * import. + */ + int isModuleImport; +} CXIdxIncludedFileInfo; + +/** + * Data for IndexerCallbacks#importedASTFile. + */ +typedef struct { + /** + * Top level AST file containing the imported PCH, module or submodule. + */ + CXFile file; + /** + * The imported module or NULL if the AST file is a PCH. + */ + CXModule module; + /** + * Location where the file is imported. Applicable only for modules. + */ + CXIdxLoc loc; + /** + * Non-zero if an inclusion directive was automatically turned into + * a module import. Applicable only for modules. + */ + int isImplicit; + +} CXIdxImportedASTFileInfo; + +typedef enum { + CXIdxEntity_Unexposed = 0, + CXIdxEntity_Typedef = 1, + CXIdxEntity_Function = 2, + CXIdxEntity_Variable = 3, + CXIdxEntity_Field = 4, + CXIdxEntity_EnumConstant = 5, + + CXIdxEntity_ObjCClass = 6, + CXIdxEntity_ObjCProtocol = 7, + CXIdxEntity_ObjCCategory = 8, + + CXIdxEntity_ObjCInstanceMethod = 9, + CXIdxEntity_ObjCClassMethod = 10, + CXIdxEntity_ObjCProperty = 11, + CXIdxEntity_ObjCIvar = 12, + + CXIdxEntity_Enum = 13, + CXIdxEntity_Struct = 14, + CXIdxEntity_Union = 15, + + CXIdxEntity_CXXClass = 16, + CXIdxEntity_CXXNamespace = 17, + CXIdxEntity_CXXNamespaceAlias = 18, + CXIdxEntity_CXXStaticVariable = 19, + CXIdxEntity_CXXStaticMethod = 20, + CXIdxEntity_CXXInstanceMethod = 21, + CXIdxEntity_CXXConstructor = 22, + CXIdxEntity_CXXDestructor = 23, + CXIdxEntity_CXXConversionFunction = 24, + CXIdxEntity_CXXTypeAlias = 25, + CXIdxEntity_CXXInterface = 26, + CXIdxEntity_CXXConcept = 27 + +} CXIdxEntityKind; + +typedef enum { + CXIdxEntityLang_None = 0, + CXIdxEntityLang_C = 1, + CXIdxEntityLang_ObjC = 2, + CXIdxEntityLang_CXX = 3, + CXIdxEntityLang_Swift = 4 +} CXIdxEntityLanguage; + +/** + * Extra C++ template information for an entity. This can apply to: + * CXIdxEntity_Function + * CXIdxEntity_CXXClass + * CXIdxEntity_CXXStaticMethod + * CXIdxEntity_CXXInstanceMethod + * CXIdxEntity_CXXConstructor + * CXIdxEntity_CXXConversionFunction + * CXIdxEntity_CXXTypeAlias + */ +typedef enum { + CXIdxEntity_NonTemplate = 0, + CXIdxEntity_Template = 1, + CXIdxEntity_TemplatePartialSpecialization = 2, + CXIdxEntity_TemplateSpecialization = 3 +} CXIdxEntityCXXTemplateKind; + +typedef enum { + CXIdxAttr_Unexposed = 0, + CXIdxAttr_IBAction = 1, + CXIdxAttr_IBOutlet = 2, + CXIdxAttr_IBOutletCollection = 3 +} CXIdxAttrKind; + +typedef struct { + CXIdxAttrKind kind; + CXCursor cursor; + CXIdxLoc loc; +} CXIdxAttrInfo; + +typedef struct { + CXIdxEntityKind kind; + CXIdxEntityCXXTemplateKind templateKind; + CXIdxEntityLanguage lang; + const char *name; + const char *USR; + CXCursor cursor; + const CXIdxAttrInfo *const *attributes; + unsigned numAttributes; +} CXIdxEntityInfo; + +typedef struct { + CXCursor cursor; +} CXIdxContainerInfo; + +typedef struct { + const CXIdxAttrInfo *attrInfo; + const CXIdxEntityInfo *objcClass; + CXCursor classCursor; + CXIdxLoc classLoc; +} CXIdxIBOutletCollectionAttrInfo; + +typedef enum { CXIdxDeclFlag_Skipped = 0x1 } CXIdxDeclInfoFlags; + +typedef struct { + const CXIdxEntityInfo *entityInfo; + CXCursor cursor; + CXIdxLoc loc; + const CXIdxContainerInfo *semanticContainer; + /** + * Generally same as #semanticContainer but can be different in + * cases like out-of-line C++ member functions. + */ + const CXIdxContainerInfo *lexicalContainer; + int isRedeclaration; + int isDefinition; + int isContainer; + const CXIdxContainerInfo *declAsContainer; + /** + * Whether the declaration exists in code or was created implicitly + * by the compiler, e.g. implicit Objective-C methods for properties. + */ + int isImplicit; + const CXIdxAttrInfo *const *attributes; + unsigned numAttributes; + + unsigned flags; + +} CXIdxDeclInfo; + +typedef enum { + CXIdxObjCContainer_ForwardRef = 0, + CXIdxObjCContainer_Interface = 