[01:03:01] <csteipp>	 Hey AaronSchulz / TimStarling, join Brad and I in the oauth channel when you get the chance
[08:08:14] <gerrit-wm>	 New patchset: Zfilipin; "Fixed link to Commons" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58733
[08:39:10] <gerrit-wm>	 New patchset: Zfilipin; "Updated Ruby gems" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58862
[09:38:37] <gerrit-wm>	 New patchset: Hashar; "(bug 46723) uses ZUUL_COMMIT as a git refspec to build" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/58865
[09:40:32] <hashar>	 qchris: hello :-]  I found a bug in the bugzilla gerrit plugin \O/
[09:40:42] <qchris>	 Oh ...
[09:40:50] <hashar>	 which component are you using in bugzilla?
[09:40:51] <qchris>	 Already in bz?
[09:41:01] <hashar>	 na about to fill it
[09:41:17] <qchris>	 Not sure ... "git/gerrit"? (No idea honestly)
[09:41:22] <hashar>	 yeah will do that
[09:41:46] <qchris>	 Hopefully it's not a nasty bug ...
[09:43:18] <hashar>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=47164
[09:43:25] * qchris  takes a look

[09:43:28] <hashar>	 if a bug summary has:  bug 1   bug 1  bug 1
[09:43:36] <hashar>	 the plugin reports the change 3 time
[09:43:58] <hashar>	 seems the bug # are appended to a list when it should be a unique hash / tuple whatever it is called in java
[09:44:16] <hashar>	 that is not worth interrupting whatever you are doing right now heh
[09:44:42] <qchris>	 oh, yes. That's true. It extracts all ids and reports on all of them. Phew ... it's not a hooks-bugzilla plugin :-) It's hooks-its.
[09:45:03] <qchris>	 But either way, I'll fix that. Thanks. Good catch :-D
[09:45:19] <qchris>	 s/plugin/bug/
[09:45:33] <qchris>	 Thanks!
[09:48:11] <qchris>	 btw ... How did you happen to have to provide hooks-bugzilla support to Daniel (Some emails in the last few days)?
[09:48:32] <qchris>	 He just found that you work for WMF, or are you listed somewhere in the plugin?
[09:55:31] <gerrit-wm>	 New patchset: Hashar; "(bug 46723) uses ZUUL_COMMIT as a git refspec to build" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/58865
[09:56:06] <hashar>	 qchris: he mailed gerritadmin@wikimedia.org  which is an alias including Chad and I (at least)
[09:56:13] <hashar>	 no idea where that alias is setup though
[09:57:33] <qchris>	 :-) Ok. Cannot find that when grepping through hooks-bugzilla either.
[11:22:41] <gerrit-wm>	 New patchset: Zfilipin; "wmf-selenium-bot should not report SCM changes in IRC notifications" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58872
[11:24:51] <zeljkof>	 hashar: fixed the irc bot: https://bugzilla.wikimedia.org/show_bug.cgi?id=47167
[11:25:51] <hashar>	 zeljkof: great! :-]
[11:26:23] <hashar>	 zeljkof: ohhh you have jenkins templates? :-]
[11:26:30] <hashar>	 how does that work?
[11:26:42] <hashar>	 I am using a python script to generate the other jobs :D
[11:32:03] <hashar>	 snack time brb
[11:47:30] <zeljkof>	 hashar: templates are great :)
[11:47:52] <zeljkof>	 I have updated all jobs with just one line change
[11:48:00] <hashar>	 zeljkof: definitely. I am using http://ci.openstack.org/jenkins-job-builder/configuration.html#job :D
[11:48:36] <zeljkof>	 hm
[11:48:48] <zeljkof>	 templates in yaml instead of xml? interested
[12:02:43] <hashar>	 :-D
[12:03:02] <hashar>	 so yeah you define everything in yaml and the python script takes care of generating the XML and updating the jobs for you
[12:03:26] <hashar>	 the drawback is that the python script does not support every plugins/options so you sometime have to hack it  :]
[12:03:43] <hashar>	 our doc is at http://www.mediawiki.org/wiki/Continuous_integration/Jenkins_job_builder
[12:03:58] <hashar>	 upstream is http://ci.openstack.org/jenkins-job-builder/installation.html
[12:31:28] <hashar>	 Note that Jenkins is going down in roughly half an hour for maintenance
[12:35:02] <hashar>	 Abraham_WMDE: :-]
[12:35:13] <hashar>	 so you need some people to be able to send tag on a git repository?
