[Bug c++/107539] New: internal compiler error: same canonical type node for different types

[Bug c++/107539] New: internal compiler error: same canonical type node for different types

[franckbehaghel_gcc at protonmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

            Bug ID: 107539
           Summary: internal compiler error: same canonical type node for
                    different types
           Product: gcc
           Version: 13.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: franckbehaghel_gcc at protonmail dot com
  Target Milestone: ---

Created attachment 53835
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53835&action=edit
preprocessed source

When compiling gimp 2.10 with a fresh gcc version :
3ad2167bbac8ae83b1e91305b105ab5287bdac55 , I get the following error :

gimppaintcore-loops.cc: In lambda function:
gimppaintcore-loops.cc:472:61: internal compiler error: same canonical type
node for different types ‘AlgorithmTemplate<typename decltype
(algorithm)::type>’ and ‘AlgorithmTemplate<typename decltype
(algorithm)::type>’
  472 |             visitor (identity<AlgorithmTemplate<NewAlgorithm>> ());
      |                                                             ^~
0xc4306c comptypes(tree_node*, tree_node*, int)
        /GIT_REPO/build/gcc/gcc/gcc/cp/typeck.cc:1717
0xb99717 template_args_equal(tree_node*, tree_node*, bool)
        /GIT_REPO/build/gcc/gcc/gcc/cp/pt.cc:9284
0xb993fc template_args_equal(tree_node*, tree_node*, bool)
        /GIT_REPO/build/gcc/gcc/gcc/cp/pt.cc:9247
0xb993fc comp_template_args(tree_node*, tree_node*, tree_node**, tree_node**,
bool)
        /GIT_REPO/build/gcc/gcc/gcc/cp/pt.cc:9329
0xba42c3 spec_hasher::equal(spec_entry*, spec_entry*)
        /GIT_REPO/build/gcc/gcc/gcc/cp/pt.cc:1668
0xbecf73 hash_table<spec_hasher, false,
xcallocator>::find_with_hash(spec_entry* const&, unsigned int)
        /GIT_REPO/build/gcc/gcc/gcc/hash-table.h:926
0xbd4dac lookup_template_class(tree_node*, tree_node*, tree_node*, tree_node*,
int, int)
        /GIT_REPO/build/gcc/gcc/gcc/cp/pt.cc:9945
0xc044ed finish_template_type(tree_node*, tree_node*, int)
        /GIT_REPO/build/gcc/gcc/gcc/cp/semantics.cc:3755
0xb6a808 cp_parser_template_id
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:18368
0xb6aaab cp_parser_class_name
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:25826
0xb60f6e cp_parser_qualifying_entity
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:7183
0xb60f6e cp_parser_nested_name_specifier_opt
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:6865
0xb7a848 cp_parser_simple_type_specifier
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:19809
0xb5e98c cp_parser_postfix_expression
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:7640
0xb469bc cp_parser_binary_expression
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:10091
0xb47792 cp_parser_assignment_expression
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:10431
0xb4c9ee cp_parser_parenthesized_expression_list_elt
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:8054
0xb4cfb1 cp_parser_parenthesized_expression_list
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:8523
0xb5ef71 cp_parser_postfix_expression
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:7798
0xb469bc cp_parser_binary_expression
        /GIT_REPO/build/gcc/gcc/gcc/cp/parser.cc:10091
Please submit a full bug report, with preprocessed source (by using
-freport-bug).
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.


To reproduce : g++ -x c++ ccYvosm6.out1

Or copy&paste ccYvosm6.out1 in compiler explorer ( https://godbolt.org/ ).
Similar output is generated.


Regards,
Franck

[Bug c++/107539] internal compiler error: same canonical type node for different types

[franckbehaghel_gcc at protonmail dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

--- Comment #1 from Franck Behaghel <franckbehaghel_gcc at protonmail dot com> ---
Created attachment 53836
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53836&action=edit
main.cpp

reduced testcase :

g++ main.cpp 
main.cpp: In lambda function:
main.cpp:222:61: internal compiler error: same canonical type node for
different types ‘AlgorithmTemplate<typename decltype (algorithm)::type>’ and
‘AlgorithmTemplate<typename decltype (algorithm)::type>’
  222 |             visitor (identity<AlgorithmTemplate<NewAlgorithm>> ());
      |                                                             ^~

[Bug c++/107539] internal compiler error: same canonical type node for different types

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

--- Comment #2 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Note the ICE only shows up with checking enabled.  It is definitely a bug but
one which most users won't see.

[Bug c++/107539] internal compiler error: same canonical type node for different types

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

--- Comment #3 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Created attachment 53837
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=53837&action=edit
reduced further

Here is even more reduced testcase.

[Bug c++/107539] [13 Regression] ICE, same canonical type node for different types with decltype inside an generic lambda inside a templated method of a templated class using a templated class argument

[pinskia at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
     Ever confirmed|0                           |1
           Keywords|                            |needs-bisection
   Last reconfirmed|                            |2022-11-06
            Summary|internal compiler error:    |[13 Regression] ICE, same
                   |same canonical type node    |canonical type node for
                   |for different types         |different types with
                   |                            |decltype inside an generic
                   |                            |lambda inside a templated
                   |                            |method of a templated class
                   |                            |using a templated class
                   |                            |argument
   Target Milestone|---                         |13.0
             Status|UNCONFIRMED                 |NEW

--- Comment #4 from Andrew Pinski <pinskia at gcc dot gnu.org> ---
Confirmed.
I am not 100% of the summary.
Also I am not 100% sure of it being a 13 regression but GCC 12 compiles this
fine even with -fchecking (GCC 13 compiles it also fine with -fno-checking).

[Bug c++/107539] [13 Regression] ICE, same canonical type node for different types with decltype inside an generic lambda inside a templated method of a templated class using a templated class argument since r13-737-gd0ef9e06197d14d7

[marxin at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[13 Regression] ICE, same   |[13 Regression] ICE, same
                   |canonical type node for     |canonical type node for
                   |different types with        |different types with
                   |decltype inside an generic  |decltype inside an generic
                   |lambda inside a templated   |lambda inside a templated
                   |method of a templated class |method of a templated class
                   |using a templated class     |using a templated class
                   |argument                    |argument since
                   |                            |r13-737-gd0ef9e06197d14d7
           Keywords|needs-bisection             |
                 CC|                            |marxin at gcc dot gnu.org,
                   |                            |ppalka at gcc dot gnu.org

--- Comment #5 from Martin Liška <marxin at gcc dot gnu.org> ---
Started with r13-737-gd0ef9e06197d14d7.

[Bug c++/107539] [13 Regression] ICE, same canonical type node for different types with decltype inside an generic lambda inside a templated method of a templated class using a templated class argument since r13-737-gd0ef9e06197d14d7

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

Patrick Palka <ppalka at gcc dot gnu.org> changed:

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

[Bug c++/107539] [13 Regression] ICE, same canonical type node for different types with decltype inside an generic lambda inside a templated method of a templated class using a templated class argument since r13-737-gd0ef9e06197d14d7

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

--- Comment #6 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The master branch has been updated by Patrick Palka <ppalka@gcc.gnu.org>:

https://gcc.gnu.org/g:a4e577b044d69977f93b2cb7769dc991eadf2cf0

commit r13-4453-ga4e577b044d69977f93b2cb7769dc991eadf2cf0
Author: Patrick Palka <ppalka@redhat.com>
Date:   Thu Dec 1 20:11:53 2022 -0500

    c++: comptypes ICE with BOUND_TEMPLATE_TEMPLATE_PARMs [PR107539]

    Here we end up giving the two BOUND_TEMPLATE_TEMPLATE_PARMs
    C<decltype(f::t)> and C<decltype(g::t)> the same TYPE_CANONICAL because
    the hash table that interns TYPE_CANONICAL for template type parameters
    doesn't set the comparing_specializations flag which controls how
    PARM_DECLs from different contexts compare equal.

    Later, from spec_hasher::equal for the corresponding two specializations
    A<C<decltype(f::t)>> and A<C<decltype(g::t)>>, we compare the two bound
    ttps with comparing_specializations set hence they now (structurally)
    compare different despite having the same TYPE_CANONICAL, and so we get
    the error:

      internal compiler error: same canonical type node for different types
        'C<decltype (t)>' and 'C<decltype (t)>'

    This suggests that we should be setting comparing_specializations from
    ctp_hasher::equal to match spec_hasher::equal.  But doing so introduces
    a separate ICE in cpp2a/concepts-placeholder3.C:

      internal compiler error: canonical types differ for identical types
        'auto [requires ::same_as<<placeholder>, decltype(f::x)>]' and
        'auto [requires ::same_as<<placeholder>, decltype(g::x)>]'

    because norm_hasher::equal doesn't set comparing_specializations either.

    I'm not sure when exactly we need to set comparing_specializations given
    what it controls (TYPENAME_TYPE equality/hashing and PARM_DECL equality)
    but it seems to be the conservative choice to set the flag wherever we
    have a global hash table that relies on type equality.  To that end this
    patch sets comparing_specializations in ctp_hasher and norm_hasher, as
    well as in atom_hasher and sat_hasher for good measure.  This turns out
    to be a compile time win of about 2% in some concepts tests, probably
    because of the improved TYPENAME_TYPE hashing enabled by the flag.

            PR c++/107539

    gcc/cp/ChangeLog:

            * constraint.cc (norm_hasher::hash, norm_hasher::equal): Set
            comparing_specializations.
            (sat_hasher::hash, sat_hasher::equal): Likewise.
            * cp-tree.h (atom_hasher::hash, atom_hasher::equal): Likewise.
            * pt.cc (ctp_hasher::hash, ctp_hasher::equal): Likewise.

    gcc/testsuite/ChangeLog:

            * g++.dg/template/canon-type-19.C: New test.

[Bug c++/107539] [13 Regression] ICE, same canonical type node for different types with decltype inside an generic lambda inside a templated method of a templated class using a templated class argument since r13-737-gd0ef9e06197d14d7

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=107539

Patrick Palka <ppalka at gcc dot gnu.org> changed:

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

--- Comment #7 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Should be fixed, thanks for the bug report.