https://gcc.gnu.org/bugzilla/show_bug.cgi?id=114685
Jason Merrill <jason at gcc dot gnu.org> changed:
What |Removed |Added
----------------------------------------------------------------------------
Resolution|--- |FIXED
CC| |jason at gcc dot gnu.org
Status|NEW |RESOLVED
--- Comment #5 from Jason Merrill <jason at gcc dot gnu.org> ---
(In reply to Nathaniel Shead from comment #2)
> This fails on
>
> if (data->hash->add_namespace_entities (decl, data->partitions))
> {
> /* It contains an exported thing, so it is exported. */
> gcc_checking_assert (DECL_MODULE_PURVIEW_P (decl));
> DECL_MODULE_EXPORT_P (decl) = true;
> }
>
> ultimately for a very similar reason to PR114683: non-functions brought in
> by using-decls currently just bind the value directly which means that
> there's nothing to indicate that they should be exported from this scope
> (and not their original one).
I just hacked around this in r15-1223 by changing the assert to setting
DECL_MODULE_PURVIEW_P.
I take your point that this shouldn't be happening; the exported decl is the
using-decl, not the original one, so add_namespace_entities shouldn't return
true for __gnu_cxx. But since it does, I think my change is a reasonable
workaround. I'll adjust the comment.
