Bug#587764: libept: FTBFS on ia64, sparc, mipsen

2010-07-10 Thread Adam D. Barratt
On Tue, July 6, 2010 18:50, Enrico Zini wrote:
> On Tue, Jul 06, 2010 at 04:52:39PM +0100, Julien Cristau wrote:
>
> [...]
>> > /build/buildd-libept_1.0.1-ia64-wbAihS/libept-1.0.1/ept/apt/apt.test.h:
>> 117: assertion `record.find("Package: sp") != string::npos' failed;
>> > (6/10) Apt: ..--> FAILED: (6/13) rawRecord (caught signal 6)
> [...]
>> I've been unable to reproduce the failures locally, but the package was
>> retried on some arches and failed again.  I'm not really comfortable
>> patching the tests without understanding why they failed or at least
>> reproducing the failure...
>
> I can look into it on the 13th of July.
>
> If someone would like to look at it before the 13th while I'm away,
> adding "cerr << record" before line 117 would show what did apt find
> instead of sp.

So far, I also haven't managed to reproduce the failure in order to debug it.

The build failed twice on mips, on corelli and lucatelli; however, it
succeeded on the porter box, gabrielli, which is the same hardware as the
buildds on which it failed. I tried builds using -j[1248] and
DEB_BUILD_OPTIONS=parallel=4 and all succeeded.

Adam




-- 
To UNSUBSCRIBE, email to debian-bugs-rc-requ...@lists.debian.org
with a subject of "unsubscribe". Trouble? Contact listmas...@lists.debian.org



Bug#587764: libept: FTBFS on ia64, sparc, mipsen

2010-07-06 Thread Enrico Zini
On Tue, Jul 06, 2010 at 04:52:39PM +0100, Julien Cristau wrote:

[...]
> > /build/buildd-libept_1.0.1-ia64-wbAihS/libept-1.0.1/ept/apt/apt.test.h: 
> > 117: assertion `record.find("Package: sp") != string::npos' failed;
> > (6/10) Apt: ..--> FAILED: (6/13) rawRecord (caught signal 6)
[...]
> I've been unable to reproduce the failures locally, but the package was
> retried on some arches and failed again.  I'm not really comfortable
> patching the tests without understanding why they failed or at least
> reproducing the failure...

I can look into it on the 13th of July.

If someone would like to look at it before the 13th while I'm away,
adding "cerr << record" before line 117 would show what did apt find
instead of sp.

It could be possible to test the libept test environment using apt only.
Currently, things are set up with these commands:

  mkdir -p test-env/etc test-env/state/lists/partial
   test-env/cache test-env/debtags
   test-env/cache/archives/partial test-env/desktop
  sed -e s,i386,${ARCH}, < ${datadir}/packagelist > 
test-env/state/lists/${listfile}
  cp -a ${datadir}/etc/sources.list test-env/etc/
  sed -e s,i386,${ARCH}, < ${datadir}/dpkg-status > test-env/dpkg-status
  cp -a ${datadir}/desktop/*.desktop test-env/desktop/
  cp ${datadir}/debtags/package-tags test-env/debtags/test.tag
  cp ${datadir}/debtags/vocabulary test-env/debtags/test.voc
  mkdir -p test-env/debtags/empty
  mkdir -p test-env/debtags/user
  mkdir -p test-env/xapian/
  mkdir -p test-env/popcon/
  mkdir -p test-env/popcon/empty
  cp -a ${datadir}/popcon/all-popcon-results.txt.gz test-env/popcon/
  cp -a ${datadir}/popcon/popularity-contest test-env/popcon/

And these custom apt settings:

  _config->Set("Initialized", 1);
  _config->Set("Dir", TEST_ENV_DIR);
  _config->Set("Dir::Cache", "cache");
  _config->Set("Dir::State", "state");
  _config->Set("Dir::Etc", "etc");
  _config->Set("Dir::State::status", TEST_ENV_DIR "dpkg-status");

So with a bit of fiddling and several -o options passed to apt, it
should be possible to run "apt-get update" and "apt-cache show sp" to
see if the bug is reproduced with libapt alone.