[12:35:19] <hashar>	 what is the repo?
[12:36:01] <hashar>	 by default I don't think anyone has access to refs/tags/*
[12:37:19] <^demon>	 Owners have rights on refs/tags/*
[12:37:54] <hashar>	 ah they got rights to refs/*
[12:37:55] <hashar>	 nice
[12:38:07] <hashar>	 Abraham_WMDE: so which repo and which user is attempting to push a tag?
[12:38:13] <^demon>	 https://gerrit.wikimedia.org/r/#/admin/projects/mediawiki,access
[12:38:19] <^demon>	 That's for all of mediawiki/*
[12:43:38] <Abraham_WMDE>	 hashar: tgritschacher
[12:43:51] <hashar>	 Abraham_WMDE: and the repo?
[12:43:56] <Abraham_WMDE>	 hashar: repo, Wikibase, DataValues, Diff
[12:44:30] <Abraham_WMDE>	 push rights for tags
[12:44:58] <Abraham_WMDE>	 that's missing for Tobias
[12:45:00] <Abraham_WMDE>	 thx!
[12:50:36] <hashar>	 Abraham_WMDE: I guess he needs to be added as an owner of those repos
[12:51:16] <Abraham_WMDE1>	 hashar: that's not a problem from my side
[12:51:19] <hashar>	 so for Wikibase
[12:51:28] <hashar>	 any member of 'extension-Wikibase' group is allowed to push tags
[12:51:36] <hashar>	 https://gerrit.wikimedia.org/r/#/admin/groups/155,members
[12:51:41] <hashar>	 that is Jeroen + the wikidata group
[12:51:53] <hashar>	 and here is the list of of wikidata members:  https://gerrit.wikimedia.org/r/#/admin/groups/32,members
[12:51:53] <Abraham_WMDE1>	 that's what I'm wondering
[12:51:59] <Abraham_WMDE1>	 why it was not possible for him
[12:52:05] <Abraham_WMDE1>	 but he got the error "rejected"
[12:52:11] <hashar>	 which list Tobias Gritschacher	in it
[12:52:16] <Abraham_WMDE1>	 yea
[12:52:57] <hashar>	 maybe it was a signed tag?
[12:53:09] <^demon>	 project owners have push annotated tag on refs/tags/*
[12:53:19] <hashar>	 ah
[12:53:48] <^demon>	 "rejected" isn't very descriptive. Can we get a full pastebin?
[12:57:19] * hashar  ** Jenkins is going down for scheduled maintenance. Will be back in 2 hours ~ **

[13:39:32] <gerrit-wm>	 New patchset: Zfilipin; "Fix "Please choose a different, descriptive title." error at commons.wikimedia.beta.wmflabs.org" [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58886
[13:52:24] <wmf-selenium-bot>	 Yippie, build fixed!