1, + CXIdxObjCContainer_Implementation = 2 +} CXIdxObjCContainerKind; + +typedef struct { + const CXIdxDeclInfo *declInfo; + CXIdxObjCContainerKind kind; +} CXIdxObjCContainerDeclInfo; + +typedef struct { + const CXIdxEntityInfo *base; + CXCursor cursor; + CXIdxLoc loc; +} CXIdxBaseClassInfo; + +typedef struct { + const CXIdxEntityInfo *protocol; + CXCursor cursor; + CXIdxLoc loc; +} CXIdxObjCProtocolRefInfo; + +typedef struct { + const CXIdxObjCProtocolRefInfo *const *protocols; + unsigned numProtocols; +} CXIdxObjCProtocolRefListInfo; + +typedef struct { + const CXIdxObjCContainerDeclInfo *containerInfo; + const CXIdxBaseClassInfo *superInfo; + const CXIdxObjCProtocolRefListInfo *protocols; +} CXIdxObjCInterfaceDeclInfo; + +typedef struct { + const CXIdxObjCContainerDeclInfo *containerInfo; + const CXIdxEntityInfo *objcClass; + CXCursor classCursor; + CXIdxLoc classLoc; + const CXIdxObjCProtocolRefListInfo *protocols; +} CXIdxObjCCategoryDeclInfo; + +typedef struct { + const CXIdxDeclInfo *declInfo; + const CXIdxEntityInfo *getter; + const CXIdxEntityInfo *setter; +} CXIdxObjCPropertyDeclInfo; + +typedef struct { + const CXIdxDeclInfo *declInfo; + const CXIdxBaseClassInfo *const *bases; + unsigned numBases; +} CXIdxCXXClassDeclInfo; + +/** + * Data for IndexerCallbacks#indexEntityReference. + * + * This may be deprecated in a future version as this duplicates + * the \c CXSymbolRole_Implicit bit in \c CXSymbolRole. + */ +typedef enum { + /** + * The entity is referenced directly in user's code. + */ + CXIdxEntityRef_Direct = 1, + /** + * An implicit reference, e.g. a reference of an Objective-C method + * via the dot syntax. + */ + CXIdxEntityRef_Implicit = 2 +} CXIdxEntityRefKind; + +/** + * Roles that are attributed to symbol occurrences. + * + * Internal: this currently mirrors low 9 bits of clang::index::SymbolRole with + * higher bits zeroed. These high bits may be exposed in the future. + */ +typedef enum { + CXSymbolRole_None = 0, + CXSymbolRole_Declaration = 1 << 0, + CXSymbolRole_Definition = 1 << 1, + CXSymbolRole_Reference = 1 << 2, + CXSymbolRole_Read = 1 << 3, + CXSymbolRole_Write = 1 << 4, + CXSymbolRole_Call = 1 << 5, + CXSymbolRole_Dynamic = 1 << 6, + CXSymbolRole_AddressOf = 1 << 7, + CXSymbolRole_Implicit = 1 << 8 +} CXSymbolRole; + +/** + * Data for IndexerCallbacks#indexEntityReference. + */ +typedef struct { + CXIdxEntityRefKind kind; + /** + * Reference cursor. + */ + CXCursor cursor; + CXIdxLoc loc; + /** + * The entity that gets referenced. + */ + const CXIdxEntityInfo *referencedEntity; + /** + * Immediate "parent" of the reference. For example: + * + * \code + * Foo *var; + * \endcode + * + * The parent of reference of type 'Foo' is the variable 'var'. + * For references inside statement bodies of functions/methods, + * the parentEntity will be the function/method. + */ + const CXIdxEntityInfo *parentEntity; + /** + * Lexical container context of the reference. + */ + const CXIdxContainerInfo *container; + /** + * Sets of symbol roles of the reference. + */ + CXSymbolRole role; +} CXIdxEntityRefInfo; + +/** + * A group of callbacks used by #clang_indexSourceFile and + * #clang_indexTranslationUnit. + */ +typedef struct { + /** + * Called periodically to check whether indexing should be aborted. + * Should return 0 to continue, and non-zero to abort. + */ + int (*abortQuery)(CXClientData client_data, void *reserved); + + /** + * Called at the end of indexing; passes the complete diagnostic set. + */ + void (*diagnostic)(CXClientData client_data, CXDiagnosticSet, void *reserved); + + CXIdxClientFile (*enteredMainFile)(CXClientData client_data, CXFile mainFile, + void *reserved); + + /** + * Called when a file gets \#included/\#imported. + */ + CXIdxClientFile (*ppIncludedFile)(CXClientData client_data, + const CXIdxIncludedFileInfo *); + + /** + * Called when a AST file (PCH or module) gets imported. + * + * AST files will not get indexed (there will not be callbacks to index all + * the entities in an AST file). The recommended action is that, if the AST + * file is not already indexed, to initiate a new indexing job specific to + * the AST file. + */ + CXIdxClientASTFile (*importedASTFile)(CXClientData client_data, + const CXIdxImportedASTFileInfo *); + + /** + * Called at the beginning of indexing a translation unit. + */ + CXIdxClientContainer (*startedTranslationUnit)(CXClientData client_data, + void *reserved); + + void (*indexDeclaration)(CXClientData client_data, const CXIdxDeclInfo *); + + /** + * Called to index a reference of an entity. + */ + void (*indexEntityReference)(CXClientData client_data, + const CXIdxEntityRefInfo *); + +} IndexerCallbacks; + +CINDEX_LINKAGE int clang_index_isEntityObjCContainerKind(CXIdxEntityKind); +CINDEX_LINKAGE const CXIdxObjCContainerDeclInfo * +clang_index_getObjCContainerDeclInfo(const CXIdxDeclInfo *); + +CINDEX_LINKAGE const CXIdxObjCInterfaceDeclInfo * +clang_index_getObjCInterfaceDeclInfo(const CXIdxDeclInfo *); + +CINDEX_LINKAGE +const CXIdxObjCCategoryDeclInfo * +clang_index_getObjCCategoryDeclInfo(const CXIdxDeclInfo *); + +CINDEX_LINKAGE const CXIdxObjCProtocolRefListInfo * +clang_index_getObjCProtocolRefListInfo(const CXIdxDeclInfo *); + +CINDEX_LINKAGE const CXIdxObjCPropertyDeclInfo * +clang_index_getObjCPropertyDeclInfo(const CXIdxDeclInfo *); + +CINDEX_LINKAGE const CXIdxIBOutletCollectionAttrInfo * +clang_index_getIBOutletCollectionAttrInfo(const CXIdxAttrInfo *); + +CINDEX_LINKAGE const CXIdxCXXClassDeclInfo * +clang_index_getCXXClassDeclInfo(const CXIdxDeclInfo *); + +/** + * For retrieving a custom CXIdxClientContainer attached to a + * container. + */ +CINDEX_LINKAGE CXIdxClientContainer +clang_index_getClientContainer(const CXIdxContainerInfo *); + +/** + * For setting a custom CXIdxClientContainer attached to a + * container. + */ +CINDEX_LINKAGE void clang_index_setClientContainer(const CXIdxContainerInfo *, + CXIdxClientContainer); + +/** + * For retrieving a custom CXIdxClientEntity attached to an entity. + */ +CINDEX_LINKAGE CXIdxClientEntity +clang_index_getClientEntity(const CXIdxEntityInfo *); + +/** + * For setting a custom CXIdxClientEntity attached to an entity. + */ +CINDEX_LINKAGE void clang_index_setClientEntity(const CXIdxEntityInfo *, + CXIdxClientEntity); + +/** + * An indexing action/session, to be applied to one or multiple + * translation units. + */ +typedef void *CXIndexAction; + +/** + * An indexing action/session, to be applied to one or multiple + * translation units. + * + * \param CIdx The index object with which the index action will be associated. + */ +CINDEX_LINKAGE CXIndexAction clang_IndexAction_create(CXIndex CIdx); + +/** + * Destroy the given index action. + * + * The index action must not be destroyed until all of the translation units + * created within that index action have been destroyed. + */ +CINDEX_LINKAGE void clang_IndexAction_dispose(CXIndexAction); + +typedef enum { + /** + * Used to indicate that no special indexing options are needed. + */ + CXIndexOpt_None = 0x0, + + /** + * Used to indicate that IndexerCallbacks#indexEntityReference should + * be invoked for only one reference of an entity per source file that does + * not also include a declaration/definition of the entity. + */ + CXIndexOpt_SuppressRedundantRefs = 0x1, + + /** + * Function-local symbols should be indexed. If this is not set + * function-local symbols will be ignored. + */ + CXIndexOpt_IndexFunctionLocalSymbols = 0x2, + + /** + * Implicit function/class template instantiations should be indexed. + * If this is not set, implicit instantiations will be ignored. + */ + CXIndexOpt_IndexImplicitTemplateInstantiations = 0x4, + + /** + * Suppress all compiler warnings when parsing for indexing. + */ + CXIndexOpt_SuppressWarnings = 0x8, + + /** + * Skip a function/method body that was already parsed during an + * indexing session associated with a \c CXIndexAction object. + * Bodies in system headers are always skipped. + */ + CXIndexOpt_SkipParsedBodiesInSession = 0x10 + +} CXIndexOptFlags; + +/** + * Index the given source file and the translation unit corresponding + * to that file via callbacks implemented through #IndexerCallbacks. + * + * \param client_data pointer data supplied by the client, which will + * be passed to the invoked callbacks. + * + * \param index_callbacks Pointer to indexing callbacks that the client + * implements. + * + * \param index_callbacks_size Size of #IndexerCallbacks structure that gets + * passed in index_callbacks. + * + * \param index_options A bitmask of options that affects how indexing is + * performed. This should be a bitwise OR of the CXIndexOpt_XXX flags. + * + * \param[out] out_TU pointer to store a \c CXTranslationUnit that can be + * reused after indexing is finished. Set to \c NULL if you do not require it. + * + * \returns 0 on success or if there were errors from which the compiler could + * recover. If there is a failure from which there is no recovery, returns + * a non-zero \c CXErrorCode. + * + * The rest of the parameters are the same as #clang_parseTranslationUnit. + */ +CINDEX_LINKAGE int clang_indexSourceFile( + CXIndexAction, CXClientData client_data, IndexerCallbacks *index_callbacks, + unsigned index_callbacks_size, unsigned index_options, + const char *source_filename, const char *const *command_line_args, + int num_command_line_args, struct CXUnsavedFile *unsaved_files, + unsigned num_unsaved_files, CXTranslationUnit *out_TU, unsigned TU_options); + +/** + * Same as clang_indexSourceFile but requires a full command line + * for \c command_line_args including argv[0]. This is useful if the standard + * library paths are relative to the binary. + */ +CINDEX_LINKAGE int clang_indexSourceFileFullArgv( + CXIndexAction, CXClientData client_data, IndexerCallbacks *index_callbacks, + unsigned index_callbacks_size, unsigned index_options, + const char *source_filename, const char *const *command_line_args, + int num_command_line_args, struct CXUnsavedFile *unsaved_files, + unsigned num_unsaved_files, CXTranslationUnit *out_TU, unsigned TU_options); + +/** + * Index the given translation unit via callbacks implemented through + * #IndexerCallbacks. + * + * The order of callback invocations is not guaranteed to be the same as + * when indexing a source file. The high level order will be: + * + * -Preprocessor callbacks invocations + * -Declaration/reference callbacks invocations + * -Diagnostic callback invocations + * + * The parameters are the same as #clang_indexSourceFile. + * + * \returns If there is a failure from which there is no recovery, returns + * non-zero, otherwise returns 0. + */ +CINDEX_LINKAGE int clang_indexTranslationUnit( + CXIndexAction, CXClientData client_data, IndexerCallbacks *index_callbacks, + unsigned index_callbacks_size, unsigned index_options, CXTranslationUnit); + +/** + * Retrieve the CXIdxFile, file, line, column, and offset represented by + * the given CXIdxLoc. + * + * If the location refers into a macro expansion, retrieves the + * location of the macro expansion and if it refers into a macro argument + * retrieves the location of the argument. + */ +CINDEX_LINKAGE void clang_indexLoc_getFileLocation(CXIdxLoc loc, + CXIdxClientFile *indexFile, + CXFile *file, unsigned *line, + unsigned *column, + unsigned *offset); + +/** + * Retrieve the CXSourceLocation represented by the given CXIdxLoc. + */ +CINDEX_LINKAGE +CXSourceLocation