[Bug c++/103598] [12 regression] __PRETTY_FUNCTION_ lost namespaces in types since r12-5783-gf78eaffd1538efb4

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103598

Jason Merrill <jason at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |UNCONFIRMED
     Ever confirmed|1                           |0

--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> ---
(In reply to Andrew Pinski from comment #2)
> (In reply to Martin Liška from comment #1)
> > Confirmed, started with r12-5783-gf78eaffd1538efb4.
> 
> Hmm, actually it makes sense to some extend since llvm::getTypeName is in
> the llvm namespace printing llvm:: here is kinda of pointless really.

Exactly, this was the intent of the change.  Though I suppose it's debatable
whether the template bindings should be affected by it, since they aren't C++
syntax and so not clearly in the declarator scope.

But we don't make any guarantees about the stability of __PRETTY_FUNCTION__,
and it is affected by -fno-pretty-templates.  And it looks like clang leaves
out the "with", so it's unequal either way.