[00:01:31] New patchset: Jdlrobson; "apply media query to tleft rules as well" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7693 [00:01:32] New patchset: Jdlrobson; "add vendor prefixes for box-sizing" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7694 [00:10:37] [WikipediaMobile] jdlrobson pushed 1 new commit to master: http://git.io/sxKWcw [00:10:37] [WikipediaMobile/master] Load Main Page on application startup - YuviPanda [00:14:32] what were the regressions that we hit with todays deployment ? [00:23:37] jdlrobson: --^ ? [00:23:55] Hi tfinc - well lots of problems with tables [00:23:59] k [00:24:12] I think also Patrick and awjr had some issues so had to leave early [00:24:13] do you guys have an etherpad/bug/wiki page tracking the issues ? [00:25:42] Not that I know of [00:27:48] k [00:33:35] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7693 [00:33:37] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7693 [00:33:56] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7694 [00:33:58] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7694 [00:35:35] New patchset: Jdlrobson; "give up on css transitions" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7698 [00:39:48] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7698 [00:39:50] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7698 [00:50:26] New patchset: Jdlrobson; "remove all attribute notation" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7700 [00:50:26] New patchset: Jdlrobson; "merge table definitions in hacks.css" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7701 [01:12:09] New patchset: Jdlrobson; "first version of main menu navigation" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7668 [01:12:18] right i'm off to bed [01:12:20] see you tomorrow [01:48:28] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7700 [01:48:29] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7700 [01:49:02] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7701 [01:49:04] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7701 [03:54:41] [04:21:50] Amgine: whats up? [04:22:18] You're bouncing a fair bit. About every 3 minutes, apparently. [04:26:24] give or take a minute. [04:27:45] yea bad wifi.. will log out... sorry ;) [09:43:04] New patchset: Jdlrobson; "center small tables (bug 36678)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7715 [09:43:57] New patchset: Jdlrobson; "center small tables (bug 36678)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7715 [11:05:21] New patchset: Jdlrobson; "force multiple images in thumbinner on to their own line (bug 36030)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7718 [11:13:29] New patchset: Jdlrobson; "stop scrubbing coordinates (bug 36651)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7720 [11:40:23] New patchset: Jdlrobson; "add canonical url to mobile site (bug 34240)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7722 [12:14:46] New patchset: Jdlrobson; "don't resize images inside an element with class noresize (bug 20030)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7723 [12:16:25] * jdlrobson laments his lack of admin powers https://bugzilla.wikimedia.org/show_bug.cgi?id=20030#c4 http://en.wikipedia.org/wiki/Wikipedia_talk:Contact_us#Bug_31591 [12:37:00] New review: MaxSem; "See inline comment, also there are two typical locations where the links are inserted: lede or infob..." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/7720 [12:38:54] New review: MaxSem; "(no comment)" [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/7722 [12:39:06] Sorry;) [13:05:19] New patchset: Jdlrobson; "add canonical url to mobile site (bug 34240)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7722 [13:07:17] New review: Jdlrobson; "I see that as a separate bug - that we don't allow trailing text after the last section. This is a m..." [mediawiki/extensions/MobileFrontend] (master) C: 0; - https://gerrit.wikimedia.org/r/7720 [13:11:34] New review: MaxSem; "Also, I think we still should remove them for WAP view." