Bug#587463: libforks-perl: FTBFS on armel: Failed test 'Check if thread 1 is deadlocked'

[Nobuhiro Iwamatsu]

Hi, Niko.

2010/9/4 Niko Tyni <ntyni at debian.org>:
> On Mon, Jun 28, 2010 at 10:34:10PM +0200, Salvatore Bonaccorso wrote:
>> Package: libforks-perl
>> Version: 0.34-1
>> Severity: serious
>> Justification: FTBFS
>>
>> This package is not migrating to testint as it fails to build on armel
>> architecture [1].
>
> The failure may well be because of #595397 / #595403 : POSIX::sigaction
> seems to be somehow broken on armel, and as a result libsys-sigaction-perl
> (a dependency of libforks-perl) fails its own test suite.
>
> As I said in #595397, one possible outcome is that we have to remove
> libsys-sigaction-perl/armel and therefore also libforks-perl/armel,
> indirectly solving this bug too :)
>
> Nobuhiro: does the libsys-sigaction-perl test suite pass for you on sh4?

Yes, I tested libsys-sigaction-perl on sh4.
Test was clear.I attached buildlog of sh4.

-----
Looks good
'CREATE_PACKLIST' is not a known MakeMaker parameter name.
Writing Makefile for Sys::SigAction
   dh_auto_build
make[1]: Entering directory
`/home/iwamatsu/build-area/libsys-sigaction-perl/libsys-sigaction-perl-0.11'
cp lib/Sys/SigAction.pm blib/lib/Sys/SigAction.pm
Manifying blib/man3/Sys::SigAction.3pm
make[1]: Leaving directory
`/home/iwamatsu/build-area/libsys-sigaction-perl/libsys-sigaction-perl-0.11'
   dh_auto_test
make[1]: Entering directory
`/home/iwamatsu/build-area/libsys-sigaction-perl/libsys-sigaction-perl-0.11'
PERL_DL_NONLAZY=1 /usr/bin/perl "-MExtUtils::Command::MM" "-e"
"test_harness(0, 'blib/lib', 'blib/arch')" t/*.t
t/mask.t ..... ok
t/name.t ..... ok
t/nested.t ... ok
t/number.t ... ok


         NOTE: Setting safe=>1... with masked signals does not seem to work.
         The problem is that the masked signals are not masked, but when
         safe=>0 they are.

         If you have an application for safe=>1 and can come up with
         a test that works in the context of this module's installation
         please send me a patch to safe.t that tests it.

         See the block below this one... which if executed would test safe mode
         with masked signals... it is a clone of part of mask.t that proves this
         is broken.

         Lincoln

t/safe.t ..... ok
t/timeout.t .. ok
All tests successful.
Files=6, Tests=32, 19 wallclock secs ( 1.71 usr  0.13 sys + 12.81 cusr
 0.19 csys = 14.84 CPU)
Result: PASS
make[1]: Leaving directory
`/home/iwamatsu/build-area/libsys-sigaction-perl/libsys-sigaction-perl-0.11'
-----

Hmm, The problem of sh4 may be different from armel in a reason.

Best regards,
  Nobuhiro
-- 
Nobuhiro Iwamatsu
   iwamatsu at {nigauri.org / debian.org}
   GPG ID: 40AD1FA6
-------------- next part --------------
A non-text attachment was scrubbed...
Name: libsys-sigaction-perl_0.11-1_sh4.build.gz
Type: application/x-gzip
Size: 1969 bytes
Desc: not available
URL: <http://lists.alioth.debian.org/pipermail/pkg-perl-maintainers/attachments/20100904/8063e698/attachment.bin>