[Bug c++/103994] New: Module ICE in mark_by_value, at cp/module.cc:4772

[Bug c++/103994] New: Module ICE in mark_by_value, at cp/module.cc:4772

[unlvsur at live dot com]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

            Bug ID: 103994
           Summary: Module ICE in mark_by_value, at cp/module.cc:4772
           Product: gcc
           Version: 12.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: unlvsur at live dot com
  Target Milestone: ---

Created attachment 52172
  --> https://gcc.gnu.org/bugzilla/attachment.cgi?id=52172&action=edit
module ICE

module;

#include <gtirb/gtirb.hpp>

export module cfg;

namespace cfg
{
export struct cfg_vertex
{
        gtirb::CfgNode const* cfg_node_ptr{};
        struct edge
        {
                std::size_t to{};
                gtirb::CfgNode const* cfg_node_ptr{};
        };
        std::vector<edge> edges;
};

export inline std::vector<cfg_vertex> get_cfg()
{
        return {};
}
}


g++ -c cfg.cc -O2 -std=c++20 -fmodules-ts

cfg.cc:5:8: internal compiler error: in mark_by_value, at cp/module.cc:4772
    5 | export module cfg;
      |        ^
0x7f8092e45fcf __libc_start_call_main
        ../sysdeps/nptl/libc_start_call_main.h:58
0x7f8092e4607c __libc_start_main_impl
        ../csu/libc-start.c:409

[Bug c++/103994] Module ICE in write_var_def with global variable in global module fragment

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |ppalka at gcc dot gnu.org
            Summary|Module ICE in               |Module ICE in write_var_def
                   |mark_by_value, at           |with global variable in
                   |cp/module.cc:4772           |global module fragment

--- Comment #1 from Patrick Palka <ppalka at gcc dot gnu.org> ---
With trunk, it seems we now crash from write_var_def.  Reduced:

$ cat pr103994.C
module;
# 0 "" 1
struct A { A(int) { } };
const A a = 0;
# 3 "" 2

export module pr103994;

export inline A f() { return a; }

$ g++ -fmodules-ts pr103994.C
pr103994.C:4:8: internal compiler error: in write_var_def, at
cp/module.cc:11649
    4 | const A a = 0;
      |        ^~~~~~

[Bug c++/103994] Module ICE in write_var_def with global variable in global module fragment

[Melven.Roehrig-Zoellner at DLR dot de]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

--- Comment #2 from Melven.Roehrig-Zoellner at DLR dot de ---
Can confirm this with current trunk (see https://godbolt.org/z/16K1jhrsW).

I stumbled upon this while trying to include the C++ library Eigen in the
global module fragment (same stacktrace)...

Thanks to anyone looking into this!

[Bug c++/103994] Module ICE in write_var_def with global variable in global module fragment

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

--- Comment #3 from Patrick Palka <ppalka at gcc dot gnu.org> ---
We accept the comment #1 testcase after r14-6979, but still ICE on the original
testcase:

cfg.cc:5:9: internal compiler error: in insert, at cp/module.cc:4924
0x77357d trees_out::insert(tree_node*, walk_kind)
        /home/patrick/gcc/gcc/cp/module.cc:4924
0xb55c50 trees_out::decl_value(tree_node*, depset*)
        /home/patrick/gcc/gcc/cp/module.cc:7725
0xb5631a trees_out::decl_node(tree_node*, walk_kind)
        /home/patrick/gcc/gcc/cp/module.cc:8764
0xb57522 trees_out::tree_node(tree_node*)
        /home/patrick/gcc/gcc/cp/module.cc:9326
0xb5c9b5 module_state::write_cluster(elf_out*, depset**, unsigned int,
depset::hash&, unsigned int*, unsigned int*)
        /home/patrick/gcc/gcc/cp/module.cc:14898
0xb5e82e module_state::write_begin(elf_out*, cpp_reader*, module_state_config&,
unsigned int&)
        /home/patrick/gcc/gcc/cp/module.cc:18084
0xb5f544 finish_module_processing(cpp_reader*)
        /home/patrick/gcc/gcc/cp/module.cc:20294
0xaeaea1 c_parse_final_cleanups()
        /home/patrick/gcc/gcc/cp/decl2.cc:5313
0xd31290 c_common_parse_file()
        /home/patrick/gcc/gcc/c-family/c-opts.cc:1319
\

[Bug c++/103994] Module ICE in write_var_def with global variable in global module fragment

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
   Last reconfirmed|                            |2024-01-27
             Status|UNCONFIRMED                 |NEW
     Ever confirmed|0                           |1

--- Comment #4 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Reduced testcase exhibiting the latest ICE:

$ cat 103994.H
template <typename> struct TransparentSupport {
  template <typename> using key_arg = int;
};
template <typename Key> struct Map {
  template <typename LookupKey>
  using key_arg = typename TransparentSupport<Key>::template
key_arg<LookupKey>;
};
struct MapKey {
  Map<MapKey> map_;
};

$ g++ -fmodule-header 103994.H
103994.H: internal compiler error: in insert, at cp/module.cc:4924

[Bug c++/103994] Module ICE in write_var_def with global variable in global module fragment

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

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++/103994] Module ICE in write_var_def with global variable in global module fragment

[cvs-commit at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

--- Comment #5 from GCC 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:f5c1224708a0cf9cc2770c44bbbe7d0c883942be

commit r14-9374-gf5c1224708a0cf9cc2770c44bbbe7d0c883942be
Author: Patrick Palka <ppalka@redhat.com>
Date:   Thu Mar 7 16:39:20 2024 -0500

    c++/modules: member alias tmpl partial inst [PR103994]

    Alias templates are weird in that their specializations can appear in
    both decl_specializations and type_specializations.  They're always in
    the decl table, and additionally appear in the type table only at parse
    time via finish_template_type.  There seems to be no good reason for
    them to appear in both tables, and the code paths end up stepping over
    each other in particular for a partial instantiation such as
    A<B>::key_arg<T> in the below modules testcase: the type code path
    (lookup_template_class) wants to set TI_TEMPLATE to the most general
    template whereas the decl code path (tsubst_template_decl called during
    instantiation of A<B>) already set TI_TEMPLATE to the partially
    instantiated TEMPLATE_DECL.  This TI_TEMPLATE change ends up confusing
    modules which decides to stream the logically equivalent TYPE_DECL and
    TEMPLATE_DECL for this partial instantiation separately.

    This patch fixes this by making lookup_template_class dispatch to
    instantiate_alias_template early for alias template specializations.
    In turn we now add such specializations only to the decl table.  This
    admits some nice simplification in the modules code which otherwise has
    to cope with such specializations appearing in both tables.

            PR c++/103994

    gcc/cp/ChangeLog:

            * cp-tree.h (add_mergeable_specialization): Remove second
            parameter.
            * module.cc (depset::disc_bits::DB_ALIAS_TMPL_INST_BIT): Remove.
            (depset::disc_bits::DB_ALIAS_SPEC_BIT): Remove.
            (depset::is_alias_tmpl_inst): Remove.
            (depset::is_alias): Remove.
            (merge_kind::MK_tmpl_alias_mask): Remove.
            (merge_kind::MK_alias_spec): Remove.
            (merge_kind_name): Remove entries for alias specializations.
            (trees_out::core_vals) <case TEMPLATE_DECL>: Adjust after
            removing is_alias_tmpl_inst.
            (trees_in::decl_value): Adjust add_mergeable_specialization
            calls.
            (trees_out::get_merge_kind) <case depset::EK_SPECIALIZATION>:
            Use MK_decl_spec for alias template specializations.
            (trees_out::key_mergeable): Simplify after MK_tmpl_alias_mask
            removal.
            (depset::hash::make_dependency): Adjust after removing
            DB_ALIAS_TMPL_INST_BIT.
            (specialization_add): Don't allow alias templates when !decl_p.
            (depset::hash::add_specializations): Remove now-dead code
            accomodating alias template specializations in the type table.
            * pt.cc (lookup_template_class): Dispatch early to
            instantiate_alias_template for alias templates.  Simplify
            accordingly.
            (add_mergeable_specialization): Remove alias_p parameter and
            simplify accordingly.

    gcc/testsuite/ChangeLog:

            * g++.dg/modules/pr99425-1_b.H: s/alias/decl in dump scan.
            * g++.dg/modules/tpl-alias-1_a.H: Likewise.
            * g++.dg/modules/tpl-alias-2_a.H: New test.
            * g++.dg/modules/tpl-alias-2_b.C: New test.

    Reviewed-by: Jason Merrill <jason@redhat.com>

[Bug c++/103994] Module ICE in write_var_def with global variable in global module fragment

[ppalka at gcc dot gnu.org]

https://gcc.gnu.org/bugzilla/show_bug.cgi?id=103994

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|ASSIGNED                    |RESOLVED
   Target Milestone|---                         |14.0
         Resolution|---                         |FIXED

--- Comment #6 from Patrick Palka <ppalka at gcc dot gnu.org> ---
Fixed for GCC 14.