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/7720 [13:12:55] New patchset: Jdlrobson; "stop scrubbing coordinates (bug 36651)" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7720 [13:49:06] YuviPanda: any reason you want to keep the initial commit on 244 - what does it win us? [13:49:28] jdlrobson: history? [13:49:35] that we can do git log and see the actual process? [13:49:46] add mf-* patch, then move it to MobileFrontend [13:51:01] but https://gerrit.wikimedia.org/r/#/c/7617/ links to https://github.com/wikimedia/WikipediaMobile/pull/244 which will always have the discussion right - regardless of whether you force amend the commit? [13:51:23] (I'm just interested more than anything in case you know something I don't) [13:51:29] a discussion that refers to code that won't be there [13:51:54] Will it not? [13:52:10] jdlrobson: wouldn't force pushing essentially kill prior commits? [13:52:26] mmm well you could create a new pull request with just the squashed commits [13:52:29] you're essentially 'losing' a commit (without git reflog, etc) when you force push one over another [13:52:36] that way this pull request gets closed and retained right? [13:52:36] true, but that'd still lose history [13:52:50] what would get lost? [13:53:11] the fact that we first had a patch, then moved it to MF after discussion about where the callback should be [13:53:16] it'll be gone from git [13:53:40] this is essentially the 'how faithful should we be to history' discussion :) [13:58:59] yeh I'm just always interested in it - I guess I'm just not understanding why it would be gone from git [14:00:04] i don't like breaking links so if we are breaking links we shouldn't force amend it - i guess i'm just not convinced we do break them if we send a new pull request [14:01:44] jdlrobson: it would be 'gone' because if you do a fresh clone, it won't have that commit that was force pushed over [14:02:52] i'm thinking git checkout -b newbranch / git cherry-pick x / send pull request - that doesn't effect the original branch right? [14:03:02] (and thus doesn't effect the original pull) [14:03:42] hmm, that would preserve the original branch+commit [14:04:14] jdlrobson: what reasons for removing that? Squashing does make it look like it was gotten 'right' on first try [14:05:45] It's just the only time I use history is when I do git blame to find out of the origins of code - so I'd rather have one commit that fixes the reference code then have it split over 2 commits - as that makes it harder to fully understand the change [14:06:05] To be concrete.. [14:07:37] If I was doing a git blame on onClickReference I get commit 0bbc3ccd419c76316bd51865455f191d9ca57850 - however I lose the valuable information that chrome.initContentLinkHandlers was added precisely for this purpose (unless I dig deeper) - as it is in the commit before [14:08:07] if that makes sense YuviPanda [14:09:08] This is all speculative thinking though [14:09:36] hmm, that is a more pragmatic line of thought than my 'history must be preserved!' [14:09:40] :) [14:11:09] devgeeks: around? [14:11:54] jdlrobson: I'll do it in a bit. Will be off for dinner in a while. [14:12:14] jdlrobson: I'm also planning on adding a small js library in. A 'modal blocking dialog' type library [14:12:18] jdlrobson: bad idea? [14:12:25] the other option is to farm that out to native code [14:12:33] What purpose does the modal dialog srve? [14:12:33] which would be faster, but more OS specific code [14:12:44] Would it only apply to OS users? [14:12:48] jdlrobson: migration. Our saved pages are in different formats from older versions. [14:13:02] jdlrobson: so there has to be a 'migration' step that'd be quite long [14:13:07] jdlrobson: if done at startup. [14:13:08] eek :-/ [14:13:34] jdlrobson: I'm adding 'versions' to saved pages now, so that migrations can happen cleanly [14:13:55] jdlrobson: can't 'convert on read' since the user may not have network access when they try to read [14:14:24] saved pages are annoying :) [14:14:50] ummm so essentially you want to say to the user 'hey you've got some saved files we need to do some magic to make them compatible with this app" [14:15:02] user says yes and we give them a please wait screen while we do stuff [14:15:06] when done the app loads [14:16:14] Hey YuviPanda [14:17:43] You were looking for me? [14:18:30] jdlrobson: I'm still thinking it through. [14:18:36] jdlrobson: