[Bug c++/60033] New: gcc 4.9.0 20140107: ICE in retrieve_specialization while compiling recursive generic lambda

[Bug c++/60033] New: gcc 4.9.0 20140107: ICE in retrieve_specialization while compiling recursive generic lambda

[vock at cs dot uni-bonn.de]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

            Bug ID: 60033
           Summary: gcc 4.9.0 20140107: ICE in retrieve_specialization
                    while compiling recursive generic lambda
           Product: gcc
           Version: 4.9.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
          Assignee: unassigned at gcc dot gnu.org
          Reporter: vock at cs dot uni-bonn.de

Created attachment 32016
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=32016&action=edit
Preprocessed output file

I got an ICE while compiling the following minimal code example (.ii is
attached):


//#include <iostream>
#include <utility>

template <typename... Funcs>
struct composer;

template <typename Func>
struct composer<Func> {
    auto operator()(Func&& f) {
        return f;
    }
};

template <typename Func, typename... Funcs>
struct composer<Func, Funcs...> {
    auto operator()(Func&& f, Funcs&&... fs) {
        composer<Funcs...> c;
        return [&] (auto v) {
            return f(composer<Funcs...>()(std::forward<Funcs>(fs)...)(v));
        };
    }
};

template <typename... Funcs>
auto compose(Funcs&&... fs) {
    return composer<Funcs...>()(std::forward<Funcs>(fs)...);
}


int main (int argc, char const* argv[]) {
    float v = 3.5f;
    auto func = [] (auto v) { return int(v-0.5); };
    auto t = compose([] (auto v) { return v >= 3; }, func)(v);
//    std::cout << std::boolalpha << t << "\n";
    auto f = compose([] (auto v) { return v > 3; }, func)(v);
//    std::cout << std::boolalpha << f << "\n";
}


Compiler output is:

> g++-4.9 -std=c++1y -save-temps -Wall -Wextra -fno-strict-aliasing -fwrapv -fno-aggressive-loop-optimizations main.cpp

main.cpp: In instantiation of ‘composer<Func, Funcs ...>::operator()(Func&&,
Funcs&& ...)::<lambda(auto:2)> [with auto:2 = float; Func = main(int, const
char**)::<lambda(auto:3)>; Funcs = {main(int, const
char**)::<lambda(auto:4)>}]’:
main.cpp:33:88:   required from here
main.cpp:19:41: internal compiler error: in retrieve_specialization, at
cp/pt.c:1042
    return f(c(std::forward<Funcs>(fs)...)(v));
                                         ^
0x59c47f retrieve_specialization
    ./src/gcc/gcc/cp/pt.c:1039
0x5a470c tsubst_pack_expansion
    ./src/gcc/gcc/cp/pt.c:9563
0x59f404 tsubst_copy_and_build(tree_node*, tree_node*, int, tree_node*, bool,
bool)
    ./src/gcc/gcc/cp/pt.c:14461
0x59fa18 tsubst_copy_and_build(tree_node*, tree_node*, int, tree_node*, bool,
bool)
    ./src/gcc/gcc/cp/pt.c:14443
0x59f3a0 tsubst_copy_and_build(tree_node*, tree_node*, int, tree_node*, bool,
bool)
    ./src/gcc/gcc/cp/pt.c:14456
0x5a31ce tsubst_expr
    ./src/gcc/gcc/cp/pt.c:13788
0x5a31a4 tsubst_expr
    ./src/gcc/gcc/cp/pt.c:13205
0x5a316f tsubst_expr
    ./src/gcc/gcc/cp/pt.c:13400
0x5a2e34 tsubst_expr
    ./src/gcc/gcc/cp/pt.c:13195
0x5a316f tsubst_expr
    ./src/gcc/gcc/cp/pt.c:13400
0x5a2038 instantiate_decl(tree_node*, int, bool)
    ./src/gcc/gcc/cp/pt.c:19656
0x5cfbb9 mark_used(tree_node*, int)
    ./src/gcc/gcc/cp/decl2.c:4701
0x57760b build_over_call
    ./src/gcc/gcc/cp/call.c:7127
0x57a301 build_op_call_1
    ./src/gcc/gcc/cp/call.c:4161
0x57a301 build_op_call(tree_node*, vec<tree_node*, va_gc, vl_embed>**, int)
    ./src/gcc/gcc/cp/call.c:4184
0x636cf8 finish_call_expr(tree_node*, vec<tree_node*, va_gc, vl_embed>**, bool,
bool, int)
    ./src/gcc/gcc/cp/semantics.c:2298
0x5ec24b cp_parser_postfix_expression
    ./src/gcc/gcc/cp/parser.c:6165
0x5ee8b8 cp_parser_unary_expression
    ./src/gcc/gcc/cp/parser.c:7171
0x5ef48f cp_parser_binary_expression
    ./src/gcc/gcc/cp/parser.c:7875
0x5ef951 cp_parser_assignment_expression
    ./src/gcc/gcc/cp/parser.c:8113