[13:52:24] <wmf-selenium-bot>	 Project _debug-browsertests-template build #108: FIXED in 3 min 12 sec: https://wmf.ci.cloudbees.com/job/_debug-browsertests-template/108/
[13:53:17] <zeljkof>	 hashar: looks like wmf-selenium-bot really behaves now :)
[13:53:29] <zeljkof>	 no SCM messages
[14:47:50] <gerrit-wm>	 New review: Cmcmahon; "typo, my fault" [qa/browsertests] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/58733
[14:47:50] <gerrit-wm>	 Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58733
[14:48:29] <gerrit-wm>	 New review: Cmcmahon; "maintenance" [qa/browsertests] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/58862
[14:48:29] <gerrit-wm>	 Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58862
[14:50:17] <gerrit-wm>	 New review: Cmcmahon; "less IRC spam from bot" [qa/browsertests] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/58872
[14:50:18] <gerrit-wm>	 Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58872
[14:51:31] <gerrit-wm>	 New review: Cmcmahon; "accommodate different good-title checks on different target systems" [qa/browsertests] (master); V: 2 C: 2;  - https://gerrit.wikimedia.org/r/58886
[14:51:31] <gerrit-wm>	 Change merged: Cmcmahon; [qa/browsertests] (master) - https://gerrit.wikimedia.org/r/58886
[15:07:16] <hashar>	 !g 58902
[15:07:16] <wm-bot>	 https://gerrit.wikimedia.org/r/#q,58902,n,z
[15:07:35] <hashar>	 ** Jenkins is backup and operational. Should be a tiny bit faster. **
[15:07:49] <hashar>	 !g 58902
[15:07:49] <wm-bot>	 https://gerrit.wikimedia.org/r/#q,58902,n,z
[15:45:20] <Reedy>	 hashar: back up, not backup ;)
[15:45:51] <hashar>	 ah
[15:46:00] <hashar>	 Reedy: sorry :(
[15:46:12] <Reedy>	 it's alright
[15:46:21] <Reedy>	 I had to read what you said a couple of times :D
[15:50:44] * p858snake|l  backups Reedy into the cloud?

[15:52:28] <jeremyb_>	 qgil: so what's the timeline like for GSoC/OPW? April's kinda sneaking up on me! /me is now composing an email to a potential student
[15:53:47] <qgil>	 jeremyb_, looks like this: https://www.google-melange.com/gsoc/events/google/gsoc2013
[15:53:54] <qgil>	 jeremyb_, what do you want to know?
[15:54:38] <jeremyb_>	 qgil: how late is it in the process for a student to start figuring out what they want to do?
[15:55:17] * jeremyb_  is reviewing that calendar

[15:56:44] <YuviPanda>	 p858snake|l: hmm, is it 'backs up' or 'backups'?
[15:58:44] <qgil>	 jeremyb_, not late at all if they move fast. Their deadline for applying hasn't even started yet
[15:58:54] <qgil>	 jeremyb_, you are "bugzilla+org.wikimedia" in bugzilla, right?
[15:59:06] <jeremyb_>	 well there's an @...
[15:59:08] <jeremyb_>	 :-)
[15:59:20] <qgil>	 jeremyb_, sorry for respecting your privacy more than Bugizlla  ;)
[15:59:26] <jeremyb_>	 hah
[15:59:48] <qgil>	 jeremyb_, I recovered the log of that chat about MeetBot
[16:00:26] <jeremyb_>	 qgil: funny thing is recently noticed spam to that address with the bugzilla+ part missing. must be some spammer code translating that to a space?
[16:01:01] <jeremyb_>	 qgil: what about it? got stuck with nickserv registration and then never finished it... i'll work on it tomorrow
[16:01:56] <jeremyb_>	 aha, i have mail
[16:02:15] <jeremyb_>	 not just -office
[16:02:20] <jeremyb_>	 also e.g. #wikimedia-us
[16:04:48] <qgil>	 jeremyb_, not urgent or anything, I just wanted to document who/what we have so far
[16:05:03] <jeremyb_>	 yeah
[18:13:18] <gerrit-wm>	 New review: MarkTraceur; "(1 comment)" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53687
[18:19:26] <gerrit-wm>	 New review: Cscott; "(1 comment)" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53687
[18:38:23] <^demon>	 Hrm, I can't login to jenkins :\
[18:53:23] <gerrit-wm>	 New review: Krinkle; "(1 comment)" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53687
[18:54:23] <mwalker>	 csteipp_afk: am I missing some built in function in MediaWiki that allows me to get a query param and apply a filter to it? I swear I've written this helper function at least 50 times and I keep wondering why it's not in core
[19:01:27] <gerrit-wm>	 New patchset: MarkTraceur; "Update parsoid test pipelines" [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/58988
[19:04:17] <gerrit-wm>	 New patchset: MarkTraceur; "Add some more Parsoid tests for other components" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53687
[19:05:22] <marktraceur>	 Krinkle: Have time to take a gander?
