[02:00:11] Hello [02:06:44] Can someone here please help me with the config for CentralAuth? [02:07:08] Installation is already done successfully. [04:22:37] Nice title="" http://cl.ly/9MKa [04:22:51] (On securepoll) [04:24:36] that's bug 30399 [04:25:00] ah [16:43:28] 20 min to mobile triage here http://hexm.de/5w [16:50:03] guillom, she's online now [16:50:11] thanks sam [16:59:28] 1min! [16:59:50] *hexmode looks around for tfinc or preilly [17:00:00] ah, preilly! [17:00:05] I'm here [17:00:11] :) [17:02:46] tfinc: ! [17:02:51] hexmode: ! [17:02:56] greetings all [17:03:13] tfinc: I'm not sure who is here for the triage [17:03:15] but [17:03:26] http://etherpad.wikimedia.org/BugTriage-2011-08 [17:03:31] *aharoni waves at hexmode [17:03:36] link to etherpad [17:03:42] aharoni: ! [17:03:49] hexmode: i added difficulties last night to most of the bugs [17:04:09] tfinc: excellent [17:04:41] hexmode: so how would you like to run this ? [17:04:44] aharoni: since you're here for the triage, are you more interested in bugs or adding features? [17:05:32] hi all [17:05:37] tfinc: so, if we can find someone interested in new features... [17:05:42] Nikerabbit: hey! [17:05:46] I'm here but I cant do much from my phone :/ [17:06:06] I recall requests for this being WONTFIXed in the main site [17:06:13] johnduhart-phone: k, so let us know: features or bugs and we'll focus on those [17:06:20] hexmode: Unfortunately, i hardly had time to go over the bugs. I'm here, because one of the he.wp CSS/JS guys requested that i ask here about a particular mobile feature. [17:06:36] Nikerabbit: also, bugs or features? [17:06:52] aharoni: this would be the place :) [17:07:31] aharoni: you can ask now: what feature? [17:07:49] I think we should focus on bugs, shiny new things can come later IMHO [17:08:13] Bryan, I'd concur. Especiailly as we don't have it on normal wikipedia et al [17:08:21] eg. what's left holding this back [17:08:24] hexmode: just following for now [17:08:52] there are plenty of bugs in the etherapd link if you scroll to the bottom [17:09:13] starting on line 50 then [17:09:14] if i understand correctly, there are project-local CSS files such as http://he.m.wikipedia.org/stylesheets/android.css . [17:09:18] for instance http://bugzilla.wikimedia.org/30118 [17:09:19] or [17:09:24] can they be customized by the admins of a project? [17:09:30] http://bugzilla.wikimedia.org/30142 [17:10:08] preilly: tfinc: project local CSS files for android/iphone, etc? [17:10:11] or http://bugzilla.wikimedia.org/22346 [17:10:16] are just some ones that jump to mind [17:10:23] but there are lots of bugs to take [17:10:54] aharoni: i think the first step would be to support the current set of favicons already present in each project [17:11:02] scale them up to the default image size [17:11:16] next we could contact the individual projects for testing and see how well it worked for them [17:12:06] assign 30118 to me I'll do it later today [17:12:17] johnduhart-phone: :) [17:12:27] will do [17:12:47] hexmode: side note: i had to prune a lot of already closed bugs from the list. there might be others. we should look over the queries were using since they do pick up some already resolved bugs [17:12:59] thanks johnduhart-phone [17:13:25] tfinc: yeah, I might have had resolved issues on there ... my bad. [17:14:13] is there any bug about the project-local CSS? [17:14:35] ok .. another easy bug to pick up would be http://bugzilla.wikimedia.org/29869 [17:14:38] hexmode: I don't think so [17:14:50] auto hiding the url bar on phones that support it [17:15:09] it would give back valuable real estate on the users screen :) [17:15:17] aharoni: could you create a bug for CSS so we have a place to find it? [17:15:32] tfinc: I thought that was fixed already [17:15:32] hexmode: yes. http://bugzilla.wikimedia.org/22659 [17:15:49] hexmode: i probably will create a bug, but i have to understand a couple of things myself, first. [17:16:03] preilly: i haven't seen that fixed and the bug is till open. [17:16:07] aharoni: it is https://bugzilla.wikimedia.org/show_bug.cgi?id=22659 [17:16:16] it you can show it to me working then i'd love to resolve it [17:16:35] i'm assuming your talking about auto collapse [17:16:44] hexmode: thanks. [17:16:51] tfinc: is android.css only for favicons? I would think that anything can go there. [17:17:02] preilly: --^ [17:17:22] tfinc: #22659 only says handheld.css [17:17:29] aharoni: anything could [17:17:44] is there a separate one for android/iphone, etc? [17:18:23] hexmode: there are http://he.m.wikipedia.org/stylesheets/android.css and http://he.m.wikipedia.org/stylesheets/iphone.css [17:18:34] they exist; i'm not sure how exactly they are used. [17:20:04] these are the ones that we use: [17:20:05] http://bits.wikimedia.org/w/extensions-1.17/MobileFrontend/stylesheets/iphone.css [17:20:18] and http://bits.wikimedia.org/w/extensions-1.17/MobileFrontend/stylesheets/android.css [17:20:36] preilly: so those aren't over-ridable by admins, then ? [17:21:02] hexmode: not currently [17:21:14] can we do something like Mediawiki:android.css that would be loaded after that? [17:21:53] hexmode: we could add additional css as well [17:22:12] k... I'll file a bug for it so we can track it [17:22:42] aharoni: anything else re project css you think we should talk about? [17:23:20] that's the thing i was asked to the mobile devs about. [17:23:36] moving on to bugs we need some people to check [17:23:46] maybe johnduhart-phone can do this from his phone ;) [17:23:52] ha ha [17:23:57] https://bugzilla.wikimedia.org/show_bug.cgi?id=30293 [17:24:08] hah sure [17:24:25] *hexmode opens it [17:24:45] one important point .. when you guys get to testing. please make sure to opt in to our beta. i have a blog post pending about it [17:24:50] but just follow this link tinyurl.com/woptin [17:25:10] johnduhart-phone: http://tinyurl.com/woptin [17:25:22] preilly: could we get something a bit more visible on the beta so that people can easily know their in it [17:25:38] i've gotten some confused souls who don't like having to scroll all the way to the bottom [17:25:44] something at the top would be nice [17:26:04] tfinc: sure I'll add this: http://www.wp7connect.com/wp-content/uploads/2011/06/BetaTesting.jpg [17:26:18] tfinc: if you opt into the beta, then does the .m. go to the beta instead of ruby? [17:26:43] opted in [17:26:50] tfinc: okay [17:27:27] hexmode: correct [17:27:39] and the blog post explains but i had to pull it last night for some last minute changes [17:27:42] it'll go back up today [17:27:43] k, so I'll test [17:28:38] johnduhart-phone: click "view this page on regular Wikipedia" link [17:28:58] the view this page on regular link seems to work. it goes to the normal Wikipedia and then if I click a link there it sends me back to mobile [17:29:15] excellent [17:29:30] tfinc: good enough to close WORKSFORME? [17:29:53] sure, just mention that we tried it out and couldn't verify. i'm sure they will re-open if its still an issue [17:30:06] *hexmode closes [17:30:27] tfinc: while I do that ... are there other things you want people to look at [17:30:41] im about to get home [17:31:01] :( .... [17:31:06] those are tears [17:31:46] or maybe I've got something in my beard [17:32:03] Nikerabbit: got a phone you can test things on? [17:32:24] aharoni: how about you? [17:32:25] *tfinc goes back to the list [17:32:34] Tada [17:32:38] don't worry ill still be here :-) [17:32:40] does anyone have kindle here? [17:32:41] Here I am [17:32:45] johnduhart-phone: ! [17:32:48] My mum does [17:32:51] But she's 30 miles away [17:32:55] johnduhart: stop it! [17:33:04] http://bugzilla.wikimedia.org/28181 for any kindle owners [17:33:11] Reedy: time to peddle! [17:33:12] I also got the iPhone sim to play with as well [17:33:21] Peddle? My car is sat outside [17:34:06] tfinc, presumably we still need a user-agent string? [17:34:17] hexmode: phone? I have nokia E71 with built-in browser and opera mini [17:34:26] just joining now. Need someone to test with an N900 running microb as a browser? [17:35:33] Nikerabbit: it would be good for you to just opt in to our beta and give it a go [17:35:49] same goes for sumanah [17:35:55] is the e71 an s60 ? [17:36:08] link to opt in: http://tinyurl.com/woptin [17:36:09] sumanah: just visit http://tinyurl.com/woptin to optin [17:36:32] tfinc: yes it is symbian [17:36:40] tfinc: yes, it is a 2nd generation S60 [17:39:24] ! [17:40:51] Nikerabbit: could you test the mobile beta in opera mini? [17:41:06] DJHartman said it was broken there [17:42:23] johnduhart: do you have opera mini? [17:42:38] Uhh for andriod? No but I could get it [17:42:46] *hexmode realizes he installed it recently [17:43:00] :p [17:44:54] http://bugzilla.wikimedia.org/28515 is very important, but tricky. Santhosh probably knows more about this, but here's what i can say: [17:45:54] my Samsung Monte phone doesn't seem to support web fonts at all, and that is true for many other devices. [17:46:03] santosh was working on dynamic font support for desktops, right? [17:46:06] hexmode: is there something specific I should test? [17:46:40] hexmode: yes, but from what i understand the same extension can work for mobile with little modification. [17:46:58] Nikerabbit: https://bugzilla.wikimedia.org/show_bug.cgi?id=30356 [17:47:49] that is, if you disregard the fact that lots of phones don't support CSS web fonts at all and can only display one or two fonts with which they are pre-installed. [17:47:51] aharoni: do you know if santosh could *do* the mobile support? [17:47:59] that includes lots of Androids. [17:48:35] or could he advise/help preilly if he couldn't do it? [17:48:36] hexmode: I don't see those problems, not in the mobile beta at least [17:48:43] hexmode: i would guess that he can do the basic adaptation of the extension to the mobile site. [17:49:50] tfinc: preilly: I think http://bugzilla.wikimedia.org/28515 is something we should discuss with santosh [17:49:58] but it would only work on phones that support web fonts. there is another possibility - to render the site on the server as an image or something and send it to the user. Santhosh is aware of it, but it's far more complicated. [17:50:12] :P [17:50:25] aharoni: and you'd have to figure out screen size [17:50:48] and render for 100s(?) of different screen sizes [17:50:50] ugh [17:50:54] not pretty [17:51:44] sumanah: did you manage to pull up the mobile beta? [17:52:04] hexmode: no, didn't try yet. didn't know it was a now thing [17:52:04] we need to research what is the extend we can reasonably go supporting webfonts in mobile devices [17:52:10] *hexmode starts opera mini [17:52:24] sumanah: if you don't have time now, then don't worry about it [17:52:43] just trying to get more eyes while we have people here [17:53:29] hexmode: I have opted in. [17:55:00] preilly: tfinc: was there any bugs sumanah could look at? [17:55:04] just to make sure: nomad.tesla.usability.wikimedia.org is not needed for testing anymore, isn't it? if i opt in at he.m.wikipedia.org , do i become a beta tester of the newest mobile site? [17:55:14] hexmode: let me take a look give me a minute [17:55:23] I know we're 5 minutes from the end of the meeting [17:55:24] aharoni: correct [17:55:25] *hexmode goes back to look at sumanah's device time [17:55:31] type [17:55:49] aharoni: yes. it is not needed any longer [17:55:51] and just to reiterate .. for anyone confused about the new beta .. blog post coming back up later today that will explain the details and feedback pages [17:56:42] def prefer mobile version to regular in opera mini :) [17:56:51] :) [17:56:53] hurrah! [17:56:53] tfinc: cool. so i'll just use this opportunity to say that until he.m worked for me well and without RTL problems on my Samsung phone. [17:56:57] and it's nice to see the strings i translated yesterday at translatewiki are already there :) [17:57:36] aharoni: it would be awesome to get your to test he .. we actually have it on our feedback page http://www.mediawiki.org/wiki/Mobile_site_rewrite/Testing [17:57:42] i'm going to redo that page today as well [17:58:17] and clicking "view regular" then following links in opera mini WORKSFORME [18:00:41] new bug! [18:01:06] preilly: clicking disable images on mobile site disables them [18:01:26] but then, clicking enable doesn't enable them. [18:01:46] hexmode: really? [18:01:51] and shows the link to enable again on the new page [18:02:09] preilly: really... at least in opera mini [18:02:30] Works on iOS safari [18:03:02] It also just worked for me [18:03:17] tfinc: i browsed a couple of articles in he.m.wp and it worked well. is there anything in particular that i should test? [18:03:34] hexmode: thanks for much running this. i have to step out to a weekly meeting in #kiwix for our offline reader but i'm sure preilly can stick around and lead this for me [18:03:51] tfinc: yes [18:03:58] tfinc: I'm about done now... but we can continue [18:04:18] who else had opera? [18:04:40] Nikerabbit: ^^ [18:04:49] I've opera on my android phone [18:04:56] have you opt'd in, Nikerabbit? [18:04:59] yes [18:05:03] (mini) [18:05:10] if you hit disable images [18:05:22] and then try "enable" images [18:05:27] do they come back? [18:06:54] Reedy: could you test enabling and disabling images when you've put the fire out? [18:08:46] hrm... weird [18:09:03] preilly: following the link doesn't re-load the images [18:09:18] but hitting reload after that shows them [18:09:51] hexmode: Okay, that makes sense [18:09:59] I'll write this up in the bug. sounds like a caching issue [18:10:06] hexmode: yes [18:10:19] I can confirm this as well [18:10:45] johnduhart: are you using opera (I forget) [18:11:05] Safari on iOS (iPhone Simulator) [18:11:27] :) [18:12:03] preilly: I'm done here after I write the bug and type up my report... anything else? [18:12:27] hexmode: nope, I think we are good for now [18:12:39] aharoni: also, good to see you here. I'll try this time from now on for Europe [18:12:47] hexmode: I really appreciate you leading this triage [18:12:55] :) [18:13:37] you're welcome. [18:14:44] Hmmm [18:15:06] I enabled mobilefrontend on my test wiki and it slows down loading to taking 33 seconds [18:15:49] ExtMobileFrontend::beforePageDisplayHTML 1 24127.428 24127.428 66.032% 3226057 ( 24127.428 - 24127.428) [1] [18:16:21] johnduhart: that might be due to WURFL [18:16:27] Maybe [18:16:40] is that only on the first req? [18:16:47] johnduhart: is it that slow on every request? [18:17:01] Yeah it's that slow each time [18:17:06] I might have caching off, hmm [18:17:10] johnduhart: also do you have memcache installed? [18:17:35] I have xcache, but it's set to CACHE_NONE atm. Let's see if that helps [18:17:51] johnduhart: if you don't change the following file: wurfl-config.xml in library/resources/ [18:18:08] and change provider to file and params to a writeable file path [18:18:40] Oh okay [18:18:54] johnduhart: filedir=../cache [18:19:30] preilly: Much better thank you. [18:19:46] johnduhart: np [18:25:44] preilly: sounds like that could be right up front in the README ;) [18:26:00] hexmode: yes, sir [18:26:45] or you could start an FAQ... but for now README is fine. [18:27:46] hexmode: yeah, good idea [19:28:37] neilk_: Poke about UW 1.17wmf1 patch [19:28:48] I'm working on it right now [19:28:54] OK [19:28:59] There's a conflict in tests I'm trying to resolve then you'll have it [19:29:05] OK, cool [19:29:24] I was getting mildly worried since it's 30 mins from now, but it sounds like you'll have it ready really soon now [19:33:22] K4-713: Is this your primary nick? It's currently listed as "K_Horn" at http://www.mediawiki.org/wiki/Developers [19:33:22] oh wait a minute, I know why the tests are in conflict. THERE ARE NOT TESTS [19:33:22] The tests are a lie. [19:33:22] Yes, you'll get a tree conflict because tests/ doesn't exist [19:33:22] s/NOT/NO/ [19:33:22] You can just to svn [revert|resolved] tests [19:33:22] *do [19:34:16] RoanKattouw: sent [19:34:29] Thanks [19:34:50] I'm going to get lunch quickly, will be back before the hour. [19:35:11] OK [19:35:26] actually I lied, I'm doing a few more checks [19:35:34] Even better [19:44:55] RoanKattouw: just FYI, I saw a couple of db errors go by the first time I tried to test this upload, but now I can't replicate them. [19:45:07] investigating further. [19:45:29] neilk_, in the last few minutes? [19:45:41] Reedy: this is locally. [19:45:45] Ah, o [19:45:46] k [19:53:01] Ok, Roan, so I did see a database error the first time I tried uploading, but I haven't seen any since. And I can't even find anything in the db logs, apache logs, or mediawiki debug logs. [19:53:08] OK, we'll see tehn [19:53:13] I've just opened the patch to look at it [19:53:22] 2594 lines. Right [19:53:25] it might be that my db is in some unusual state, since I'm still using the same db as I do for all the other wacky stuff I do. [19:53:28] *RoanKattouw removes i18n crap [19:53:48] probably should have done that for you :( [19:53:56] Still about 800 lines left [19:54:19] we do have legitimate i18n changes in this patch though, remember [19:54:40] yes [19:54:41] we changed the name of a key [19:54:43] Wait, what is this patch against [19:54:49] 1.17wmf1 [19:54:52] - 'us_timestamp < ' . wfTimestamp( TS_MW, time() - UploadStash::REPO_AGE * 3600 ), [19:54:53] + 'us_timestamp < ' . $dbr->timestamp( time() - $wgUploadStashMaxAge ), [19:54:56] You sure? [19:55:01] Oh, nm [19:55:02] --- maintenance/cleanupUploadStash.php (revision 93074) [19:55:05] It does strange this for added files [19:55:36] neilk_: Do you have commit access to 1.17wmf1? [19:55:45] actually I'm not sure that I do [19:55:53] I think you probably don;t [19:55:58] I could try [19:56:03] Please do [19:56:07] Myra: Oops, I forgot to change that. fixing. [19:56:25] K4-713: Cool. Thanks. :-) [19:57:03] No neilk_ hasn't [19:57:07] fail [19:57:20] ^demon, mind fixing that? [19:57:33] <^demon> One moment. [19:58:14] <^demon> Done [19:59:50] demon, RoanKattouw -- committed [19:59:50] Cheers [19:59:56] thanks demon [20:00:04] RoanKattouw: r94806 [20:00:05] Yay [20:00:07] He does have access [20:00:16] I keep forgetting that you need to do yet another patch of the patch [20:00:42] ? [20:00:56] well, after 1.17wmf1 isn't there yet more deploy hacks [20:00:58] ? [20:01:01] No [20:01:02] or is this it [20:01:06] This is it [20:01:09] k [20:01:13] Local mods on the cluster are rare and discouraged [20:01:45] unless termporary [20:01:52] or temporary [20:01:56] Yes [20:02:11] Well you know what 'temporary' means around here, right? ;) [20:02:32] neilk_: DB-wise I just apply the patch .sql file, right? [20:02:44] <^demon> He does have access [20:02:45] ...which is just a CREATE TABLE so that'll be fine [20:02:45] RoanKattouw: I believe so. [20:02:49] Yes. [20:02:53] <^demon> He does /now/ ;-) [20:02:56] haha [20:02:59] When did he get it? [20:03:03] <^demon> 5 minutes ago [20:03:06] whaha [20:03:15] <^demon> Reedy: ^demon, mind fixing that? [20:03:16] <^demon> ^demon: One moment. [20:03:17] <^demon> ^demon: Done [20:03:17] You gave him access just in time for his commit to go through or...? [20:03:22] Oh [20:03:30] No neilk_ hasn't [20:03:30] fail [20:03:32] :P [20:03:40] Meh that was here [20:03:42] I missed that ,fail [20:03:52] heh [20:04:01] "Previously, on #wikimedia-dev..." [20:05:07] Alright [20:05:14] So you have shell on fenari, right? [20:05:30] RoanKattouw: yes [20:05:32] cd to /home/wikipedia/common/php , run svn up, and make coffee [20:05:39] (it takes 1.5-2 minutes to run) [20:06:02] Pffft [20:06:04] Be sensible [20:06:09] svn up extensions/UploadWizard [20:06:24] Reedy: unfortunately it's core + extensions this time [20:06:31] do both seperately [20:06:40] ok [20:06:45] having it iterate over all the other extensions will just increase the time stupidly [20:06:55] whoever is using vim in this directory needs to config their temp directory [20:07:30] hurr do i need to be in group wikidev? everything here is owned by catrope... [20:07:38] <^demon> You should be in that group [20:07:52] You should be, yes [20:07:55] I own most of the source tree [20:08:01] But it's all g+w wikidev [20:08:10] So check that you're in wikidev and that your umask is 0022 [20:08:17] Ahm [20:08:19] 0002 I mean [20:08:40] sure [20:08:44] If it's 0022 and you run svn up, we will hate you forever [20:08:47] also I need to put some keys on this system [20:08:57] You can use agent forwarding [20:09:05] if you have an agent running on your local box [20:09:08] I do [20:09:16] Do you even have SSH access to the Apaches? [20:09:29] RoanKattouw: I don't know their hostnames [20:09:37] srvNNN [20:09:46] But no, you don't have access [20:09:50] There is no neilk on srv200 [20:10:16] So you can do almost everything, I'll have to do the last 5 keystrokes for you [20:10:20] s c a p \n [20:10:24] Ticket is still open [20:10:38] Might just need to poke m ark to give it his ok, then someone can action it [20:11:02] It'll take hours to take effect anyway [20:11:58] svn is updating, yay [20:12:25] It'll be slow [20:12:28] *RoanKattouw stabs NFS [20:12:40] hence suggesting not to svn update the whole damn checkout :P [20:12:40] nope, it was fast, I limited it to the directories I knew were affected. [20:12:48] extensions/UploadWizard includes maintenance [20:13:13] Yeah makes sense [20:13:15] however, I notice that there were several other commits even after the commit I did like a minute ago. [20:13:41] Not to wmf1 there weren't [20:13:51] https://secure.wikimedia.org/wikipedia/mediawiki/wiki/Special:Code/MediaWiki/?path=/branches/wmf/1.17wmf1 [20:13:55] oh right, the count of commits is just global [20:13:56] nm [20:14:19] we're still in california zip code commit territory [20:14:43] <^demon> Not for long :) [20:14:49] Yeah that's gonna go away [20:14:50] ok so fenari is updated, now what [20:14:51] We're movign North [20:15:18] no, it's going south [20:15:32] northern cal is 94xxx. San Jose is 95xxx [20:15:52] but there's very little logic to american zip codes [20:15:55] But Portland is 97xxx [20:16:02] *^demon makes 5000 useless commits [20:16:06] And LA is 90xxx [20:16:24] So on the scale of the entire West Coast I guess the numbers rise going North [20:16:43] RoanKattouw: there's just no logic to US zip codes, give up [20:16:53] <^demon> Yeahhh [20:16:53]