A running version compiled using clang++ 3.5 is here:
http://coliru.stacked-crooked.com/a/acc691b9a407d6f2




System:

Arch Linux
Linux xxx 3.12.6-1-ARCH #1 SMP PREEMPT Fri Dec 20 19:39:00 CET 2013 x86_64
GNU/Linux

Build Command:

g++-4.9 -std=c++1y -save-temps -Wall -Wextra -fno-strict-aliasing -fwrapv
-fno-aggressive-loop-optimizations main.cpp


Version:

Using built-in specs.
COLLECT_GCC=g++-4.9
COLLECT_LTO_WRAPPER=/usr/lib/gcc/x86_64-unknown-linux-gnu/4.9.0/lto-wrapper
Target: x86_64-unknown-linux-gnu
Configured with: ./src/gcc/configure --prefix=/usr --libdir=/usr/lib
--libexecdir=/usr/lib --mandir=/usr/share/man --infodir=/usr/share/info
--datadir=/usr/share/gcc-4.9
--with-bugurl='http://aur.archlinux.org/packages.php?ID=16045'
--enable-languages=c,c++,lto --enable-shared --enable-threads=posix
--with-system-zlib --enable-__cxa_atexit --disable-libunwind-exceptions
--enable-clocale=gnu --disable-libstdcxx-pch --enable-gnu-unique-object
--enable-linker-build-id --enable-cloog-backend=isl
--disable-cloog-version-check --enable-lto --enable-gold --enable-ld=default
--enable-plugin --with-plugin-ld=ld.gold --with-linker-hash-style=gnu
--disable-install-libiberty --disable-multilib --disable-libssp
--disable-werror --enable-checking=release --program-suffix=-4.9
--enable-version-specific-runtime-libs
Thread model: posix
gcc version 4.9.0 20140107 (experimental) (GCC)
>From gcc-bugs-return-442326-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org Sun Feb 02 19:57:36 2014
Return-Path: <gcc-bugs-return-442326-listarch-gcc-bugs=gcc.gnu.org@gcc.gnu.org>
Delivered-To: listarch-gcc-bugs@gcc.gnu.org
Received: (qmail 5894 invoked by alias); 2 Feb 2014 19:57:35 -0000
Mailing-List: contact gcc-bugs-help@gcc.gnu.org; run by ezmlm
Precedence: bulk
List-Id: <gcc-bugs.gcc.gnu.org>
List-Archive: <http://gcc.gnu.org/ml/gcc-bugs/>
List-Post: <mailto:gcc-bugs@gcc.gnu.org>
List-Help: <mailto:gcc-bugs-help@gcc.gnu.org>
Sender: gcc-bugs-owner@gcc.gnu.org
Delivered-To: mailing list gcc-bugs@gcc.gnu.org
Received: (qmail 5877 invoked by uid 48); 2 Feb 2014 19:57:32 -0000
From: "vock at cs dot uni-bonn.de" <gcc-bugzilla@gcc.gnu.org>
To: gcc-bugs@gcc.gnu.org
Subject: [Bug c++/60033] gcc 4.9.0 20140107: ICE in retrieve_specialization while compiling recursive generic lambda
Date: Sun, 02 Feb 2014 19:57:00 -0000
X-Bugzilla-Reason: CC
X-Bugzilla-Type: changed
X-Bugzilla-Watch-Reason: None
X-Bugzilla-Product: gcc
X-Bugzilla-Component: c++
X-Bugzilla-Version: 4.9.0
X-Bugzilla-Keywords:
X-Bugzilla-Severity: normal
X-Bugzilla-Who: vock at cs dot uni-bonn.de
X-Bugzilla-Status: UNCONFIRMED
X-Bugzilla-Priority: P3
X-Bugzilla-Assigned-To: unassigned at gcc dot gnu.org
X-Bugzilla-Target-Milestone: ---
X-Bugzilla-Flags:
X-Bugzilla-Changed-Fields:
Message-ID: <bug-60033-4-CkTS5cETMR@http.gcc.gnu.org/bugzilla/>
In-Reply-To: <bug-60033-4@http.gcc.gnu.org/bugzilla/>
References: <bug-60033-4@http.gcc.gnu.org/bugzilla/>
Content-Type: text/plain; charset="UTF-8"
Content-Transfer-Encoding: 7bit
X-Bugzilla-URL: http://gcc.gnu.org/bugzilla/
Auto-Submitted: auto-generated
MIME-Version: 1.0
X-SW-Source: 2014-02/txt/msg00083.txt.bz2
Content-length: 201

