[00:00:23] MaxSem: could you do me a favor and take a look at https://gerrit.wikimedia.org/r/5394 [00:00:48] i'd like some feedback on the sanity of the approach before i go much further [00:08:39] preilly, you too ^ [00:13:27] awjr: preilly isn't at his desk right now [00:13:46] tfinc no worries, we communicated about it already [00:13:51] k [00:14:42] tfinc: i've got feedback form submissions posting to a remote wiki article working [00:14:51] woot! [00:15:34] tfinc: http://www.mediawiki.org/wiki/MobileFrontend_Feedback_Test [00:15:40] tfinc: I see you pointed jwishnie and co in our direction. [00:16:07] hmm ? [00:16:16] We're already working with them on other projects. [00:16:21] :-) [00:16:24] woot [00:16:31] MaxSem: (cur | http://www.mediawiki.org/w/index.php?title=MobileFrontend_Feedback_Test&diff=527720&oldid=527719) http://www.mediawiki.org/w/index.php?title=MobileFrontend_Feedback_Test&oldid=527720‎ http://www.mediawiki.org/wiki/User:MaxSem (http://www.mediawiki.org/wiki/User_talk:MaxSem | http://www.mediawiki.org/wiki/Special:Contributions/MaxSem)‎ . . (695 bytes) (-92)‎ . . (kthxno) (http://www.mediawiki.org/w/index.php?title=MobileFron [00:16:32] hahaah [00:16:46] hehe [00:16:48] I just saw that [00:17:31] yawn [00:17:38] time to sleep for me [00:17:48] Me too. [00:18:14] gnight jerith [00:18:28] You too, MaxSem. [00:23:19] New patchset: preilly; "Fix for clicking on Wikipedia logo to non-whitelisted language provides no warning" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5396 [00:28:10] awjr: https://gerrit.wikimedia.org/r/#change,5355 [00:28:23] awjr: https://gerrit.wikimedia.org/r/#change,5398 [00:32:25] awjr: thanks! [00:33:19] preilly is there a good way for me to actually test the zero rated changes? [00:33:58] awjr: you can put, " RequestHeader set X-CARRIER "Wikimedia" " in your .htaccess file [00:34:21] awjr: and configure that extension with require_once("extensions/ZeroRatedMobileAccess/ZeroRatedMobileAccess.php"); [00:34:31] awjr: and $wgEnableZeroRatedMobileAccessTesting = true; [00:34:47] awjr: and $wgZeroDisableImages = 1; if you want [00:35:22] groovy [00:35:23] thanks [00:37:56] awjr: do you think that this https://gerrit.wikimedia.org/r/#change,5361 is really supposed to be in master? [00:38:24] preilly im not sure yet i haven't fully reviewed the changes [00:38:39] preilly but considering the commit message, probably not until the rewrite is finished [00:38:51] that kind of thing would be better to do in a remote branch, i think [00:40:05] awjr: yeah, that is the point that I was trying to make [00:40:19] awjr: I wonder if he is still having problems with remote branches or something [00:40:32] awjr: seems a bit dangerous to have as it is [00:40:36] preilly yeah [00:40:42] the remote branches are a bit tricky [00:40:43] awjr: sort of tempted to abandon it [00:41:02] i've had a really hard time with merging changes from master -> remote branch, although i think ive got it dialed now [00:41:15] preilly i think that's legit with an informative comment [00:41:16] awjr: yeah, it can be a pain [00:41:35] Change abandoned: preilly; "Please use the skin-rewrite remote branch for these changes." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5361 [00:42:12] preilly which reminds me - when i merge changes from master -> remote branch, do you prefer i squash the merged commits or just submit them as individual commits? [00:42:34] awjr: squash the merged commits [00:43:09] k [00:45:46] awjr: can you also look at https://gerrit.wikimedia.org/r/#change,5396 [00:45:48] awjr: still problems with the new app pdate? [00:46:04] preilly yeah - i gotta go afk for a second and then i'll look [00:46:11] Thehelpfulone: eh? [00:46:12] awjr: okay [00:46:24] I meant iTunes reviews for wikipedia mobile on iOS [00:46:34] still 1 star reviews, people are getting problems [00:46:42] but I didn't face them when I was using the RC... [00:46:53] Thehelpfulone: i'm actually not sure - tfinc or YuviPanda would know more [00:47:04] ok [00:53:07] Thehelpfulone: some iOS5.1 issues. we haven't been able to replicate any of them [00:53:10] which is frustrating [00:53:28] I'm on iOS 5.1, I didn't see any of them either.. [00:53:45] tfinc: where is the option for me to find your contact details from itunes? [00:54:14] http://en.m.wikipedia.org/w/index.php?title=Special:MobileFeedback&returnto=Main+Page [00:54:18] [00:54:21] not a good title there [00:54:37] who broke the message cache.... [00:54:41] preilly: awjr --^ [00:54:51] i'm also seeing some weird ui glitches in the bet [00:54:52] a [00:55:53] ! [00:56:05] tfinc: we are missing the mobile-frontend-leave-feedback-page-title message key [00:56:24] damn [00:56:25] preilly: and who is fixing that? [00:56:42] i can fix it [00:57:08] preilly is there a good way to cherrypick a change for deployment in git? [00:58:03] hmm there's 'mobile-frontend-leave-feedback-title' => 'Give us feedback about your mobile site experience', [00:58:09] is that what the title used to be? [00:58:23] awjr: it's there twice [00:58:44] 'mobile-frontend-leave-feedback-title' => 'Give us feedback about your mobile site experience', [00:58:47] >_< [00:58:53] 'mobile-frontend-leave-feedback-title' => 'Mobile site feedback', [00:59:07] i presume the latter is correct? [00:59:23] awjr: I think so [00:59:40] awjr: you'll need to scap to refresh the message cache [00:59:48] yeah [01:00:16] awjr: this is a pian [01:00:20] s/pian/pain [01:00:23] preilly can i cherry pick a change? i dont feel awesome about pushing master right now [01:00:57] awjr: I can't really think of a way [01:01:05] shit [01:01:30] awjr: ask Roan [01:05:46] New patchset: awjrichards; "Fix duplicate message and missing page title" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5400 [01:05:55] preilly ^ [01:06:16] awjr: http://wikitech.wikimedia.org/view/How_to_deploy_code#Case_1b:_extension_changes [01:06:25] * awjr reading [01:07:11] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5400 [01:07:13] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5400 [01:07:58] awjr: notice the # Cherry-pick a commit from master, identified by its hash [01:07:58] $ git cherry-pick 176ffdd3b71e463d3ebaa881a6e77b82acba635d [01:08:05] awjr: section [01:08:11] awjr: i think that should do it? [01:08:26] preilly aye but it looks like we need to set up a special deployment branch first [01:09:24] awjr: I don't think so [01:09:30] oh really? [01:09:47] can i just cherrypick the change onto the existing submodule? [01:11:34] awjr: actually I'm not sure [01:11:48] blargh [01:11:51] awjr: I'm asking RoanKattouw [01:12:05] awjr: he is active in #wikimedia-dev [01:12:12] awjr: you could ask him too [01:12:23] ok im paying attn there now [01:12:53] awjr: cool [01:18:46] bah initializing the submodules is taking forever [01:21:57] ok i have the change in SVN for 1.19, still waiting on submodule init before continuing with the instructions for git [01:30:13] preilly not sure how to do this correctly with git review instead of git push [01:30:30] in the example box just above "Now that your extension has a deployment branch", i'm at this step: [01:30:30] $ git push origin wmf/1.20wmf1 [01:30:46] i tried: [01:30:47] awjrichards@virtualbot:~/Dev/wikimedia_git/core/extensions/MobileFrontend$ [wmf/1.20wmf1]$ git review [01:30:47] Creating a git remote called "gerrit" that maps to: [01:30:47] ssh://awjrichards@gerrit.wikimedia.org:29418/mediawiki/extensions/MobileFrontend.git [01:30:47] Errors running git rebase -i remotes/gerrit/wmf/1.20wmf1 [01:30:47] fatal: Needed a single revision [01:30:47] Invalid base [01:31:05] tried with -R also but that had bad results too [01:35:57] preilly you there? [01:36:35] damn. i have to leave soon - i am tempted to scap with the change on the 1.19 wikis and deal with the 1.20 wikis later since this is a nightmare [01:36:39] tfinc, preilly ^ [01:37:39] enwiki is still on 1.19 [06:13:40] YuviPanda: [07:01:39] devgeeks: [07:01:46] just saw your ping [07:02:07] i'm on my daily 'learn lua' time before heading back into the app :) [07:02:27] HAHA [07:02:29] nice [07:02:34] Lua, huh? [07:02:43] yes [07:03:04] so it's a hour or so of lua time, then some fixes/features to some of the python packages I maintain, then the app. [07:05:54] devgeeks: plus mediawiki is moving away from the horribly-hacked-together set of parser functions that people use as a 'primitive' programming language to lua, and I'd like to be able to help :) [07:06:37] hehe [11:05:29] MaxSem: what would be the easiest way to throw away all the contents of a div with class content_block in MobileFormatter.php ? [11:32:26] http://www.mediawiki.org/wiki/MobileFrontend/Dynamic_Sections < MaxSem [11:32:40] and YuviPanda [11:44:08] heya jdlrobson [11:44:09] looking [11:44:46] seems like a good first step [11:44:49] in my head anyway [11:49:19] jdlrobson: so, this is actually a network request per section [11:49:35] for non-js ye [13:24:30] jdlrobson, do you