clang_indexLoc_getCXSourceLocation(CXIdxLoc loc); + +/** + * Visitor invoked for each field found by a traversal. + * + * This visitor function will be invoked for each field found by + * \c clang_Type_visitFields. Its first argument is the cursor being + * visited, its second argument is the client data provided to + * \c clang_Type_visitFields. + * + * The visitor should return one of the \c CXVisitorResult values + * to direct \c clang_Type_visitFields. + */ +typedef enum CXVisitorResult (*CXFieldVisitor)(CXCursor C, + CXClientData client_data); + +/** + * Visit the fields of a particular type. + * + * This function visits all the direct fields of the given cursor, + * invoking the given \p visitor function with the cursors of each + * visited field. The traversal may be ended prematurely, if + * the visitor returns \c CXFieldVisit_Break. + * + * \param T the record type whose field may be visited. + * + * \param visitor the visitor function that will be invoked for each + * field of \p T. + * + * \param client_data pointer data supplied by the client, which will + * be passed to the visitor each time it is invoked. + * + * \returns a non-zero value if the traversal was terminated + * prematurely by the visitor returning \c CXFieldVisit_Break. + */ +CINDEX_LINKAGE unsigned clang_Type_visitFields(CXType T, CXFieldVisitor visitor, + CXClientData client_data); + +/** + * Visit the base classes of a type. + * + * This function visits all the direct base classes of a the given cursor, + * invoking the given \p visitor function with the cursors of each + * visited base. The traversal may be ended prematurely, if + * the visitor returns \c CXFieldVisit_Break. + * + * \param T the record type whose field may be visited. + * + * \param visitor the visitor function that will be invoked for each + * field of \p T. + * + * \param client_data pointer data supplied by the client, which will + * be passed to the visitor each time it is invoked. + * + * \returns a non-zero value if the traversal was terminated + * prematurely by the visitor returning \c CXFieldVisit_Break. + */ +CINDEX_LINKAGE unsigned clang_visitCXXBaseClasses(CXType T, + CXFieldVisitor visitor, + CXClientData client_data); + +/** + * Describes the kind of binary operators. + */ +enum CXBinaryOperatorKind { + /** This value describes cursors which are not binary operators. */ + CXBinaryOperator_Invalid, + /** C++ Pointer - to - member operator. */ + CXBinaryOperator_PtrMemD, + /** C++ Pointer - to - member operator. */ + CXBinaryOperator_PtrMemI, + /** Multiplication operator. */ + CXBinaryOperator_Mul, + /** Division operator. */ + CXBinaryOperator_Div, + /** Remainder operator. */ + CXBinaryOperator_Rem, + /** Addition operator. */ + CXBinaryOperator_Add, + /** Subtraction operator. */ + CXBinaryOperator_Sub, + /** Bitwise shift left operator. */ + CXBinaryOperator_Shl, + /** Bitwise shift right operator. */ + CXBinaryOperator_Shr, + /** C++ three-way comparison (spaceship) operator. */ + CXBinaryOperator_Cmp, + /** Less than operator. */ + CXBinaryOperator_LT, + /** Greater than operator. */ + CXBinaryOperator_GT, + /** Less or equal operator. */ + CXBinaryOperator_LE, + /** Greater or equal operator. */ + CXBinaryOperator_GE, + /** Equal operator. */ + CXBinaryOperator_EQ, + /** Not equal operator. */ + CXBinaryOperator_NE, + /** Bitwise AND operator. */ + CXBinaryOperator_And, + /** Bitwise XOR operator. */ + CXBinaryOperator_Xor, + /** Bitwise OR operator. */ + CXBinaryOperator_Or, + /** Logical AND operator. */ + CXBinaryOperator_LAnd, + /** Logical OR operator. */ + CXBinaryOperator_LOr, + /** Assignment operator. */ + CXBinaryOperator_Assign, + /** Multiplication assignment operator. */ + CXBinaryOperator_MulAssign, + /** Division assignment operator. */ + CXBinaryOperator_DivAssign, + /** Remainder