if they say 'no', then what? [14:18:51] devgeeks: was looking for that 'blocking modal window' js library you were using [14:18:53] jdlrobson: they are :) [14:18:59] ah [14:19:02] BlockUI [14:19:10] jdlrobson: with tracking in place, I hope to convince tfinc sometime that we can get rid of that 'feature' [14:19:16] http://jquery.malsup.com/block/ [14:19:31] yuvipanda Couldn't that just be a confirm which hides everything except a big loading icon and when done shows everything again? [14:19:37] if you say no you just disable saved pages [14:19:43] jdlrobson: 'disable'? [14:19:46] disable older ones? [14:19:54] yes [14:20:00] or altogether if the older ones break it [14:20:12] so the message would be "The saved pages feature will only work if you do this" [14:20:12] older ones just won't work [14:20:26] so yeh you disable it until they want it and it is convenient to upgrade [14:20:32] but you can still save new pages, etc [14:20:44] jdlrobson: perhaps have an 'upgrade saved pages' button of sorts in the saved pages overlay? [14:20:51] interesting.. [14:20:55] jdlrobson: if they say yes first, but it fails (network, perhaps) [14:20:58] so another way to do it [14:21:00] they need to find a way to redo it [14:21:07] would be on the saved pages gray out the old pages that are broken [14:21:10] New patchset: MaxSem; "Mobile script delivery via ResourceLoader" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7670 [14:21:11] put a warning triangle next to them [14:21:15] that when clicked does the upgrade [14:21:27] jdlrobson: not very descriptive, no? [14:21:32] (after a dialog saying why you are showing the warning triangle) [14:22:01] e.g. I go to saved pages my article 'San Francisco' has a warning box next to it and looks darkened. eek i think and I click the warning triangle to see what's wrong [14:22:32] It tells me the article needs to be redownloaded - I click yes and when it's done it goes back to normal [14:22:52] and the user has to do this for every saved page? [14:22:54] 20, 30 of them? [14:23:08] well on clicking one of them the option could be to fix all of them [14:23:30] I think it is acceptable whilst doing this operation to simply show a large spinner and block the ui until it is all done [14:24:13] a big spinner where? [14:24:16] in the content area? [14:24:22] 'overlayed' on top of the overlay? [14:29:33] I keep trying to screen shot mine, but I basically do that where I want them to not do anything else till I am done with some network call (logging in, etc) [14:32:21] http://cl.ly/2T3m3X382B0u1u2C0Y1B [14:35:12] jdlrobson: ^ [14:35:23] sorry YuviPanda got sidetracked [14:35:26] that is the thing I'd want to be showing when doing the upgrade [14:35:28] heh [14:35:44] so this is related to what i keep telling phil and linSmith about the spinner - I don't like the fact we use the search spinner for everything - it's bad ui [14:35:55] the spinner on search should only spin when the search is doing something [14:36:05] When I'm loading a page or doing something else I want the whole page to spin [14:36:18] (well the window where the action is occuring) [14:36:48] io the ui in http://cl.ly/2T3m3X382B0u1u2C0Y1B does a much better job imo [14:37:03] it's much more reassuring stuff is happening [14:38:16] +1 [14:38:18] jdlrobson: but it is blocking [14:38:23] jdlrobson: i'd not want to use that for a page load [14:38:30] why not? [14:38:32] You need both [14:38:41] I click a link I expect something to happen [14:38:41] a non-blocking but reassuring spinner [14:38:43] since you can do things in the time between tapping on a link (starts page load) [14:38:51] and a blocking one that says, you gotta wait [14:38:55] to the page actually loading [14:38:56] in the current setup I get no effective indication my click worked [14:39:07] devgeeks: +1 [14:39:12] I use both kinds [14:39:16] like cancelling, or clicking another link [14:39:28] a non-blocking but reassuring spinner is all that is needed [14:39:35] Images that need time to load have a background set to an animated gif spinner, etc [14:39:40] YuviPanda: you still could [14:39:54] with a non-blocking spinner,y es [14:40:01] yes that's what I'm talking about [14:40:03] I agree that what we