[Bug c++/51180] New: [C++0x][C++11] inner class alias-definition variadic template error

[Bug c++/51180] New: [C++0x][C++11] inner class alias-definition variadic template error

[trashyankes at wp dot pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

             Bug #: 51180
           Summary: [C++0x][C++11] inner class alias-definition variadic
                    template error
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: trashyankes@wp.pl


gcc version 4.7.0 20111108 (experimental) (niXman build)

code:
---------------------------
---------------------------

template<class, class>
struct t2 //error: provided for 'template<class, class> struct t2'
{

};

template<template<class...> class M>
struct m
{
template<class... B>
using inner = M<B...>; //error: wrong number of template arguments (1, should
be 2)
};

m<t2> sta2;

---------------------------
---------------------------

[Bug c++/51180] [C++0x] inner class alias-definition variadic template error

[trashyankes at wp dot pl]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

--- Comment #1 from trashyankes at wp dot pl 2011-11-23 18:34:43 UTC ---
I find easy workaround this bug.

---------------------------
---------------------------
template<template<class...> class M>
struct m
{
    template<class... B>
    struct _inner_fix
    {
        typedef M<B...> type
    }
    template<class... B>
    using inner = typename _inner_fix<B...>::type;
};
---------------------------
---------------------------

[Bug c++/51180] [C++0x] inner class alias-definition variadic template error

[dodji at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

Dodji Seketeli <dodji at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |ASSIGNED
   Last reconfirmed|                            |2011-12-01
         AssignedTo|unassigned at gcc dot       |dodji at gcc dot gnu.org
                   |gnu.org                     |
     Ever Confirmed|0                           |1

--- Comment #2 from Dodji Seketeli <dodji at gcc dot gnu.org> 2011-12-01 19:25:02 UTC ---
This is somewhat related to PR c++/51239.  I think at some point
during the substituting of t2 in to m, we should leave t2<B...>
"unbound" instead of trying to substitute B... into it.

The proper substituting of t2<B...> would happen later when B... has a
proper matching argument pack.

I am working on a joint patch for this and PR c++/51239.

[Bug c++/51180] [C++0x] inner class alias-definition variadic template error

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |jason at gcc dot gnu.org

--- Comment #3 from Jason Merrill <jason at gcc dot gnu.org> 2011-12-01 19:41:33 UTC ---
(In reply to comment #2)
> This is somewhat related to PR c++/51239.  I think at some point
> during the substituting of t2 in to m, we should leave t2<B...>
> "unbound" instead of trying to substitute B... into it.

I'm surprised that it doesn't just work already; writing t2<B...> already
works, doing the substitution ought to work the same way.

[Bug c++/51180] [C++0x] inner class alias-definition variadic template error

[dodji at seketeli dot org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

--- Comment #4 from dodji at seketeli dot org <dodji at seketeli dot org> 2011-12-02 10:15:06 UTC ---
"jason at gcc dot gnu.org" <gcc-bugzilla@gcc.gnu.org> a écrit:

> I'm surprised that it doesn't just work already; writing t2<B...> already
> works, doing the substitution ought to work the same way.

It wasn't working in this case because we are getting into the "if"
below in coerce_template_parms:

  if ((nargs > nparms && !variadic_p)
      || (nargs < nparms - variadic_p
      && require_all_args
      && (!use_default_args
          || (TREE_VEC_ELT (parms, nargs) != error_mark_node
                  && !TREE_PURPOSE (TREE_VEC_ELT (parms, nargs))))))
    {

We are getting there because the number of arguments of t2 in that
case is smaller than its number of parms.

I am still tweaking other parts of the patch to make it pass bootstrap
and testing.

[Bug c++/51180] [C++0x] inner class alias-definition variadic template error

[dodji at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

--- Comment #5 from Dodji Seketeli <dodji at gcc dot gnu.org> 2011-12-03 17:25:02 UTC ---
A candidate patch for this has been submitted to
http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00163.html

[Bug c++/51180] [C++0x] inner class alias-definition variadic template error

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=51180

--- Comment #6 from Paolo Carlini <paolo.carlini at oracle dot com> ---
Dodji, any progress on this? Looks like Part 1 of the latest patch set still
needs work? http://gcc.gnu.org/ml/gcc-patches/2011-12/msg00728.html