Rust: Generalize some models

[geoffw0]

Generalize some existing models to trait models.

• generally that means replacing a model <a as b>::c with <_ as b>::c, unless there's a good reason why that model only applies to a.
• resulting duplicate models have been removed.
• there are a few odd models, I've leaned towards preserving behaviour and thus results, rather than maximal cleanliness.
• works well in the tests I tried (i.e. finds a few more results), full CI and DCA to follow.

[Copilot]

Pull Request Overview

This PR generalizes existing Rust models by replacing specific type implementations with trait models, improving data flow detection. The generalization transforms models like <specific_type as trait>::method to <_ as trait>::method, allowing the models to apply to all types implementing the trait rather than just specific implementations.

Key changes:

• Generalized trait models in stdlib framework definitions
• Updated test expectations to reflect improved data flow detection
• Added change note documenting the improvement

Reviewed Changes

Copilot reviewed 8 out of 8 changed files in this pull request and generated no comments.

Show a summary per file

File	Description
rust/ql/test/library-tests/dataflow/strings/main.rs	Updated comment to reflect newly detected taint flow
rust/ql/test/library-tests/dataflow/strings/inline-taint-flow.expected	Updated test expectations with new generalized models and additional flow detection
rust/ql/test/library-tests/dataflow/sources/InlineFlow.expected	Updated test expectations reflecting model generalizations and renumbering
rust/ql/test/library-tests/dataflow/local/main.rs	Updated comments to reflect newly detected taint flows
rust/ql/lib/codeql/rust/frameworks/stdlib/lang-core.model.yml	Generalized iterator and conversion trait models
rust/ql/lib/codeql/rust/frameworks/stdlib/lang-alloc.model.yml	Generalized string, arithmetic, and allocator trait models
rust/ql/lib/codeql/rust/frameworks/stdlib/io.model.yml	Generalized IO trait models
rust/ql/lib/change-notes/2025-10-15-models.md	Added change note documenting the improvement

[geoffw0]

DCA LGTM.

[paldepind]

LGTM!

[paldepind]

Nice! 🎉

[hvitved]

Is this model needed?

[geoffw0]

Yes - its needed for this line in the dataflow/strings test:

sink("Hello ".to_string() + &s1); // $ hasTaintFlow=37

i.e. its a workaround for dealing with the implicit dereference on the RHS when the + is appending strings.

[hvitved]

I wonder if we should have made the model more explicit then, i.e. alloc::string::String as core::ops::arith::Add, since the implementation of + uses a &str as the RHS type instead of the default Self: https://doc.rust-lang.org/std/string/struct.String.html#impl-Add%3C%26str%3E-for-String.

[geoffw0]

I'd rather keep it general I think - I don't think there's anything stopping people implementing it that way for other types and if they do, we'll cover that.