[00:00:03] jeremylee: ok try 'vagrant enable-role mobilefrontend' [00:00:16] then, vagrant provision [00:00:18] then vagrant up [00:01:26] that tells vagrant to follow the rules for setting up mobilefrontend, which i believe has a rule to ensure that mediawiki is present - so puppet should go ahead and install everything for you [00:02:50] jeremylee: it may take a little while for everything to run [00:03:20] jeremylee: did you just want to mess around with mobilefrontend in general, or were you hoping to more closely emulate how we have things set up on the Wikimedia cluster? [00:03:45] awjr: its running. good so far. [00:04:54] awjr: hoping to closely emulates whats on the cluster [00:04:58] emulate* [00:05:17] jeremylee: ok, you can get part of the way there using existing vagrant roles [00:05:37] update. vagrant up is done [00:06:24] you'll want to enable: echo, geodata, mobilefrontend, visualeditor, and scribunto (scribunto just for quieting some errors and making all the tests run ok) [00:06:31] and there is some red text here. starting with "The following SSH command responded with a non-zero xit status." [00:06:37] jeremylee: ok, what do you see now at 127.0.0.1:8080? [00:06:38] oh [00:06:51] "Vagrant assumes that this means the command failed!" [00:07:19] still dead on 8080 [00:07:28] php is running tho [00:07:31] does it say what specifically failed? [00:07:45] you may need to scroll up through the backscroll to find the actual errors [00:09:06] alot of text but something key about MobileFrontend "skipping because of failed dependencies." [00:09:23] can i just rewind for a moment. [00:09:35] and go back to the beginning of this to make sure i didnt miss something. [00:09:45] step 1 get git. [00:09:47] got it [00:09:58] step 2 get VirtualBox [00:10:00] got it [00:10:08] step 3 get vagrant. got it [00:10:22] then git clone vagrant. [00:10:30] mw-vargant* [00:10:39] mw-vagrant* [00:10:51] I can clone into any directory i want right? [00:10:55] sure [00:11:37] jeremylee not sure how much scrollback you can see, but if you can copy/paste the output into a pastebin or something i can take a look [00:12:12] eek. using git bash on windows. Not sure how to copy and paste from it. [00:12:55] * awjr shrugs [00:13:05] can you select the text in the backscroll? [00:13:19] if you can, i imagine ctrl+c should do the trick [00:13:46] I think i got it [00:13:54] (03PS14) 10Dr0ptp4kt: WIP: SpecialMobileWebApp: Get the startup module caching [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/111485 (owner: 10Jdlrobson) [00:14:26] (03PS15) 10Dr0ptp4kt: WIP: SpecialMobileWebApp: Get the startup module caching [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/111485 (owner: 10Jdlrobson) [00:15:56] awjr: http://pastebin.com/QdPSW9yw [00:16:39] thought i was running vagrant up. those were the results of vagrant provision [00:17:44] that's ok, things didn't go well regardless :p [00:18:15] jeremylee: i recommend you delete the repo you cloned in #4 and start step #4 over again. [00:18:38] i like your idea [00:18:41] the first time you run vagrant up after that, keep an eye on the output and if things *still* don't work, copy/paste the output again and i can take a look [00:19:08] its going to be a long download [00:19:23] jeremylee: there is another opiton we haven't tried [00:19:56] what's that? [00:20:03] you can try 'vagrant destroy' [00:20:06] and then 'vagrant up' [00:20:10] i did attempt that [00:20:16] before coming here [00:20:30] jeremylee: actually that should do essentially the same thing as #4 [00:20:31] will try one more time [00:20:41] jeremylee: yeah, this time capture the output [00:21:34] the first time i run it. why does it take so long? [00:21:53] because it is provisioning your instance - it's downloading all the software needed and configuring it all for you [00:22:02] presumably something is breaking for you during that process [00:22:27] this runs on ubuntu right? [00:22:39] yeah [00:22:41] preicse pangolian [00:22:52] ok. waiting for machine to boot [00:24:03] so far so good [00:26:36] awjr: who are you on bugzilla? [00:26:44] jeremylee: awjrichards [00:27:53] you commented on bug 62614? [00:28:13] jeremylee: yes [00:28:16] oh ok jw [00:28:17] oh, yeah, 'Arthur Richards' :) [00:30:38] yea the complexity of the bug is beyond me. i just wanted to have a go at trying to reproduce something. Never done it before. [00:30:54] :D [00:31:14] that bug may be a real challenge to reproduce [00:31:27] (most of the devs on the project haven't been able to reproduce it) [00:31:59] but getting vagrant etc all set up will put you in a good place to try! and to maybe reproduce some other bugs :) [00:32:39] awjr: http://pastebin.com/d7RFNygg vagrant up output [00:32:46] ended with some red text. [00:33:04] some more "skipping because of missing dependencies" [00:33:10] jeremylee: ok, and it looks like your backscroll buffer is too small to have captured it all [00:33:58] i could have logged it. i forgot. [00:34:25] but the output turns red only after a certain point [00:34:33] jeremylee: did you run 'vagrant destroy'? [00:34:35] yes [00:34:42] before vagrant up [00:34:43] (03CR) 10JGonera: [C: 032] Delay EventLogging click track until next page load [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116035 (owner: 10Jdlrobson) [00:34:57] hmm [00:35:04] (03CR) 10jenkins-bot: [V: 04-1] Delay EventLogging click track until next page load [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116035 (owner: 