[Bug c++/98570] New: ICE: canonical types differ for identical types

[Bug c++/98570] New: ICE: canonical types differ for identical types

[acoplan at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

            Bug ID: 98570
           Summary: ICE: canonical types differ for identical types
           Product: gcc
           Version: 11.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: acoplan at gcc dot gnu.org
  Target Milestone: ---

GCC ICEs on the following (on at least aarch64 and x86). The testcase was
reduced from polymake and is accepted by clang.

template <int a> struct b { static constexpr int c = a; };
template <int> struct h { typedef int d; };
template <bool e, typename> using f = typename h<e>::d;
namespace g {
  template <typename...> using i = b<false>;
  using j = long;
}
using g::i;
using g::j;
template <typename> using k = i<>;
template <typename... l> f<!i<l...>::c, j> m();
template <typename... n> f<!i<k<n>...>::c, j> m();

On AArch64, I can reproduce the ICE back to GCC 6. To reproduce:

$ aarch64-elf-gcc -c testcase.ii
testcase.ii:12:49: internal compiler error: canonical types differ for
identical types 'h<(! g::i<k<n>...>::c)>' and 'h<(! g::i<l ...>::c)>'
   12 | template <typename... n> f<!i<k<n>...>::c, j> m();
      |                                                 ^
0x9f85e5 comptypes(tree_node*, tree_node*, int)
        /home/alecop01/toolchain/src/gcc/gcc/cp/typeck.c:1543
0x9f9900 same_type_ignoring_top_level_qualifiers_p(tree_node*, tree_node*)
        /home/alecop01/toolchain/src/gcc/gcc/cp/typeck.c:1579
0x9f7827 structural_comptypes
        /home/alecop01/toolchain/src/gcc/gcc/cp/typeck.c:1423
0x9f8570 comptypes(tree_node*, tree_node*, int)
        /home/alecop01/toolchain/src/gcc/gcc/cp/typeck.c:1532
0x7b9309 decls_match(tree_node*, tree_node*, bool)
        /home/alecop01/toolchain/src/gcc/gcc/cp/decl.c:1026
0x7b983c decls_match(tree_node*, tree_node*, bool)
        /home/alecop01/toolchain/src/gcc/gcc/cp/decl.c:1080
0x7bacf8 duplicate_decls(tree_node*, tree_node*, bool, bool)
        /home/alecop01/toolchain/src/gcc/gcc/cp/decl.c:1459
0x8aed4d do_pushdecl
        /home/alecop01/toolchain/src/gcc/gcc/cp/name-lookup.c:3734
0x8b13cc do_pushdecl_with_scope
        /home/alecop01/toolchain/src/gcc/gcc/cp/name-lookup.c:4881
0x8b1533 pushdecl_namespace_level(tree_node*, bool)
        /home/alecop01/toolchain/src/gcc/gcc/cp/name-lookup.c:6245
0x97701e push_template_decl(tree_node*, bool)
        /home/alecop01/toolchain/src/gcc/gcc/cp/pt.c:6049
0x7d1839 start_decl(cp_declarator const*, cp_decl_specifier_seq*, int,
tree_node*, tree_node*, tree_node**)
        /home/alecop01/toolchain/src/gcc/gcc/cp/decl.c:5531
0x8fc85e cp_parser_init_declarator
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:21659
0x8ff1fb cp_parser_single_declaration
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:30427
0x8ff2f8 cp_parser_template_declaration_after_parameters
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:29999
0x8ffc2d cp_parser_explicit_template_declaration
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:30265
0x8ffc2d cp_parser_template_declaration_after_export
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:30284
0x90026a cp_parser_template_declaration
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:16533
0x909dcd cp_parser_declaration
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:14009
0x908f3b cp_parser_toplevel_declaration
        /home/alecop01/toolchain/src/gcc/gcc/cp/parser.c:14107
Please submit a full bug report,
with preprocessed source if appropriate.
Please include the complete backtrace with any bug report.
See <https://gcc.gnu.org/bugs/> for instructions.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

Marek Polacek <mpolacek at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
           See Also|                            |https://gcc.gnu.org/bugzill
                   |                            |a/show_bug.cgi?id=83529
                 CC|                            |mpolacek at gcc dot gnu.org
   Target Milestone|---                         |8.5
     Ever confirmed|0                           |1
            Summary|ICE: canonical types differ |[8/9/10/11 Regression] ICE:
                   |for identical types         |canonical types differ for
                   |                            |identical types
   Last reconfirmed|                            |2021-01-06
             Status|UNCONFIRMED                 |NEW

--- Comment #1 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Confirmed, started with r226990, so not a dup of bug 83529.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

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

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

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[jakub at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

Jakub Jelinek <jakub at gcc dot gnu.org> changed:

