[Bug c++/103057] New: Internal compiler error: Error reporting routines re-entered

[Bug c++/103057] New: Internal compiler error: Error reporting routines re-entered

[plasmahh at gmx dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

            Bug ID: 103057
           Summary: Internal compiler error: Error reporting routines
                    re-entered
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: plasmahh at gmx dot net
  Target Milestone: ---

Created attachment 51728
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51728&action=edit
cvise generated testcase

Compiling the attached testcase (generated with cvise out of a "real world"
piece of code, where supressing an "unused parameter" warning makes this go
away, due to the testcase being an error now, the error will trigger it too I
guess) with -std=gnu++2b (and likely some others) will ICE out:

Internal compiler error: Error reporting routines re-entered.
0xbd474d cxx_incomplete_type_diagnostic(unsigned int, tree_node const*,
tree_node const*, diagnostic_t)
        ../../gcc/cp/typeck2.c:330
0xb6629c cxx_incomplete_type_diagnostic(tree_node const*, tree_node const*,
diagnostic_t)
        ../../gcc/cp/cp-tree.h:8016
0xb6629c cxx_incomplete_type_error(tree_node const*, tree_node const*)
        ../../gcc/cp/cp-tree.h:8025
0xb6629c instantiate_class_template_1
        ../../gcc/cp/pt.c:12106
0xb66cd2 instantiate_class_template(tree_node*)
        ../../gcc/cp/pt.c:12302
0xbb58d9 complete_type(tree_node*)
        ../../gcc/cp/typeck.c:143
0xbb5a94 complete_type(tree_node*)
        ../../gcc/cp/typeck.c:169
0xbb5a94 complete_type_or_maybe_complain(tree_node*, tree_node*, int)
        ../../gcc/cp/typeck.c:156
0xb4ddb2 tsubst(tree_node*, tree_node*, int, tree_node*)
        ../../gcc/cp/pt.c:16069
0xb4ef9b tsubst_template_args(tree_node*, tree_node*, int, tree_node*)
        ../../gcc/cp/pt.c:13411
0xb4caa2 tsubst(tree_node*, tree_node*, int, tree_node*)
        ../../gcc/cp/pt.c:15449
0xa3266c dump_template_bindings
        ../../gcc/cp/error.c:482
0xa2bbba dump_function_decl
        ../../gcc/cp/error.c:1796
0xa33e0d decl_to_string
        ../../gcc/cp/error.c:3216
0xa33e0d cp_printer
        ../../gcc/cp/error.c:4387
0x1f9317c pp_format(pretty_printer*, text_info*)
        ../../gcc/pretty-print.c:1475
0x1f76b92 diagnostic_report_diagnostic(diagnostic_context*, diagnostic_info*)
        ../../gcc/diagnostic.c:1342
0x1f770f6 diagnostic_impl
        ../../gcc/diagnostic.c:1504
0x1f7787f permerror(unsigned int, char const*, ...)
        ../../gcc/diagnostic.c:1771
0xa26072 c_parse_final_cleanups()
        ../../gcc/cp/decl2.c:5255

[Bug c++/103057] Internal compiler error: Error reporting routines re-entered

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1
           Keywords|                            |ice-on-invalid-code
                 CC|                            |marxin at gcc dot gnu.org
   Last reconfirmed|                            |2021-11-03

--- Comment #1 from Martin Liška <marxin at gcc dot gnu.org> ---
Is the original test-case a valid code?

[Bug c++/103057] Internal compiler error: Error reporting routines re-entered

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

Martin Liška <marxin at gcc dot gnu.org> changed:

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

--- Comment #2 from Martin Liška <marxin at gcc dot gnu.org> ---
Likely started with r11-291-g0f50f6daa140186a.

[Bug c++/103057] Internal compiler error: Error reporting routines re-entered

[plasmahh at gmx dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

--- Comment #3 from Dennis Lubert <plasmahh at gmx dot net> ---
Yes, the original testcase is valid code that compiles fine with -w

[Bug c++/103057] [11/12 Regression] Internal compiler error: Error reporting routines re-entered

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

Andrew Pinski <pinskia at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Target Milestone|---                         |11.3
            Summary|Internal compiler error:    |[11/12 Regression] Internal
                   |Error reporting routines    |compiler error: Error
                   |re-entered                  |reporting routines
                   |                            |re-entered

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
(In reply to Dennis Lubert from comment #3)
> Yes, the original testcase is valid code that compiles fine with -w

There might be two different bugs here I think, can you redo your reduction but
this test also testing if -w still compiles the source?