[19:06:14] <AaronSchulz>	 "yes, sir, you gandered!"
[19:08:06] <Krinkle>	 marktraceur: gander?
[19:08:21] <marktraceur>	 AaronSchulz: I'd ask him to take a drake, but ducks are notoriously more pugilant than geese
[19:08:23] <Krinkle>	 'adult male goose'
[19:08:27] * Krinkle  looks up goose

[19:08:31] <marktraceur>	 Krinkle: "brief look"
[19:08:49] <Krinkle>	 A wild-goose chase?
[19:08:52] <AaronSchulz>	 marktraceur: have you seen Sweeney Todd?
[19:08:55] <Krinkle>	 Link me and I'll check.
[19:09:14] <marktraceur>	 AaronSchulz: I saw a high school production of it, so probably "no"
[19:09:37] <AaronSchulz>	 heh
[19:16:30] <csteipp>	 mwalker: No you didn't miss it, other than getInt/getBool/etc.... Post being hired here, I've become more convinced that escaping the output really is is more reliable than filtering the input.
[19:48:10] <hashar>	 ah
[20:06:06] <wmf-selenium-bot>	 Yippie, build fixed!
[20:06:07] <wmf-selenium-bot>	 Project browsertests-commons.wikimedia.beta.wmflabs.org-linux-chrome build #4: FIXED in 4 min 11 sec: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-linux-chrome/4/
[20:54:01] <gerrit-wm>	 Change merged: MarkTraceur; [integration/zuul-config] (master) - https://gerrit.wikimedia.org/r/58988
[20:54:11] <gerrit-wm>	 Change merged: MarkTraceur; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53687
[21:10:33] <Krinkle>	 hashar: "20:45:39 tput: No value for $TERM and no -T specified"
[21:10:41] <Krinkle>	 Did something change in Jenkins?
[21:10:55] <Krinkle>	 This worked up until now
[21:12:26] <gerrit-wm>	 New review: Hashar; "(8 comments)" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/53687
[21:13:35] <hashar>	 Krinkle: "This worked up until now"  <--- that is the MOST useful message you can ever tell to someone when looking for support :-]
[21:13:58] <Krinkle>	 Well, someone or something changed it and broke it.
[21:14:33] <hashar>	 maybe related to the change of workspace for jobs
[21:14:49] <hashar>	 they are now under /srv/ssd/jenkins/workspace/<jobname>
[21:14:56] <hashar>	 but on SSD
[21:15:09] <hashar>	 Krinkle: where is that $TERM issue ?
[21:15:25] <Krinkle>	 https://integration.wikimedia.org/ci/job/mwext-VisualEditor-doc-test/304/console
[21:15:34] <Krinkle>	 https://integration.wikimedia.org/ci/job/mwext-VisualEditor-doc-test/305/console
[21:16:21] <Krinkle>	 It worked in https://integration.wikimedia.org/ci/job/mwext-VisualEditor-doc-test/296/console and beofre
[21:16:23] <hashar>	 Krinkle: also some jslint failed on a mwext : https://integration.wikimedia.org/ci/job/mwext-TemplateSandbox-jslint/25/console
[21:16:35] <hashar>	 different issue though
[21:16:37] <Krinkle>	 https://integration.wikimedia.org/ci/job/mwext-VisualEditor-doc-test/297/console the TERM was first missing
[21:17:30] <Krinkle>	 hashar: What jslint link, what are you referring to? Seems like a normal failure to me. What am I loking for?
[21:18:54] <hashar>	 Krinkle: ahh that is the jenkins bug I have been working on this week maybe
[21:19:47] <Krinkle>	 what bug? the (something) you found in jslint, or the missing TERM as of 11AM this morning?