http://gcc.gnu.org/bugzilla/show_bug.cgi?id`033

--- Comment #1 from Richard Vock <vock at cs dot uni-bonn.de> ---
I know I left a redundant composer<Funcs...> c; line in there. Just ignore
that...

[Bug c++/60033] ICE in retrieve_specialization while compiling recursive generic lambda

[paolo.carlini at oracle dot com]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

Paolo Carlini <paolo.carlini at oracle dot com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
   Last reconfirmed|                            |2014-02-03
                 CC|vock at cs dot uni-bonn.de         |adam at jessamine dot co.uk
            Summary|gcc 4.9.0 20140107: ICE in  |ICE in
                   |retrieve_specialization     |retrieve_specialization
                   |while compiling recursive   |while compiling recursive
                   |generic lambda              |generic lambda
     Ever confirmed|0                           |1

--- Comment #2 from Paolo Carlini <paolo.carlini at oracle dot com> ---
Maybe Adam can have a look.

[Bug c++/60033] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

Adam Butcher <abutcher at gcc dot gnu.org> changed:

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

--- Comment #3 from Adam Butcher <abutcher at gcc dot gnu.org> ---
The following reduced test case gives the same ICE:

---------------------------
#include <utility>

template <typename Func>
auto compose(Func&& f) {
  return f;
}

template <typename Func, typename... Funcs>
auto compose(Func&& f, Funcs&&... fs) {
  return [&] (auto v) {
    return f(compose(std::forward<Funcs>(fs)...)(v));
  };
}

int main()
{
  float v = 3.5f;
  auto func = [] (auto v) { return int(v-0.5); };
  auto t = compose([] (auto v) { return v >= 3; }, func)(v);
  auto f = compose([] (auto v) { return v > 3; }, func)(v);
  return (t == true && f == false)? 0 : 1;
}
---------------------------

'retrieve_specialization' is failing an assertion due to a mismatch between the
template argument depth and template decl depth when considering the pack
expansion of the captured 'fs' within the generic lambda:

---------------------------
|| args =>  <tree_vec 0x7ffd49d54ed8
||     elt 0 <tree_vec 0x7ffd49d54eb0
||         elt 0 <record_type 0x7ffd49d48c78 __lambda2 needs-constructing
type_5 type_6 QI
||             size <integer_cst 0x7ffd49df1280 constant 8>
||             unit size <integer_cst 0x7ffd49df12a0 constant 1>
||             align 8 symtab 0 alias set -1 canonical type 0x7ffd49d48c78
fields <type_decl 0x7ffd49d41b80 __lambda2> context <function_decl
0x7ffd49d2bc00 main>
||             full-name "struct main()::<lambda(auto:3)>"
||             needs-constructor X() has-type-conversion X(constX&) this=(X&)
n_parents=0 use_template=0 interface-unknown
||             pointer_to_this <pointer_type 0x7ffd49d4b9d8> reference_to_this
<reference_type 0x7ffd49d510a8> chain <type_decl 0x7ffd49d41ac8 __lambda2>>
||         elt 1 <type_argument_pack 0x7ffd49d56a80
||             type <tree_vec 0x7ffd49d53a20 elt 0 <reference_type
0x7ffd49d48930>>
||             VOID
||             align 8 symtab 0 alias set -1 canonical type 0x7ffd49d56a80>>
||     elt 1 <tree_vec 0x7ffd49d53a40
||         elt 0 <real_type 0x7ffd49e02348 float type_6 SF
||             size <integer_cst 0x7ffd49df1440 constant 32>
||             unit size <integer_cst 0x7ffd49df1460 constant 4>
||             align 32 symtab 0 alias set -1 canonical type 0x7ffd49e02348
precision 32
||             pointer_to_this <pointer_type 0x7ffd49e02540>>>>
|| tmpl =>  <field_decl 0x7ffd49d34ab0 __fs
||     type <type_pack_expansion 0x7ffd49d450a8
||         type <reference_type 0x7ffd49d45000 type <decltype_type
0x7ffd49d40f18>
||             unsigned DI
||             size <integer_cst 0x7ffd49df10c0 constant 64>
||             unit size <integer_cst 0x7ffd49df10e0 constant 8>
||             align 64 symtab 0 alias set -1 structural equality>
||         tree_0 VOID
||         align 8 symtab 0 alias set -1 structural equality>
||     used decl_7 VOID file gcc/testsuite/g++.dg/cpp1y/pr60033.C line 15 col
42
||     align 1 offset_align 1 context <record_type 0x7ffd49d40000 __lambda0>
chain <type_decl 0x7ffd49d410b8 __lambda0>>
|| TMPL_ARGS_DEPTH (args) => 2
|| TREE_CODE (tmpl) == TEMPLATE_DECL => 0
|| template_class_depth (DECL_CONTEXT (tmpl)) => 1
---------------------------

