[00:44:22] 3Wikimedia / 3Continuous integration: Figure out how to test for database backwards incompatibilities - 10https://bugzilla.wikimedia.org/66684#c2 (10Bryan Davis) (In reply to Greg Grossmeier from comment #0) > I am trying to figure out a way to catch these types of mistakes before we > have outages in produc... [08:42:06] (03PS3) 10Nikerabbit: beta: update job for cxserver [integration/jenkins-job-builder-config] - 10https://gerrit.wikimedia.org/r/139336 (owner: 10Hashar) [09:48:10] 3Wikimedia / 3Continuous integration: Add jenkins jobs for BounceHandler extension - 10https://bugzilla.wikimedia.org/66703 (10Kunal Mehta (Legoktm)) 3NEW p:3Unprio s:3normal a:3None Initial code was added in I02f4750cc37e7b4b928e4f20882382915e465be8, would be good to have jenkins jobs for it. [11:04:56] zeljkof:hi Zeljko! are you busy now [11:05:28] Jagori: hi [11:05:32] no, not really busy [11:05:43] Will it be possible for you to have session tom? [11:06:36] I wanted to learn few things like the git diff..am not claer with that [11:06:43] Jagori: sure [11:06:50] the same time tomorrow? [11:07:00] ya should be fine 6-7 pm? [11:07:06] sydney time [11:07:14] Jagori: will create the event right now [11:07:21] thank you so much.. [11:07:57] I had a chat with Chris over why the visualeditor toolbar tests were failing [11:08:02] its working now on betalabs [11:08:48] I could see some steps like page.links_diff_view.should match Regexp.escape("[http://www.example.com Editing] ") in links_step definition file [11:09:01] am not clear with the use of Regexp.escape [11:09:08] probably would discuss tomorrow on this [11:11:09] Jagori: created the event [11:11:12] and send you the invite [11:11:20] let me know if it should be earlier or later [11:11:30] sure [11:11:55] http://ruby-doc.org/core-2.1.2/Regexp.html#method-c-escape [11:12:13] yes this time is fine [11:12:24] I will have a look aat the link as well [11:15:06] @zeljkof:Regexp is class then [11:15:18] but do we not need to include this class to access its methods? [11:15:56] will discuss tomorrow,sorry I got to go dinner [11:16:31] Jagori: sure [11:16:34] see you [12:32:25] 3Wikimedia / 3Continuous integration: Add jenkins jobs for BounceHandler extension - 10https://bugzilla.wikimedia.org/66703 (10Sam Reed (reedy)) s:5normal>3enhanc [13:23:34] thank you manybubbles for reviewing :) [13:23:42] thanks for working on it! [14:08:41] hello zeljkof [14:08:56] hi vikas [14:10:02] zeljkof: I got the code reviewed by manybubbles and got +1, when can I pair up with you ? [14:11:35] vikas: tomorrow? [14:11:46] or friday? [14:12:33] zeljkof: at what time ? [Let me look into weather forecasts :P ] [14:14:07] vikas: I guess early in my day would be late in your day :) [14:15:29] vikas: 3pm your time tomorrow? [14:15:57] zeljkof: yup ! that should be fine :) [14:16:12] vikas: great, I have created the event in the calendar and invited you [14:16:25] zeljkof: thank you :) I will accept it. [14:16:35] vikas: see you tomorrow then [14:16:38] zeljkof: Also is hashar going to come today ? [14:16:57] vikas: not sure, did not see him today, he is probably still not feeling good [14:17:32] zeljkof: oh ! ok thank you for the information :) [16:04:52] zeljkof: this is odd: https://integration.wikimedia.org/ci/job/browsertests-VisualEditor-test2.wikipedia.org-linux-firefox-sauce/1/testReport/junit/(root)/VisualEditor/Basic_edit/ [16:05:10] zeljkof: I thought HTTP::Persistent was gone from that build [16:05:45] is VE repo updated to the latest mediawiki selenium gem? [16:05:52] chrismcmahon: ^ [16:06:15] zeljkof: I thought I'd done that, I'll check for sure [16:08:33] maybe Net::HTTP::Persistent is now a part of ruby :) [17:27:52] chrismcmahon: i'm seeing just 3 outstanding failures for the ve browsertests. i can't tell if they're failing correctly [17:28:08] chrismcmahon: links.feature:10,18,26 [17:30:25] marxarelli: all in links.feature, huh? I have an idea of what might cause that-- iirc, the browser needs to have focus on the desktop for the lookahead link selector to appear [17:40:13] chrismcmahon: actually, stand by on that. i may have been running ve off an old branch [17:41:51] chrismcmahon: nm. links.feature is working fine against the latest master. i'll rerun the suite to verify the rest [17:42:02] marxarelli: right now I expect one failure in VE, the multiedit-workflow test. VE master branch has a fairly hairy bug right now I worked around this morning: https://bugzilla.wikimedia.org/show_bug.cgi?id=66717 [17:43:01] marxarelli: I will probably revert this when the bug is fixed: https://gerrit.wikimedia.org/r/#/c/140120/ [17:56:06] chrismcmahon: so, i'm getting the multiedit failure, but i'm also getting failures across the board for edit_utf8_logged_in.feature [17:57:13] marxarelli: OK. I'm checking, but that is likely to be legit. [18:02:43] chrismcmahon: great. i'll stop bugging you and move on to mobilefrontend :) [19:30:35] (03PS1) 10Zfilipin: WIP new jobs [integration/jenkins-job-builder-config] (cloudbees) - 10https://gerrit.wikimedia.org/r/140184 [21:41:05] chrismcmahon: i have the mobilefrontend browser tests running under the new vagrant setup [21:41:18] marxarelli: nice! [21:41:19] chrismcmahon: however, there are many failing tests due to what seem to be missing dependencies and i'm wondering how to proceed in fixing that [21:41:45] marxarelli: got an example? [21:41:55] chrismcmahon: e.g. account creation tests fail due to missing captcha [21:42:29] OK, that stands to reason [21:42:31] chrismcmahon: i've figured out that installing the ConfirmEdit extension fixes that [21:43:23] chrismcmahon: however, given that MF works fine without CE installed (just without captcha at sign-up) should?: [21:44:18] chrismcmahon: 1) we refactor the tests to handle the case where captcha isn't there; or 2) we make sure the extra dependencies are installed by vagrant/puppet [21:46:45] marxarelli: I'd lean toward installing the dependencies to make the tests pass, but I'd defer to what the Mobile team wants to do. Want to ask on their mail list? https://lists.wikimedia.org/mailman/listinfo/mobile-l [21:47:47] chrismcmahon: more generally, should we refactor browser tests to work without optional dependencies, or should we install the most featureful set of dependencies in environments where we're running browser tests? [21:47:56] got it [21:48:08] chrismcmahon: i'll ask [21:48:13] marxarelli: yeah, I think the answer is going to vary by project [21:48:42] marxarelli: also, we could do a thing with @tags in Cucumber like we do for the Jenkins builds [21:49:04] chrismcmahon: i was thinking the same. i'll propose that as one option [21:49:20] marxarelli: we started using a tag @clean for tests that should pass in a bare wiki, but it didn't catch on so far [21:49:39] chrismcmahon: what about tags for extension dependencies in certain tests? [21:49:50] chrismcmahon: e.g. @ext-ConfirmEdit [21:50:04] marxarelli: we haven't done that, but no reason we couldn't [21:50:20] chrismcmahon: along with a cucumber profile for vagrant where only the installed extensions are given as tags [21:52:00] chrismcmahon: and maybe further upstream (jenkins) we could test against a full-featured environment (another cuc profile for jenkins perhaps) [21:55:29] chrismcmahon: anyway, i won't get ahead of myself. we'll see what mobile wants to do first [21:55:37] sounds good