have is suboptimal [14:40:12] just give it a bit of opacity and put it in the middle of the screen with the highest z-index [14:40:23] you can still see through it, still scroll but it's clear something is happening [14:40:26] wouldn't that be confusing? [14:40:31] clicking the spinner could even confuse it [14:40:37] *cancel it [14:40:47] I don't think it would be confusing [14:41:23] You could always set a div that is actually loading to have a background of a spinner [14:41:26] jdlrobson: i'll open up a bug. Don't want to do it for this release, let's try variants out [14:41:40] jdlrobson: I do agree that having it only on the search bar is suboptimal [14:42:13] New patchset: MaxSem; "Mobile script delivery via ResourceLoader" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7670 [14:42:26] yeh YuviPanda I think you'll need assistance from heatherw and linSmith [14:42:41] yeah, agreed. [14:42:50] Should be part of the new nav as well [14:43:00] New review: MaxSem; "jQuery needs to be made raw, too." [mediawiki/extensions/MobileFrontend] (master) C: -1; - https://gerrit.wikimedia.org/r/7670 [14:44:09] jdlrobson: for this particular one (migrating saved pages), I suppose we could do 1. Grey out with a red triangle 2. Tapping gives them option to 'update' that page only or all pages [14:44:26] or, 1. On startup, tell them upgrades are taking place and just upgrade everything [14:44:39] well I guess the latter is okay [14:44:44] since they will have just upgraded [14:44:48] so should expect this sort of thing [14:44:50] yeah [14:44:53] and it's nice to get it out of the way [14:45:01] question is what happens on network issues, etc on first upgrade [14:45:13] hmm [14:45:14] I think [14:45:15] disable save pages altogether in ui [14:45:27] 1. Grey out the saved pages, 2. Let it run on next startup as well [14:45:36] so they can just quit and re-open [14:45:37] yup I think so [14:45:43] jdlrobson: that's not necessary. They can still save new pages [14:45:46] and open those [14:45:55] versioning will tell me which are problematic, which are not [14:46:00] sorry that's what I meant - gray out the old ones [14:46:26] ok [14:46:32] sounds good. Will do tonight. [14:47:00] off for dinner, brb! [14:47:02] Night guys, I better sleep [14:47:08] night devgeeks [14:47:16] thanks for the link :) [14:47:22] Heh, my pleasure [14:47:36] If you want the CSS I used for mine, just let me know [14:48:02] devgeeks: thanks :) [14:48:24] Though animated GIFs don't work in Android, so ya gotta find a different way there anyway [14:48:30] :/ [14:49:28] night devgeeks ! [14:49:34] night [15:05:12] New patchset: Jdlrobson; "first version of main menu navigation" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7668 [17:09:30] * jdlrobson laments his lack of admin powers again in hope someone will bestow him some pity https://bugzilla.wikimedia.org/show_bug.cgi?id=20030#c4 http://en.wikipedia.org/wiki/Wikipedia_talk:Contact_us#Bug_31591 [17:10:01] jdlrobson what do you need? [17:10:15] it's all there - i need a wikipedia admin to make some changes [17:10:35] jdlrobson, I am an admin [17:11:01] though a few months of consistent editing and you might become one, too;) [17:11:11] so this one is a bit more complicated = https://bugzilla.wikimedia.org/show_bug.cgi?id=20030#c4- as I'm not sure where the template comes from [17:11:18] but http://en.wikipedia.org/wiki/Wikipedia_talk:Contact_us#Bug_31591 is trivial [17:11:19] i thought i was but i am not [17:11:21] *should be [17:14:13] done [17:14:45] thanks MaxSem will check in a bit - just have my 20% checkin [17:33:31] awjr: so I need to go now [17:33:45] but if you were planning a deployment today I think the table issues should be somewhat sorted with my latest push [17:34:11] jdlrobson: can you send out a quick mail to mobile-tech explaining the issue(s) and solution(s)? [17:34:39] just so we know what to look for as we're testing [17:34:46] Sure.. although will have to remind myself :) [17:34:49] it was rather confusing [17:46:43] no preilly today? [17:47:32] awjr: sent [17:47:37] jdlrobson thanks a ton [17:47:44] are you giong to be back on later tonight? [17:48:10] awjr, are you doing anything with ExtMobileFrontend ATM? [17:48:19] MaxSem: nope [17:48:33] great, so we can't collide [17:48:51] awjr: maybe - but don't rely on it [17:48:54] on mail for anything major [17:48:55] MaxSem: i think we should be ok [17:49:00] jdlrobson ok sounds goo [17:49:01] d [17:49:02] thanks again [17:49:35] cause when I finish, there will be no ExtMobileFrontend XD [17:51:16] greetings all [17:51:45] hello tfinc [17:56:10] philinje: I thought the WLM meeting was on Wednesday. Is there one happening today? [17:56:12] pchang: ^ [17:57:49] preilly: should we set up a remote branch for the core migration stuff? [17:59:14] YuviPanda: yes, team meeting in a half hour [17:59:23] ok [17:59:32] talking with Elke before [18:04:43] awjr: preilly: how are we looking for todays make up deployment ? [18:04:50] if anything we need to get the android fixes out [18:05:01] tfinc: I think we are good [18:05:35] k [18:06:21] tfinc: im starting to test jon's changes, i'll let you know how things are looking in a bit. if they're still not prod-ready, i can investigate picking up just the android fixes [18:09:18] New patchset: preilly; "Update minified javascript" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7736 [18:10:19] wow the new firefox beta is actually not horrible [18:10:57] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7715 [18:10:59] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7715 [18:11:30] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7718 [18:11:32] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7718 [18:11:44] awjr: they've been making great progress on it thanks to brions qa efforts :D [18:11:49] he does file all of our firebox bugs [18:11:51] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7723 [18:11:53] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7723 [18:11:55] haha awesome [18:12:32] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7736 [18:12:34] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7736 [18:13:18] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7720 [18:13:20] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7720 [18:14:32] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7722 [18:14:34] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7722 [18:24:03] New review: awjrichards; "This is certainly up for debate, but I'd way rather see HTML generation happen inside of HEREDOCs ra..." [mediawiki/extensions/MobileFrontend] (master); V: 0 C: -1; - https://gerrit.wikimedia.org/r/7668 [18:24:57] awjr: +1 for HEREDOCs [18:25:20] \o/ [18:25:28] here here! [18:27:04] did the RSA host key for mobile-geo change for some reason? [18:27:28] preilly ^ [18:33:14] YuviPanda: are you on skype [18:33:26] philinje: now am [18:34:01] awjr: mobile-geo is no longer [18:34:08] awjr: it's now mobile-testing [18:35:03] awjr: but, for compatibility reasons I keep the mobile-geo host name [18:35:13] awjr: anything is cloned from the other instance [18:36:23] s/anything/everything [18:38:24] preilly ok cool thanks [18:46:27] preilly: do we really have 16G of RAM on the new instance? [18:46:51] and 8 cores [18:46:52] wow. [18:47:00] awjr: Mem: 16050 [18:47:17] nicely done. [18:49:14] QEMU Virtual CPU version 0.13.91 [18:49:14] cpu MHz : 2659.644 [18:49:15] cache size : 4096 KB [18:49:24] there's something in the san francisco article on mobile-testing that's causing a fatal error but im not seeing it on other random pages [18:49:55] actually also happening here: [18:49:55] http://mobile-geo.wmflabs.org/w_awjr/index.php/Eric_(novel) [18:50:06] Fatal error: Maximum function nesting level [18:50:31] YuviPanda: we lost you as you were likely crawling through tunnels of doom [18:52:00] tfinc: http://yuvi.in/POSM/ [18:52:02] philinje: http://yuvi.in/POSM/ [18:52:05] preilly i think it's realted to expansion of templates? [18:52:12] awjr: okay, it should be fixed now [18:52:16] o [18:52:29] sho nuff [18:52:31] what was it? [18:52:35] core bug? [18:52:54] or php ini config change [18:53:09] the infinite recursion protection was set too low [18:53:25] yeah, 100 is hardly infinite [18:53:42] awjr: how does the speed of http://mobile-geo.wmflabs.org/w/index.php/San_Francisco seem? [18:53:52] a gazillion times better [18:54:32] a big part of the nasty load time for that article is the