olestrohm added a comment. Essentially, in each pass the type for the variable starts off being extract from the SourceTypeInfo, but since the address space isn't specified, this type is then missing an address space. Adding the address space deduction then reintroduces the address space to the type.
This is then done at least three times (pre-specialization, in the middle of specialization and when completing specialization), but since this is a template it may not be possible to correctly deduce the type earlier in case the template type introduces an address space in template specialization? Repository: rG LLVM Github Monorepo CHANGES SINCE LAST ACTION https://reviews.llvm.org/D82781/new/ https://reviews.llvm.org/D82781 _______________________________________________ cfe-commits mailing list firstname.lastname@example.org https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits