[Bug c++/18733] New: [4.0 Regression] friend rejected

[Bug c++/18733] New: [4.0 Regression] friend rejected

[pinskia at gcc dot gnu dot org]

I don't know if this is valid code or not, to me this looks like valid code but I could be wrong.

: Search converges between 2004-07-20-trunk (#491) and 2004-07-21-trunk (#492).

This is reduced from PR 15855:

typedef int GVID;
class VG;
template <class T> struct VDifferential
{
  typedef T Summary;
  Summary summarize(const VG*, GVID, GVID) const;
};
class VG {
    friend class VDifferential<char>;
    friend VDifferential<char>::Summary VDifferential<char>::summarize(const VG*, GVID, GVID) const;
};

-- 
           Summary: [4.0 Regression] friend rejected
           Product: gcc
           Version: 4.0.0
            Status: UNCONFIRMED
          Keywords: rejects-valid
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: pinskia at gcc dot gnu dot org
                CC: gcc-bugs at gcc dot gnu dot org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |giovannibajo at gcc dot gnu
                   |                            |dot org, pinskia at gcc dot
                   |                            |gnu dot org
   Target Milestone|---                         |4.0.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-11-30 04:51 -------
Comeau C/C++'s online tester accepts this.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-11-30 10:21 -------
Confirmed, this is valid code. I have a feeling that this was exposed with one 
of my patches on July 21st...

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-11-30 10:21:12
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[reichelt at gcc dot gnu dot org]

------- Additional Comments From reichelt at gcc dot gnu dot org  2004-11-30 11:00 -------
More compact testcase:

================================
template<int> struct A
{
    void foo();
};

struct B
{
    friend void A<0>::foo();
};
================================


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |reichelt at gcc dot gnu dot
                   |                            |org
           Severity|normal                      |critical
           Keywords|                            |monitored


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-12-01 07:11 -------
Kriang, this looks like a latent bug exposed by my cleanup of template 
specialization headers. Since it affects friends, you may want to take a look. 
Otherwise, I'll get to it sooner or later.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |lerdsuwa at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[lerdsuwa at gcc dot gnu dot org]

------- Additional Comments From lerdsuwa at gcc dot gnu dot org  2004-12-01 14:32 -------
The problem is in check_explicit_specialization and/or
current_tmpl_spec_kind functions.  They don't handle 
specializations in friends correctly.
I still have a few other PRs to deal with and won't get to 
this bug until other patches are finished.  So feel free to
take it.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-12-02 09:28 -------
Right, the whole template header check code is a mess across many files and 
executed at different times. This works as a stop-gap:


Index: pt.c
===================================================================
RCS file: /cvs/gcc/gcc/gcc/cp/pt.c,v
retrieving revision 1.945
diff -c -3 -p -r1.945 pt.c
*** pt.c        12 Nov 2004 21:47:08 -0000      1.945
--- pt.c        2 Dec 2004 09:25:16 -0000
*************** check_explicit_specialization (tree decl
*** 1739,1745 ****
    tree dname = DECL_NAME (decl);
    tmpl_spec_kind tsk;

!   tsk = current_tmpl_spec_kind (template_count);

    switch (tsk)
      {
--- 1739,1753 ----
    tree dname = DECL_NAME (decl);
    tmpl_spec_kind tsk;

!   if (is_friend)
!     {
!       if (!processing_specialization)
!       tsk = tsk_none;
!       else
!       tsk = tsk_excessive_parms;
!     }
!   else
!     tsk = current_tmpl_spec_kind (template_count);

    switch (tsk)
      {


I'm testing it.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-12-02 13:41 -------
Patch here:
http://gcc.gnu.org/ml/gcc-patches/2004-12/msg00144.html

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|giovannibajo at gcc dot gnu |
                   |dot org                     |
         AssignedTo|unassigned at gcc dot gnu   |giovannibajo at libero dot
                   |dot org                     |it
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-12-17 12:07 -------
Kriang, can you comment on my patch? Does it look correct to you?

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |nathan at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[lerdsuwa at gcc dot gnu dot org]

------- Additional Comments From lerdsuwa at gcc dot gnu dot org  2004-12-17 16:44 -------
The patch looks correct although the code is messy.  With the patch,
when is_friend is true and processing_specialization is false, we
no longer count the number of template header to see if there are 
too many - but we've already done that in 
cp_parser_check_template_parameters.
The code duplication in cp_parser_check_template_parameters and 
current_tmpl_spec_kind could be removed in GCC 4.1.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-12-17 23:48 -------
Yes, that's the problem. The whole point is that the parser does not have 
enough context information to fully check for the number of template headers: 
there is an off-by-one possible error it has to allow. For instance:

template <> 
void A<int>::foo(void);

The correctness of this depends on whether A<int> uses an explicit 
specialization or not. So yes, we could unify this for 4.1.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2004-12-23 01:49 -------
Subject: Bug 18733

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	giovannibajo@gcc.gnu.org	2004-12-23 01:49:39

Modified files:
	gcc/cp         : ChangeLog pt.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/g++.dg/template: friend33.C 

Log message:
	PR c++/18733
	* pt.c (check_explicit_specialization): Use special logic to validate
	befriended specializations.
	
	PR c++/18733
	* g++.dg/template/friend33.C: New testcase.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&r1=1.4553&r2=1.4554
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/pt.c.diff?cvsroot=gcc&r1=1.959&r2=1.960
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.4800&r2=1.4801
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/template/friend33.C.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733

[Bug c++/18733] [4.0 Regression] friend rejected

[giovannibajo at libero dot it]

------- Additional Comments From giovannibajo at libero dot it  2004-12-23 01:50 -------
Fixed in GCC 4.0.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|                            |FIXED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=18733