10Jdlrobson) [00:35:25] jeremylee: can you do me a favor and run 'vagrant destroy' followed by 'vagrant list-roles' and tell me if any of the roles are enabled (have an * next to them) [00:36:57] out of many roles listed only "mobilefrontend" has a * [00:37:29] interesting i thought vagrant destroy would disabled that [00:38:14] what are these roles? and if you have no other ideas. I think ill try from scratch and reclone [00:38:28] jeremylee: ok, if you're up to try provisining an instance again - try 'vagrant destroy', 'vagrant disable-role mobilefrontend', 'vagrant up' [00:38:36] ok [00:38:37] please capture all of the output from vagrant up if you can [00:39:34] agh. how do u log stdout? [00:40:48] jeremylee: do you want to write all output to a particular file? [00:41:01] $ somecommand &> /log/file [00:41:14] that'll send stderr and stdout to the specified file [00:41:21] at least with bash [00:41:23] im not sure about gitbash [00:41:42] script [00:42:21] doesn't work on git bash :[ [00:43:35] running disable-role prompts me to run vagrant provision [00:43:55] yeah ok, run vagrant provision before continuing [00:44:45] jeremylee: what happens when you try your git bash command with '&> /some/log/file' appended to it? [00:45:08] it works as expected [00:45:17] but gitbash doesn't have script support [00:46:59] single line of output from vagrant provision [00:47:13] ==> default: VM not created. Moving on... [00:50:41] jeremylee: ok; i have to go in ~10mins. if things still aren't working, you can try talking to ori livneh. he sometimes hangs out in this channel, but not always. he's currently hanging out in the #mediawiki channel. his nick is 'ori' or 'ori-l' [00:51:44] alright thanks for your help [00:51:57] np; i'll still be around for a bit but wanted to give you fair warning :) [00:52:16] ok [00:52:26] vagrant up still running [00:53:57] awjr: before you go, can you explain the purpose of roles? [00:54:27] these belong to the vagrant system right? [00:54:31] jeremylee: 'roles' are a way of easily turning on/off specific configurations/types of functionality [00:54:50] jeremylee: the 'role' concept is a vagrant thing; the roles that we have defined in mw-vagrant are wikimedia/mediawiki-specific [00:55:09] so enabling mobilefrontend does what? [00:55:33] I mean the mobilefrontend role. [00:55:57] im assuming that install the mobilefrontend extension on the mediawiki instance [00:56:01] installs* [00:56:17] jeremylee: in your mw-vagrant checkout, open up the following file: puppet/manifests/roles/mobilefrontend.pp [00:57:15] that's the file that tells puppet (configuration management software) what to do when the mobilefrontend role is enabled [00:57:16] ok [00:57:40] line 5: it says include the mediawiki role (which is its own list of instructions) [00:57:47] line 6: same thing but for EventLogging [00:58:02] then it downloads/enables the TextExtracts and PageImages extensions [00:58:21] then it gets the MObileFrontend extension, and sets a few configuration directives for the extension [00:58:29] i see [00:59:09] so installing the items you mentioned on bugzilla (echo, parsoid, etc) are as simple as enabling roles here ? [00:59:44] jeremylee: if we wanted to have those things be set up by default when the mobilefrontend role is enabled, yes [01:00:02] but at least for now, the mobielfrontend role is pretty bare-bones [01:00:27] the intent is to just get up and running minimally to develop with MobileFrontend, rather than to try and create an envrionment that resembles production [01:00:32] (03PS1) 10Jdlrobson: WIP: Story 1747 Notification drawer [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120733 [01:00:37] http://pastebin.com/3njLrpNa [01:00:56] I know you got to go. [01:01:00] o_O [01:01:13] 1.   default: SSH address: 127.0.0.1:2222 [01:01:13] 2.     default: SSH username: vagrant [01:01:13] 3.     default: SSH auth method: private key [01:01:13] 4.     default: Error: Connection timeout. Retrying... [01:01:14] 5.     default: Error: Remote connection disconnect. Retrying... [01:01:17] that's the problem [01:01:31] ? [01:01:33] it is ? [01:01:45] i think so [01:01:58] i thought that was just it waiting to boot up [01:02:10] it eventually gets to machine booted and ready [01:02:15] jeremylee: i gotta take off; talk to ori - tell him we chatted a bit and show up this pastebin [01:02:22] sure thing [01:02:23] thanks [01:03:03] actually now that im reading further that's probably not the problem [01:03:22] i think it must just sit there trying to log in over and over until it suceeds (when the machine is booted) [01:04:17] right [01:04:33] ok im not seeing anything blatantly obvious but hopefully ori will have better insight - he's the one who originally put together mw-vagrant ;) [01:04:38] anyway, im off - see ya and good luck [01:05:01] sweet [01:05:05] thanks for the pointers [01:05:07] later [01:07:23] (03CR) 10Jdlrobson: [C: 04-2] WIP: Story 1747 Notification drawer [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120733 (owner: 10Jdlrobson) [01:07:44] (03PS16) 10Jdlrobson: Delay EventLogging click track until next page load [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116035 [01:08:47] (03CR) 10Jdlrobson: [C: 032] "Rebased" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116035 (owner: 10Jdlrobson) [01:09:26] (03Merged) 10jenkins-bot: Delay EventLogging click track until next page load [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116035 (owner: 10Jdlrobson) [02:07:59] (03PS1) 