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

--- Comment #2 from Jakub Jelinek <jakub at gcc dot gnu.org> ---
I don't see how that commit could have changed the behavior.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[mpolacek at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

--- Comment #3 from Marek Polacek <mpolacek at gcc dot gnu.org> ---
Maybe it was r226992 and our bisect binaries are broken :/.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

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
                 CC|                            |jason at gcc dot gnu.org

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

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

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

--- Comment #4 from Patrick Palka <ppalka at gcc dot gnu.org> ---
FWIW, here's a slightly more reduced ICE-on-valid testcase:

template <bool> struct h {};
struct b { static constexpr int c = true; };
template <class...> using i = b;
template <class... Ts> h<i<Ts...>::c> m();

template <class> using k = bool;
template <class... Ts> h<i<k<Ts>...>::c> m();

Bisection to r7-7375 as the commit that introduced the ICE for this particular
testcase.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

--- Comment #5 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:6e0a231a4aa2407bb7167daf98a37795a67364d8

commit r11-7011-g6e0a231a4aa2407bb7167daf98a37795a67364d8
Author: Jason Merrill <jason@redhat.com>
Date:   Wed Jan 27 17:15:39 2021 -0500

    c++: alias in qualified-id in template arg [PR98570]

    template_args_equal has handled dependent alias specializations for a
while,
    but in this testcase the actual template argument is a SCOPE_REF, so we
    called cp_tree_equal, which doesn't handle aliases specially when we get to
    them.

    This patch generalizes this by setting a flag so structural_comptypes will
    check for template alias equivalence (if we aren't doing partial ordering).
    The existing flag, comparing_specializations, was too broad; in particular,
    when we're doing decls_match, we want to treat corresponding parameters as
    equivalent, so we need to separate that from alias comparison.  So I
    introduce the comparing_dependent_aliases flag.

    From looking at other uses of comparing_specializations, it seems to me
that
    the new flag is what modules wants, as well.

    The other use of comparing_specializations in structural_comptypes is a
hack
    to deal with spec_hasher::equal not calling push_to_top_level, which we
    also don't want to tie to the alias comparison semantics.

    This patch also changes how we get to structural comparison of aliases from
    checking TYPE_CANONICAL in comptypes to marking the aliases as getting
    structural comparison when they are built, which is more consistent with
how
    e.g. typename is handled.

    As I mention in the comment for comparing_dependent_aliases, I think the
    default should be to treat different dependent aliases for the same type as
    distinct, only treating them as equal during deduction (particularly
partial
    ordering).  But that's a matter for the C++ committee, to try in stage 1.

    gcc/cp/ChangeLog:

            PR c++/98570
            * cp-tree.h: Declare it.
            * pt.c (comparing_dependent_aliases): New flag.
            (template_args_equal, spec_hasher::equal): Set it.
            (dependent_alias_template_spec_p): Assert that we don't
            get non-types other than error_mark_node.
            (instantiate_alias_template): SET_TYPE_STRUCTURAL_EQUALITY
            on complex alias specializations.  Set TYPE_DEPENDENT_P here.
            (tsubst_decl): Not here.
            * module.cc (module_state::read_cluster): Set
            comparing_dependent_aliases instead of
            comparing_specializations.
            * tree.c (cp_tree_equal): Remove comparing_specializations
            module handling.
            * typeck.c (structural_comptypes): Adjust.
            (comptypes): Remove comparing_specializations handling.

    gcc/testsuite/ChangeLog:

            PR c++/98570
            * g++.dg/cpp0x/alias-decl-targ1.C: New test.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[jason at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
         Resolution|---                         |FIXED
   Target Milestone|8.5                         |11.0

--- Comment #6 from Jason Merrill <jason at gcc dot gnu.org> ---
Fixed for GCC 11.  Since older releases only ICE with checking enabled, there
seems no need to backport.

[Bug c++/98570] [8/9/10/11 Regression] ICE: canonical types differ for identical types

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=98570

--- 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:5c62e4f255bfac65e18213fd93ee1c9908b4a750

commit r11-7089-g5c62e4f255bfac65e18213fd93ee1c9908b4a750
Author: Jason Merrill <jason@redhat.com>
Date:   Wed Feb 3 00:29:00 2021 -0500

    c++: Fix alias comparison [PR98926]

    The comparison of dependent aliases wasn't working here because
    processing_template_decl wasn't set, so dependent_alias_template_spec_p was
    always returning false.

    gcc/cp/ChangeLog:

            PR c++/98926
            PR c++/98570
            * pt.c (spec_hasher::equal): Set processing_template_decl.
            * Make-lang.in (check-g++-strict-gc): Add --param
            hash-table-verification-limit=10000.

    gcc/testsuite/ChangeLog:

            * g++.dg/cpp0x/alias-decl-dr1558.C: Pass --param
            hash-table-verification-limit=10000.