[00:07:08] <chrismcmahon>	 RoanKattouw: thanks for breaking VE on beta!  :-)
[00:07:14] <RoanKattouw>	 lol
[00:07:17] <RoanKattouw>	 Should be fixed by now
[00:07:31] <chrismcmahon>	 RoanKattouw: I'm a little behind
[00:08:34] <chrismcmahon>	 RoanKattouw: at some point I'd like to ask you a really obscure question about ResourceLoader, but no big hurry
[00:08:38] <RoanKattouw>	 OK
[00:10:53] <chrismcmahon>	 RoanKattouw: basically the last comment on https://bugzilla.wikimedia.org/show_bug.cgi?id=56567 .  What would cause a client in Chrome but not Firefox to issue multiple GET requests to RL for the same resource?
[00:11:35] <RoanKattouw>	 I have no idea
[00:11:45] <RoanKattouw>	 Have you been able to reproduce this on a desktop outside of browser tests?
[00:13:19] <chrismcmahon>	 RoanKattouw: no, but we have seen Chrome do other strange things with javascript in Chrome, like load up the HotCat gadget https://bugzilla.wikimedia.org/show_bug.cgi?id=43335
[00:14:13] <RoanKattouw>	 Hmm
[00:14:24] <RoanKattouw>	 So those elements are expected in the iframe but should be in the head, not in the body
[00:15:02] <RoanKattouw>	 WTF: 		newDoc.body.appendChild( styleNode );
[00:15:05] <RoanKattouw>	 OK that's just a bug
[00:15:11] <RoanKattouw>	 Adding that to my existing "rewrite this crap" bug
[00:15:12] <chrismcmahon>	 RoanKattouw:  I wonder if this odd stuff in Chrome not FF is the root of the anecdotal sentiment that UploadWizard, VisualEditor, etc. are fragile.  I do not have a clean repro though.
[00:15:54] <RoanKattouw>	 I've never encountered this in Chrome
[00:16:03] <chrismcmahon>	 RoanKattouw: I continue to look for one :-)
[00:16:14] <RoanKattouw>	 And I would be more concerned if you could reproduce it outside of Saucelabs
[00:17:14] <chrismcmahon>	 RoanKattouw: I've been getting an education about all the places it is possible to stash javascript in Mediawiki
[00:19:07] <RoanKattouw>	 Yeah there are a lot
[09:40:26] <grrrit-wm>	 (PS1) Hashar: beta-code-update: use mwdeploy home for wmf-beta-autoupdate [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/93941
[09:44:09] <grrrit-wm>	 (CR) Hashar: [C: 2] beta-code-update: use mwdeploy home for wmf-beta-autoupdate [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/93941 (owner: Hashar)
[09:44:28] <grrrit-wm>	 (Merged) jenkins-bot: beta-code-update: use mwdeploy home for wmf-beta-autoupdate [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/93941 (owner: Hashar)
[17:12:08] <James_F>	 chrismcmahon: Is https://gerrit.wikimedia.org/r/#/c/93597/ good to go?
[17:15:40] <chrismcmahon>	 James_F: should be.  or we can wait to merge it until we have a modern Parsoid.
[17:16:12] <James_F>	 chrismcmahon: "Modern"?
[17:16:21] <chrismcmahon>	 interesting?
[17:16:54] <James_F>	 Whether or not Parsoid round-trips normal UTF8 characters without corruption is interesting whatever version is being run. :-)
[17:17:10] <chrismcmahon>	 merge at will then :-)
[17:17:30] <James_F>	 Done. :-)
[17:25:07] <chrismcmahon>	 James_F: also, dunno if you noticed, but the test is set up to add any interesting other strings you might like.  the only reason I chose ÀÈÌÒÙ ÄËÏÖÜ ÄËÏÖÜ ÂÊÎÔÛ ÂÊÎÔÛ áéíóú was because it was in my google search results
[17:26:07] <James_F>	 chrismcmahon: Sure. We could round-trip the entire Unicode basic plane, but that's hardly risk-targeted testing. :-)
[17:26:57] <chrismcmahon>	 let's not
[18:46:45] <James_F>	 chrismcmalunch: Occasional "Unable to pick a platform for the provided browser" errors are just to be expected, right? We got two in https://wmf.ci.cloudbees.com/job/VisualEditor-en.wikipedia.beta.wmflabs.org-linux-firefox/127/testReport/ and I don't think it's due to a change in VE?
