shafik added inline comments.
================ Comment at: clang/docs/ReleaseNotes.rst:571 (`#61758 <https://github.com/llvm/llvm-project/issues/61758>`_) +- Correcly diagnose jumps into statement expressions. + (`#63682 <https://github.com/llvm/llvm-project/issues/63682>`_) ---------------- Maybe remark this conforms with gcc behavior. ================ Comment at: clang/lib/Sema/JumpDiagnostics.cpp:489 + Scopes.push_back(GotoScope(ParentScope, + diag::note_enters_statement_expression, 0, + SE->getBeginLoc())); ---------------- I guess we don't have a variable that stands or no-diagnostic. ================ Comment at: clang/lib/Sema/SemaExpr.cpp:16468 PushExpressionEvaluationContext(ExprEvalContexts.back().Context); + setFunctionHasBranchProtectedScope(); } ---------------- I see why we need this but I am not sure how someone would know that need to do this in addition to the change in `BuildScopeInformation(...)` ================ Comment at: clang/test/Sema/scope-check.c:251 + { + (void)sizeof (int){({ goto O; 1; })}; // ok (not evaluated) + O: ; ---------------- I am sorry for this example: ``` int main() { sizeof(({goto label;}), 0); return 3; ({label:1;}); } ``` see godbolt: https://godbolt.org/z/aeb6TbsoG Note gcc's behavior Vs clangs. Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D154696/new/ https://reviews.llvm.org/D154696 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
