asavonic added a comment. In https://reviews.llvm.org/D49723#1178127, @Anastasia wrote:
> In https://reviews.llvm.org/D49723#1174837, @asavonic wrote: > > > In https://reviews.llvm.org/D49723#1173352, @Anastasia wrote: > > > > > Btw, has this restriction been removed from CL 2.0? > > > > > > No, it applies for CL2.0 as well. > > > It seems however the restriction on pointer to pointer was removed (see > s6.9.a last item) in CL2.0. Right, and it seems that pointers in struct arguments should also be legal in CL2.0. I'll submit another patch to remove this check for CL2.0. ================ Comment at: lib/Sema/SemaDecl.cpp:8187 + // walk around RecordDecl::fields(). + assert((PT->isArrayType() || PT->isRecordType()) && "Unexpected type."); + const Type *FieldRecTy = Field->getType()->getPointeeOrArrayElementType(); ---------------- Anastasia wrote: > Do we need to assert PT here too? It doesn't seem to be modified in this > loop... We should check for a field instead. Thanks! Repository: rC Clang https://reviews.llvm.org/D49723 _______________________________________________ cfe-commits mailing list cfe-commits@lists.llvm.org http://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
