[00:00:49] jgonera: the name isn't final. do mention it to Zeljko, or send email to the QA list [00:02:37] chrismcmahon, ok, I'll talk to him tomorrow. just FYI, I'm working on this too but would like to show a different approach. I'll try to implement this single functionality (creating an article) without going much deeper for tomorrow to show what I mean [00:03:52] jgonera: I did a quick review of mediawiki API clients in Ruby and didn't find anything I much like. What we have now is a POC that is readable but needs refining. [00:05:50] chrismcmahon, I know, I went through it, however I don't feel it's very... Rubyish. I'd model it after something like https://github.com/futuresimple/dropbox-api. As I said, I'm not planning to invest too much time in it, but just prepare another POC to show you guys what I mean [02:38:49] (03PS1) 10Legoktm: Revert "mwext-GlobalPreferences-testextensions-master non voting" [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/111401 [02:39:09] (03PS2) 10Legoktm: Revert "mwext-GlobalPreferences-testextensions-master non voting" [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/111401 [08:20:12] good morning [08:23:05] zeljkof: so I think I more or less managed to setup the VisualEditor job with a headless firefox [08:23:14] but watir complains it can not find it :D [08:23:40] http://integration.wikimedia.org/ci/job/Visualeditor-doitall/36/console [08:23:44] hashar: morning :) [08:23:45] unable to obtain stable firefox connection in 60 seconds (127.0.0.1:7055) (Selenium::WebDriver::Error::WebDriverError) [08:23:46] let me see [08:24:09] do you have firefox installed on the server? via apt-get? [08:24:19] yes sir [08:24:30] I think the easiest way would be for you to connect on the instance that is running the browser tests [08:24:37] hashar: probably [08:24:40] so you can run cucumber yourself :D [08:25:09] do you have the time to pair on that right now? [08:25:20] had my coffee and cigarette [08:25:22] so yeah :-D [08:25:35] I need 5 minutes or so [08:25:39] take your time [08:25:40] I will ping you in a few minutes [08:25:45] ping me whenever you are ready [08:25:57] I got a bunch of mails to triage [08:34:31] hashar: I am ready [08:34:52] hangout? or here? [08:35:45] hangout [08:35:55] so we can screen share if need be [08:36:01] hashar: calling [09:17:23] (03PS3) 10Hashar: mwext-GlobalPreferences-testextensions-master is now voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/111401 (owner: 10Legoktm) [09:18:20] (03CR) 10Hashar: [C: 032] "Thank you! Deploying :-)" [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/111401 (owner: 10Legoktm) [09:18:24] (03Merged) 10jenkins-bot: mwext-GlobalPreferences-testextensions-master is now voting [integration/zuul-config] - 10https://gerrit.wikimedia.org/r/111401 (owner: 10Legoktm) [10:48:32] zeljkof: well done :-] [10:48:50] hashar: a lot of work done :) [10:48:56] yup [11:01:14] hashar: https://help.github.com/articles/working-with-ssh-key-passphrases#platform-mac [11:01:29] ssh-keygen -p changes the passphrase [11:03:20] \O/ [11:03:23] glad you found out [11:18:55] (03CR) 10Pastakhov: "Dependence does not work here, please help me." [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/110164 (owner: 10Pastakhov) [11:20:27] zeljkof: I found out!!!!!!!!!!!!!!! :-D [11:20:39] zeljkof: turns out some VisualEditor javascript is in a submodule which was not being pulled [11:20:40] :D [11:20:44] :) [11:23:33] damn git submodules [11:25:17] (03PS1) 10Hashar: multigit: refresh submodules after checkout [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111433 [11:25:30] (03CR) 10Hashar: [C: 032] multigit: refresh submodules after checkout [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111433 (owner: 10Hashar) [11:25:33] (03Merged) 10jenkins-bot: multigit: refresh submodules after checkout [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111433 (owner: 10Hashar) [11:27:00] zeljkof: I was looking for a way to point bundler to a shared gems cache [11:27:07] aka not the vendor/cache nor the system gems [11:27:15] apparently that is not supported so I filled a bug report for it [11:27:35] zeljkof: https://github.com/bundler/bundler/issues/2856 [11:27:57] that would speed up bundle install hopefully [11:28:24] hashar: great [11:28:38] I am watching the issue [11:29:54] Uncaught TypeError: Cannot read property 'status' of null ve.init.mw.ViewPageTarget.js:252 [11:29:55] :( [11:31:21] hashar: good luck [11:41:41] and ve is not enable on user pages :D [11:41:43] $wgVisualEditorNamespaces = $wgContentNamespaces; [11:44:18] (03PS14) 10Hashar: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 [11:50:05] YEAHHH PROGRESS [11:50:13] http://integration.wikimedia.org/ci/job/Visualeditor-doitall/42/console [11:50:21] it apparently managed to edit the page [12:18:51] zeljkof: I have to reach out with the parsoid folks :/ [12:18:55] no clue why it does not work [13:09:39] (03CR) 10Hashar: [C: 032] multigit: set -e [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111217 (owner: 10Hashar) [13:10:16] (03PS3) 10Hashar: multigit: set -e [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111217 [13:10:22] (03CR) 10Hashar: [C: 032] multigit: set -e [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111217 (owner: 10Hashar) [13:10:25] (03Merged) 10jenkins-bot: multigit: set -e [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111217 (owner: 10Hashar) [13:45:53] (03CR) 10Hashar: "deployed, all jobs are passing." [integration/jenkins] - 10https://gerrit.wikimedia.org/r/111217 (owner: 10Hashar) [14:50:56] hashar: zeljkof: is there some shell script you're using to setup ruby, gem, bundler, etc.. on the wmf jenkins? [14:52:56] Tobi_WMDE: hello [14:53:01] Tobi_WMDE: nothing usable right now :( [14:53:26] Tobi_WMDE: in JJB-config repo look in macro.yaml for the "browsertest" macro [14:53:49] err [14:53:53] there is a builder "browsertests" [14:54:17] where can I find it? [14:54:17] it injects a bunch of $wg var, install mw, run bundler and then cucumber [14:54:20] it is veryyy ugly [14:55:02] Tobi_WMDE: https://git.wikimedia.org/blob/integration%2Fjenkins-job-builder-config/master/macro.yaml#L70 [14:55:55] it hardcodes a bunch of settings such as the BROWSER_LABEL [14:55:59] currently set to phantomjs [14:56:11] but some job (i.e. forVisualEditor) will need to use firefox [14:56:50] hashar: thx, that helps me a bit. foy my use case it would be simpler anyhow [14:57:22] hopefully [14:57:39] I will make it cleaner later on [15:01:35] Tobi_WMDE: let me know if you need any more information [15:08:15] thx [15:16:26] Tobi_WMDE: there is, let me find the link to the docs [15:17:14] Tobi_WMDE: here it is [15:17:15] https://github.com/wikimedia/mediawiki-selenium/blob/master/docs/template.md#jelly-based-transformation [15:17:29] tag contains all setup [15:32:16] hi zeljkof I have a strange issue: I want to use both a css: selector and an :index selector on an element, but when I use both, Watir gives me an error: invalid attribute: :css (Watir::Exception::MissingWayOfFindingObjectException) [15:32:50] zeljkof: this seems to work for every type of locator except css [15:33:21] chrismcmahon: I do not think it is supported [15:33:21] https://github.com/watir/watir-webdriver/issues/241 [15:35:17] aha, thanks zeljkof. it's a bug, but it seems I can work around it [15:36:54] zeljkof: if you talk to Jari, tell him to fix that one :-) [15:38:18] chrismcmahon: I think jari and alex are talking about how to fix that [15:38:39] hmmm: invalid element state: Failed to execute query: 'div.tipsy-inner > div.flow-tipsy-flyout > ul > li.flow-action-edit-title > a.mw-ui-button.flow-edit-topic-link[1]' is not a valid selector. [15:44:40] hashar: did you see this? https://github.com/bundler/bundler/issues/2856#issuecomment-34168418 [15:44:51] yeah [15:44:55] not sure what to think about it [15:45:00] hashar: same here :) [15:45:07] I am not sure I understood the answer [15:45:10] would it mean bundler will have to download all gems there ? [15:45:12] we can try :-] [15:45:22] you should ask for clarifications [15:45:29] or just try doing what is suggested [15:45:43] I am not sure I understood the answer, like I already said [15:46:09] basically he is saying to point GEM_HOME to a directory where the user can write [15:46:17] which probably mean that ruby will no more be using system gems at all [15:46:33] and thus gems will fail to find any gems requested, thus download them from ruby gems and write them to GEM_HOME [15:46:53] on the next run of bundler, it would find them all in that GEM_HOME directory, and should thus not have to download anything [15:46:59] unless some new dependencies have been added [15:47:21] i will try :D [15:50:14] gah, this is still not going to work [15:52:11] zeljkof: export GEM_HOME="$WORKSPACE/../gems" [15:52:11] :D [15:53:23] hashar: works? [15:53:50] nop :( [15:54:03] ahh [16:05:47] ah where is zeljkof [16:28:36] (03PS15) 10Hashar: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 [16:43:05] (03PS16) 10Jforrester: (WIP) (WIP) visualeditor-doitall (WIP) (WIP) [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/100800 (owner: 10Hashar) [16:44:54] James_F: Zeljkof and I paired this morning for the VE browser tests [16:45:07] James_F: and I ended up being stuck with parsoid not responding :/ [16:45:43] hashar, James_F: I tried to test visualeditor locally with mediawiki-vagrant [16:45:58] but I found this: https://bugzilla.wikimedia.org/show_bug.cgi?id=60905 [16:47:52] zeljkof: How do you have MW 1.22alpha? [16:48:08] zeljkof: do git pull in the mediawiki directory and it should work fine. [16:48:17] hashar: Not good. :-( [16:48:38] James_F: I do not know how I got 1.22, I have created a clean vm with vagrant [16:48:53] James_F: and thanks, will try with git pull [16:48:54] is it broken with mediawiki master ? [16:49:12] Not here. [16:50:09] zeljkof: the GEM_HOME seems to be working, it speed up bundle install a bunch (1+minute saved) [16:50:47] zeljkof: that would install gems there. I am not sure how 'bundle exec cucumber' would manage to point to the proper version referenced in Gemfile.lock though [16:51:07] zeljkof: I am afraid require Something would simply point to the last version of Something. Though bundle check seems to list the proper versions [16:51:29] James_F: do you have any idea how to debug the "parsoid server" could not be contacted? [16:51:31] hashar: I think bundle checks the gemfile.lock and then requires the correct version [16:51:39] zeljkof: hopefully :-] [16:51:55] zeljkof: will leave as is for now [16:52:07] James_F: maybe I should fill a bug with all details : -] [17:02:35] hashar: Hmm. [17:02:38] hashar: No idea, sorry. [17:05:55] James_F: will get some traces / informations and ping gabriel [17:06:23] hashar: Right. [17:12:51] hashar: I'm late to the conversation, but are you trying to get parsoid to run on mediawiki-vagrant? [17:14:10] JeffHall: not at all, on a labs instance :-] [17:14:34] hashar: OK, I'm probably not going to be much help then :( [17:14:41] but for some reason my VisualEditor can't find the Parsoid daemon :/ [17:15:24] hashar: Is it definitely running? [17:15:38] hashar: Yeah, I had the same problem with vagrant, and had to correct the default $wgVisualEditorParsoidURL value in VisualEditor.php [17:15:39] hashar: And is LocalSettings.php's config line for where to find Parsoid correct? [17:16:20] I think I will add a few wfDebug() message :-] [17:18:35] $wgVisualEditorParsoidURL = string(21) "http://127.0.0.1:8000"; [17:25:40] hashar: Do you get a valid HTP 200 response when you curl that $wgVisualEditorParsoidURL value? [17:27:23] yup [17:29:15] when saving [17:29:22] the API yield back a 200: {"error":{"code":"parsoidserver","info":"Error contacting the Parsoid server"}} [17:29:23] :( [17:29:30] nothing obvious in the debug log [17:52:53] giving up for tonight, will look at it again tomorrow [17:52:54] *wave* [20:13:26] if beta labs ain't happy, ain't nobody happy [20:29:59] And beta labs just doesn't seem to be very happy much these days - must be a case of winter blahs [21:10:28] JeffHall: go ahead and remove line 65 completely, no need for the comment there https://gerrit.wikimedia.org/r/#/c/111531/1/modules/ve-mw/test/browser/features/support/pages/visual_editor_page.rb [21:11:09] chrismcmahon: Aaagh, thought I had removed that! Will get rid of that now.... [21:17:31] chrismcmahon: At some point, I'd like to chat with you about this new test for VisualEditor: https://gerrit.wikimedia.org/r/#/c/111142/ [21:18:03] chrismcmahon: It's a test a production wikis, so perhaps should not be running in CloudBees with other VE browser tests [21:18:57] JeffHall: yeah, I'm not so sure that is a good idea [21:19:39] chrismcmahon: Might be worth a Google Hangout pairing sesssion? I can try to grab a slot on your calendar [21:22:29] JeffHall: yeah, I think I now have Flow and Mobile tests back running, or pretty close. In general, my experience of browser tests on production envs is not super good. I'll have to think about it a little more [21:22:44] in the meantime I'm starving and I have to run an errand, back in a bit [22:58:53] whee, Flow builds are all back to green now [22:59:26] spage: ^^ (I took the liberty of updating your branch and self-merging today) [23:00:41] chrismcmahon ++: that is fantastic! We've been deploying updates every day, no time [23:01:30] spage: yes, I like the new list'o'actions, had to think a little about how to test that, and I'll probably write some more tests soon now that that is looking a bit more stable [23:05:17] I love youuu man!