10Kaldari: Story 1732: WIP - Make licensing agreement customizable [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 [02:08:20] (03CR) 10Kaldari: [C: 04-2] "WIP" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 (owner: 10Kaldari) [02:09:03] (03CR) 10jenkins-bot: [V: 04-1] Story 1732: WIP - Make licensing agreement customizable [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 (owner: 10Kaldari) [07:14:09] (03CR) 10Siebrand: Custom pull-to-refresh for articles. (031 comment) [apps/ios/wikipedia] - 10https://gerrit.wikimedia.org/r/120500 (owner: 10Mhurd) [07:38:28] (03CR) 10Siebrand: [C: 04-1] "i18n/L10n reviewed." (032 comments) [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 (owner: 10Kaldari) [09:33:05] mobil team staff araound (mobileweb)? [09:33:30] kinda [09:38:04] MaxSem: means you are one? :) [09:38:20] probably:) [13:34:38] MaxSem: Steinsplitter was scared of you? :P [13:57:59] (03CR) 10Siebrand: "Jon, ping?" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119484 (owner: 10Siebrand) [15:33:52] (03CR) 10MaxSem: [C: 032] Avoid redirects when visiting URLS in browser tests [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [15:36:09] (03Merged) 10jenkins-bot: Avoid redirects when visiting URLS in browser tests [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [15:43:51] (03CR) 10Cmcmahon: "I think this should not be merged, for a few reasons." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [15:46:26] hi MaxSem I sort of wish you hadn't merged that. sorry I was late with my comments. [15:46:41] chrismcmahon, feel free to revert then:) [15:46:48] NOW REVERT AND HAVE A WAR AS THE GODS MEANT FOR IT TO BE! [15:46:49] :) [15:46:58] MaxSem: I'm in enough trouble with you guys right now [15:47:31] MaxSem: it took me a while to come to a whole conclusion about what the right thing to do is [15:47:34] whaaa? we love you! [15:49:55] :\ trying to run a test on editor_ve.feature over beta labs. And I keep running into some errors. [15:50:00] MaxSem: I did have lunch with awjr_away yesterday. It is very nice to talk face to face with an expert on the mobile front [15:50:27] jeremylee: it is a bug https://bugzilla.wikimedia.org/show_bug.cgi?id=63074 I reported it this morning [15:50:41] jeremylee: see, these tests do find bugs ! [15:51:18] chrismcmahon: O_O that is so cool [15:51:30] well i wonder if the error im getting is related. [15:52:04] can i show you what im getting ? [15:54:57] chrismcmahon: http://pastebin.com/yi9FGPVU [16:03:36] jeremylee: that is almost certainly a different error https://bugzilla.wikimedia.org/show_bug.cgi?id=62614 and it's very perplexing [16:05:09] is it likely that a different browser/browser version could affect the outcome of this test? [16:05:30] jeremylee: out of curiosity, are you running Linux or something else? [16:05:43] chrismcmahon: running ff out of windows [16:06:12] jeremylee: awesome, so far the only repro we have is on Linux, not MacOSX, now you have a repro on Windows [16:06:26] should i post this output? [16:06:36] on bug 62614 ? [16:10:38] jeremylee: if you watch the test run, do you see the red error box that says "Error saving settings. Please make sure that you have cookies enabled"? If you do, please add a comment that you can reproduce the error on Windows [16:10:45] Nope. [16:11:09] That is the one distinct difference. [16:11:51] There is no
error. [16:12:08] jeremylee: when you watch the test run, do you see the ON/OFF toggle being set properly? in the bug there, what we see is a) the browser turns the beta toggle on b) the browser shows the error about cookies c) the ON/OFF toggle for beta goes back to OFF [16:12:50] mine only does a) and b) [16:12:59] I mean... [16:13:02] a) and c) [16:13:04] not b) [16:18:50] does that make sense? [16:23:28] jeremylee: it does make sense, and it's kind of weird. :-) do please add a comment on the bug. [16:24:39] can i run this same test without cucumber? manually? [16:24:57] like. open up firefox, navigate to the pages and do the same thing? [16:25:22] because when I do this manually, everything works. Is the automated test doing something different? [16:25:26] jeremylee: I have not been able to reproduce the issue manually, but I have not tried all that hard either. [16:25:46] chrismcmahon: I mean, how does the automated test trigger the save button ? [16:26:06] jeremylee: the automated test is moving very quickly is what I think. When the test clicks "Save", the cookie or session information for the browser is not completed [16:28:41] Jon mentioned introducing a delay in the test. Does that seem like something i could try? [16:29:24] it seems you think the test is running too fast. [16:31:37] jeremylee: sure. put a "sleep" line just before the test clicks Save and see what happens [16:32:51] chrismcmahon: im a little new here, do you mind showing me how to do that? seems pretty simple? [16:33:08] jeremylee: sure, one moment [16:33:15] sweet thanks [16:35:24] jeremylee: I'm bringing up the files in my editor so I can show you exactly what's going on, just a moment... [16:36:00] chrismcmahon: ok. [16:37:12] jeremylee: so if we look at a test failure like the one at https://wmf.ci.cloudbees.com/job/MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-firefox/443/testReport/(root)/VisualEditor/Ensure_we_load_the_correct_section/ [16:37:54] ok [16:37:59] jeremylee: down at the bottom of that stack trace we see that the step that failed is "Given I am in alpha mode" -- make