Ciao,

Enrico

-- 
GPG key: 4096R/E7AD5568 2009-05-08 Enrico Zini 


signature.asc
Description: Digital signature


Bug#587764: libept: FTBFS on ia64, sparc, mipsen

2010-07-06 Thread Julien Cristau
On Thu, Jul  1, 2010 at 14:41:00 +0100, Julien Cristau wrote:

> libept FTBFS on various archs:
> https://buildd.debian.org/status/package.php?p=libept
> 
> Build logs end with:
> 
> (1/10) Expression: . 1/1 ok
> (2/10) AptPackagerecord: . 1/1 ok
> (3/10) AptVersion: .. 6/6 ok
> (4/10) Vocabulary: ... 19/19 ok
> (5/10) Popcon: . 5/5 ok
> /build/buildd-libept_1.0.1-ia64-wbAihS/libept-1.0.1/ept/apt/apt.test.h: 117: 
> assertion `record.find("Package: sp") != string::npos' failed;
> (6/10) Apt: ..--> FAILED: (6/13) rawRecord (caught signal 6)
> (6/10) Apt: .--> FAILED: (8/13) recordIteration (caught signal 11)
> (6/10) Apt: --> FAILED: (9/13) recordIteration2 (caught signal 11)
> (6/10) Apt: .--> FAILED: (11/13) stlRecordIteration (caught signal 11)
> (6/10) Apt: . 9/13 ok
> (7/10) Debtags: . 5/5 ok
> (8/10) AptRecordparser:  12/12 ok
> (9/10) PopconLocal: . 1/1 ok
> (10/10) Axi: . 1/1 ok
> overall 60/64 ok
> 
I've been unable to reproduce the failures locally, but the package was
retried on some arches and failed again.  I'm not really comfortable
patching the tests without understanding why they failed or at least
reproducing the failure...

Cheers,
Julien


signature.asc
Description: Digital signature


Bug#587764: libept: FTBFS on ia64, sparc, mipsen

2010-07-01 Thread Julien Cristau
Package: libept
Version: 1.0.1
Severity: serious

Hi Enrico,

libept FTBFS on various archs:
https://buildd.debian.org/status/package.php?p=libept

Build logs end with:

(1/10) Expression: . 1/1 ok
(2/10) AptPackagerecord: . 1/1 ok
(3/10) AptVersion: .. 6/6 ok
(4/10) Vocabulary: ... 19/19 ok
(5/10) Popcon: . 5/5 ok
/build/buildd-libept_1.0.1-ia64-wbAihS/libept-1.0.1/ept/apt/apt.test.h: 117: 
assertion `record.find("Package: sp") != string::npos' failed;
(6/10) Apt: ..--> FAILED: (6/13) rawRecord (caught signal 6)
(6/10) Apt: .--> FAILED: (8/13) recordIteration (caught signal 11)
(6/10) Apt: --> FAILED: (9/13) recordIteration2 (caught signal 11)
(6/10) Apt: .--> FAILED: (11/13) stlRecordIteration (caught signal 11)
(6/10) Apt: . 9/13 ok
(7/10) Debtags: . 5/5 ok
(8/10) AptRecordparser:  12/12 ok
(9/10) PopconLocal: . 1/1 ok
(10/10) Axi: . 1/1 ok
overall 60/64 ok

Also in the future please try to coordinate SONAME bumps with
debian-release to minimize impact on other transitions.  E.g. there are
scheduled apt and xapian transitions for squeeze, which this bug will
block.

Cheers,
Julien

-- System Information:
Debian Release: squeeze/sid
  APT prefers unstable
  APT policy: (500, 'unstable'), (500, 'testing'), (500, 'stable'), (101, 
'experimental')
Architecture: amd64 (x86_64)

Kernel: Linux 2.6.32-5-amd64 (SMP w/2 CPU cores)
Locale: LANG=C, LC_CTYPE=en_US.UTF-8 (charmap=UTF-8)
Shell: /bin/sh linked to /bin/bash


signature.asc
Description: Digital signature