need to throw it away unconditinally? [13:24:55] yes - see http://www.mediawiki.org/wiki/MobileFrontend/Dynamic_Sections for more background [13:25:09] there's also a few questions I have around the api [13:27:29] jdlrobson, have you seen the section information action=mobileview provides to you? [13:27:46] it contains all the information about section hierarchy [13:28:13] sure.. but I don't want to have to make an initial request to do this [13:28:41] so currently I'd load http://m.wikipedia.org/wiki/San%20Francisco with empty sections [13:28:53] then I'd have to call the api to find out what the id's of the first level sections are [13:29:19] I then on loading a section pass the id found to get its html [13:29:40] ideally I don't want to use the api to find out what the first level sections are- I want MobileFormatter.php to annotate the headings with this information so I can just use it straight away [13:29:42] but you'll need to load the lead anyway, so mobileview gives you the ability to request sections with it [13:29:58] yeh but the lead would be provided by the server not the api [13:30:07] how? [13:30:29] well http://www.mediawiki.org/wiki/MobileFrontend/Dynamic_Sections#Proposed_Solution [13:30:36] summary = lead [13:31:04] I object to screen-scraping [13:31:11] I'm not screen scraping.. [13:32:24] no you are - you're proposing an extra infrastructure to load just sections [13:32:58] also, section titles are unreliable and there may be several section with the same title on the page [13:33:15] well we allow editing of individual sections e.g. http://www.mediawiki.org/w/index.php?title=MobileFrontend/Dynamic_Sections&action=edit§ion=1 [13:33:26] all I'm saying is we would provide a way of loading sections individually [13:33:42] e.g. http://www.mediawiki.org/w/index.php?title=MobileFrontend/Dynamic_Sections§ion=1 [13:34:06] and change the meaning of http://www.mediawiki.org/w/index.php?title=MobileFrontend/Dynamic_Sections on the mobile site to mean load the 'lead text' and links to section headings [13:34:12] your proposal contains links to sections with names [13:34:47] well that was just an implementation suggestion - it could be done via querystring or whatever else makes sense [13:35:22] It can be done via querystring now. [13:35:31] (unparsed) [13:36:21] how? [13:36:23] https://www.mediawiki.org/w/index.php?title=MobileFrontend/Dynamic_Sections§ion=1&action=raw [13:36:47] hmm, this proposal has some problems with Varnish purging [13:37:07] well this is why I put it on a wiki page.. for discussion of issues :-) [13:37:21] Amgine: that link seems to download [13:37:56] and is just the wikitext ? [13:38:10] mmm, yah. raw, unparsed. [13:38:11] MW purges HTTP cache on edits, however if you makeits content available via numerous section links, you have no sane way of purging all those links on edit [13:38:31] (parsoid?) [13:38:38] so yeh what I was looking for was a rendered section e.g. http://www.mediawiki.org/wiki/File:MobileFrontendSectionPage.png [13:39:49] I'll think of ways to implement this reasonably [13:40:12] Cool. It was just the idea I wanted to get across as I think it would be a good first step to a more performant mobile site [13:40:13] but no annotations in HTML, plz [13:40:16] ;) [13:40:57] well, did you mean "mobile site that serves content in smaller chunks"? [13:41:28] cuz no wai servers will work faster this way [13:41:39] I see a two-step process can do it, but you'll be transferring the content 3x so possibly more expensive in bandwidth. [13:41:57] why 3x ? [13:42:28] i meant in terms of how much the user downloads - I think currently a typical page is just over 1mb [13:42:58] action=raw -> mobile, mobile -> API:Parsing wikitext, rendered -> mobile. [13:44:22] ok i see - but i'd imagine that most visitors would never invoke that as most visitors will have javascript and use the api instead [13:44:50] Amgine, there's API that returns parsed sections [13:45:13] Oh, then just use that! it'd be silly else. Where is it? [13:46:05] action=mobileview [13:46:40] Undocumented? [13:49:02] Extension:Mobile Frontend. How odd. [13:55:22] MaxSem: is this new API implemented via MF now, or part of 1.20? [13:55:32] MF [13:55:44] Thanks! [14:08:35] New review: MaxSem; "(no comment)" [mediawiki/extensions/MobileFrontend] (contact-us-redesign) C: 0; - https://gerrit.wikimedia.org/r/5394 [14:09:50] brb [16:09:45] jdlrobson, what do you think about https://bugzilla.wikimedia.org/show_bug.cgi?id=36018 - would it make sense to always avoid close tags, or it will make stuff even worse? [16:18:48] greetings all [16:18:52] jdlrobson: are you chatting with karen? [16:18:58] yep [16:19:21] do you want to join us? [16:19:31] yes please [16:19:40] conf room, Skype, other ? [16:19:42] skype [16:44:06] MaxSem why would it make stuff worse? [16:44:32] for example,