[21:19:58] <hashar>	 jslint sorry
[21:20:08] <Krinkle>	 So what bug?
[21:21:05] <hashar>	 ok not important, lets look at the TERM issue first
[21:22:30] <hashar>	 http://integration.wikimedia.org/ci/job/Test%20-%20TERM/1/console
[21:22:32] <hashar>	 echo $TERM
[21:22:35] <hashar>	 dumb
[21:22:36] <hashar>	 :-]
[21:23:08] <Krinkle>	 hashar: bash -xe is the default in jenkins
[21:23:13] <Krinkle>	 Trying tou
[21:23:16] <Krinkle>	 tput
[21:24:05] <hashar>	 http://integration.wikimedia.org/ci/job/Test%20-%20TERM/2/console :(
[21:25:15] <Krinkle>	 has something been upgraded?
[21:25:33] <Krinkle>	 (this morning)
[21:25:48] <Krinkle>	 well, this afternoon (10 hours ago)
[21:25:51] <hashar>	 we had a new drive inserted at 3pm
[21:26:05] <hashar>	 restarted jenkins completely
[21:27:25] <hashar>	 I am not sure why it worked previously
[21:27:44] <Krinkle>	 shouldn't you ask why it doesn't work now?
[21:27:57] <Krinkle>	 I am assuming that jenkins (simulates) a tty, right?
[21:28:09] <hashar>	 na it is not a term / interactive
[21:28:18] <Krinkle>	 I know it isn't (obviously)
[21:28:30] <Krinkle>	 but doesn't it simulate it?
[21:28:48] <Krinkle>	 since it takes colors and such (with the ansii plugin)
[21:29:04] <hashar>	 if you know, why do you ask? :-D
[21:29:11] <Krinkle>	 I don't know
[21:29:21] <hashar>	 I am puzzled :-D
[21:29:23] <Krinkle>	 I know it isn't a real interactive terminal
[21:29:26] <Krinkle>	 that's obvious
[21:31:49] <Krinkle>	 hashar: Did you restart zuul in the last 40 minutes?
[21:31:55] <hashar>	 nop
[21:32:00] <Krinkle>	 https://gerrit.wikimedia.org/r/#/c/53803/ isn't triggered on zuul
[21:32:53] <hashar>	 hmm
[21:33:10] <hashar>	 one possibility is that when restarting jenkins it inherit the environment of the user restarting it
[21:33:17] <hashar>	 that would sound a bit weird
[21:33:56] <Krinkle>	 https://github.com/wayneeseguin/rvm/issues/758
[21:34:38] <Krinkle>	 nevermind
[21:36:24] * Krinkle  adds export TERM=xterm to the macro

[21:37:35] <hashar>	 Krinkle: yeah that would work
[21:37:42] <hashar>	 should :-D
[21:37:51] <hashar>	 I am not sure how it ended up having a $TERM defined
[21:41:34] <Krinkle>	 hashar: the jshint error is for whitespace errors
[21:41:41] <Krinkle>	 and since that extension has no jshintrc file it is using the defaults
[21:41:58] <Krinkle>	 presumably somewhere in our stack something is defaulting to crockford's white:true
[21:43:35] <Krinkle>	 hashar: The i18n commit of 2 days ago succeeded
[21:43:43] <Krinkle>	 the i18n commit that his commit is depending on https://gerrit.wikimedia.org/r/#/c/58969/ also failed
[21:43:59] <Krinkle>	 (though obscured by l10n bot self merge)
[21:44:58] <Krinkle>	 hashar: Is it (as far as you know) only happening to that repository?
[21:45:34] <hashar>	 ahhh
[21:45:51] <hashar>	 Krinkle: so http://integration.wikimedia.org/ci/job/test-mediawiki-dumpenv/60/console  build on march 15th did have a $TERM set
[21:46:02] <Krinkle>	 Yes, of course
[21:46:18] <Krinkle>	 So what changed :D
[21:46:30] <Krinkle>	 Perhaps it was set by ZUUL ?
