Merge branch 'ly/changed-paths-traversal'

Lift the limitation to use changed-path filter in "git log" so that
it can be used for a pathspec with multiple literal paths.

* ly/changed-paths-traversal:
  bloom: optimize multiple pathspec items in revision
  revision: make helper for pathspec to bloom keyvec
  bloom: replace struct bloom_key * with struct bloom_keyvec
  bloom: rename function operates on bloom_key
  bloom: add test helper to return murmur3 hash

Author: gitster

blame.c

@@ -1311,7 +1311,7 @@ static void add_bloom_key(struct blame_bloom_data *bd,
 	}
 
 	bd->keys[bd->nr] = xmalloc(sizeof(struct bloom_key));
-	fill_bloom_key(path, strlen(path), bd->keys[bd->nr], bd->settings);
+	bloom_key_fill(bd->keys[bd->nr], path, strlen(path), bd->settings);
 	bd->nr++;
 }
 

bloom.c

@@ -107,7 +107,7 @@ int load_bloom_filter_from_graph(struct commit_graph *g,
  * Not considered to be cryptographically secure.
  * Implemented as described in https://en.wikipedia.org/wiki/MurmurHash#Algorithm
  */
-uint32_t murmur3_seeded_v2(uint32_t seed, const char *data, size_t len)
+static uint32_t murmur3_seeded_v2(uint32_t seed, const char *data, size_t len)
 {
 	const uint32_t c1 = 0xcc9e2d51;
 	const uint32_t c2 = 0x1b873593;
@@ -221,9 +221,7 @@ static uint32_t murmur3_seeded_v1(uint32_t seed, const char *data, size_t len)
 	return seed;
 }
 
-void fill_bloom_key(const char *data,
-		    size_t len,
-		    struct bloom_key *key,
+void bloom_key_fill(struct bloom_key *key, const char *data, size_t len,
 		    const struct bloom_filter_settings *settings)
 {
 	int i;
@@ -243,7 +241,7 @@ void fill_bloom_key(const char *data,
 		key->hashes[i] = hash0 + i * hash1;
 }
 
-void clear_bloom_key(struct bloom_key *key)
+void bloom_key_clear(struct bloom_key *key)
 {
 	FREE_AND_NULL(key->hashes);
 }
@@ -280,6 +278,55 @@ void deinit_bloom_filters(void)
 	deep_clear_bloom_filter_slab(&bloom_filters, free_one_bloom_filter);
 }
 
+struct bloom_keyvec *bloom_keyvec_new(const char *path, size_t len,
+				      const struct bloom_filter_settings *settings)
+{
+	struct bloom_keyvec *vec;
+	const char *p;
+	size_t sz;
+	size_t nr = 1;
+
+	p = path;
+	while (*p) {
+		/*
+		 * At this point, the path is normalized to use Unix-style
+		 * path separators. This is required due to how the
+		 * changed-path Bloom filters store the paths.
+		 */
+		if (*p == '/')
+			nr++;
+		p++;
+	}
+
+	sz = sizeof(struct bloom_keyvec);
+	sz += nr * sizeof(struct bloom_key);
+	vec = (struct bloom_keyvec *)xcalloc(1, sz);
+	if (!vec)
+		return NULL;
+	vec->count = nr;
+
+	bloom_key_fill(&vec->key[0], path, len, settings);
+	nr = 1;
+	p = path + len - 1;
+	while (p > path) {
+		if (*p == '/') {
+			bloom_key_fill(&vec->key[nr++], path, p - path, settings);
+		}
+		p--;
+	}
+	assert(nr == vec->count);
+	return vec;
+}
+
+void bloom_keyvec_free(struct bloom_keyvec *vec)
+{
+	if (!vec)
+		return;
+	for (size_t nr = 0; nr < vec->count; nr++)
+		bloom_key_clear(&vec->key[nr]);
+	free(vec);
+}
+
 static int pathmap_cmp(const void *hashmap_cmp_fn_data UNUSED,
 		       const struct hashmap_entry *eptr,
 		       const struct hashmap_entry *entry_or_key,
@@ -500,9 +547,9 @@ struct bloom_filter *get_or_compute_bloom_filter(struct repository *r,
 
 		hashmap_for_each_entry(&pathmap, &iter, e, entry) {
 			struct bloom_key key;
-			fill_bloom_key(e->path, strlen(e->path), &key, settings);
+			bloom_key_fill(&key, e->path, strlen(e->path), settings);
 			add_key_to_filter(&key, filter, settings);
-			clear_bloom_key(&key);
+			bloom_key_clear(&key);
 		}
 