sense? [16:38:46] yes [16:39:14] jeremylee: now we need to hack on whatever it means that "Given I am in alpha mode". You can find that step in the steps files any way you like. I use grep a lot myself. [16:39:54] under step_definitions? [16:40:05] jeremylee: so however you do it, you'll find the step for "Given I am in alpha mode" in the file /step_definitions/common_steps.rb [16:40:28] jeremylee: and the step starts on line 77 in that file [16:40:38] yes i do . [16:40:44] awesome :-) [16:41:58] OK, now in that step you see the lines about "page.save_settings_element.when_present.click" . what you'll want to do is to insert a line just above those that say "sleep 5" to pause for 5 seconds. Run the test and see what happens. replace with "sleep 1", run the test, see what happens. [16:42:05] jeremylee: *^^ [16:42:13] awjr_away: heads up that i moved up our 1:1 to this morning to catch up [16:42:46] jeremylee: it is almost always a bad idea to use sleep() in a real test that runs in a real build, but it is useful when figuring things out [16:42:56] i see. [16:43:07] there are two lines which read page.save_settings. [16:43:15] could you explain this to me? [16:46:39] I mean to ask, if lines 77-83 were commented, what would they say? [16:47:57] also. the sleep works, I tested with 10, 5 and 1. But the same behavior results. [16:49:32] jeremylee: did you get vagrant working? [16:49:41] I did. [16:49:44] \o/ [16:49:47] what was the problem? [16:49:58] tfinc thanks for the head's up [16:50:03] no idea. i just waited a little longer for everything to get set up [16:50:07] tfinc and welcome back :) [16:50:32] well im glad you got it working jeremylee [16:50:38] awjr: thanks [16:51:02] awjr: but it finally worked. although it runs so slow. Page requests take forever [16:51:32] jeremylee: yeah; vagrant is sharing resources with the same machine you're running it from; it indeed tends to run fairly slowly [16:51:37] you can allocate more resources to the VM [16:51:47] which you may or may not want to do, depending on your set up [16:52:14] awjr: I may want to do that. [16:52:14] jeremylee: you are starting to get into the territory of "questions it would be good to ask on the QA mail list" :-) [16:52:40] chrismcmahon: oh :O [16:53:53] not to sound pushy but chrismcmahon, is this because the list is more public? Or is it because I should just be more deligent in my own research? [16:54:37] I should probably take this as a sign that I need to review cucumber documentation. [16:54:41] jeremylee: because we have a lot of inexperienced people on the QA list that would be interested in the question and in the answer, and also because the answers are starting to get a bit complicated. [16:55:14] I see. [16:55:46] chrismcmahon: the sleep doesn't seem to resolve the issue. [16:56:03] jeremylee: there is a subsection at https://www.mediawiki.org/wiki/Vagrant#Advanced_usage called 'adjust the resources allocated to the VM' [16:56:52] jeremylee: have you looked at page files yet, e.g. features/support/pages/mobile_options_page.rb ? That's where you'll find how the Save button is identified. [16:57:36] chrismcmahon: I haven't. And that is exactly what I was looking for. [16:58:03] awjr: I had configured it earlier to run in GUI mode. I'll have to check out these configurations options as well. [16:58:29] jeremylee: and if you haven't read it yet, this is a big part of what's going on in these tests: https://github.com/cheezy/page-object/wiki [17:01:04] in short, before I dive into this. PageObjects are like handles for elements on a page? [17:01:20] Am i walking in with the right intuition chrismcmahon? [17:02:27] jeremylee: yes, page objects are a design pattern for browser tests, you can google it. that code on github is a particular implementation of Page Objects that takes advantage of a lot of cool features available in Ruby [17:03:29] ok sweet :] [17:04:33] Well, I have to head off to school. But I'm sure I won't learn more today than what I've learned in the last hour here. [17:06:26] thanks jeremylee we appreciate you learning this and doing this [17:09:06] chrismcmahon: The pleasure is mine. Thanks for walking me through it. [17:10:35] hi,jgonera [17:22:56] (03CR) 10Kaldari: "Why don't we just make the article shorter? There's no reason the San Francisco article has to be 20 pages long on beta labs and test2." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119557 (owner: 10Jdlrobson) [17:25:20] hmm, no brion [17:36:52] (03CR) 10Kaldari: "I shortened the article on beta labs as a test. It now takes 5 seconds to reload from save instead of 15+." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119557 (owner: 10Jdlrobson) [17:37:02] brb.... [17:39:56] (03CR) 10Jdlrobson: "Even so I think the default timeout is way too short, and sometimes beta labs is just plain slow. Giving the test steps more time to run s" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119557 (owner: 10Jdlrobson) [17:40:11] (03CR) 10Jdlrobson: "@chrismcmahon poke - please reconsider this -2" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119557 (owner: 10Jdlrobson) [17:50:19] kaldari_afk: yo do we need to make changes to typography refresh [18:12:10] jdlrobson: Yeah, sometime this week [18:12:22] kaldari: well we should do it before thursday.. [18:15:35] dr0ptp4kt: with brion not around, think you'll have time to merge a couple of small changes so I can push out an alpha? :) [18:16:01] dr0ptp4kt: I'll merge the zero change :) [18:16:03] yeah, add