[21:46:57] <hashar>	 I don't think so :D
[21:47:15] <hashar>	 I must have started jenkins manually instead of via  the init script
[21:47:22] <hashar>	 the env has   06:22:56 declare -x MYSQL_PS1="mysql:\\u@\\h [\\d]> "
[21:47:29] <Krinkle>	 and xterm-256color
[21:47:34] <Krinkle>	 which is not what I have
[21:47:34] <hashar>	 which is I think from my bashrc
[21:47:38] <Krinkle>	 presumably your setting
[21:48:18] <hashar>	 so we know the cause.  TERM was set, it is no more set
[21:48:26] <Krinkle>	 k
[21:48:34] <hashar>	 TERM being set was a bug :-]
[21:48:47] <hashar>	 as for the exact root cause, I can only suspect I started jenkins manually
[21:49:36] <Krinkle>	 hashar: https://integration.wikimedia.org/ci/job/mwext-VisualEditor-doc-test/jobConfigHistory/showDiffFiles?timestamp1=2013-04-12_21-36-43&timestamp2=2013-04-12_21-36-47&name=mwext-VisualEditor-doc-test&isJob=true
[21:49:39] <Krinkle>	 What is that?
[21:49:47] <hashar>	 an url too long?
[21:49:55] <hashar>	 ohh
[21:49:56] <Krinkle>	 No, look at it
[21:50:26] <hashar>	 yeah so you are using the branch parameter
[21:50:32] <hashar>	 but it is not in job name
[21:50:52] <Krinkle>	 https://integration.wikimedia.org/ci/job/mwext-VisualEditor-doc-test/jobConfigHistory/showDiffFiles?timestamp1=2013-04-12_21-36-37&timestamp2=2013-04-12_21-36-47&name=mwext-VisualEditor-doc-test&isJob=true
[21:51:13] <hashar>	 so in a job group, JJB will do a cartesian products of all the parameter values
[21:51:21] <hashar>	 and expand the templates using each of the rows
[21:52:40] <hashar>	 like it will expand the template '{name}-{ext-name}-doc-test' with the parameters (mwext,VisualEditor,REL1_19)
[21:52:50] <hashar>	 then do it again with (mwext,VisualEditor,REL1_20)
[21:52:57] <hashar>	 and again with  (mwext,VisualEditor,REL1_21)
[21:53:00] <Krinkle>	 I know
[21:53:09] <hashar>	 hehe
[21:53:17] <hashar>	 I need to print T-Shirts for Amsterdam
[21:53:29] <hashar>	 you would be  <krinkle> I know   and mark would  be  <mark> yes
[21:53:30] <hashar>	 :-]
[21:53:43] <hashar>	 ^^^ -- that is what would be on mine:   <hashar> :-]
[21:53:51] <hashar>	 so
[21:53:52] <Krinkle>	 Indeed
[21:54:06] <hashar>	 that you got  REL1_21 (or whatever was the last value for the branch parameter)
[21:54:08] <Krinkle>	 ^^ would be Teal'c's
[21:54:14] <Krinkle>	 ("Indeed")
[21:54:33] <Krinkle>	 I was hoping it would not mess with jobs that don't use the {branch} parameter
[21:55:08] <hashar>	 I did a change in JJB to prevent it from generating the same template over and over
[21:55:09] <hashar>	 https://review.openstack.org/#/c/26839/
[21:55:12] <hashar>	 might be a side effect
[21:55:26] <hashar>	 not that one hmm
[21:55:28] <hashar>	 gr wrong change
[21:55:49] <hashar>	 maybe the branch values are released in a random order
[21:56:46] <marktraceur>	 hashar: Seen this before? https://integration.wikimedia.org/ci/job/mwext-Parsoid-merge/1750/console
[21:57:42] <hashar>	 Krinkle: http://dpaste.com/1056414/
[21:57:47] <hashar>	 Krinkle: the last seems to be REL1_19
[21:58:38] <hashar>	 marktraceur: oops
[21:59:31] <hashar>	 21:10:26 stderr: error: refs/remotes/origin/master does not point to a valid object!
