Perl 6 - the future is here, just unevenly distributed

IRC log for #perl11, 2017-03-08

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary

All times shown according to UTC.

Time Nick Message
00:36 travis-ci perl11/cperl#2373 (smoke/gh259-do-warn - b9f9d83 : Tony Cook): The build has errored. https://travis-ci.org/perl11/cperl/builds/208700577
02:57 travis-ci perl11/cperl#2374 (smoke/gh259-do-warn - 029d3e1 : Tony Cook): The build passed. https://travis-ci.org/perl11/cperl/builds/208760315
06:23 anton_p joined #perl11
09:36 travis-ci perl11/cperl#2375 (smoke/master - 2685432 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/208884440
11:03 travis-ci perl11/cperl#2376 (master - 2685432 : Reini Urban): The build was broken. https://travis-ci.org/perl11/cperl/builds/208905872
11:36 lizmat joined #perl11
12:22 travis-ci perl11/cperl#2376 (master - 2685432 : Reini Urban): The build was broken. https://travis-ci.org/perl11/cperl/builds/208905872
12:26 travis-ci RPerl build errored. Will Braswell says 'RPerl Front-End Script, Add Magic & Code Options, Part 3'
12:26 travis-ci https://travis-ci.org/wbraswell/rperl/builds/208925252 https://github.com/wbraswell/rperl/compare/e0b288606b7b...f2403c3a49cf
12:35 willthechill can anyone please help me figure out why travis has never started working again after the recent Amazon failure?
12:51 willthechill ribasushi: any ideas on why travis is still borked?
12:52 ribasushi willthechill: "borked" how? got a link to a failure?
12:53 willthechill my travis has been borked since the recent amazon failure, it started erroring on builds which had passed before the amazon failure, and now it errors on every build after trying to run for much longer than it should...
12:53 willthechill https://travis-ci.org/wbraswell/rperl/builds
12:54 willthechill scroll down to the last green check mark passing build, then look at the one right above that "Learning RPerl, Typo Fix; OO Classes Cont'd, Part 1"
12:54 willthechill that now-erroring build was passing before the amazon failure, then I clicked "restart build" when travis came back online but it failed, and all builds since then have failed
12:55 willthechill oddly, travis _is_ working for rurban's cperl builds, but I can't figure out why his is working but rperl is not
12:59 ribasushi looking...
13:01 willthechill thanks!  :-)
13:01 willthechill the RPerl build jobs should only run for about 25 mins each, instead they are running for 2 - 3 times that long before erroring
13:05 ribasushi willthechill: it's not aws related at all just a coincidence
13:06 ribasushi in the working part you have: http://paste.scsys.co.uk/557413
13:07 ribasushi in the breaking part you have: http://paste.scsys.co.uk/557414
13:07 willthechill checking
13:07 ribasushi from there on things diverge, and it *looks* like Test2::Suite just hangs on one of the tests, and then the job times out
13:08 ribasushi willthechill: I would start by trying to `cpanm Test2::Suite` before the hanging `cpanm --installdeps .`
13:08 ribasushi and pushing a job to see whether this is indeed the problem
13:08 willthechill okay so it is because Alien::base upgraded and now requires Test2::Suite
13:08 willthechill *sigh* yay for other people breaking stuff
13:08 willthechill :-P
13:08 ribasushi *looks* like that
13:09 ribasushi there is a lot of output, and I only looked briefly
13:09 ribasushi it may be a red herring
13:09 ribasushi i.e. something inbetween took way too long, and then by the time we got to test2 things timed out
13:09 willthechill will modify travis config now and try again
13:09 willthechill just put `cpanm Test2::Suite` ?  not `cpanm -n Test2::Suite` or something?
13:10 ribasushi no -n: you want to run the tests
13:10 ribasushi ( as they are what hangs )
13:10 ribasushi willthechill: in all logs the last line is
13:10 ribasushi t/modules/Tools/Compare.t ....................... 7/?
13:11 ribasushi so 99% certainty that it is Test2::Suite that is the problem
13:11 willthechill okay cool, trying now
13:17 willthechill ribasushi: okay now we see what happens!  :-)   https://travis-ci.org/wbraswell/rperl/builds/208965769
13:18 ribasushi willthechill: you want --verbose
13:18 ribasushi sorry I wasn't explicit, I assumed you would copy what you did below
13:19 ribasushi as it is now if it hangs you won't see anything
13:19 willthechill OH okay, changing now
13:20 willthechill so what I'm supposed to add to .travis.yml is this?   `cpanm --verbose Test2::Suite`
13:20 ribasushi yessir
13:26 willthechill okay great!  :-)
13:26 willthechill changing now
13:27 willthechill ribasushi: take 2   https://travis-ci.org/wbraswell/rperl/builds/208968754
13:34 ribasushi willthechill: and there you have it: it hangs
13:35 willthechill okay so what does that teach us?
13:36 ribasushi willthechill: actually no - it continued, hang on
13:36 willthechill okay brb
13:40 anton joined #perl11
13:43 ribasushi willthechill: so the problem is that T2::S' tests are shoddy - it's been spinning for 10 minutes already doing something not terribly complicated
13:44 ribasushi willthechill: it does advance *very* slowly, but you have an upper bound of 50 minutes, and this causes you to hit it
