shafik wrote: > > Why remove the support for ParenListExpr entirely? Could you not fallback > > on looking at the init list expression only when ParenListExpr is null? > > We could keep `ParenListExpr`, but I'm not sure the benefit (IMO, removing it > seems like an API improvement, and make code simpler). The `ParenListExpr` is > redundant with the existing `Inits` parameter to some degree, the only > benefit to use it is for the precise `()` location information (as we deduct > the template type here, I think it is fine to use a less-precise source > location). > > Looking at all usages (total 4) of > `DeduceTemplateSpecializationFromInitializer` in clang, only 1 > [place](https://github.com/llvm/llvm-project/blob/main/clang/lib/Sema/SemaDecl.cpp#L12959) > passes the `ParenListExpr`.
As long as it does not effect the quality of the diagnostics then this makes sense. https://github.com/llvm/llvm-project/pull/75779 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
