[Bug c++/21645] New: G++ ICE while building post-html.cpp in groff

[Bug c++/21645] New: G++ ICE while building post-html.cpp in groff

[ahaas at airmail dot net]

CVS g++ currently fails while building CVS groff:

g++ -I. -I/devel/gnu/gnu_cvs/groff/src/devices/grohtml
-I/devel/gnu/gnu_cvs/groff/objdir-0328/src/include
-I/devel/gnu/gnu_cvs/groff/src/include -DHAVE_CONFIG_H -v -O2 -march=pentium-mmx
-finline-limit=100000 -fno-check-new -fno-exceptions -save-temps -DNDEBUG -c 
/devel/gnu/gnu_cvs/groff/src/devices/grohtml/post-html.cpp
Using built-in specs.
Target: i586-pc-linux-gnu
Configured with: /usr/src/gcc_3_5_branch/gcc/configure --prefix=/opt/gnu
--enable-shared --enable-threads=posix --enable-__cxa_atexit
--enable-languages=c,c++,objc --disable-checking --with-system-zlib --with-gc=page
Thread model: posix
gcc version 4.1.0 20050517 (experimental)
 /opt/gnu/libexec/gcc/i586-pc-linux-gnu/4.1.0/cc1plus -E -quiet -v -I.
-I/devel/gnu/gnu_cvs/groff/src/devices/grohtml
-I/devel/gnu/gnu_cvs/groff/objdir-0328/src/include
-I/devel/gnu/gnu_cvs/groff/src/include -D_GNU_SOURCE -DHAVE_CONFIG_H -DNDEBUG
/devel/gnu/gnu_cvs/groff/src/devices/grohtml/post-html.cpp -march=pentium-mmx
-finline-limit=100000 -fno-check-new -fno-exceptions -O2 -fpch-preprocess -o
post-html.ii
#include "..." search starts here:
#include <...> search starts here:
 .
 /devel/gnu/gnu_cvs/groff/src/devices/grohtml
 /devel/gnu/gnu_cvs/groff/objdir-0328/src/include
 /devel/gnu/gnu_cvs/groff/src/include
 /opt/gnu/lib/gcc/i586-pc-linux-gnu/4.1.0/../../../../include/c++/4.1.0
 /opt/gnu/lib/gcc/i586-pc-linux-gnu/4.1.0/../../../../include/c++/4.1.0/i586-pc-linux-gnu
 /opt/gnu/lib/gcc/i586-pc-linux-gnu/4.1.0/../../../../include/c++/4.1.0/backward
 /usr/local/include
 /opt/gnu/include
 /opt/gnu/lib/gcc/i586-pc-linux-gnu/4.1.0/include
 /opt/gnu/lib/gcc/i586-pc-linux-gnu/4.1.0/../../../../i586-pc-linux-gnu/include
 /usr/include
End of search list.
 /opt/gnu/libexec/gcc/i586-pc-linux-gnu/4.1.0/cc1plus -fpreprocessed
post-html.ii -quiet -dumpbase post-html.cpp -march=pentium-mmx -auxbase
post-html -O2 -version -finline-limit=100000 -fno-check-new -fno-exceptions -o
post-html.s
GNU C++ version 4.1.0 20050517 (experimental) (i586-pc-linux-gnu)
        compiled by GNU C version 4.1.0 20050517 (experimental).
GGC heuristics: --param ggc-min-expand=42 --param ggc-min-heapsize=24060
/devel/gnu/gnu_cvs/groff/src/devices/grohtml/post-html.cpp: In constructor
'style::style(font*, int, int, int, int, color)':
/devel/gnu/gnu_cvs/groff/src/devices/grohtml/post-html.cpp:5054: internal
compiler error: Segmentation fault
Please submit a full bug report.

The failures began on May 14; previous CVS GCC builds were able to compile groff
without problem.