[21:59:31] <hashar>	 sounds weird
[21:59:48] <Krinkle>	 I've seen that before
[21:59:59] <Krinkle>	 when creating a new job and triggering it manually for the first time
[22:00:12] <Krinkle>	 but that's not the case here
[22:02:30] <hashar>	 marktraceur: no idea sorry :(
[22:03:43] <marktraceur>	 Blargh.
[22:07:36] <gerrit-wm>	 New patchset: Hashar; "(bug 46723) uses ZUUL_COMMIT as a git refspec to build" [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/58865
[22:08:17] <gerrit-wm>	 New review: Hashar; "Changed the other occurences of ZUUL_BRANCH to ZUUL_COMMIT" [integration/jenkins-job-builder-config] (master) C: -2;  - https://gerrit.wikimedia.org/r/58865
[22:08:36] <hashar>	 that one is the most nasty bug ever
[22:10:19] <Krinkle>	 hashar: Aha, I almost figured that out last week
[22:10:21] <Krinkle>	 Nice catch
[22:10:28] <Krinkle>	 I was rebuilding a job manually
[22:10:38] <Krinkle>	 and it ignored the ZUUL_COMMIT completely
[22:10:44] <hashar>	 yup
[22:10:51] <Krinkle>	 instead it triggered for the last sha1 checked out in that branch
[22:11:02] <hashar>	 that is all because of git plugin and how our Zuul git repo has a master that changes
[22:11:11] <hashar>	 so master might point to a different commit
[22:11:28] <hashar>	 and Git plugin tries to be helpful by building the current origin/master instead of the refspec :-D
[22:11:44] <hashar>	 took me a while to understand how the plugin works. Mind you, I had to read some java code
[22:12:29] <hashar>	 for a long time I thought that the ref spec was what was going to be build
[22:12:41] <hashar>	 it is just used to specify what to git fetch
[22:12:57] <hashar>	 the revision to build has its own logic based on the branch specifier
[22:12:59] <hashar>	 that is nasty
[22:13:03] <hashar>	 https://bugzilla.wikimedia.org/show_bug.cgi?id=46723#c11  summarize it
[22:13:16] <hashar>	 so I will have to update all the jobs on monday
[22:13:17] <hashar>	 :(
[22:13:30] <hashar>	 and check that it is working fine with the European folks
[22:13:36] <hashar>	 + with -i18n team on monday evening
[22:19:37] <hashar>	 and I am off
[22:19:39] <hashar>	 have a good weekend
[22:25:36] <YuviPanda>	 marktraceur: kaldari merged it :)
[22:25:39] <Krinkle>	 saper: you there?
[22:25:43] <marktraceur>	 Whoop whoop
[22:25:48] <YuviPanda>	 though yeah, the individual photos thing needs fixing
[22:25:58] <kaldari>	 although it's a much rarer case
[22:26:00] <Krinkle>	 Anyone know a bash test that returns true for any kind of file?
[22:26:05] <YuviPanda>	 kaldari: true.
[22:26:07] <Krinkle>	 (file, symlink, dead symlink)
[22:26:14] <YuviPanda>	 the set with the same photos is probably the more common case
[22:26:27] <Krinkle>	 -e returns false if there is a symlink but target does not exist locally
[22:52:56] <gerrit-wm>	 New patchset: Krinkle; "Set export TERM=xterm for bash macros that use tput." [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/59018
[22:53:15] <gerrit-wm>	 Change merged: Krinkle; [integration/jenkins-job-builder-config] (master) - https://gerrit.wikimedia.org/r/59018
[22:56:47] <awight>	 This is probably a FAQ: Is there already a way to attach a message-id to MWExceptions so that the error string can be translated and presented?  Or do I have to catch specific exceptions, and build the translation there?
[23:29:50] <ori-l>	 going to sync a config change for test/test2 ( https://gerrit.wikimedia.org/r/#/c/57649/ )
[23:31:01] <ori-l>	 done