[Bug c++/50611] New: Error reporting routines re-entered

[Bug c++/50611] New: Error reporting routines re-entered

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

             Bug #: 50611
           Summary: Error reporting routines re-entered
    Classification: Unclassified
           Product: gcc
           Version: 4.6.1
            Status: UNCONFIRMED
          Keywords: diagnostic, ice-on-invalid-code
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: jakub@gcc.gnu.org
                CC: jason@gcc.gnu.org


namespace std class multiset;
namespace blue_sky template < class > class smart_ptr;
typedef unsigned ulong;
typedef
blue_sky::smart_ptr < >
  spv_float;
struct wpi_strategy_3d
{
  enum
  {
    D;
    typedef ulong vertex_pos_i
      template <
      class
      strat_t > struct wpi_impl:
      public
    {
      strat_t::D;
      typedef
      std::multiset < >
intersect_path
class
intersect_action
      {
      public:typedef
{
} remove_dups2 ()
{
  typedef int
    dirvec_t[D];
  struct top_surv
  {
  top_surv (dirvec_t &, intersect_path &,)};
  dirvec_t
    dir
  {
top_surv (dir, x_, m_size_)}} intersect_path &
  x_
  vertex_pos_i
  m_size_;
      }
      well_path_ident_d (
  {
  intersect_action A (;
      A.remove_dups2 ()}
      spv_float well_path_ident
      {
      wpi_impl <
      wpi_strategy_3d >::
      well_path_ident_d}
}
}
}

after issuing loads of errors ICEs with
Internal compiler error: Error reporting routines re-entered.
This is delta reduced from https://bugzilla.redhat.com/show_bug.cgi?id=738516
which ICEs both on the 4.6 branch and on the trunk the same way.

[Bug c++/50611] Error reporting routines re-entered

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2011-10-11
         AssignedTo|unassigned at gcc dot       |paolo.carlini at oracle dot
                   |gnu.org                     |com
     Ever Confirmed|0                           |1

--- Comment #1 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-11 02:05:08 UTC ---
On it.

[Bug c++/50611] Error reporting routines re-entered

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

--- Comment #2 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-11 02:24:56 UTC ---
Hummm. This patchlet avoids the ICE and apparently passes the testsuite
(already past g++.dg), but the amount of diagnostics in mainline seems in any
case much more than in 4.6, in particular the final "In instantiation..."
notes.

I don't know if we want to deal with that, and as part of this PR?!? Jason?

////////////////

Index: pt.c
===================================================================
--- pt.c        (revision 179778)
+++ pt.c        (working copy)
@@ -13026,7 +13026,11 @@
        if (error_msg)
          error (error_msg);
        if (!function_p && TREE_CODE (decl) == IDENTIFIER_NODE)
-         decl = unqualified_name_lookup_error (decl);
+         {
+           if (complain & tf_error)
+             unqualified_name_lookup_error (decl);
+           decl = error_mark_node;
+         }
        return decl;
       }

[Bug c++/50611] Error reporting routines re-entered

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|4.6.1                       |4.7.0

--- Comment #3 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-11 11:18:49 UTC ---
By the way, I can see this only in mainline, Jakub.

(which indeed makes sense given that my patchlet touches pt.c and the crash
happens in the middle of "In instantiation...")

[Bug c++/50611] Error reporting routines re-entered

[jakub at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

--- Comment #4 from Jakub Jelinek <jakub at gcc dot gnu.org> 2011-10-11 11:20:47 UTC ---
The reduced testcase yes.  But please try the original testcase in 4.6, whether
your patch fixes it.

[Bug c++/50611] Error reporting routines re-entered

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

--- Comment #5 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-11 11:35:42 UTC ---
I see. It does, anyway.

[Bug c++/50611] Error reporting routines re-entered

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

--- Comment #6 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-11 11:40:38 UTC ---
I mean, fixes the ICE both in mainline and in 4_6-branch. Indeed, the latter
seems more serious, because otherwise for the original testcase we produce no
useful diagnostics at all.

[Bug c++/50611] Error reporting routines re-entered

[paolo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

--- Comment #7 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> 2011-10-11 13:07:55 UTC ---
Author: paolo
Date: Tue Oct 11 13:07:52 2011
New Revision: 179802

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=179802
Log:
2011-10-11  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/50611
    * pt.c (tsubst_copy_and_build): If (complain & tf_error) is false
    do not call unqualified_name_lookup_error.

Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/pt.c

[Bug c++/50611] Error reporting routines re-entered

[paolo at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

--- Comment #8 from paolo at gcc dot gnu.org <paolo at gcc dot gnu.org> 2011-10-11 13:08:09 UTC ---
Author: paolo
Date: Tue Oct 11 13:08:05 2011
New Revision: 179803

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=179803
Log:
2011-10-11  Paolo Carlini  <paolo.carlini@oracle.com>

    PR c++/50611
    * pt.c (tsubst_copy_and_build): If (complain & tf_error) is false
    do not call unqualified_name_lookup_error.

Modified:
    branches/gcc-4_6-branch/gcc/cp/ChangeLog
    branches/gcc-4_6-branch/gcc/cp/pt.c

[Bug c++/50611] Error reporting routines re-entered

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=50611

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|                            |FIXED
   Target Milestone|---                         |4.6.2

--- Comment #9 from Paolo Carlini <paolo.carlini at oracle dot com> 2011-10-11 13:09:04 UTC ---
Fixed mainline and 4.6.2.