[Bug c++/95678] New: [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[Bug c++/95678] New: [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[doko at debian dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95678

            Bug ID: 95678
           Summary: [9 Regression] ICE in dependent_type_p, at
                    cp/pt.c:25610
           Product: gcc
           Version: 9.3.1
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: doko at debian dot org
  Target Milestone: ---

seen with the gcc-9 branch 20200518, not with the gcc-10 branch.

$ cat unique.ii
# 7 "" 3
template <class b> struct c {
  using d = b;
  template <class bb, class e> auto operator()(bb l, e) -> decltype(d{l});
};
template <class f, template <class> class g> h() {
  using i = g<f>;
  int a[]{};
  unsigned j;
  i{}(a, j);
}
k() { h<int *, c>(); }

$ g++ -c -std=c++14 -Wall -fconcepts unique.ii
: In instantiation of ‘decltype (c<b>::d{l}) c<b>::operator()(bb, e) [with bb =
int*; e = unsigned int; b = int*]’:
:9:37: internal compiler error: in dependent_type_p, at cp/pt.c:25610
0x7f56f3effe0a __libc_start_main
        ../csu/libc-start.c:308
Please submit a full bug report,
with preprocessed source if appropriate.

[Bug c++/95678] [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[rguenth at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95678

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to work|                            |9.3.0
   Target Milestone|---                         |9.4
           Keywords|                            |ice-on-valid-code

--- Comment #1 from Richard Biener <rguenth at gcc dot gnu.org> ---
I can't reproduce it.  What version works for you?

[Bug c++/95678] [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[doko at debian dot org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95678

--- Comment #2 from Matthias Klose <doko at debian dot org> ---
the gcc-9 branch from 20200408 works for me. 20200615 also fails.

: In instantiation of ‘decltype (c<b>::d{l}) c<b>::operator()(bb, e) [with bb =
int*; e = unsigned int; b = int*]’:
:9:37: internal compiler error: in dependent_type_p, at cp/pt.c:25619
0x5abade dependent_type_p(tree_node*)
        ../../src/gcc/cp/pt.c:25619
0x5abade dependent_type_p(tree_node*)
        ../../src/gcc/cp/pt.c:25610
0xd9e253 zero_init_expr_p(tree_node*)
        ../../src/gcc/cp/tree.c:4387
0x82b8a7 write_expression
        ../../src/gcc/cp/mangle.c:3177
0x11f7673 write_type
        ../../src/gcc/cp/mangle.c:2307
0xeaef6a write_encoding
        ../../src/gcc/cp/mangle.c:2688
0xeaef6a write_mangled_name
        ../../src/gcc/cp/mangle.c:789
0xeaef6a mangle_decl_string
        ../../src/gcc/cp/mangle.c:3839
0xeacbbf get_mangled_id
        ../../src/gcc/cp/mangle.c:3861
0xeacbbf mangle_decl(tree_node*)
        ../../src/gcc/cp/mangle.c:3899
0xe9b7ef decl_assembler_name(tree_node*)
        ../../src/gcc/tree.c:701
0xe9b7ef symbol_table::insert_to_assembler_name_hash(symtab_node*, bool)
        ../../src/gcc/symtab.c:174
0xe9b7ef symbol_table::insert_to_assembler_name_hash(symtab_node*, bool)
        ../../src/gcc/symtab.c:161
0xe9b7ef symbol_table::symtab_initialize_asm_name_hash()
        ../../src/gcc/symtab.c:266
0xe9b7ef symbol_table::symtab_initialize_asm_name_hash()
        ../../src/gcc/symtab.c:259
0xead95b analyze_functions
        ../../src/gcc/cgraphunit.c:1298
0xeac067 symbol_table::finalize_compilation_unit()
        ../../src/gcc/cgraphunit.c:2837
Please submit a full bug report,
with preprocessed source if appropriate.

[Bug c++/95678] [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95678

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ppalka at gcc dot gnu.org
   Last reconfirmed|                            |2020-06-16
           Assignee|unassigned at gcc dot gnu.org      |ppalka at gcc dot gnu.org
     Ever confirmed|0                           |1
             Status|UNCONFIRMED                 |ASSIGNED

--- Comment #3 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Judging by the backtrace, I suspect this is ICE is due to r9-8526.  I think
r10-7933 needs to be backported to the 9 branch.

[Bug c++/95678] [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95678

--- Comment #4 from CVS Commits <cvs-commit at gcc dot gnu.org> ---
The releases/gcc-9 branch has been updated by Patrick Palka
<ppalka@gcc.gnu.org>:

https://gcc.gnu.org/g:d9867034e033c08fdabaebfe2f0914d7d9bb25c6

commit r9-8680-gd9867034e033c08fdabaebfe2f0914d7d9bb25c6
Author: Patrick Palka <ppalka@redhat.com>
Date:   Tue Jun 16 12:16:02 2020 -0400

    c++: zero_init_expr_p of dependent expression [PR95678]

    gcc/cp/ChangeLog:

            PR c++/95678
            * tree.c (zero_init_expr_p): Use uses_template_parms instead of
            dependent_type_p.

    gcc/testsuite/ChangeLog:

            PR c++/95678
            * g++.dg/cpp0x/dependent3.C: New test.

    (cherry picked from commit 9a453da5cac7b6610cd342a7bea256acba5e61e1)

[Bug c++/95678] [9 Regression] ICE in dependent_type_p, at cp/pt.c:25610

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=95678

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

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

--- Comment #5 from Patrick Palka <ppalka at gcc dot gnu.org> ---
This should be fixed now.