mrcvtl wrote: > I think we should revert to the previous approach, what do you think? > > ```c++ > if (IsGslPtrValueFromGslTempOwner && DiagLoc.isValid()) { > if (const auto *VD = > dyn_cast_if_present<VarDecl>(ExtendingEntity->getDecl()); > SemaRef.getLangOpts().CPlusPlus23 && VD && > VD->isCXXForRangeImplicitVar()) > return true; > SemaRef.Diag(DiagLoc, diag::warn_dangling_lifetime_pointer) > << DiagRange; > return false; > } > ```
Thanks for debugging that! I agree btw, didn't think about this side effect before. https://github.com/llvm/llvm-project/pull/145164 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits