On Apr 2, 2018, Jason Merrill <ja...@redhat.com> wrote:
> On Sat, Mar 31, 2018 at 4:23 AM, Alexandre Oliva <aol...@redhat.com> wrote:
>> On Mar 30, 2018, Jason Merrill <ja...@redhat.com> wrote:
>> template <typename T>
>> void foo(T t) {
>> typename T::template C<auto> u = t;
>> T::template C<auto> (t);
>> T::template C<auto>::f (t, u);
>> }
> We should be able to distinguish those cases based on tag_type.
Uhh... I don't see how.
I mean, yeah, if we see typename or any other tag first, tag_type will
tell us we must find types for anything template-like, be it an
intermediate scope or the final typename.
However, if we see 'T::template C<auto> (t)', we can't tell whether T::C
is a template class and we're "calling a constructor" to create a
temporary out of t, or T::C is a static member template function that's
being called with t as an argument. (ok, we know from the typename line
that C is a type, but nevermind that, it could be a different name in
the testcase, and then we wouldn't; heck, couldn't it even be a template
function hiding the template class of the same name?)
And then, while we're parsing "template C<auto>", we haven't yet reached
the '::' after the closing angle bracket that would tell us to regard
the id necessarily as a typename, so I don't see how we'd get a
scope_type in tag_type for the third case.
--
Alexandre Oliva, freedom fighter http://FSFLA.org/~lxoliva/
You must be the change you wish to see in the world. -- Gandhi
Be Free! -- http://FSFLA.org/ FSF Latin America board member
Free Software Evangelist|Red Hat Brasil GNU Toolchain Engineer