[16:47:59] jdlrobson, ^ [16:49:29] how would that happen? If someone puts it in wikitext or can wikitext output that? [16:51:13] jdlrobson: hi Jon, did you see my email from about 10 hours ago? [16:51:44] hi philinje yeh it wasn't so clear so was hoping to have a chat over skype about it [16:52:05] ok, after the stand-up? [16:52:12] yup [16:52:29] great [16:52:49] jdlrobson, empty linescan potentially result in empty

s [16:53:06] s/linescan/lines can/ [16:54:05] mm that's annoying [16:57:55] is that common MaxSem [16:58:10] what, empty lines? :) [16:58:10] if it's not common I'd say don't worry about it [16:58:16] has the standup started, etc? [16:58:24] no outputting

instead of
[16:58:26] YuviPanda: 2min [16:58:33] YuviPanda: you just made it [16:58:34] congrats [16:58:35] ;) [16:58:57] :) [17:00:10] all right … lets do our stand up [17:06:11] http://www.mediawiki.org/wiki/MobileFrontend/Dynamic_Sections [17:07:03] jdlrobson: thanks [17:09:31] philinje: please give me 10 mins to finish up something :) [17:13:01] sure [17:13:10] YuviPanda: so we've been bad about one thing in the apps [17:13:24] YuviPanda: links to attribution … and we need to fix it [17:13:40] current versions - fix in MF will carry over [17:13:56] YuviPanda: philinje suggested this http://www.mediawiki.org/wiki/Mobile_design/Contributors but i'm wondering if we shouldn't just stash it in the menu for now [17:14:19] not a fan of stashing page-specific info in the menu - putting it in the footer would be good enough? [17:14:21] tfinc: this is fixed right? https://bugzilla.wikimedia.org/show_bug.cgi?id=32596 [17:14:21] * YuviPanda looks at link [17:14:21] we have a good debating going with pete forsyth (sp?) about it [17:14:47] YuviPanda: the footer is not a dumping ground for all things known to man [17:14:51] thats what were trying to avoid [17:14:55] neither is the menu [17:14:58] plus the foots has no context of the current aritcle [17:14:59] plus, on iOS? [17:15:04] yes but i need to find a place for it [17:15:15] and we need it before the new nav shows up [17:15:20] plus I'd say that a menu item is *too* prominent [17:15:23] agreed. [17:15:44] YuviPanda: i'm open to alternatives if you have them [17:16:41] tfinc: http://www.mediawiki.org/wiki/File:Mobile_Contributors_A.png [17:16:43] i like that [17:16:56] New patchset: Jdlrobson; "move problematic css line into iphone.css" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5442 [17:17:11] jdlrobson: take a quick look and let me know what you think too of http://www.mediawiki.org/wiki/Mobile_design/Contributors too [17:17:21] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5442 [17:17:23] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5442 [17:17:26] looks ugly as a link [17:17:56] YuviPanda: cool [17:17:57] jdlrobson: yes. which is why i'm discussion alternatives like in the app menus [17:18:03] I'd expect an 'about this page' link [17:18:13] which when clicked would give me information on contributors, license [17:18:18] jdlrobson: Lindsey's original idea was to make it look like a section heading but it just opens a page [17:18:38] awjr, or MaxSem: https://gerrit.wikimedia.org/r/#change,5396 [17:18:42] jdlrobson: whatever text we put will *only* take us to the desktop article history page [17:18:57] jdlrobson: do you mean the word "Contributors" is not right? [17:19:28] correct - it just looks like we couldn't find the right place to put it so just chucked it down there [17:19:41] I guess long term this would go in the main menu [17:19:57] hard to find a right place in the new UI [17:20:04] I'd expect a page authors or about page link to take me to a page which contains the contributors [17:20:21] but article-specific [17:20:47] another option... [17:20:49] so lets keep things organized we need to [17:20:57] 1) we need to put this into the current beta nav [17:21:03] 2) we need to put it in the future nav [17:21:07] lets not worry about #2 now [17:21:08] put a page flip [17:21:09] http://marianagness.blogg.se/images/2010/page-turn-3d_113630597.jpg [17:21:13] (without the man) [17:21:18] in the bottom right [17:21:19] just