[19:00:55] <Nikerabbit>	 James_F: I also had one for ULS
[19:01:00] <Nikerabbit>	 went away in the next run
[19:01:12] <James_F>	 Nikerabbit: Hopefully it's just transient, then. :-(
[19:01:17] <James_F>	 Nikerabbit: (Thanks!)
[19:16:23] <greg-g>	 chrismcmalunch: maybe a stupid question: We have these need-[unit|integration|parser]-test keywords in bugzilla. They are either not used very much, or we only need 3 tests written currently for issues reported to BZ: http://ur1.ca/fzsfm
[19:16:37] <greg-g>	 the search includes all bug states, resolved or otherwise.
[19:17:13] <greg-g>	 oh, so the question: should I start using the keywords and review them periodically, maybe with you/zeljko?
[19:48:53] <chrismcmahon>	 greg-g: I've been using a keyword 'browser-test-bug' to identify issues found by way of the automated browser tests, those others don't look like they're used much.
[19:49:37] <greg-g>	 right right
[19:50:01] <greg-g>	 so, should we start using something like them to keep track of issues that were reported, resolved, and need a test to prevent future breakages?
[19:50:14] <greg-g>	 I personally like the "fix a bug? write a test." mantra
[20:07:29] <grrrit-wm>	 (PS1) Hashar: qa-browsertests-postmerge to create a tarball [integration/zuul-config] - https://gerrit.wikimedia.org/r/93994
[20:07:46] <grrrit-wm>	 (PS1) Hashar: qa-browsertests-postmerge to create a tarball [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/93995
[20:08:22] <grrrit-wm>	 (PS2) Hashar: qa-browsertests-postmerge to create a tarball [integration/zuul-config] - https://gerrit.wikimedia.org/r/93994
[20:08:32] <grrrit-wm>	 (CR) Hashar: [C: 2] qa-browsertests-postmerge to create a tarball [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/93995 (owner: Hashar)
[20:08:59] <grrrit-wm>	 (CR) Hashar: [C: 2] qa-browsertests-postmerge to create a tarball [integration/zuul-config] - https://gerrit.wikimedia.org/r/93994 (owner: Hashar)
[20:09:08] <grrrit-wm>	 (Merged) jenkins-bot: qa-browsertests-postmerge to create a tarball [integration/zuul-config] - https://gerrit.wikimedia.org/r/93994 (owner: Hashar)
[20:09:17] <grrrit-wm>	 (Merged) jenkins-bot: qa-browsertests-postmerge to create a tarball [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/93995 (owner: Hashar)
[20:16:19] <grrrit-wm>	 (Restored) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292 (owner: Hashar)
[20:16:23] <grrrit-wm>	 (PS3) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292
[20:28:09] <grrrit-wm>	 (PS1) Hashar: rename qa-browsertests-{build,run} [integration/zuul-config] - https://gerrit.wikimedia.org/r/94000
[20:30:12] <grrrit-wm>	 (PS1) Hashar: qa-browsertests: always build before running tests [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94001
[20:30:19] <grrrit-wm>	 (CR) Hashar: [C: 2] rename qa-browsertests-{build,run} [integration/zuul-config] - https://gerrit.wikimedia.org/r/94000 (owner: Hashar)
[20:30:28] <grrrit-wm>	 (Merged) jenkins-bot: rename qa-browsertests-{build,run} [integration/zuul-config] - https://gerrit.wikimedia.org/r/94000 (owner: Hashar)
[20:30:59] <grrrit-wm>	 (PS4) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292
[20:32:44] <grrrit-wm>	 (PS1) Hashar: qa-browsertests-run is now non voting [integration/zuul-config] - https://gerrit.wikimedia.org/r/94002
[20:32:55] <grrrit-wm>	 (CR) Hashar: [C: 2] qa-browsertests-run is now non voting [integration/zuul-config] - https://gerrit.wikimedia.org/r/94002 (owner: Hashar)
[20:33:05] <grrrit-wm>	 (Merged) jenkins-bot: qa-browsertests-run is now non voting [integration/zuul-config] - https://gerrit.wikimedia.org/r/94002 (owner: Hashar)
[20:34:05] <grrrit-wm>	 (CR) jenkins-bot: [V: -1] Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292 (owner: Hashar)
[20:34:35] <grrrit-wm>	 (PS5) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292
[20:35:06] <hashar>	 chrismcmahon: around ?