13:44 ribasushi willthechill: file a bug and add as a workaround `cpanm -n Test2::Suite` to keep going
13:44 ribasushi willthechill: welcome to CPAN :/
13:51 travis-ci perl11/cperl#2376 (master - 2685432 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/208905872
14:13 travis-ci perl11/cperl#2376 (master - 2685432 : Reini Urban): The build passed. https://travis-ci.org/perl11/cperl/builds/208905872
14:14 willthechill ribasushi: okay thank you for helping figure it out!  :-)
14:15 ribasushi willthechill: by the way I was wrong - it does eventually hang: it never advanced past test 7
14:15 ribasushi willthechill: I think someone else is digging into it, but not sure: in any case you should file a bug
14:15 willthechill someone else?
14:15 ribasushi this is a rather big problem
14:16 willthechill okay I kinda thought it was a big problem
14:16 willthechill I just mentioned it to jberger & plicease on #native
14:16 willthechill they own Alien
14:17 willthechill so now I do not know how to proceed?
14:18 willthechill I guess I will start with the `cpanm -n Test2::Suite` workaround?
14:18 ribasushi willthechill: that will solve it for you, yes: you already have this for cppbuilder higher up
14:20 willthechill new build with workaround   https://travis-ci.org/wbraswell/rperl/builds/208986206
14:23 travis-ci Parse::Yapp build passed. Will Braswell says 'Francois Desarmenien Copyright Assignment Doc & E-Mail'
14:23 travis-ci https://travis-ci.org/wbraswell/parse-yapp/builds/208971143 https://github.com/wbraswell/parse-yapp/compare/10f0ce06f9b0...b16177cad8ce
14:24 willthechill wow I didn't even know I had travis enabled for that repo!  :-P
14:24 willthechill (Parse::Yapp is not related to this current Test2::Suite problem)
14:36 willthechill ribasushi: the workaround seems to work, so that's one tiny step forward at least!  :-)
14:37 willthechill now I'm going to try to do a manual build of Test2::Suite, per haarg's recommendation in #toolchain
14:39 willthechill I canceled the workaround build, it was in the middle of the RPerl tests, so at least it made it that far
15:29 willthechill ribasushi: haarg says that Test2::Suite buffers its own output, so it seems to me like it is not hanging but is just taking too long?
15:34 willthechill ribasushi: it is not hanging, it is just taking too long    https://s3.amazonaws.com/archive.travis-ci.org/jobs/208992405/log.txt
15:55 rurban joined #perl11
16:03 rurban I'm also very suspicious of all those 777 intermixed with term sequences in STDOUT/ERR. Looks like Term::Table to me
16:04 rurban Anyway. I banned Test2 from cperl until they can provide a proper Test-Simple replacement. Using Test2 is ok, but overwriting the existing test framework with a not-working 20% slower variant is not acceptable.
16:04 rurban And even more people who enforce Test2 onto me.
16:19 travis-ci RPerl build errored. Will Braswell says 'Travis-CI, Force Install Of Test2::Suite As Required By Alien::base v0.036, Part 4'
16:19 travis-ci https://travis-ci.org/wbraswell/rperl/builds/208992404 https://github.com/wbraswell/rperl/compare/5b5295a59bfd...d3e948335f7f
16:27 rurban will: they (toolchainers) are all installing their deps without testing. so it should be fair to do that with rperl also
16:28 willthechill rurban: yes I understand that point-of-view, but I do not agree with it, ribasushi and I both believe it is best to run rigorous tests in order to catch bugs, because if we don't catch the bugs then one of our downstream users will!
16:29 rurban hmm, but this is costing us weeks
16:31 willthechill it will cost us months if we just wait for somebody else to find the problem and report at a later date
16:31 willthechill bugs in CPAN hurt everyone
16:31 willthechill and they directly hurt RPerl
16:31 willthechill anyway, it is not really costing much, I only just started to debug the issue a few hours ago
16:31 rurban Locally everything works fine
16:31 willthechill and it IS a bug of some kind, even if just a travis bug
16:32 rurban agreed. but not a rperl bug
16:32 willthechill agreed
16:33 willthechill also, originally I thought this was due to the recent Amazon failure, but riba says it was just a timing coincidence, and it was only due to T2::S not Amazon
16:43 rurban that's my docker attempt: https://travis-ci.org/perl11/rperl/jobs/207054137
16:45 ribasushi note that the docker env is underpowered, it just spins up faster
16:45 ribasushi so for something like rperl it may actually make building take more time than the saving of the startup
16:47 willthechill rurban: RPerl needs sudo and update-alternatives, won't work with docker
18:59 rurban I'm not convinced. pluto can be at ~/ and g++ is fine
19:55 rurban1 joined #perl11
20:15 rurban2 joined #perl11
20:59 travis-ci perl11/cperl#2377 (smoke/master - ba11e57 : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/209065669
23:57 travis-ci perl11/cperl#2378 (smoke/master - 1a855a9 : Reini Urban): The build has errored. https://travis-ci.org/perl11/cperl/builds/209078265

| Channels | #perl11 index | Today | | Search | Google Search | Plain-Text | summary