On 7/9/24 9:44 AM, Nathaniel Shead wrote:
On Mon, Jul 08, 2024 at 12:26:41PM -0400, Jason Merrill wrote:
For a using-decl in the same scope as the original decl, won't this replace
it so only the using-decl is visible to lookup? I had expected to omit the
USING_DECL in that case.
Yup it will; I think I'd originally done that so that more recent
(re-)declaration would be the one referred to by diagnostics, but on
retrospect that seems unhelpful; fixed. (Though need to keep the
replacement for CONST_DECLs, because the modules handling otherwise only
handles them in the context of their containing enumeration type, which
isn't what we want here; I've added a new test for this as well.)
Ah, using-25, sure. I would think we could still tell the difference by
comparing PURVIEW/EXPORT on the CONST_DECL to those of its type?
Or perhaps have add_binding_entity skip implicitly inserted enumerators,
and instead insert them again when reading the enum, which should also
save a bit of space.
Jason
