[Bug c++/52510] New: [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[Bug c++/52510] New: [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

             Bug #: 52510
           Summary: [4.8 regression] libitm/config/posix/rwlock.cc doesn't
                    compile
    Classification: Unclassified
           Product: gcc
           Version: 4.7.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P3
         Component: c++
        AssignedTo: unassigned@gcc.gnu.org
        ReportedBy: ro@gcc.gnu.org
                CC: jason@gcc.gnu.org
              Host: *-*-solaris2*
            Target: *-*-solaris2*
             Build: *-*-solaris2*


Between 20120302 and 20120306, mainline doesn't bootstrap on Solaris any
longer:
compiling libitm/config/posix/rwlock.cc fails like this:

/vol/gcc/src/hg/trunk/local/libitm/config/posix/rwlock.cc: In constructor
'GTM::gtm_rwlock::gtm_rwlock()':
/vol/gcc/src/hg/trunk/local/libitm/config/posix/rwlock.cc:40:17: error: no
matching function for call to '_pthread_mutex::_pthread_mutex(<brace-enclosed
initializer list>)'
/vol/gcc/src/hg/trunk/local/libitm/config/posix/rwlock.cc:40:17: note:
candidates are:
In file included from /usr/include/sys/wait.h:20:0,
                 from /usr/include/stdlib.h:22,
                 from /vol/gcc/src/hg/trunk/local/libitm/libitm_i.h:36,
                 from
/vol/gcc/src/hg/trunk/local/libitm/config/posix/rwlock.cc:25:
/usr/include/sys/types.h:381:16: note: _pthread_mutex::_pthread_mutex()
/usr/include/sys/types.h:381:16: note:   candidate expects 0 arguments, 1
provided
/usr/include/sys/types.h:381:16: note: constexpr
_pthread_mutex::_pthread_mutex(const _pthread_mutex&)
/usr/include/sys/types.h:381:16: note:   no known conversion for argument 1
from '<brace-enclosed initializer list>' to 'const _pthread_mutex&'
/usr/include/sys/types.h:381:16: note: constexpr
_pthread_mutex::_pthread_mutex(_pthread_mutex&&)
/usr/include/sys/types.h:381:16: note:   no known conversion for argument 1
from '<brace-enclosed initializer list>' to '_pthread_mutex&&'

and several more.  Compiling the preprocessed with with g++ 4.7 works just
fine.

The issue can be reproduced with the attached testcase:

$ cc1plus -fpreprocessed init.ii -quiet -g -O2 -std=gnu++11 -o init.s
init.ii: In constructor 'gtm_rwlock::gtm_rwlock()':
init.ii:24:46: error: no matching function for call to
'_pthread_cond::_pthread_cond(<brace-enclosed initializer list>)'
init.ii:24:46: note: candidates are:
init.ii:7:16: note: _pthread_cond::_pthread_cond()
init.ii:7:16: note:   candidate expects 0 arguments, 1 provided
init.ii:7:16: note: constexpr _pthread_cond::_pthread_cond(const
_pthread_cond&)
init.ii:7:16: note:   no known conversion for argument 1 from '<brace-enclosed
initializer list>' to 'const _pthread_cond&'
init.ii:7:16: note: constexpr _pthread_cond::_pthread_cond(_pthread_cond&&)
init.ii:7:16: note:   no known conversion for argument 1 from '<brace-enclosed
initializer list>' to '_pthread_cond&&'

I suspect that this patch

2012-03-03  Jason Merrill  <jason@redhat.com>

        * init.c (perform_member_init): Cope with uninstantiated NSDMI.

        Core 1270
        * call.c (build_aggr_conv): Call reshape_init.
        (convert_like_real): Likewise.
        * typeck2.c (process_init_constructor): Clear TREE_CONSTANT if
        not all constant.

is the culprit.

  Rainer

[Bug c++/52510] [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

--- Comment #1 from Rainer Orth <ro at gcc dot gnu.org> 2012-03-06 13:59:33 UTC ---
Created attachment 26840
  --> http://gcc.gnu.org/bugzilla/attachment.cgi?id=26840
preprocessed input

[Bug c++/52510] [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[rguenth at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

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

           What    |Removed                     |Added
----------------------------------------------------------------------------
            Version|4.7.0                       |4.8.0
   Target Milestone|---                         |4.8.0

[Bug c++/52510] [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[ro at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

--- Comment #2 from Rainer Orth <ro at gcc dot gnu.org> 2012-03-07 17:49:20 UTC ---
Nothing Solaris-specific here, fails in the same way on
x86_64-unknown-linux-gnu
as of r185060.

  Rainer

[Bug c++/52510] [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[ro at CeBiTec dot Uni-Bielefeld.DE]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

--- Comment #3 from ro at CeBiTec dot Uni-Bielefeld.DE <ro at CeBiTec dot Uni-Bielefeld.DE> 2012-03-07 20:34:57 UTC ---
A reghunt revealed that this patch (r184876) is indeed the culprit:

2012-03-03  Jason Merrill  <jason@redhat.com>

       Core 1270
       * call.c (build_aggr_conv): Call reshape_init.
       (convert_like_real): Likewise.
       * typeck2.c (process_init_constructor): Clear TREE_CONSTANT if
       not all constant.

Jason, could you please have a look?

Thanks.
        Rainer

[Bug c++/52510] [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

--- Comment #4 from Jason Merrill <jason at gcc dot gnu.org> 2012-03-20 19:14:33 UTC ---
Author: jason
Date: Tue Mar 20 19:14:29 2012
New Revision: 185587

URL: http://gcc.gnu.org/viewcvs?root=gcc&view=rev&rev=185587
Log:
    PR c++/52510
    * decl.c (reshape_init_class): Handle repeated reshaping.
    * search.c (lookup_field_1): Add sanity check.

Added:
    trunk/gcc/testsuite/g++.dg/cpp0x/initlist66.C
Modified:
    trunk/gcc/cp/ChangeLog
    trunk/gcc/cp/decl.c
    trunk/gcc/cp/search.c
    trunk/gcc/testsuite/ChangeLog

[Bug c++/52510] [4.8 regression] libitm/config/posix/rwlock.cc doesn't compile

[jason at gcc dot gnu.org]

http://gcc.gnu.org/bugzilla/show_bug.cgi?id=52510

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

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

--- Comment #5 from Jason Merrill <jason at gcc dot gnu.org> 2012-07-03 04:03:21 UTC ---
Fixed.