[Bug c++/103057] [11/12 Regression] Internal compiler error: Error reporting routines re-entered

[plasmahh at gmx dot net]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

--- Comment #5 from Dennis Lubert <plasmahh at gmx dot net> ---
Created attachment 51759
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=51759&action=edit
cvise generated compling testcase

This one is pretty big but compiles with -w but ICEs without it. Differently to
my original code, here there is no unused argument that when being used
silences the warning and so circumvents the ICE. So I am not totally sure if
there maybe a third bug lurking somewhere...

Currently our code to be able to not ICE disables Wunused-parameter and
Wunused-value for the functions in question.

[Bug c++/103057] [11/12 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

Martin Liška <marxin at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[11/12 Regression] Internal |[11/12 Regression] Internal
                   |compiler error: Error       |compiler error: Error
                   |reporting routines          |reporting routines
                   |re-entered                  |re-entered since
                   |                            |r11-291-g0f50f6daa140186a
           Keywords|ice-on-invalid-code         |ice-on-valid-code

--- Comment #6 from Martin Liška <marxin at gcc dot gnu.org> ---
Thank you for the test-case. It started with r11-291-g0f50f6daa140186a with
-std=gnu2a.

[Bug c++/103057] [11/12 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

Richard Biener <rguenth at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Priority|P3                          |P2

[Bug c++/103057] [11/12 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
           Assignee|unassigned at gcc dot gnu.org      |jason at gcc dot gnu.org
             Status|NEW                         |ASSIGNED

[Bug c++/103057] [11/12 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

--- Comment #7 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Jason Merrill <jason@gcc.gnu.org>:

https://gcc.gnu.org/g:f3e6ef7d873885ffaa0db7f4185364bbd7a70505

commit r12-6872-gf3e6ef7d873885ffaa0db7f4185364bbd7a70505
Author: Jason Merrill <jason@redhat.com>
Date:   Tue Jan 25 15:12:16 2022 -0500

    c++: alias template and typename [PR103057]

    Usually we handle DR1558 substitution near the top of tsubst, but in this
    case while substituting TYPENAME_TYPE we were passing an alias
    specialization to tsubst_aggr_type, which ignored its aliasness.

            PR c++/103057

    gcc/cp/ChangeLog:

            * pt.cc (tsubst_aggr_type): Call tsubst for alias template
            specialization.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/alias-decl-void1.C: New test.

[Bug c++/103057] [11 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[11/12 Regression] Internal |[11 Regression] Internal
                   |compiler error: Error       |compiler error: Error
                   |reporting routines          |reporting routines
                   |re-entered since            |re-entered since
                   |r11-291-g0f50f6daa140186a   |r11-291-g0f50f6daa140186a

--- Comment #8 from Jason Merrill <jason at gcc dot gnu.org> ---
The second testcase isn't actually valid code, either.  Fixed for 12 so far.

[Bug c++/103057] [11 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

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

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

--- Comment #9 from Jason Merrill <jason at gcc dot gnu.org> ---
*** Bug 104184 has been marked as a duplicate of this bug. ***

[Bug c++/103057] [11 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

--- Comment #10 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-11 branch has been updated by Jason Merrill
<jason@gcc.gnu.org>:

https://gcc.gnu.org/g:b3411d7a36ee1d38923dcf9db26a1dd252827b64

commit r11-9682-gb3411d7a36ee1d38923dcf9db26a1dd252827b64
Author: Jason Merrill <jason@redhat.com>
Date:   Tue Jan 25 15:12:16 2022 -0500

    c++: alias template and typename [PR103057]

    Usually we handle DR1558 substitution near the top of tsubst, but in this
    case while substituting TYPENAME_TYPE we were passing an alias
    specialization to tsubst_aggr_type, which ignored its aliasness.

            PR c++/103057

    gcc/cp/ChangeLog:

            * pt.c (tsubst_aggr_type): Call tsubst for alias template
            specialization.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/alias-decl-void1.C: New test.

[Bug c++/103057] [11 Regression] Internal compiler error: Error reporting routines re-entered since r11-291-g0f50f6daa140186a

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103057

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

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

--- Comment #11 from Jason Merrill <jason at gcc dot gnu.org> ---
Fixed for 11.3.