[20:35:17] <chrismcmahon>	 hi hashar
[20:35:20] <hashar>	 chrismcmahon: I have enabled the trigger of browser tests when a change is sent on qa/browsertests.git
[20:35:29] <hashar>	 but they take a loooooong time to run :(
[20:35:57] <hashar>	 https://gerrit.wikimedia.org/r/#/c/57292/ is my test change
[20:36:00] <chrismcmahon>	 hashar: they do take a long time to run.  you are only running the tests out of the /qa/browsertests repo, yes?
[20:36:05] <hashar>	 the run is processing on https://integration.wikimedia.org/ci/job/qa-browsertests-run/24/console
[20:36:13] <hashar>	 yeah the qa/browsertests one
[20:37:20] <chrismcmahon>	 hashar: I think it is pretty amazing that they run under WMF Jenkins on WMF hosts at all.  I was told 18 months ago or so that would never, ever happen.
[20:38:38] <hashar>	 yeah you keep saying that :D
[20:39:42] <chrismcmahon>	 hashar: ah, that's running PhantomJS
[20:39:54] <hashar>	 AHHH
[20:40:01] <hashar>	 it is running them against beta :D
[20:40:12] <hashar>	 so
[20:40:23] <hashar>	 I should probably not run them on each patchset change
[20:40:28] <chrismcmahon>	 hashar: "
[20:40:28] <chrismcmahon>	 20:35:44 + export BROWSER_LABEL=phantomjs 20:35:44 + BROWSER_LABEL=phantomjs"
[20:40:30] <hashar>	 that is going to be too long for you
[20:40:37] <hashar>	 yeah
[20:40:47] <hashar>	 it is not using sauce labs at all
[20:40:59] <chrismcmahon>	 hashar: a whole lot of these tests are going to fail on phantomjs
[20:41:07] <hashar>	 so I shuold revert
[20:41:16] <hashar>	 and just attempt to install using bundler
[20:43:33] <grrrit-wm>	 (PS1) Hashar: Revert "rename qa-browsertests-{build,run}" [integration/zuul-config] - https://gerrit.wikimedia.org/r/94005
[20:43:41] <grrrit-wm>	 (CR) Hashar: [C: 2] Revert "rename qa-browsertests-{build,run}" [integration/zuul-config] - https://gerrit.wikimedia.org/r/94005 (owner: Hashar)
[20:44:17] <grrrit-wm>	 (Merged) jenkins-bot: Revert "rename qa-browsertests-{build,run}" [integration/zuul-config] - https://gerrit.wikimedia.org/r/94005 (owner: Hashar)
[20:47:16] <chrismcmahon>	 hashar: have we investigated running a headless Firefox?
[20:47:29] <grrrit-wm>	 (PS2) Hashar: qa-browsertests: always build before running tests [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94001
[20:47:53] <hashar>	 chrismcmahon: never heard of it before :-]
[20:48:08] <hashar>	 I first want to have browser tests to run for ULS
[20:49:38] <grrrit-wm>	 (CR) Hashar: [C: 2] qa-browsertests: always build before running tests [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94001 (owner: Hashar)
[20:51:19] <chrismcmahon>	 hashar: FF will run headless under Xvfb https://www.google.com/search?q=run+firefox+headless+selenium
[20:51:48] <hashar>	 if one of you figure out the exact command, we could add it :-]
[20:52:07] <hashar>	 the reason I went with phantomJs is that we already use it for qunit tests
[20:54:55] <grrrit-wm>	 (Merged) jenkins-bot: qa-browsertests: always build before running tests [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94001 (owner: Hashar)
[20:55:25] <chrismcmahon>	 hashar:  I'm not sure exactly what Zeljko has in mind.  we could probably run Firefox headless, but talking to Sauce Labs would probably be best.
