Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
> > Disagree on almost all points but I find debates about bug severities > > so utterly demotivating I will defer. > > I'm sorry to hear that. I certainly didn't want to demotivate you. > On the contrary, I very much appreciate your work on Debian. Many thanks > for that! Oh, you weren't being demotivating, I was just explaining why I wouldn't be entering the discussion. :) > FWIW I'd have fixed this bug last night by marking the failing test as > TODO Neat, and thanks. Kindest regards, -- ,''`. : :' : Chris Lamb `. `'` la...@debian.org / chris-lamb.co.uk `-
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
On Mon, Dec 21, 2015 at 08:35:29PM +0100, gregor herrmann wrote: > On Mon, 21 Dec 2015 20:49:50 +0200, Niko Tyni wrote: > > > While I'm not going to start a severity war (and agree that the package > > should be fixed), this is not the traditional interpretation. FTBFS bugs > > have been routinely downgraded to 'important' in the past when the build > > failures were nondeterministic and the build succeeded part of the time. > > That's my understanding as well. > Just yesterday I downgraded a bug of this type to important after > consulting with a member of the release team. I agree with Niko and gregoa. I don't see a practical benefit of making such bugs RC in the general case. I'm sure there are circumstances where it would be justified, but based on my experience with perl and some of the more esoteric architectures, I think such a policy would do more harm than good. Dominic.
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
Disagree on almost all points but I find debates about bug severities so utterly demotivating I will defer. -lamby On Mon, 21 Dec 2015, at 08:19 PM, Dominic Hargreaves wrote: > On Mon, Dec 21, 2015 at 08:35:29PM +0100, gregor herrmann wrote: > > On Mon, 21 Dec 2015 20:49:50 +0200, Niko Tyni wrote: > > > > > While I'm not going to start a severity war (and agree that the package > > > should be fixed), this is not the traditional interpretation. FTBFS bugs > > > have been routinely downgraded to 'important' in the past when the build > > > failures were nondeterministic and the build succeeded part of the time. > > > > That's my understanding as well. > > Just yesterday I downgraded a bug of this type to important after > > consulting with a member of the release team. > > I agree with Niko and gregoa. I don't see a practical benefit of making > such bugs RC in the general case. I'm sure there are circumstances > where it would be justified, but based on my experience with perl and > some of the more esoteric architectures, I think such a policy would > do more harm than good. > > Dominic. Regards, -- ,''`. : :' : Chris Lamb `. `'` la...@debian.org / chris-lamb.co.uk `-
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
severity 799718 serious thanks > This package failed to build in the reproducible.debian.net CI setup, > and I can reproduce this locally by running the test in a loop while > loading the host. Apparently there's a race condition in the test suite. Given that it makes the package unreliably build from source, this warrants an RC severity. Regards, -- ,''`. : :' : Chris Lamb `. `'` la...@debian.org / chris-lamb.co.uk `-
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
On Mon, 21 Dec 2015 20:49:50 +0200, Niko Tyni wrote: > While I'm not going to start a severity war (and agree that the package > should be fixed), this is not the traditional interpretation. FTBFS bugs > have been routinely downgraded to 'important' in the past when the build > failures were nondeterministic and the build succeeded part of the time. That's my understanding as well. Just yesterday I downgraded a bug of this type to important after consulting with a member of the release team. Cheers, gregor -- .''`. Homepage: http://info.comodo.priv.at/ - OpenPGP key 0xBB3A68018649AA06 : :' : Debian GNU/Linux user, admin, and developer - https://www.debian.org/ `. `' Member of VIBE!AT & SPI, fellow of the Free Software Foundation Europe `- NP: R.E.M.: Bang and Blame signature.asc Description: Digital Signature
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
On Mon, Dec 21, 2015 at 06:35:33PM +, Chris Lamb wrote: > severity 799718 serious > thanks > > > This package failed to build in the reproducible.debian.net CI setup, > > and I can reproduce this locally by running the test in a loop while > > loading the host. Apparently there's a race condition in the test suite. > > Given that it makes the package unreliably build from source, this warrants > an RC severity. While I'm not going to start a severity war (and agree that the package should be fixed), this is not the traditional interpretation. FTBFS bugs have been routinely downgraded to 'important' in the past when the build failures were nondeterministic and the build succeeded part of the time. My understanding is that the baseline requirement from the release team is that 'buildds need to be able to build the packages'. It may be that the awesome continuous integration that reproducible.d.n is doing has changed this picture and we should start treating all such bugs as release critical, but discussing that first with the release team and/or on the debian-devel list would be appropriate IMHO. -- Niko Tyni nt...@debian.org
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
On Mon, Dec 21, 2015 at 09:05:43PM +, Chris Lamb wrote: > Disagree on almost all points but I find debates about bug severities > so utterly demotivating I will defer. I'm sorry to hear that. I certainly didn't want to demotivate you. On the contrary, I very much appreciate your work on Debian. Many thanks for that! FWIW I'd have fixed this bug last night by marking the failing test as TODO, but was caught out by #808560 which took priority by preventing me to upload :) -- Niko
Bug#799718: libthread-pool-perl: nondeterministic test failure in t/Pool01.t
Package: libthread-pool-perl Version: 0.33-1 Severity: important This package failed to build in the reproducible.debian.net CI setup, and I can reproduce this locally by running the test in a loop while loading the host. Apparently there's a race condition in the test suite. t/Pool01.t .. 1..42 ok 1 - use Thread::Pool; # Test general functionality ok 2 - 'check object type' isa 'Thread::Pool' ok 3 - Thread::Pool->can(...) ok 4 - check number of workers ok 5 - check \# jobs todo, \#1 ok 6 - check number of workers, \#1 ok 7 - check first jobid ok 8 - check second jobid ok 9 - check tid of 2nd worker thread ok 10 - check number of workers, \#2 ok 11 - check number of workers, \#3 ok 12 - check number of workers, \#4 ok 13 - check number of removed, \#1 not ok 14 - check \# jobs todo, \#2 # Failed test 'check \# jobs todo, \#2' # at t/Pool01.t line 77. ok 15 - check result_dontwait ok 16 - check result ok 17 - check third jobid ok 18 - check result remove ok 19 - check number of workers, \#5 ok 20 - check number of removed, \#2 ok 21 - check for remaining threads ok 22 - check number of workers, \#6 ok 23 - check number of removed, \#3 ok 24 - check \# jobs todo, \#3 ok 25 - check \# jobs done, \#3 ok 26 - check not-used threads, \#1 ok 27 - 'check object type' isa 'Thread::Pool' ok 28 - check fourth jobid ok 29 - check number of workers, \#7 ok 30 - check result after add ok 31 - check result waitfor ok 32 - check fifth jobid ok 33 - check result after add ok 34 - check whether job id found ok ok 35 - check sixth jobid ok 36 - check result remove_me ok 37 - check \# jobs todo, \#4 ok 38 - check \# jobs done, \#4 ok 39 - check number of workers, \#7 ok 40 - check number of removed, \#4 ok 41 - check for remaining threads ok 42 - check not-used threads, \#2 # Looks like you failed 1 test of 42. Dubious, test returned 1 (wstat 256, 0x100) Failed 1/42 subtests -- Niko Tyni nt...@debian.org