me to 'em then paste gerrit urls here [18:16:10] no quid pro quo, but thanks :) [18:16:11] dr0ptp4kt: https://gerrit.wikimedia.org/r/#/c/120532/ [18:16:24] dr0ptp4kt: and https://gerrit.wikimedia.org/r/#/c/120545/ [18:16:27] dr0ptp4kt: hehe :) [18:17:37] YuviPanda: starting [18:17:42] dr0ptp4kt: sweet [18:30:02] (03CR) 10Dr0ptp4kt: [C: 032 V: 032] Fix namespaces missing on displaying titles + add tests [apps/android/wikipedia] - 10https://gerrit.wikimedia.org/r/120532 (owner: 10Yuvipanda) [18:30:09] woot [18:32:42] YuviPanda: joining us on hangout ? [18:32:50] tfinc: yeah, chrome issues. logging in on ff [18:33:23] (03CR) 10Awjrichards: "I agree with Chris on this one, although this has now been merged. I suggest reverting this change and resolving this by changing the expe" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [18:33:48] tfinc: bah, that crashed firefox. restarting machine. [18:34:47] (03CR) 10Jdlrobson: "I would argue "Main Page" is more customer friendly than "Main_Page" and we should encourage the former." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [18:35:33] (03CR) 10Jdlrobson: ""Main Page" will always redirect to "Main_Page". This is how MediaWiki works." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [18:36:21] (03CR) 10Dr0ptp4kt: [C: 032 V: 032] Direct feedback email to the OTRS queue [apps/android/wikipedia] - 10https://gerrit.wikimedia.org/r/120545 (owner: 10Yuvipanda) [18:37:23] YuviPanda: done. this test devices uses that testa email address. i requested that we rename it to 'send app feedback', although i get that may not be practical for smaller screens [18:37:49] so that otrs ticket was proof of concept for testing, not strictly a feature request [18:39:46] MaxSem: Hello [18:39:52] have a qq for you [18:39:53] hey [18:39:55] hey! [18:40:17] (03CR) 10Jdlrobson: "I should also note I am responsible for all the instances of "Main Page". To me it is more user friendly to put "Main Page" but I sometime" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120564 (owner: 10Jdlrobson) [18:40:18] sure [18:40:19] So question is what happens when you use the extract feature to querry a red link? [18:55:20] MaxSem: Any idea? [18:55:34] ah [18:55:35] Will the extract feature return empty? [18:55:51] article is a red link article [18:55:56] dr0ptp4kt: oh no, I like 'send app feedback' [18:55:59] dr0ptp4kt: much better! :) [18:56:06] dr0ptp4kt: but in production app we'll hide it in about or somewhre [18:56:15] so what do you mean, exactly? api.php?action=query&prop=extracts&titles=somenonexistentpage? [18:56:46] vbamba, https://en.wikipedia.org/w/api.php?action=query&prop=extracts&titles=somenonexistentpage [18:56:47] YuviPanda: yeah. i guess there's at least a slightly better chance of not getting feedback on articles with that qualifier. i'll post a small patchset to tweak the verbiage, ok? [18:58:00] YuviPanda: i realized i hadn't used submit on the otrs email. so i just did that. [18:58:16] dr0ptp4kt: ok! :) [19:00:46] (03CR) 10Cmcmahon: "I really think we need to fix these test-by-test, in order to understand the underlying issues as to why they fail across browsers." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119557 (owner: 10Jdlrobson) [19:00:56] (03PS1) 10Dr0ptp4kt: Change 'Send feedback' to 'Send app feedback'. [apps/android/wikipedia] - 10https://gerrit.wikimedia.org/r/120852 [19:01:35] YuviPanda: ^ [19:01:41] yeah looking :) [19:01:55] (03CR) 10Yuvipanda: [C: 032 V: 032] Change 'Send feedback' to 'Send app feedback'. [apps/android/wikipedia] - 10https://gerrit.wikimedia.org/r/120852 (owner: 10Dr0ptp4kt) [19:02:02] ty dr0ptp4kt [19:02:09] ty YuviPanda [19:02:14] dr0ptp4kt: :) [19:02:43] dr0ptp4kt: hmm, think we can find someone who is on an actual zero rated carrier who can test for us? [19:02:49] I'm going to merge the zero change before I cut the alpha [19:05:14] YuviPanda: we're emailing with an operator on that. good thing is if we add a cellular ip of our own, too. [19:05:30] dr0ptp4kt: ah, right. [19:05:33] dr0ptp4kt: 'cellular ip'? [19:06:51] dr0ptp4kt: pm [19:06:54] YuviPanda: that is, an ip address on a cellular-based obile network [19:21:47] ragesoss: are you a OTRS vol? [19:21:55] YuviPanda: yes. [19:22:02] nominally. [19:22:04] ragesoss: we're switching app feedback mechanism to an OTRS queue :D [19:22:07] I almost never actually use it. [19:22:27] but I'm happy to at least *say* I'll help with mobile. [19:22:38] hahahaha :D [19:22:53] I have rather more interest in that than I do in commons image permissions and en.wiki general inquiries. [19:23:08] :D [19:23:11] I'd figured. [19:23:33] ragesoss: I'll ask around [19:23:42] ragesoss: username? [19:23:47] ragesoss [19:30:47] someone know where this jdlrobson person is [19:31:05] violetto: he's right here! [19:31:10] :P [19:31:25] YuviPanda: not responding though [19:31:39] violetto: did you try turning it off and on? [19:32:15] YuviPanda: he's responding to you? [19:32:41] violetto: no [19:32:54] violetto: last was an hour or so ago [19:41:42] mhurd, this is the email address for wikipedia for ios that sends to otrs: mobile-ios-wikipedia@wikimedia.org. commons for ios is mobile-ios-commons@wikimedia.org. [19:41:55] no mhurd, will email him [19:44:45] jdlrobson: are you online? [20:24:04] violetto: is there a 'log out' icon? [20:24:17] vbamba: zz_prtksxna chicken is alive again! Andrew Bogott helped me rescue chicken. [20:24:54] (03PS1) 10Kaldari: Hygiene: Handle fatal errors when loading talk page [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120864 [20:26:19] MaxSem: Seeing this link [20:26:59] can't understand it thought [20:27:08] is that what the extrcat feature returns? [20:27:17] sorry for asking rubbish questions [20:28:34] vbamba: so if I read that link correctly, MaxSem says that if there is no such page, then the API will return an appropriate response saying that no such apge exists [20:28:41] and if popups does not already handle it properly it should [20:30:20] jdlrobson, what is https://wikimedia.mingle.thoughtworks.com/projects/mobile/cards/1767 blocked on? [20:30:41] Read comments jgonera [20:31:02] I see, thanks [20:33:56] so right now popups shows nothing [20:34:15] vbamba: should be a trivial to fix bug for zz_prtksxna when he wakes up, I reckon [20:34:18] YuviPanda: ^ [20:34:41] Actually I have a bit of a DEEEsign for it [20:34:52] :^) [20:35:07] wondering if we can insert something there instead of saying 'This page does not exist' [20:35:17] which I wanted to discuss with Dan and Prateek [20:35:30] vbamba: right. [20:35:53] vbamba: it should be easy to detect when that is happening and do whatever you want to do with it [20:36:05] vbamba, this is not an API problem [20:36:17] this is up to popups to decide [20:36:30] vbamba: if you have questions along the lines of 'is this possible to do?' feel free to ask me until prtk wakes up :) [20:36:44] vbamba: or email prateek :) He has a pretty good sense of these thinges. [20:37:34] YuviPanda: dont think we need one for log out [20:37:42] violetto: hmm, so no icon? [20:37:51] violetto: that might feel weird since everything else has icons [20:38:09] YuviPanda: is actually the reason why i wanted to not have one [20:38:17] oh? [20:38:23] well, I'll take it out :) [20:38:36] YuviPanda: so it's easy to find [20:38:43] violetto: you want log out to be easy to find? [20:38:50] violetto: nobody uses logout :| [20:38:54] YuviPanda: i want to be evil but not that evil [20:39:08] violetto: I am so confused atm [20:39:09] YuviPanda: kewl! thank you offering your esteemed time sir [20:39:13] YuviPanda: basically not placing the same emphasis [20:39:25] with other links [20:39:31] violetto: hmm, I still think we should hide logout under settings :P but fine. [20:40:43] YuviPanda: i know, i know we want to make it not so obvious but at the same time i'd hate to hear someone saying with frustration "why would someone put log out in settings" [20:41:00] it makes more sense that im logging out of my account, hence it's under my username [20:41:18] YuviPanda: under as in within or literally listed under it [20:41:23] violetto: or we could put it in the drop down 'user' menu that expands when you tap your username (I showed you a prototype) along with your page, your talk page, contributions :P [20:41:32] YuviPanda: yeah [20:41:40] violetto: want me to do that expanding menu again? :) [20:41:45] I have tha code saved up [20:42:10] YuviPanda: what do we have currently in there again? [20:42:17] i forgot what we decided last [20:42:57] violetto: I said 'My contributions, Log out'. Then soon 'My watchlist, My notifications'. We can also have a quick language switcher later on. [20:44:30] violetto: also 'my user page', 'my talk page' [20:44:37] violetto: and 'my editing sandbox' [20:44:46] the last 3 we can add now itself if we want [20:49:46] YuviPanda: https://www.dropbox.com/s/npzd9kl2ldfod7z/Screenshot%202014-03-25%2013.48.38.png we can put log out and editing sandbox in the user menu [20:50:44] violetto: but not talk page, watchlist or my contributions? [20:50:46] that split makes no sense to me [20:53:03] YuviPanda: im unsure how much editing sandbox is being used. I'm thinking the username should be a link to the user page. If there's a menu it should be about that account. My contribs, talk page, watchlist, to read, history, is all about you [20:53:11] YuviPanda: they should stick together [20:53:24] YuviPanda: if they go in the menu, they should go in the menu together [20:54:15] YuviPanda: they should stick together, especially how we'd want to surface my contribs and talk page more than it already has [20:54:26] violetto: hmm, fine. [20:55:05] violetto: in that case everything in that list is related to the user, so I guess we should just leave it as it is. Log out is in the wrong grouping, so maybe I can move it to be *above* settings, so it is in the same grouping but separate so you can not accidentally tap it? [20:55:16] violetto: also remember 'send feedback' will not be there in production release. it would be hidden more [20:56:01] YuviPanda: where will send feedback be? [20:56:34] violetto: I suppose in 'About' [20:56:53] violetto: the reason for that being it was fairly well hidden in the old app and even then the volume of feedback was so overwhelming we had to shut it down. [20:56:57] YuviPanda: so in place of "send feedback" is an about link [20:57:17] violetto: about is usually under 'Settings' in Android [20:57:37] violetto: so less clutter on main nav, less cramping when in landscape and also more familiar for users [20:57:51] to send feedback Settings > About > Send feedback? [20:58:16] violetto: yes. [20:58:42] violetto: 99% of the feedback we got was also 'this article is wrooooong!!!11' and 'you are hiding the truth about energy healing!' as well. [20:58:53] along with a number of 'justin bieber is gay lol' [20:59:02] really, it was almost youtube commentsish [20:59:07] YuviPanda: im unfamiliar with this issue, I'm okay with that if you think it makes sense. i wonder what kenan thinks [20:59:14] is kenan on irc? [20:59:26] YuviPanda: hahaha i see [20:59:36] violetto: hehe :) [21:00:03] violetto: there were also a lot of empty emails being received. apparently a lot of people get confused when they hit 'send feedback' and their email client opens up, so instead of going back they just hit.... 