[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-24 Thread Michael Casadevall
Here's a debdiff that works around the issue. This package has been successfully built in a devirtualized PPA on armel. ** Attachment added: libipc.debdiff http://launchpadlibrarian.net/23052582/libipc.debdiff -- Shared memory operations on very fast ARM hardware suffer from non-atomic

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-24 Thread Michael Casadevall
After much debugging, the issue is a race issue. On non-armel platforms, shared memory operations appear to be atomic, but on ARM, on very fast machines, these shared memory operations have a slight lag to the, which causes the test suite failures. I'm trying to see if I can cobble together a test

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-19 Thread Michael Casadevall
** Changed in: libipc-sharelite-perl (Ubuntu Jaunty) Assignee: Oliver Grawert (ogra) = Michael Casadevall (mcasadevall) -- armel build failure (without ignoring testsuite results) https://bugs.launchpad.net/bugs/299847 You received this bug notification because you are a member of Ubuntu

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-17 Thread Michael Casadevall
Testing on rimu with lamont's help, it seems to be an alignment issue. Setting /proc/cpu/alignment to 2 (fixup) causes the following failures: # Failed test 'num_segments' # at t/sharelite.t line 40. # got: '1' # expected: '3' t/shareliteNOK 11/14

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-13 Thread Oliver Grawert
tested on the babbage board where it finishes to build with no issues at all ... i will do a test on qemu and on the porter machine in parallel, my suspicion is that the buildd/porter HW or the used kernel is at fault here -- armel build failure (without ignoring testsuite results)

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-13 Thread Oliver Grawert
same issues on the porter box: /usr/bin/make test make[1]: Entering directory `/home/ogra/perltest/libipc-sharelite-perl-0.13' PERL_DL_NONLAZY=1 /usr/bin/perl -MExtUtils::Command::MM -e test_harness(0, 'blib/lib', 'blib/arch') t/*.t t/00-load..ok 1/1# Testing IPC::ShareLite 0.13

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-13 Thread Oliver Grawert
buildlog excerpt on the babbge: /usr/bin/make test make[1]: Entering directory `/home/ogra/perltest/libipc-sharelite-perl-0.13' PERL_DL_NONLAZY=1 /usr/bin/perl -MExtUtils::Command::MM -e test_harness(0, 'blib/lib', 'blib/arch') t/*.t t/00-load..ok 1/1# Testing IPC::ShareLite 0.13

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-13 Thread Oliver Grawert
buildlog excerpt on qemu /usr/bin/make test make[1]: Entering directory `/home/ogra/perltest/libipc-sharelite-perl-0.13' PERL_DL_NONLAZY=1 /usr/bin/perl -MExtUtils::Command::MM -e test_harness(0, 'blib/lib', 'blib/arch') t/*.t t/00-load..ok 1/1# Testing IPC::ShareLite 0.13 t/00-load..ok

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-02-12 Thread Loïc Minier
** Changed in: libipc-sharelite-perl (Ubuntu Jaunty) Assignee: (unassigned) = Oliver Grawert (ogra) -- armel build failure (without ignoring testsuite results) https://bugs.launchpad.net/bugs/299847 You received this bug notification because you are a member of Ubuntu Bugs, which is

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-01-30 Thread Matthias Klose
reopened. just builds because the testsuite results are ignored on armel for the last upload. ** Summary changed: - armel build failure (package not yet in the archive) + armel build failure (without ignoring testsuite results) ** Changed in: libipc-sharelite-perl (Ubuntu Jaunty)

[Bug 299847] Re: armel build failure (without ignoring testsuite results)

2009-01-30 Thread Loïc Minier
So it failed on the buildds again. It could be that the SHM segment is already used, but that would fail more tests. I suspect it's a board specific issue in the support of SHM features. I couldn't reproduce on my evm board either. -- armel build failure (without ignoring testsuite results)