 	cleanup:
@@ -540,3 +587,26 @@ int bloom_filter_contains(const struct bloom_filter *filter,
 
 	return 1;
 }
+
+int bloom_filter_contains_vec(const struct bloom_filter *filter,
+			      const struct bloom_keyvec *vec,
+			      const struct bloom_filter_settings *settings)
+{
+	int ret = 1;
+
+	for (size_t nr = 0; ret > 0 && nr < vec->count; nr++)
+		ret = bloom_filter_contains(filter, &vec->key[nr], settings);
+
+	return ret;
+}
+
+uint32_t test_bloom_murmur3_seeded(uint32_t seed, const char *data, size_t len,
+				   int version)
+{
+	assert(version == 1 || version == 2);
+
+	if (version == 2)
+		return murmur3_seeded_v2(seed, data, len);
+	else
+		return murmur3_seeded_v1(seed, data, len);
+}

bloom.h

@@ -74,24 +74,40 @@ struct bloom_key {
 	uint32_t *hashes;
 };
 
+/*
+ * A bloom_keyvec is a vector of bloom_keys, which
+ * can be used to store multiple keys for a single
+ * pathspec item.
+ */
+struct bloom_keyvec {
+	size_t count;
+	struct bloom_key key[FLEX_ARRAY];
+};
+
 int load_bloom_filter_from_graph(struct commit_graph *g,
 				 struct bloom_filter *filter,
 				 uint32_t graph_pos);
 
+void bloom_key_fill(struct bloom_key *key, const char *data, size_t len,
+		    const struct bloom_filter_settings *settings);
+void bloom_key_clear(struct bloom_key *key);
+
 /*
- * Calculate the murmur3 32-bit hash value for the given data
- * using the given seed.
- * Produces a uniformly distributed hash value.
- * Not considered to be cryptographically secure.
- * Implemented as described in https://en.wikipedia.org/wiki/MurmurHash#Algorithm
+ * bloom_keyvec_new - Allocate and populate a bloom_keyvec with keys for the
+ * given path.
+ *
+ * This function splits the input path by '/' and generates a bloom key for each
+ * prefix, in reverse order of specificity. For example, given the input
+ * "a/b/c", it will generate bloom keys for:
+ *   - "a/b/c"
+ *   - "a/b"
+ *   - "a"
+ *
+ * The resulting keys are stored in a newly allocated bloom_keyvec.
  */
-uint32_t murmur3_seeded_v2(uint32_t seed, const char *data, size_t len);
-
-void fill_bloom_key(const char *data,
-		    size_t len,
-		    struct bloom_key *key,
-		    const struct bloom_filter_settings *settings);
-void clear_bloom_key(struct bloom_key *key);
+struct bloom_keyvec *bloom_keyvec_new(const char *path, size_t len,
+				      const struct bloom_filter_settings *settings);
+void bloom_keyvec_free(struct bloom_keyvec *vec);
 
 void add_key_to_filter(const struct bloom_key *key,
 		       struct bloom_filter *filter,
@@ -137,4 +153,18 @@ int bloom_filter_contains(const struct bloom_filter *filter,
 			  const struct bloom_key *key,
 			  const struct bloom_filter_settings *settings);
 
+/*
+ * bloom_filter_contains_vec - Check if all keys in a key vector are in the
+ * Bloom filter.
+ *
+ * Returns 1 if **all** keys in the vector are present in the filter,
+ * 0 if **any** key is not present.
+ */
+int bloom_filter_contains_vec(const struct bloom_filter *filter,
+			      const struct bloom_keyvec *v,
+			      const struct bloom_filter_settings *settings);
+
+uint32_t test_bloom_murmur3_seeded(uint32_t seed, const char *data, size_t len,
+				   int version);
+
 #endif

line-log.c

@@ -1172,12 +1172,13 @@ static int bloom_filter_check(struct rev_info *rev,
 		return 0;
 
 	while (!result && range) {
-		fill_bloom_key(range->path, strlen(range->path), &key, rev->bloom_filter_settings);
+		bloom_key_fill(&key, range->path, strlen(range->path),
+			       rev->bloom_filter_settings);
 
 		if (bloom_filter_contains(filter, &key, rev->bloom_filter_settings))
 			result = 1;
 
-		clear_bloom_key(&key);
+		bloom_key_clear(&key);
 		range = range->next;
 	}