the current beta design thats been in testing for a while [17:21:27] that takes you to information about that specific page [17:21:28] along with our apps [17:21:44] ohh … that'd be interesting [17:21:46] tfinc: i'll note that it should be easy to pick up whatever design the beta gets into the App. [17:21:47] New review: MaxSem; "Instead of using escapeAppendQuery() and then fixing double escaping, just use escapeAppend()" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/5396 [17:21:48] we could have information such as revisions [17:21:50] contributors [17:21:52] history etc. [17:22:16] sure. i just don't want us to worry about the new nav attribution right now. we can solve that problem later [17:22:22] for now we just need something for the current beta design and apps [17:22:49] jdlrobson: nice idea, but still need the fallback version [17:22:50] or some kind of watermark if we don't like the page turn [17:23:00] fallback version? [17:23:05] feature phones [17:23:19] if the link was smaller, more like the text size in the footer? [17:23:30] it would work the same just not animate... [17:23:39] jdlrobson: i see [17:23:45] i like it because it's not too imposing [17:23:51] yeah [17:23:54] but it's playful enough that someone will click it to see what happens [17:24:31] hmm, liking that [17:24:38] lets mock it up [17:24:39] I think an icon on one of the corners is what I'm getting at fundamentally [17:24:45] yes [17:25:06] let's ask Heather to mock that up [17:25:11] ready for the call? [17:25:34] next deployment is Monday right? [17:25:34] jdlrobson: yes [17:25:36] MaxSem: what is escapeAppend? [17:26:14] bleh, I meant appendQuery() [17:27:08] so yep philinje ready now [17:27:22] ok, give me a sec [17:28:46] jdlrobson: https://bugzilla.wikimedia.org/show_bug.cgi?id=32596 is fixed. [17:28:57] \o/ srikanthlogic [17:29:04] thanks for sanity checking me :) [17:29:09] as for communication, people who are bothered about mobile hompage would have made the change by now (6 months) and gerard blogged about it too! http://ultimategerardm.blogspot.in/2012/02/have-mobile-have-menu.html [17:29:26] ah thanks hadn't seen that! [17:33:09] New patchset: preilly; "Fix for clicking on Wikipedia logo to non-whitelisted language provides no warning" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5396 [17:34:27] jdlrobson: you know of a way where I could put a section only on mobile homepage not on the regular one ? [17:35:03] i'm not seeing us on the deployment calendar for next week [17:35:09] do we want mon or tues? [17:35:17] i assumed we were keeping monday [17:35:21] so that jon could be around [17:35:53] New patchset: preilly; "Fix for clicking on Wikipedia logo to non-whitelisted language provides no warning" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5396 [17:36:27] MaxSem: can you take a look at: https://gerrit.wikimedia.org/r/#change,5396 again [17:36:34] awjr: --^ [17:36:51] tfinc one sec im working with roan i'll take a look when im done [17:36:55] k [17:41:01] tfinc we're still on the calendar for mondays as per usual under 'Ongoing' [17:41:20] MaxSem: can you take a look at: https://gerrit.wikimedia.org/r/#change,5396 again [17:41:41] awjr: there it is [17:41:47] thanks [17:42:04] no problem [17:42:41] New review: MaxSem; "(no comment)" [mediawiki/extensions/MobileFrontend] (master) C: 1; - https://gerrit.wikimedia.org/r/5396 [17:46:02] preilly binasher: what would it take to get a graph of varnish 503's for mobile ? [17:46:13] i just got one but can't replicate it [17:50:50] preilly & awjr, any commenta regarding my rewrite? [17:52:45] sorry MaxSem not yet, i've still got my hands full [17:56:10] awjr: can you approve https://gerrit.wikimedia.org/r/#change,5396 [17:56:21] MaxSem: no comments yet [17:56:35] * preilly New patchset: preilly; "Add ACL for new carriers and redirect support for carriers landing page on m. domain" [operations/puppet] (production) - https://gerrit.wikimedia.org/r/5447 [17:57:01] preilly in a bit, just wrapping up this push to the 1.19 wikis [17:57:13] awjr: okay [17:58:32] awjr, not the best time as there are problems on cluster [17:58:41] HTTPS is down [17:58:50] MaxSem i was given the go-ahead [17:58:53] preilly: thoughts on my 503 question ? [17:59:26] tfinc: I'm not sure [17:59:33] albeit i think i was given the goahead and started scap just before the HTTPS problems showed up [17:59:37] tfinc: I'd need to investigate [18:00:07] generally i wouldn't be worried but i can't reproduce it which leaves me unsure about the state [18:00:09] i'll file a bug [18:01:31] tfinc: okay [18:02:14] tfinc: so, it was a white screen HTTP Error 503 - Service unavailable error from Varnish right? [18:02:56] tfinc: did you catch the Guru Meditation: XID? [18:05:27] LieFI [18:05:30] preilly: yes [18:07:19] tfinc: If the error happened just a short time ago the transaction might still be in the shared memory log segment. [18:07:19] so, we might be able to: [18:07:20] $ varnishlog -d -c -m TxStatus:503 [18:08:04] awjr: so before i disappear into the weekend i'm going to look into the contact us weirdness with you [18:08:16] jdlrobson sure [18:08:25] good idea [18:09:50] and member:srikanthlogic: sorry been on a call - I'm not sure it's possible to put a section only on mobile homepage but not on regular one... [18:10:23] I might be wrong though srikanthlogic .. [18:11:01] jdlrobson sorry my attn is a bit bifurcated [18:11:20] srikanthlogic > might be worth mentioning in discussions on https://bugzilla.wikimedia.org/show_bug.cgi?id=30405 [18:11:25] ok awjr [18:11:27] jut fetching [18:12:42] my local instance has got very very slow for some reason [18:13:36] * srikanthlogic clicks [18:14:56] jdlrobson: yeah, it is really bad [18:15:01] so interestingly login reponse awjr is html error code help [18:15:10] ! [18:15:15] anyi ideas preilly what happening? [18:15:18] that is weird [18:15:20] no [18:16:43] awjr ! [18:16:50] wow. jdlrobson it is working for me [18:17:10] oh wait maybe not [18:17:11] one sec [18:17:33] jdlrobson: if I disable memcache this is how long it takes: [18:17:34] 13 requests ❘ 107.97KB transferred ❘ 1.0min (onload: 1.0min, DOMContentLoaded: 1.0min) [18:17:42] :-O [18:18:13] just the page latency is 50.27s [18:20:06] ack jdlrobson sorry my instance is out of whack [18:21:30] :( [18:21:57] a by product of git deployment stuff [18:22:05] it'll just take me a sec to get this sorted [18:24:08] er or more [18:24:28] yuck crap network speeds [18:24:59] jdlrobson in the meantime, can you private pastebin me your pertinent local settings and what you're getting from the var_dumps? [18:27:23] jdlrobson do you know which api call was resulting in   ? [18:27:33] to login [18:28:05] heh that is conveniently not listed as a possible error code in the docs [18:28:05] http://www.mediawiki.org/wiki/API:Login#Errors [18:29:11] help is a general error code [18:29:22] MaxSem what does it mean? [18:29:22] not login-specific [18:30:07] awjr, so what are your API problems? [18:30:36] MaxSem not entirely sure yet, trying to debug jdlrobson's problems with the remote wiki writing changeset [18:31:27] we need to reach him PHP XD [18:31:44] s/reach/teach/ [18:32:17] jdlrobson: pull req sent, look into it when you can. [18:32:20] * YuviPanda goes off for the night [18:32:23] good night everyone [18:33:44] jdlrobson it is 100% working for me [18:33:54] how strange.. [18:33:54] mm [18:34:03] jdlrobson can you show me your configuration? [18:34:21] yeh don't think there is anything private in there [18:34:31] jdlrobson actually [18:34:36] username/pass for our test account [18:34:50] on mw.org [18:35:03] ls [18:35:05] oops ww [18:35:38] true i'll comment that out [18:35:48] sorry not comment [18:35:49] change [18:35:55] actually would be good tos ee that too [18:35:58] maybe just email me what you've got [18:39:02] jdlrobson i just swapped my pertinent config stuff with yours and it works for me as well [18:42:21] jdlrobson while you're getting food, im going to run and grab a sandiwch real quick. if you get back before me, make sure your MW is up to date and try again i guess. if it still doesnt work, it would be useful to see var_dumps of $req and response vars [18:45:07] after some poking, error=help happens when no action has been provided [18:58:26] MaxSem yeah it seems to just pull up the api help documentation [19:03:22] New patchset: L10n-bot; "Localisation updates from http://translatewiki.net." [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5471 [19:03:26] New review: gerrit2; "Auto-approving/merging l10n updates" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5471 [19:03:26] Change merged: gerrit2; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5471 [19:21:36] jdlrobson: back yet? [19:22:41] yup [19:22:52] yo [19:23:14] jdlrobson: what version of mw core are you using? [19:24:02] dev version [19:24:20] is it up to date? [19:24:29] i'll do a git pull now [19:24:48] cool then try again and lmk what happens [19:24:57] yep still same problem [19:26:16] jdlrobson can you add "var_dump( $req );" on line 247 as well as a var_dump( $loginResponse ); after $loginResponse = json_decode( $loginResponse, true ); [19:26:28] and pastebin the oubput? [19:27:10] sure [19:28:21] https://raw.github.com/gist/7ccfeda133a94ff36092/698f5a7179527ffe8a4a95b85bb7e2245676f9a8/gistfile1.txt [19:29:18] anything suspect there? [19:29:31] just the null [19:30:07] jdlrobson can you try the same again but move the var_dump( $loginResponse ) to just after $loginResponse = $req->getContent() [19:30:48] that was the help stuff i had before.. [19:31:12] jdlrobson do you have php compiled with json support? [19:31:29] i have whatever preilly gave me :) so possibly not... [19:31:43] how do i check if I have json support? [19:32:03] i would expect php complain if you didn't because of the json_decode() function [19:33:24] jdlrobson: from the command line, php -i | grep json [19:33:35] should see something like this: [19:33:35] awjrichards@virtualbot:~/websrv/mobile_dev$ [master]$ php -i | grep json [19:33:36] json [19:33:36] json support => enabled [19:33:36] json version => 1.2.1 [19:33:53] yeh it's enabled [19:34:06] jdlrobson iirc you were getting xml back [19:34:26] rather than json? [19:35:38] seems like it [19:35:51] but it's acting like the wrong parameters are being passed [19:36:24] yeah [19:38:41] jdlrobson what happens if you just go to http://www.mediawiki.org/w/api.php ? [19:38:46] in your browser i mean [19:38:54] same [19:39:02] it's like the query parameters are not passed.. [19:40:21] MaxSem or preilly, can one of you checkout the changeset and see if it works for you? [19:40:46] linSmith: http://touchstoneclimbing.blogspot.com/2011/10/project-bandaloop.html seek to about 5:00 minutes in [19:40:53] linSmith: my class will be indoors today [19:41:02] but it'll give you an idea of whats its like [19:41:56] jdlrobson just for kicks can you do "php -l > phpinfo.txt" and mail it to me? i have a hunch there's a loser level configuration issue at play [19:42:19] im gonna have to take off in a minute and prep for an interview [19:42:22] but that way i can keep digign [19:44:22] jdlrobson and quickly can you show me output of "php -i | grep curl" [19:44:34] awjr, what revision, API editing? [19:44:52] MaxSem: https://gerrit.wikimedia.org/r/#change,5394 [19:45:05] MaxSem: i sent you pertinent config settings yesterday for it [19:45:22] im going to go prep for this interview - i will likely be unresponsive for a while [19:51:39] k awjr now it is saying 'thanks for your feedback' but not actually saving to your mediawiki [19:51:41] this is just weird [19:53:40] heads up that i'm still seeing 503's sporadically . preilly is investigating but if you see them let us know [19:55:26] jdlrobson: does MF still seem really slow to you? [19:55:35] jdlrobson: sigh. hopefully your output from php -i will be illuminating [19:55:45] it's very slow.. [19:56:15] jdlrobson: on your local instance right? [19:56:45] i do not have curl set up with my version of php - which means my reuqests are going through fopen. but if you have curl set up with php, it will use curl instead. i wonder if that's where the issue is. i'll have to test against curl too because im pretty sure that's what in use on the cluster [19:57:06] yup preilly [19:57:09] even with memcached [19:58:36] jdlrobson: is your checkout of core current? [19:58:51] yupp [19:59:16] awjr, jdlrobson, https://www.mediawiki.org/w/index.php?title=MobileFrontend_Feedback_Test&curid=86459&diff=527986&oldid=527720 [20:00:06] OuKB: can you do 'php -i | grep curl' and post your output? [20:00:28] ok going into interview, bbiab [20:00:43] awjr: might not be here when you are back but will mail you what you asked fo [20:00:53] awjr, emptiness ;) [21:01:46] jdlrobson still with us? [21:01:47] oh [21:01:48] no [21:01:53] you just signed off. [21:02:16] MaxSem: i think the problem is related to php curl [21:02:38] you and i did not have php configured with curl support, jdlrobson did [21:02:53] i just set up php with curl locally and am now encountering the same problem that jdlrobson did [21:06:40] chrismcmahon, flooder! :P [21:12:11] New review: awjrichards; "These changes do not work as expected when PHP is configured with cURL." [mediawiki/extensions/MobileFrontend] (contact-us-redesign); V: -1 C: -1; - https://gerrit.wikimedia.org/r/5394 [21:19:00] MaxSem: sorry, I don't know why this IRC client has such an issue with freenode. don't know whether to shoot pidgin, freenode, Comcast, or all three [21:19:18] or possibly Linux [21:26:28] New patchset: awjrichards; "Modest cleanup of html handling; getFeedbackHtml() now returns html rather than piping it directly to getOutput. This now handled in execute() method." [mediawiki/extensions/MobileFrontend] (contact-us-redesign) - https://gerrit.wikimedia.org/r/5394 [21:27:49] MaxSem: found the issue. it's because i was setting 'method' => 'post' rather than 'method' => 'POST' for the MWHttpRequests, which was causing cURL to make get requests instead of posts [21:27:54] dumb. [21:28:10] and a bug [21:30:05] aye [21:37:17] MaxSem yeah the HttpRequest code has a lot of $this->method == 'POST' [21:37:25] i am going to fix that now. [21:50:23] philinje: want to tell me about the mockup? [21:52:06] also, not sure what has been decided but the footer is specifically not tied to any article [22:18:40] jerith: ping [22:18:54] preilly: Pong. [22:19:13] jerith: how do I start vumi on the vumi-gw1 instance? [22:19:33] Is that the new one? [22:19:42] jerith: the old one [22:19:51] Ah. [22:20:13] jerith: I actually can't get the new one to work [22:20:20] jerith: supervisor starts it [22:20:30] jerith: but, the jabber accounts don't seem active [22:20:31] MaxSem: i submitted a fix. [22:20:44] preilly: That's odd. [22:21:01] Let me log into vumi-gw1 and check the details. [22:21:44] jerith: actually can you check the vumi instance [22:21:53] jerith: not the gw1 one [22:22:31] How do I get to it? [22:24:20] same way [22:24:31] just use vumi name instead of vumi-gw1 [22:24:40] Cool. [22:26:49] Hrm. [22:27:08] Looks like supervisord's running, but not listening on the control port. [22:27:43] jerith: okay, want to kill it? [22:28:37] Actually, it looks like it's the system supervisord. [22:28:51] And we can probably use that instead of running another instance. [22:29:30] Is it safe for me to edit those configs? [22:29:42] jerith: yes [22:29:44] The ones in /etc/vumi, that is. [22:29:58] jerith: yes [22:30:04] They aren't going to be replaced out from under me by puppet or anything? [22:30:12] no [22:30:19] Cool. [22:33:32] Seems to be running now. Do you see the XMPP? [22:34:07] * jerith fires up Adium to check. [22:35:28] Yay, looks happy. \o/ [22:35:34] So, here's what I did. [22:36:05] I commented out all the non-program block in the vumi supervisord config. [22:36:32] Then I symlinked it into /etc/supervisor/conf.d/ [22:37:10] The "sudo supervisorctl" to get into the control shell. [22:37:32] "reload" to tell it to reread all its configs and restart all supervised processes. [22:37:53] Since the only config there is the vumi one, that means it started all the vumi processes. [22:39:04] So you probably want to update puppet to deploy a smaller supervisord.wikipedia.conf directly into /etc/supervisor/conf.d or something. [22:39:28] And then issue a "supervisorctl reload" if it's changed or something. [22:39:55] preilly: Make sense? [22:40:33] Um. There's something else that needs fixing. [22:40:48] System supervisord runs as root. [22:40:56] So does everything it starts. [22:42:46] You probably want to set "user=someuser" and "autostart=true" for each of the program blocks as well. [22:43:21] http://supervisord.org/configuration.html#program-x-section-settings for more config options. [22:47:33] jerith: yes, I see the XMPP [22:48:02] preilly: I need to pack up and head home now. [22:48:20] If you want to puppet that up or something, I'll probably be back online in about 15-20 minutes. [22:48:31] jerith: okay [22:48:34] jerith: thanks [22:48:56] preilly: Not for very long, though. It's bedtime here. :-) [22:51:24] jerith: it's okay we can do it some other time [22:51:41] preilly: Cool. Enjoy! [22:56:48] awjr: can you approve https://gerrit.wikimedia.org/r/#change,5396 [23:02:10] preilly: looking [23:03:45] New review: awjrichards; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5396 [23:03:47] Change merged: awjrichards; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/5396 [23:04:22] awjr: thanks! [23:24:13] * preilly New review: preilly; "(no comment)" [mediawiki/core] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/5498 [23:24:13] Change merged: preilly; [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/5498 [23:24:51] thanks preilly [23:25:05] awjr: np [23:53:01] doh coworking space closes at 5 gonna head home. have a great weekend everyone