mizvekov added inline comments.
================
Comment at: clang/lib/Sema/SemaTemplate.cpp:1534-1538
+ if (TInfo->getType()->getContainedAutoType()) {
Diag(D.getIdentifierLoc(),
diag::warn_cxx14_compat_template_nontype_parm_auto_type)
<< QualType(TInfo->getType()->getContainedAutoType(), 0);
}
----------------
shafik wrote:
> shafik wrote:
> > mizvekov wrote:
> > > I think checking that the deduced type is undeduced was really
> > > unnecessary though, as I don't think we update the type source infos
> > > after deduction in any case.
> > I am not sure about the `DeducedTemplateSpecializationType` I was looking
> > at the change that brought in that type and it is not clear to me what a
> > good test would be. I believe outside of C++17 mode they would be hard
> > errors. Can you provide an example that would work outside of C++17 mode?
> The `DeducedTemplateSpecializationType` also feels a bit outside the bug I am
> fixing so I might just choose to address it in a follow-up.
Well the bit on your test below, which you marked with no diagnostic expected,
is an example.
This is something that works on C++20 but not on C++17, which is what this
warning is supposed to warn, right?
CHANGES SINCE LAST ACTION
https://reviews.llvm.org/D132990/new/
https://reviews.llvm.org/D132990
_______________________________________________
cfe-commits mailing list
cfe-commits@lists.llvm.org
https://lists.llvm.org/cgi-bin/mailman/listinfo/cfe-commits