assignment operator. */ + CXBinaryOperator_RemAssign, + /** Addition assignment operator. */ + CXBinaryOperator_AddAssign, + /** Subtraction assignment operator. */ + CXBinaryOperator_SubAssign, + /** Bitwise shift left assignment operator. */ + CXBinaryOperator_ShlAssign, + /** Bitwise shift right assignment operator. */ + CXBinaryOperator_ShrAssign, + /** Bitwise AND assignment operator. */ + CXBinaryOperator_AndAssign, + /** Bitwise XOR assignment operator. */ + CXBinaryOperator_XorAssign, + /** Bitwise OR assignment operator. */ + CXBinaryOperator_OrAssign, + /** Comma operator. */ + CXBinaryOperator_Comma +}; + +/** + * Retrieve the spelling of a given CXBinaryOperatorKind. + */ +CINDEX_LINKAGE CXString +clang_getBinaryOperatorKindSpelling(enum CXBinaryOperatorKind kind); + +/** + * Retrieve the binary operator kind of this cursor. + * + * If this cursor is not a binary operator then returns Invalid. + */ +CINDEX_LINKAGE enum CXBinaryOperatorKind +clang_getCursorBinaryOperatorKind(CXCursor cursor); + +/** + * Describes the kind of unary operators. + */ +enum CXUnaryOperatorKind { + /** This value describes cursors which are not unary operators. */ + CXUnaryOperator_Invalid, + /** Postfix increment operator. */ + CXUnaryOperator_PostInc, + /** Postfix decrement operator. */ + CXUnaryOperator_PostDec, + /** Prefix increment operator. */ + CXUnaryOperator_PreInc, + /** Prefix decrement operator. */ + CXUnaryOperator_PreDec, + /** Address of operator. */ + CXUnaryOperator_AddrOf, + /** Dereference operator. */ + CXUnaryOperator_Deref, + /** Plus operator. */ + CXUnaryOperator_Plus, + /** Minus operator. */ + CXUnaryOperator_Minus, + /** Not operator. */ + CXUnaryOperator_Not, + /** LNot operator. */ + CXUnaryOperator_LNot, + /** "__real expr" operator. */ + CXUnaryOperator_Real, + /** "__imag expr" operator. */ + CXUnaryOperator_Imag, + /** __extension__ marker operator. */ + CXUnaryOperator_Extension, + /** C++ co_await operator. */ + CXUnaryOperator_Coawait +}; + +/** + * Retrieve the spelling of a given CXUnaryOperatorKind. + */ +CINDEX_LINKAGE CXString +clang_getUnaryOperatorKindSpelling(enum CXUnaryOperatorKind kind); + +/** + * Retrieve the unary operator kind of this cursor. + * + * If this cursor is not a unary operator then returns Invalid. + */ +CINDEX_LINKAGE enum CXUnaryOperatorKind +clang_getCursorUnaryOperatorKind(CXCursor cursor); + +/** + * @} + */ + +/** + * @} + */ + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Platform.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Platform.h new file mode 100644 index 0000000..67c1fff --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Platform.h @@ -0,0 +1,53 @@ +/*===-- clang-c/Platform.h - C Index platform decls -------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*| +|* *| +|* This header provides platform specific macros (dllimport, deprecated, ...) *| +|* *| +\*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_PLATFORM_H +#define LLVM_CLANG_C_PLATFORM_H + +#include "clang-c/ExternC.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +/* Windows DLL import/export. */ +#ifndef CINDEX_NO_EXPORTS + #define CINDEX_EXPORTS +#endif +#ifdef _WIN32 + #ifdef CINDEX_EXPORTS + #ifdef _CINDEX_LIB_ + #define CINDEX_LINKAGE __declspec(dllexport) + #else + #define CINDEX_LINKAGE __declspec(dllimport) + #endif + #endif +#elif defined(CINDEX_EXPORTS) && defined(__GNUC__) + #define CINDEX_LINKAGE __attribute__((visibility("default"))) +#endif + +#ifndef CINDEX_LINKAGE + #define CINDEX_LINKAGE +#endif + +#ifdef __GNUC__ + #define CINDEX_DEPRECATED __attribute__((deprecated)) +#else + #ifdef _MSC_VER + #define CINDEX_DEPRECATED __declspec(deprecated) + #else + #define CINDEX_DEPRECATED + #endif +#endif + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Rewrite.h