Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
On 2018-03-30 4:56 AM, Eugene V. Lyubimkin wrote: Thank you for taking a look. On 30.03.2018 01:01, John David Anglin wrote: [...] On hurd-i386, I tracked it down to a misbehaviour of timeout(1), reported as #894379. For HPPA, I didn't see any available porterboxes for HPPA on db.debian.org, so I could not reproduced it. Did I miss any? I'll take a look at the timeout problem. Access to a box can be arranged. I find timeout works correctly on hppa. I see, so it's a different issue then. Please let me know if/when there is a box I could reproduce the test case failure on. The attached change fixed the build failure on my rp3440. Don't know if 60s is enough for the slower buildds. Dave -- John David Anglin dave.ang...@bell.net --- ./test/t/manage/worker/packages/hooks/pre-install-pkgs/feeding-input.t.save 2018-03-30 11:11:02.697750034 -0400 +++ ./test/t/manage/worker/packages/hooks/pre-install-pkgs/feeding-input.t 2018-03-30 11:11:41.437969834 -0400 @@ -21,7 +21,7 @@ my ($comment, $hook_command, $checker) = @_; my $hook_options = "-o dpkg::pre-install-pkgs::='$hook_command' "; - my $command = get_worker_command("timeout 30s $cupt", "install '*xyz*' $hook_options", 'simulate'=>0); + my $command = get_worker_command("timeout 60s $cupt", "install '*xyz*' $hook_options", 'simulate'=>0); my $output; subtest $comment => sub {
Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
Hi, Thank you for taking a look. On 30.03.2018 01:01, John David Anglin wrote: [...] >>> On hurd-i386, I tracked it down to a misbehaviour of timeout(1), reported >>> as #894379. >>> For HPPA, I didn't see any available porterboxes for HPPA on db.debian.org, >>> so I could not >>> reproduced it. Did I miss any? >> I'll take a look at the timeout problem. Access to a box can be arranged. > I find timeout works correctly on hppa. I see, so it's a different issue then. Please let me know if/when there is a box I could reproduce the test case failure on. Regards, -- Eugene V. Lyubimkin aka JackYF C++ GNU/Linux userspace developer, Debian Developer
Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
Control: reassign 882238 cupt Control: reassign 894379 libc0.3 Control: tags 894379 + pending Samuel Thibault, on ven. 30 mars 2018 02:23:04 +0200, wrote: > Control: reassign -1 libc0.3 Ah sorry, fixing this. Samuel
Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
Control: reassign -1 libc0.3 Hello, John David Anglin, on jeu. 29 mars 2018 13:54:23 -0400, wrote: > On 2018-03-29 11:53 AM, Eugene V. Lyubimkin wrote: > > On hurd-i386, I tracked it down to a misbehaviour of timeout(1), reported > > as #894379. > > For HPPA, I didn't see any available porterboxes for HPPA on db.debian.org, > > so I could not > > reproduced it. Did I miss any? > I'll take a look at the timeout problem. Access to a box can be arranged. On hurd-i386, it is an unfortunate confluence of borderline behaviors. Basically, what happens is that the SIGCHLD signal is catched by the thread started by timer_create(). I'll give a try at masking signals so main() does catch the signal. Samuel
Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
On 2018-03-29 1:54 PM, John David Anglin wrote: On 2018-03-29 11:53 AM, Eugene V. Lyubimkin wrote: On hurd-i386, I tracked it down to a misbehaviour of timeout(1), reported as #894379. For HPPA, I didn't see any available porterboxes for HPPA on db.debian.org, so I could not reproduced it. Did I miss any? I'll take a look at the timeout problem. Access to a box can be arranged. I find timeout works correctly on hppa. Dave -- John David Anglin dave.ang...@bell.net
Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
On 2018-03-29 11:53 AM, Eugene V. Lyubimkin wrote: On hurd-i386, I tracked it down to a misbehaviour of timeout(1), reported as #894379. For HPPA, I didn't see any available porterboxes for HPPA on db.debian.org, so I could not reproduced it. Did I miss any? I'll take a look at the timeout problem. Access to a box can be arranged. Dave -- John David Anglin dave.ang...@bell.net
Bug#882238: [Cupt-devel] Bug#882238: cupt: FTBFS on hppa and hurd-i386: feeding-input.t fails
Hi Aaron, Hurd and HPPA porters, Thank you for the report. Sorry that it took so long to get it to it. On hurd-i386, I tracked it down to a misbehaviour of timeout(1), reported as #894379. For HPPA, I didn't see any available porterboxes for HPPA on db.debian.org, so I could not reproduced it. Did I miss any? Otherwise, if you know simple-to-use substitutes for timeout(1) or know what's going on in #894379, your input is welcome. Regards, -- Eugene V. Lyubimkin aka JackYF C++ GNU/Linux userspace developer, Debian Developer