[Bug target/14663] New: [3.5 regression] bootstrap failure in libobjc (first .m file)

[Bug target/14663] New: [3.5 regression] bootstrap failure in libobjc (first .m file)

[debian-gcc at lists dot debian dot org]

In file included from /build/packages/gcc/exp/gcc-experimental-20040320/src/libo
bjc/linking.m:28:
/build/packages/gcc/exp/gcc-experimental-20040320/src/libobjc/objc/Object.h:43: 
internal compiler error: Segmentation fault
Please submit a full bug report,
with preprocessed source if appropriate.
See <URL:http://gcc.gnu.org/bugs.html> for instructions.
make[4]: *** [linking.lo] Error 1
make[4]: Leaving directory `/build/packages/gcc/exp/gcc-experimental-20040320/bu
ild/hppa-linux/libobjc'


Program received signal SIGSEGV, Segmentation fault.
0x0002a210 in hash_lookup ()
(gdb) bt
#0  0x0002a210 in hash_lookup ()
#1  0x0002fd88 in add_method_to_hash_list ()
#2  0x0002ffa8 in objc_add_method ()
#3  0x0001f374 in yyparse ()
#4  0x00021928 in c_parse_file ()
#5  0x0007f19c in c_common_parse_file ()
#6  0x00255cf8 in toplev_main ()
#7  0x4004845c in ?? ()

-- 
           Summary: [3.5 regression] bootstrap failure in libobjc (first .m
                    file)
           Product: gcc
           Version: 3.5.0
            Status: UNCONFIRMED
          Severity: critical
          Priority: P2
         Component: target
        AssignedTo: unassigned at gcc dot gnu dot org
        ReportedBy: debian-gcc at lists dot debian dot org
                CC: gcc-bugs at gcc dot gnu dot org
 GCC build triplet: hppa-linux
  GCC host triplet: hppa-linux
GCC target triplet: hppa-linux


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[pinskia at gcc dot gnu dot org]

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
           Keywords|                            |build, ice-on-valid-code,
                   |                            |wrong-code
   Target Milestone|---                         |3.5.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-20 17:45 -------
>From <http://gcc.gnu.org/ml/gcc-testresults/2004-03/msg01045.html>, it does not look like 
it is failing at all.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-20 17:51 -------
But based on the next one: <http://gcc.gnu.org/ml/gcc-testresults/2004-03/msg01046.html>.  It 
does look there is a problem.

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[danglin at gcc dot gnu dot org]

------- Additional Comments From danglin at gcc dot gnu dot org  2004-03-20 20:27 -------
> From <http://gcc.gnu.org/ml/gcc-testresults/2004-03/msg01045.html>, it does not look like 
> it is failing at all.

This run was done at D2004.03.18.20.58.00 (i.e., just before the Kenner
patch that caused other breakage because parts of the patch were missed).

The following run was done at D2004.03.19.14.19.00.  This is just after
the last of the fixes applied by Kenner to his original fix:

2004-03-19  Richard Kenner  <kenner@vlsi1.ultra.nyu.edu>

        * cxx-pretty-print.c (pp_cxx_unqualified_id): Use
        TEMPLATE_TYPE_PARM_INDEX instead of TYPE_FIELDS.
        * search.c (dfs_unuse_fields): Add two more TREE_CODES that mean
        the field is named TEMPLATE_TYPE_PARM_INDEX.

http://gcc.gnu.org/ml/gcc-testresults/2004-03/msg01046.html

There are a huge number of objc test fails in this test run.

At a somewhat latter time, we get the error compiling linking.m.
I haven't identified which change caused this regression.


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[danglin at gcc dot gnu dot org]

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-21 03:17 -------
Here are the patches which could effect the testsuite (meaning not including libjava patches and no 
C++ ones) went 
inbetween the time which David mentions:
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00853.html
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00854.html
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00856.html
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00857.html
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00869.html (math builtins, most likely not)
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00870.html (math builtins, most likely not)
http://gcc.gnu.org/ml/gcc-cvs/2004-03/msg00875.html (this patch caused other failures so it is most 
likely one here)

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
      Known to fail|                            |3.5.0
      Known to work|                            |3.4.0


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[dave at hiauly1 dot hia dot nrc dot ca]