'send' [21:00:23] YuviPanda: on ios we're not even thinking about a setting link so i guess if we're sure we want stuff in there we should have settings on ios [21:00:47] YuviPanda: yeah makes sense, has to be more of an intended action [21:00:56] violetto: :) This has happened 3 different times in the past, so hopefully there won't be a 4th time, but I wouldn't put it past us :P [21:01:16] violetto: yeah. Having it in about and with verbiage that makes it extremely clear that it is only about issues about the app itself, and nothing more [21:01:23] violetto: and we'll build a kill switch for that too [21:07:38] YuviPanda: https://www.dropbox.com/s/eirkj2llq7m95ab/Screenshot%202014-03-25%2014.07.09.png [21:08:05] violetto: nah, let's just let it go for now. I like the way things are now, and we can experiment with location of log out later. [21:08:14] violetto: and you are right - I don't know if many people will use the sandbox. [21:08:53] YuviPanda: okay, fine with me [21:08:56] violetto: :) [21:15:07] (03PS1) 10Jdlrobson: QA: Make sure watchlist tests are setup correctly [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120938 [21:17:37] (03CR) 10Jdlrobson: [C: 04-1] Hygiene: Handle fatal errors when loading talk page (032 comments) [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120864 (owner: 10Kaldari) [21:19:51] kaldari: jgonera so i need to rebase https://gerrit.wikimedia.org/r/#/c/116912/ for a 4th time and it's been open since the beginning of the month - if i rebase it now will one of you commit to reviewing it? [21:21:05] I can commit to reviewing later this afternoon [21:23:22] thanks jgonera i will rebase it now then [21:26:56] jdlrobson, did you talk to Trevor/Roan about GSoC? [21:29:50] (03PS2) 10Kaldari: Hygiene: Handle fatal errors when loading talk page [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120864 [21:32:23] awjr_away, is someone trying to fix the bug with cookies that prevents switching to mobile on port different than 80? [21:32:36] it would be great to get this fixed before GSoC... [21:35:08] violetto: heya! [21:35:55] mhurd: hey back atcha [21:36:34] violetto: have a sec to link 2 more icons to left drawer design card? [21:36:44] What were they [21:36:55] violetto: random and save [21:37:36] Save for later? mhurd [21:38:22] yeah, but one for taking the action of saving - like a little disk or something - i have the other save icon for the "Saved articles" entry [21:39:28] mhurd: https://gerrit.wikimedia.org/r/120944 should let us tweak config params for things like 'isAnonEditingAllowed' from the server easily :) [21:39:43] YuviPanda: nice! [21:40:01] (03PS7) 10Jdlrobson: Proper edit handling in case of redirects where page does not exist [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116912 [21:43:06] violetto: ? [21:44:16] mhurd: give me 15 mins [21:44:51] violetto: nevermind - i can make them faster [21:46:03] (03CR) 10Jdlrobson: "The tests now do not seem to be running for me. Are they running anywhere? See https://bugzilla.wikimedia.org/show_bug.cgi?id=63090" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119802 (owner: 10Cmcmahon) [21:51:42] (03PS1) 10Yuvipanda: [WIP] Show option to log in or save anonymously when not logged in [apps/android/wikipedia] - 10https://gerrit.wikimedia.org/r/120949 [21:56:58] (03PS2) 10Kaldari: Story 1732: Make licensing agreement customizable [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 [21:58:30] (03CR) 10Kaldari: "Can't merge until WikimediaMessages changes are merged." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 (owner: 10Kaldari) [22:04:39] (03PS3) 10Kaldari: Story 1732: Make licensing agreement customizable [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 [22:07:27] jdlrobson, jgonera : BTW, the license changes are ready for review (https://gerrit.wikimedia.org/r/#/c/120745/). They can't be merged though until my changes to the WikimediaMessages extension are merged (which unfortunately are sitting on my computer at work as I forgot to submit them to gerrit before I left yesterday). [22:08:55] jdlrobson: I'll start working on the Typography changes now unless you've already started. [22:09:03] ergh sorry everyone, just realized my chat client has been closed for the last few hours :| [22:09:31] awjr: Does that mean we can't talk about you any more? [22:09:34] jgonera: i was starting to fix it but haven't had the chance to take it all the way; i have some ideas about how to better approach it if you want to tackle it [22:09:59] kaldari: nono, carry on. i see all the conversation thanks to my chat bouncer anyways :) [22:10:11] mhurd: :P [22:11:14] i was surprised at how quiet everything was seeming... [22:12:25] YuviPanda: hehe ? [22:13:02] mhurd: err, it was supposed to be just :P in response to kaldari and awjr, but somehow your name got stuck before [22:13:19] damn you autocomplete! [22:13:45] YuviPanda: ha! [22:14:33] awjr: heh [22:20:30] awjr, what was the problem with the fix we reverted? [22:20:52] I'm gonna get some lunch [22:20:53] jgonera: it was still possible to enter an infinite loop [22:21:01] jgonera: which apparently happened on foundationwiki [22:21:29] MaxSem: try to knock out https://bugzilla.wikimedia.org/show_bug.cgi?id=61868 when you can? Should be mostly trivial for you.... [22:21:59] awjr, if I have some time I'll ask you how far you've got [22:22:05] coo jgonera [22:22:08] *cool [22:22:38] jgonera: i thought of a relatively easy fix that i just haven't implemented, but suspect there may be a better way to do it [22:27:56] (03PS4) 10Jdlrobson: Story 1732: Make licensing agreement customizable [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120745 (owner: 10Kaldari) [22:28:58] ok night time [22:54:33] chrismcmahon: are you there? could do with your help testing something [22:56:02] jdlrobson: in the middle of something, but what's up? [22:56:47] (03PS1) 10Jdlrobson: See if sleeping helps avoid token issues on Special:MobileOptions [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120959 [22:57:13] ^ chrismcmahon could you see if that gets round the problem with tokens on special mobile options? I can't replicate but I thought it was worth exploring [22:58:27] jdlrobson: I want to see what happens in the builds overnight. We saw no token problems today, I'm wondering if maybe it's an artifact of porting beta from pmtpa to eqiad. [23:03:53] chrismcmahon: mmm that's promising :-) could you update the bug report with what you're seeing and about the eqiad port? [23:07:40] (03CR) 10Jdlrobson: [C: 04-1] "It would be useful to see if this helps anyone who's been experiencing this problem but note this latest update from Chris" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120959 (owner: 10Jdlrobson) [23:08:19] (03CR) 10Jdlrobson: [C: 04-1] "See my patchset 1 comments - not sure if you've commented but I haven't seen any response." [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120864 (owner: 10Kaldari) [23:10:17] (03CR) 10Jdlrobson: Hygiene: Handle fatal errors when loading talk page (031 comment) [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120864 (owner: 10Kaldari) [23:16:59] heh. I just realized I wrote something really dumb in an email to mobile-l . I'll apologize when you bust me. :-) [23:20:33] (03CR) 10Jdlrobson: [C: 032] "Sorry I got confused.. there are instances which should be using this code which are not. We might need to restore this later to include t" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119484 (owner: 10Siebrand) [23:20:39] (03PS3) 10Jdlrobson: Hygiene: Remove unused local variables [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119484 (owner: 10Siebrand) [23:20:43] (03CR) 10jenkins-bot: [V: 04-1] Hygiene: Remove unused local variables [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119484 (owner: 10Siebrand) [23:20:47] (03CR) 10Jdlrobson: [C: 032] Hygiene: Remove unused local variables [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119484 (owner: 10Siebrand) [23:21:45] (03Merged) 10jenkins-bot: Hygiene: Remove unused local variables [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/119484 (owner: 10Siebrand) [23:34:40] jdlrobson, so, https://gerrit.wikimedia.org/r/#/c/116912/7/tests/browser/features/editor_wikitext_nosave.feature [23:34:53] nosave but it looks like it does save, doesn't it? [23:34:59] (03Abandoned) 10Jdlrobson: See if sleeping helps avoid token issues on Special:MobileOptions [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120959 (owner: 10Jdlrobson) [23:38:42] (03CR) 10JGonera: [C: 04-1] "It seems like we're saving but we shouldn't." (031 comment) [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116912 (owner: 10Jdlrobson) [23:45:10] (03CR) 10Jdlrobson: "Hi Kelson, are you still able to make these amendments?" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/117699 (owner: 10Kelson) [23:45:44] (03CR) 10Jdlrobson: "Hey Theo are you still working on this? We miss you and would be great to work with you again!!" [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/108300 (owner: 10Theopolisme) [23:45:57] (03PS6) 10Jdlrobson: Enable Special:MobileMenu for Vector skin [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/118118 [23:46:43] (03PS7) 10Jdlrobson: Enable Special:MobileMenu for Vector skin [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/118118 [23:49:56] (03PS1) 10Jdlrobson: Ignore WIP patches when evaluating [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/120966 [23:53:30] MaxSem, how about using LIMIT of say 100 in that watchlist count SQL? would that make it acceptable? [23:53:43] (talking about https://gerrit.wikimedia.org/r/#/c/116920/4) [23:54:16] I suspect that it will still depend on table hotness [23:54:35] because the original query depended on it HEAVILY [23:55:05] what do you mean by hotness? [23:55:10] so this one will just be faster but worst case performance will be not very shiny [23:55:27] hot row=in memory, cold=on disk [23:56:03] you could ask Sean if this is doable [23:56:12] considering we're querying for recent changes, the chance is pretty good they're "hot" [23:56:22] what's his IRC nickname? [23:57:02] springle. that table is very large so parts of it will be constantly unloaded [23:58:57] (03CR) 10Jdlrobson: Proper edit handling in case of redirects where page does not exist (031 comment) [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116912 (owner: 10Jdlrobson) [23:59:03] (03PS8) 10Jdlrobson: Proper edit handling in case of redirects where page does not exist [extensions/MobileFrontend] - 10https://gerrit.wikimedia.org/r/116912