-- 
           Summary: G++ ICE while building post-html.cpp in groff
           Product: gcc
           Version: 4.1.0
            Status: UNCONFIRMED
          Severity: normal
          Priority: P2
         Component: c++
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: ahaas at airmail dot net
                CC: gcc-bugs at gcc dot gnu dot org
  GCC host triplet: i586-pc-linux-gnu


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] G++ ICE while building post-html.cpp in groff

[ahaas at airmail dot net]

------- Additional Comments From ahaas at airmail dot net  2005-05-18 13:46 -------
Created an attachment (id=8922)
 --> (http://gcc.gnu.org/bugzilla/attachment.cgi?id=8922&action=view)
Preprocessed file triggering bug

Here's the file, compressed with 'gzip' as it is large.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] G++ ICE while building post-html.cpp in groff

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-18 14:00 -------
Confirmed, reduced testcase:
struct color {
  ~color();
};
struct style {
  color col;
  style (color);
};

style::style(color c)
  : col(c)
{
}

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
           Keywords|                            |ice-on-valid-code
   Last reconfirmed|0000-00-00 00:00:00         |2005-05-18 14:00:27
               date|                            |
            Summary|G++ ICE while building post-|[4.1 Regression] G++ ICE
                   |html.cpp in groff           |while building post-html.cpp
                   |                            |in groff
   Target Milestone|---                         |4.1.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] G++ ICE while building post-html.cpp in groff

[janis at gcc dot gnu dot org]

------- Additional Comments From janis at gcc dot gnu dot org  2005-05-18 22:19 -------
This is the same problem that causes cc1plus to segfault building SPEC CPU2000
test eon for powerpc-linux at any level of optimization including -O0.  The
segfault happens when maybe_fold_offset_to_component_ref calls
cxx_types_compatible_p with y==0.  The failure shows up with this patch from
Jan Hubicka:

  http://gcc.gnu.org/ml/gcc-cvs/2005-05/msg00624.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-19 00:40 -------
Hmm, we have an INDIRECT_REF with no type, I wonder how that could.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
            Summary|[4.1 Regression] G++ ICE    |[4.1 Regression] C++ ICE in
                   |while building post-html.cpp|cxx_types_compatible_p
                   |in groff                    |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-19 00:48 -------
Hmm, the problem is we don't have a reference type at the point we clone the constructor.  The reason 
why it worked before Honza's patch is because we did not call build_fold_indirect_ref before.  This is a 
bug in the C++ front-end.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |pinskia at gcc dot gnu dot
                   |                            |org


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-19 01:02 -------
I think I have a patch, the problem is that the type for the cloned parm does not get updated.  I wonder 
why this has not showed up before.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
         AssignedTo|unassigned at gcc dot gnu   |pinskia at gcc dot gnu dot
                   |dot org                     |org
             Status|NEW                         |ASSIGNED


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-20 22:16 -------
Patch summitted:
http://gcc.gnu.org/ml/gcc-patches/2005-05/msg02088.html

I noticed I forgot to attach a testcase, oh well.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                URL|                            |http://gcc.gnu.org/ml/gcc-
                   |                            |patches/2005-
                   |                            |05/msg02088.html
           Keywords|                            |patch


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[ahaas at airmail dot net]

------- Additional Comments From ahaas at airmail dot net  2005-05-21 15:39 -------
I applied the patch, rebuilt, installed, and can confirm the patch works as I
had a successful groff CVS build.

Thanks!

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[cvs-commit at gcc dot gnu dot org]

------- Additional Comments From cvs-commit at gcc dot gnu dot org  2005-05-24 19:06 -------
Subject: Bug 21645

CVSROOT:	/cvs/gcc
Module name:	gcc
Changes by:	pinskia@gcc.gnu.org	2005-05-24 19:06:20

Modified files:
	gcc/cp         : ChangeLog optimize.c 
	gcc/testsuite  : ChangeLog 
Added files:
	gcc/testsuite/g++.dg/init: ctor5.C 

Log message:
	2005-05-24  Andrew Pinski  <pinskia@physics.uc.edu>
	
	PR C++/21645
	* g++.dg/init/ctor5.C: New test.
	
	2005-05-24  Andrew Pinski  <pinskia@physics.uc.edu>
	
	PR C++/21645
	* optimize.c (update_cloned_parm): Copy the TYPE also from the
	original one.

Patches:
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/ChangeLog.diff?cvsroot=gcc&r1=1.4751&r2=1.4752
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/cp/optimize.c.diff?cvsroot=gcc&r1=1.114&r2=1.115
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/ChangeLog.diff?cvsroot=gcc&r1=1.5510&r2=1.5511
http://gcc.gnu.org/cgi-bin/cvsweb.cgi/gcc/gcc/testsuite/g++.dg/init/ctor5.C.diff?cvsroot=gcc&r1=NONE&r2=1.1



-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645

[Bug c++/21645] [4.1 Regression] C++ ICE in cxx_types_compatible_p

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2005-05-24 19:06 -------
Fixed.

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=21645