[clang] [clang][Sema] Diagnose unsatisfied constraints on rejected partial specializations (PR #239106)
Oliver Hunt via cfe-commits
cfe-commits at lists.llvm.org
Fri Nov 6 01:43:31 PST 2026
ojhunt wrote:
> Thank you for the detailed review and guidance, @ojhunt!
>
> I have addressed all the feedback in the latest commit:
>
> * **Doxygen comments:** Replaced the comments in `clang/include/clang/Sema/Sema.h` with concise, standard Clang documentation.
> * **Code deduplication:** Unified the class template and variable template partial specialization diagnostic logic into a single `NoteUnsatisfiedPartialSpecializationsImpl` template helper, removing the duplicated traversal, argument deduction, and binding formatting logic.
> * **`dyn_cast_or_null`:** Replaced all redundant `dyn_cast_or_null` calls with `dyn_cast` across `SemaDeclCXX.cpp`, `SemaExpr.cpp`, `SemaExprMember.cpp`, and `SemaTemplate.cpp` where the context pointer is already verified non-null.
> * **Assertion:** Added `assert(Templated && "Expected non-null templated declaration");` in `getTemplateParameters` in `SemaOverload.cpp`.
> * **Null check on `Selected`:** Clarified with comments that `Selected` is intentionally null when the primary template was instantiated (as opposed to a partial specialization).
> * **Test formatting:** Updated `concepts-class-template-partial-spec-notes.cpp` to use `#label` line tags (e.g., `// #case1_spec`) and placed the corresponding `// expected-note@#case1_spec` directives at the instantiation sites.
>
> All tests have been verified with `lit` and pass cleanly.
Very quick question - were these bullet points written by you or an llm?
https://github.com/llvm/llvm-project/pull/239106
More information about the cfe-commits
mailing list