b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Rewrite.h new file mode 100644 index 0000000..ce1b055 --- /dev/null +++ b/libraries/rythe/reflection-core/third_party/clang-c/include/clang-c/Rewrite.h @@ -0,0 +1,63 @@ +/*===-- clang-c/Rewrite.h - C CXRewriter --------------------------*- C -*-===*\ +|* *| +|* Part of the LLVM Project, under the Apache License v2.0 with LLVM *| +|* Exceptions. *| +|* See https://llvm.org/LICENSE.txt for license information. *| +|* SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception *| +|* *| +|*===----------------------------------------------------------------------===*/ + +#ifndef LLVM_CLANG_C_REWRITE_H +#define LLVM_CLANG_C_REWRITE_H + +#include "clang-c/CXString.h" +#include "clang-c/ExternC.h" +#include "clang-c/Index.h" +#include "clang-c/Platform.h" + +LLVM_CLANG_C_EXTERN_C_BEGIN + +typedef void *CXRewriter; + +/** + * Create CXRewriter. + */ +CINDEX_LINKAGE CXRewriter clang_CXRewriter_create(CXTranslationUnit TU); + +/** + * Insert the specified string at the specified location in the original buffer. + */ +CINDEX_LINKAGE void clang_CXRewriter_insertTextBefore(CXRewriter Rew, CXSourceLocation Loc, + const char *Insert); + +/** + * Replace the specified range of characters in the input with the specified + * replacement. + */ +CINDEX_LINKAGE void clang_CXRewriter_replaceText(CXRewriter Rew, CXSourceRange ToBeReplaced, + const char *Replacement); + +/** + * Remove the specified range. + */ +CINDEX_LINKAGE void clang_CXRewriter_removeText(CXRewriter Rew, CXSourceRange ToBeRemoved); + +/** + * Save all changed files to disk. + * Returns 1 if any files were not saved successfully, returns 0 otherwise. + */ +CINDEX_LINKAGE int clang_CXRewriter_overwriteChangedFiles(CXRewriter Rew); + +/** + * Write out rewritten version of the main file to stdout. + */ +CINDEX_LINKAGE void clang_CXRewriter_writeMainFileToStdOut(CXRewriter Rew); + +/** + * Free the given CXRewriter. + */ +CINDEX_LINKAGE void clang_CXRewriter_dispose(CXRewriter Rew); + +LLVM_CLANG_C_EXTERN_C_END + +#endif diff --git a/libraries/rythe/reflection-core/third_party/lib/clangAST.lib b/libraries/rythe/reflection-core/third_party/lib/clangAST.lib new file mode 100644 index 0000000..da4a9cb Binary files /dev/null and b/libraries/rythe/reflection-core/third_party/lib/clangAST.lib differ diff --git a/libraries/rythe/reflection-core/third_party/lib/clangASTMatchers.lib b/libraries/rythe/reflection-core/third_party/lib/clangASTMatchers.lib new file mode 100644 index 0000000..4255461 Binary files /dev/null and b/libraries/rythe/reflection-core/third_party/lib/clangASTMatchers.lib differ diff --git a/libraries/rythe/reflection-core/third_party/lib/libclang.dll b/libraries/rythe/reflection-core/third_party/lib/libclang.dll new file mode 100644 index 0000000..3953a6d Binary files /dev/null and b/libraries/rythe/reflection-core/third_party/lib/libclang.dll differ diff --git a/libraries/rythe/reflection-core/third_party/lib/libclang.lib b/libraries/rythe/reflection-core/third_party/lib/libclang.lib new file mode 100644 index 0000000..2e3456b Binary files /dev/null and b/libraries/rythe/reflection-core/third_party/lib/libclang.lib differ diff --git a/libraries/rythe/rythe-standard-library b/libraries/rythe/rythe-standard-library index a49594e..61d4cf1 160000 --- a/libraries/rythe/rythe-standard-library +++ b/libraries/rythe/rythe-standard-library @@ -1 +1 @@ -Subproject commit a49594ec811413b3a644d33af46c4e1c1f06585e +Subproject commit 61d4cf1487856324a5ff487536d56371a5474a79 diff --git a/rythe-build-system b/rythe-build-system index 922fafb..cd473d0 160000 --- a/rythe-build-system +++ b/rythe-build-system @@ -1 +1 @@ -Subproject commit 922fafbee5e4dc9b26d6d959a025812a9ef26634 +Subproject commit cd473d038bfaa851c8d54828b9c5c60f63d027d4