bug-libtool
[Top][All Lists]
Advanced

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: Re: libtool-2.2.11a on AIX 5.3 current git master 2010-08-25 /status


From: Rainer Tammer
Subject: Re: Re: libtool-2.2.11a on AIX 5.3 current git master 2010-08-25 /status
Date: Sat, 18 Sep 2010 10:37:44 +0200
User-agent: Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.8.1.3) Gecko/20070326 Thunderbird/2.0.0.0 Mnenhy/0.7.5.0

 Hello,
I just did a bisect and unfortunately the problem did not show up.
After that I tested the same thing with a fresh check out and the
problem is back.

This is how I run my test:

export CC=cc_r
export CXX=xlC_r
./configure
gmake
gmake check  TESTSUITEFLAGS=-V TESTS='$(F77_TESTS) $(FC_TESTS)'

...
gmake[2]: Entering directory `/daten/source/libtool-2.2.11a'
gmake  check-TESTS check-local
gmake[3]: Entering directory `/daten/source/libtool-2.2.11a'
gmake[4]: Entering directory `/daten/source/libtool-2.2.11a'
PASS: tests/f77demo-static.test
PASS: tests/f77demo-static-make.test
PASS: tests/f77demo-static-exec.test
PASS: tests/f77demo-conf.test
PASS: tests/f77demo-make.test
FAIL: tests/f77demo-exec.test
PASS: tests/f77demo-shared.test
PASS: tests/f77demo-shared-make.test
FAIL: tests/f77demo-shared-exec.test
PASS: tests/fcdemo-static.test
PASS: tests/fcdemo-static-make.test
PASS: tests/fcdemo-static-exec.test
PASS: tests/fcdemo-conf.test
PASS: tests/fcdemo-make.test
FAIL: tests/fcdemo-exec.test
PASS: tests/fcdemo-shared.test
PASS: tests/fcdemo-shared-make.test
FAIL: tests/fcdemo-shared-exec.test
====================================
4 of 18 tests failed
See ./test-suite.log
Please report to address@hidden
====================================

The CC/CXX  env vars do not change the outcome.

>From my run:

# dump -vn lt-fprogram  | head -n 20

lt-fprogram:

                        ***Loader Section***
                      Loader Header Information
VERSION#         #SYMtableENT     #RELOCent        LENidSTR
0x00000001       0x0000001c       0x0000004b       0x000000fe

#IMPfilID        OFFidSTR         LENstrTBL        OFFstrTBL
0x00000006       0x00000644       0x00000136       0x00000742


                        ***Import File Strings***
INDEX  PATH                          BASE                MEMBER
0     
/daten/source/libtool-2.2.11a/tests/f77demo/.libs:/daten/source/libtool-2.2.11a/_inst-f77demo/lib:/usr/vac/lib:/usr/lib/threads:/usr/lib:/lib
1                                    libc.a              shr.o
2                                    libgfortran.a       libgfortran.so.2
3                                    libfoo.a            libfoo.so.0
4                                    libfoo3.a           libfoo3.so.0
5                                    libgcc_s.a          shr.o


Fom the bisect run:

# dump -vn lt-fprogram  | head -n 20

lt-fprogram:

                        ***Loader Section***
                      Loader Header Information
VERSION#         #SYMtableENT     #RELOCent        LENidSTR
0x00000001       0x0000001c       0x0000004b       0x0000016b

#IMPfilID        OFFidSTR         LENstrTBL        OFFstrTBL
0x00000006       0x00000644       0x00000136       0x000007af


                        ***Import File Strings***
INDEX  PATH                          BASE                MEMBER
0     
/daten/source/libtool-master.bisect/libtool/tests/f77demo/.libs:/daten/source/libtool-master.bisect/libtool/_inst-f77demo/lib:/opt/freeware/lib/gcc/powerpc-ibm-aix5.3.0.0/4.2.4:/opt/freeware/lib/gcc/powerpc-ibm-aix5.3.0.0/4.2.4/../../..:/usr/lib:/lib
1                                    libc.a              shr.o
2                                    libgfortran.a       libgfortran.so.2
3                                    libfoo.a            libfoo.so.0
4                                    libfoo3.a           libfoo3.so.0
5                                    libgcc_s.a          shr.o


I have absolutely no idea why this is different...

Ralf: Please could you provide me a already bootstrapped libtool master ??

Bye
  Rainer

On 20:59, Ralf Wildenhues wrote:
> Hello Rainer,
>
> * Rainer Tammer wrote on Wed, Aug 25, 2010 at 10:26:06AM CEST:
>> this is the current status of libtool (git master, 2010-08-25) on AIX 5.3:
>>
>> Old test suite:
>>
>> FAIL: tests/f77demo-exec.test
>> FAIL: tests/f77demo-exec.test
>> FAIL: tests/fcdemo-exec.test
>> FAIL: tests/fcdemo-exec.test
>>
>> -> path in exec missing (path to the fortran libs)
> I'm sorry but I have to ask for your help on this part, because on the
> AIX 5.3 system I can test on, these tests pass for me with gfortran 4.2.
> Specifically, can I ask you to try a 'git bisect' on this issue in order
> to find out which commit introduced the breakage?  Do you know how git
> bisect works?  If not, looking at the git-bisect(1) manpage quickly
> can't hurt.
>
> Below is a recipe that I'm writing from memory now but that I haven't
> tested much, so you might have to fix it.  The 'bisect start' command
> gets listed a bad and a good commit, I'm assuming from your earlier
> mails that v2.2.8 was a good commit?
>
> You might want to override 'export MAKE=make' depending on your setup,
> GNU make is definitely preferable for this testing.
>
> Please send all output from that, preferably gzip'ed.
>
> In addition to that, please also run
>   gmake check  TESTSUITEFLAGS=-V TESTS='$(F77_TESTS) $(FC_TESTS)'
>
> with git master checked out, and send the test-suite.log file as well as
> all tests/f*demo*.log files.
>
> Thanks,
> Ralf
>
> git clone git://git.savannah.gnu.org/libtool.git
> cd libtool
>
> cat > bisect-script <<'EOF'
> #! /bin/sh
> set -x
>
> : ${MAKE=gmake}
>
> reconfdirs='. libltdl tests/f77demo tests/fcdemo' ./bootstrap || exit 125
> ./configure || exit 125
> $MAKE || exit 125
> (
>   $MAKE check TESTSUITEFLAGS=-V TESTS='$(F77_TESTS) $(FC_TESTS)'
>   echo $?
> ) 2>&1 | tee checklog || exit 125
>
> if sed -n '$p' checklog | grep 0; then
>   exit 0
> else
>   if grep 'FAIL: f[c7]*demo-exec' checklog; then
>     exit 1
>   else
>     exit 125
>   fi
> fi
> EOF
>
> chmod +x ./bisect-script
>
> git bisect start master v2.2.8
> git bisect run ./bisect-script
>
>




reply via email to

[Prev in Thread] Current Thread [Next in Thread]