MythreyaK wrote: Apologies for the delay and thank you for the review!
> Wouldn't it make more sense to use a SmallVector<variant<ParmVarDecl, > FieldDecl>> and resolve them individually? I thought of that initially, but I couldn't think of a construct where both `ParmVarDecl` and `FieldDecl` could initialize the same structure. My understanding was that if we resolve the parameter to a `ParmVarDecl`, then the underlying struct has a constructor which prevents the struct from being direct-initialized. I opted to returning the full `CXXRecordDecl` instead of the individual `FieldDecl`s as it made it easier to inspect the base classes (at the callsite) and drop those members. > Also, this looks like it would fail if we are also initializing base class > fields, e.g. I believe that is covered in `CXX20AggregateParenInitNoCtorDerived`. Please let me know if that isn't the case! https://github.com/llvm/llvm-project/pull/176635 _______________________________________________ cfe-commits mailing list [email protected] https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