[Bug c++/60033] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

Adam Butcher <abutcher at gcc dot gnu.org> changed:

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

--- Comment #4 from Adam Butcher <abutcher at gcc dot gnu.org> ---
I think I have a fix for this but I need a sanity check.  Does this adjustment
in 'retrieve_specialization' look to be along the right lines Jason?  Certainly
the programs attached to this issue compile and run as expected; haven't gone
any further than that though.

  /* When retrieving a capture pack from a generic lambda, remove the lambda
     call op's own template argument list from ARGS.  Only the template
     arguments active for the closure type should be used to retrieve the pack
     specialization.  */
  if (TREE_CODE (tmpl) == FIELD_DECL
      && LAMBDA_FUNCTION_P (current_function_decl)
      && DECL_TEMPLATE_INFO (current_function_decl))
    args = strip_innermost_template_args (args, 1);

[Bug c++/60033] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

--- Comment #5 from Adam Butcher <abutcher at gcc dot gnu.org> ---
Turns out that the above breaks the generic lambda variadic testcase:

  auto vglambda = [](auto printer) {
    return [=](auto&& ... ts) { // OK: ts is a function parameter pack
      printer(std::forward<decltype(ts)>(ts)...);
      return [=]() {
        printer(ts ...);
      };
    };
  };

So not there yet.  Maybe using 'current_function' is wrong?

[Bug c++/60033] [c++1y] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

--- Comment #6 from Adam Butcher <abutcher at gcc dot gnu.org> ---
A further reduced testcase:

// PR c++/60033
// { dg-options -std=c++1y }

template <typename... T>
auto f(T&&... ts)
{
  return sizeof...(ts);
}

template <typename... T>
auto g(T&&... ts) {
  return [&] (int v) {
    return f(ts...);
  };
}

int main()
{
  return g(1,2,3,4)(5) == 4 ? 0 : 1;
}

[Bug c++/60033] [c++1y] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

--- Comment #7 from Adam Butcher <abutcher at gcc dot gnu.org> ---
(In reply to Adam Butcher from comment #6)
>   return [&] (int v) {
>     return f(ts...);
>   };
Should have been:

  return [&] (auto v) {
    return f(ts...);
  };

The 'int' version works as expected.

[Bug c++/60033] [c++1y] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

--- Comment #8 from Adam Butcher <abutcher at gcc dot gnu.org> ---
Author: abutcher
Date: Sat Mar  8 09:33:12 2014
New Revision: 208427

URL: http://gcc.gnu.org/viewcvs?rev=208427&root=gcc&view=rev
Log:
Fix PR c++/60033

    PR c++/60033
    * pt.c (tsubst_copy): When retrieving a capture pack from a generic
    lambda, remove the lambda's own template argument list prior to fetching
    the specialization.

    PR c++/60033
    * g++.dg/cpp1y/pr60033.C: New testcase.

Added:
    trunk/gcc/testsuite/g++.dg/cpp1y/pr60033.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/pt.c
    trunk/gcc/testsuite/ChangeLog

[Bug c++/60033] [c++1y] ICE in retrieve_specialization while compiling recursive generic lambda

[abutcher at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=60033

Adam Butcher <abutcher at gcc dot gnu.org> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|NEW                         |RESOLVED
         Resolution|---                         |FIXED
           Assignee|unassigned at gcc dot gnu.org      |abutcher at gcc dot gnu.org
   Target Milestone|---                         |4.9.0

--- Comment #9 from Adam Butcher <abutcher at gcc dot gnu.org> ---
Fixed in 4.9.