[20:55:42] <hashar>	 well we can just run a normal firefox in sauce labs :]
[20:56:24] <chrismcmahon>	 hashar: and Chrome and all the IEs also.  Chrome is my problem right now.
[20:56:53] <grrrit-wm>	 (PS6) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292
[20:58:26] <hashar>	 chrismcmahon: the idea was to run phantomJS as a touchstone / quick build
[20:58:33] <hashar>	 and then trigger all the other browsers
[20:59:49] <chrismcmahon>	 hashar: OK. in that case, we should run the suite a few times under phantomjs and figure out what does pass.  that is our first smoke test I guess.
[21:00:00] <hashar>	 yup
[21:00:12] <hashar>	 I am assuming they would be fast to run
[21:00:18] <hashar>	 faster than driving a normal browser
[21:09:41] <grrrit-wm>	 (PS7) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292
[21:17:38] <grrrit-wm>	 (PS8) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292
[21:17:58] <hashar>	 ok finally reverted my previous change
[21:22:34] <grrrit-wm>	 (Abandoned) Hashar: Jenkins job validation (DO NOT SUBMIT) [qa/browsertests] - https://gerrit.wikimedia.org/r/57292 (owner: Hashar)
[21:23:17] <hashar>	 ahh mediawiki/selenium had some activity :-]
[21:24:45] <hashar>	 he is not tagging the releases :D
[21:27:44] <grrrit-wm>	 (CR) Krinkle: "@Hashar: Ping, see irc." [qa/browsertests] - https://gerrit.wikimedia.org/r/57292 (owner: Hashar)
[21:33:48] <spagewmf>	 chrismcmahon: FWIW my manual test string is - Déjà vu 北京 <tag & "OK" 'end' . 2-byte, 3-byte, HTML entities and quotes.
[21:42:44] <grrrit-wm>	 (PS1) Hashar: mediawiki-selenium-gembuild job [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94052
[21:42:55] <grrrit-wm>	 (CR) Hashar: [C: 2] mediawiki-selenium-gembuild job [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94052 (owner: Hashar)
[21:42:57] <spagewmf>	 is there a way to automate saucelabs "Take a picture of this web page in 30 different browser+OS combinations"?
[21:43:52] <grrrit-wm>	 (Merged) jenkins-bot: mediawiki-selenium-gembuild job [integration/jenkins-job-builder-config] - https://gerrit.wikimedia.org/r/94052 (owner: Hashar)
[21:46:49] <grrrit-wm>	 (PS1) Hashar: trigger mediawiki-selenium-gembuild on postmerge [integration/zuul-config] - https://gerrit.wikimedia.org/r/94055
[21:47:05] <grrrit-wm>	 (CR) Hashar: [C: 2] trigger mediawiki-selenium-gembuild on postmerge [integration/zuul-config] - https://gerrit.wikimedia.org/r/94055 (owner: Hashar)
[21:52:03] <grrrit-wm>	 (Merged) jenkins-bot: trigger mediawiki-selenium-gembuild on postmerge [integration/zuul-config] - https://gerrit.wikimedia.org/r/94055 (owner: Hashar)
[21:54:38] <chrismcmahon>	 spagewmf: interestingly, the "北京" is working for Firefox but WebDriver for Chrome results in a " ¬" character in VE, /me investigates
[22:14:31] <chrismcmahon>	 spagewmf: aaaand, we validated a bug in ChromeDriver/Chrome with that string  https://gerrit.wikimedia.org/r/#/c/94059/