------- Additional Comments From dave at hiauly1 dot hia dot nrc dot ca  2004-03-21 03:29 -------
Subject: Re:  [3.5 regression] bootstrap failure in libobjc

This is the patch that caused the testsuite regression:

CVSROOT:        /cvs/gcc
Module name:    gcc
Changes by:     rsandifo@gcc.gnu.org    2004-03-19 09:59:01

Modified files:
        gcc            : ChangeLog expmed.c 

Log message:
	* expmed.c (choose_mult_variant, expand_mult_const): New, split from...
	(expand_mult): ...here.
	(extract_high_half): New, split out from expand_mult_highpart.
	(expand_highpart_optab): Likewise.  Don't clobber target prematurely.
	(expand_highpart): Evaluate the cost of a shift/add sequence,
	then see if any of the specialized optabs are cheaper.

I still haven't identified the change that caused the linking.m
error but possibly reverting Richard's change will fix it.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[danglin at gcc dot gnu dot org]

------- Additional Comments From danglin at gcc dot gnu dot org  2004-03-21 05:47 -------
Reverting Richard's patch resolved the bootstrap failure and testsuite
regressions.

This patch from Ulrich Weigand

http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01740.html

fixed the bootstrap error but we had two new testsuite regressions:

FAIL: gcc.c-torture/execute/941101-1.c execution,  -O3 -fomit-frame-pointer
FAIL: gcc.c-torture/execute/941101-1.c execution,  -O3 -g

As these involve division by 7, it appears there must be a problem
with Ulrich's patch.


-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
                 CC|                            |rsandifo at redhat dot com


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[danglin at gcc dot gnu dot org]

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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[pinskia at gcc dot gnu dot org]

------- Additional Comments From pinskia at gcc dot gnu dot org  2004-03-21 06:03 -------
Confirming, thanks David for investigating work.

-- 
           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|UNCONFIRMED                 |NEW
     Ever Confirmed|                            |1
   Last reconfirmed|0000-00-00 00:00:00         |2004-03-21 06:03:22
               date|                            |


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[uweigand at gcc dot gnu dot org]

------- Additional Comments From uweigand at gcc dot gnu dot org  2004-03-21 17:38 -------
Could you try my updated patch at 
http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01772.html

-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[doko at cs dot tu-berlin dot de]

------- Additional Comments From doko at cs dot tu-berlin dot de  2004-03-21 18:11 -------
Subject: Re:  [3.5 regression] bootstrap failure in libobjc (first .m file)

uweigand at gcc dot gnu dot org writes:
> Could you try my updated patch at 
> http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01772.html

the bootstrap with your first build passed the libobjc build.

	Matthias


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[dave at hiauly1 dot hia dot nrc dot ca]

------- Additional Comments From dave at hiauly1 dot hia dot nrc dot ca  2004-03-21 18:24 -------
Subject: Re:  [3.5 regression] bootstrap failure in libobjc

> uweigand at gcc dot gnu dot org writes:
> > Could you try my updated patch at 
> > http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01772.html
> 
> the bootstrap with your first build passed the libobjc build.

Could you let us know if there are any testsuite regressions?

Thanks,
Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[danglin at gcc dot gnu dot org]

------- Additional Comments From danglin at gcc dot gnu dot org  2004-03-21 21:59 -------
Fixed by patch.


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


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[dave at hiauly1 dot hia dot nrc dot ca]

------- Additional Comments From dave at hiauly1 dot hia dot nrc dot ca  2004-03-22 04:55 -------
Subject: Re:  [3.5 regression] bootstrap failure in libobjc

> Could you try my updated patch at 
> http://gcc.gnu.org/ml/gcc-patches/2004-03/msg01772.html

Thanks.  Test in progress.

Dave


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663

[Bug target/14663] [3.5 regression] bootstrap failure in libobjc (first .m file)

[doko at cs dot tu-berlin dot de]

------- Additional Comments From doko at cs dot tu-berlin dot de  2004-03-22 06:02 -------
Subject: Re:  [3.5 regression] bootstrap failure in libobjc (first .m file)

dave at hiauly1 dot hia dot nrc dot ca writes:
> Could you let us know if there are any testsuite regressions?

http://gcc.gnu.org/ml/gcc-testresults/2004-03/msg01096.html


-- 


http://gcc.gnu.org/bugzilla/show_bug.cgi?id=14663