image resources [18:54:48] philinje: http://developers.cloudmade.com/projects/web-maps-api/examples/marker-clustering [18:54:50] tfinc: http://developers.cloudmade.com/projects/web-maps-api/examples/marker-clustering [18:55:47] https://github.com/CloudMade/Leaflet/issues/442 [18:55:59] awjr: dude it's loading a 4.50MB image [18:56:17] yeah the twin peaks image [18:56:23] that is dumb. [18:56:29] preilly, does the new instance also have faster I/O? :P [18:56:52] awjr: it's loading this: http://upload.wikimedia.org/wikipedia/commons/d/da/SF_From_Marin_Highlands3.jpg [18:57:13] MaxSem: yes [18:57:22] o_0 [18:57:25] how? [18:57:26] MaxSem: it's running on the new cisco box [18:57:44] PREILLY: ALSO: http://upload.wikimedia.org/wikipedia/commons/c/cd/SanFrancisco_from_TwinPeaks_dusk_MC.jpg [18:57:48] oops sorry for yellng [18:58:28] abysmal /mnt performance is what prevented me from importing much, much more stuff from enwiki [18:58:41] is there any kind of wikipedia policy around image sizes? [18:58:48] there's no reason those images should be that large [18:58:59] at least no good one [18:59:22] awjr, are you going to train me during today's deployment? [18:59:40] MaxSem if you'll be around that would be great [19:00:24] the www related files are now in /data/project as well [19:00:33] so they are shared with all mobile instances [19:00:56] so you can fire up a new vm and auto mount /data/project and have all of your files images etc [19:01:39] and the /mnt/store is also copyable to new instances as an image [19:03:27] i might reimport the SF article from the current one on enwiki - they seem to have resolved image size issues there. [19:04:20] awjr: okay [19:11:13] hmm the reimport did not seem to pick up the image size changes [19:18:34] huh i wonder if it's a configuration thing. the articles are identical, but the en.wiki one is loading smaller images [19:19:37] awjr, no scaling software on Labs? [19:19:51] MaxSem: dunno - what is used in produciton? [19:20:04] ImageMagick [19:20:56] it's ridiculous that packaged PHP doesn't come without even GD and they call it "modularity" [19:22:35] MaxSem: we've got gd installed [19:22:55] I'm installing imagick now [19:23:01] as am i [19:23:03] ... [19:23:08] but MaxSem go ahead :) [19:23:20] nope, I'm already apt-getting [19:23:27] it has locks [19:23:30] so... [19:23:39] * MaxSem shows awjr a togue [19:23:45] :p [19:30:20] * YuviPanda enjoys the wind on the roof, which has better Internet connectivity [19:30:57] awjr, scaling now works [19:31:08] awesome MaxSem! [19:32:45] * MaxSem notices someone trolling... [19:33:08] from preliminary testing, jon's changes are looking pretty good (at least on android devices, gonna look at iOS after lunch). the one thing that is weird is that on android 4.0.2 large tables will scroll, while on android 2.3.6 a large table causes the entire article to horizontal scroll (example can be seen here: http://mobile-geo.wmflabs.org/w_awjr/index.php?title=Main_Page&mobileaction=toggle_view_mobile) [19:33:09] MaxSem: HAHAHAHAHAHAHAHA [19:33:35] philinje: I think I lost you [19:34:25] imagemagick ftw. it turned those 4/5MB images into ~200kb images [19:34:29] hmm, the rooftop gives me better access to skype but the mosquitoes are a pain [19:35:31] tfinc: btw, did I ever tell you that our app will always be slower on iOS than on Android because Apple doesn't let apps use the JIT javascript engine that safari uses? [19:35:37] MaxSem: besides installing ImageMagick what else did you change? [19:35:59] uncommented $wgUseImageMagick = treu in LS [19:36:45] YuviPanda: yeah, Nitro is not available for PhoneGap and other webapp-related solutions. [19:37:05] yeah, that sucks. [19:44:48] awjr: are you testing Template:Track listing [19:45:04] preilly i tried that as ewll - i imported the revolver article [19:45:07] at least into w_awjr [19:45:16] http://mobile-geo.wmflabs.org/w_awjr/index.php/Revolver_(album) [19:48:06] awjr: did you switch your LocalSettings file to use your DB? [19:48:13] preilly yes [19:49:25] awjr: I'm also cleaning up spam http://mobile-geo.wmflabs.org/w/index.php?title=Special:Contributions/Spam_cleanup_script&offset=&limit=500&target=Spam+cleanup+script [19:56:52] New patchset: MaxSem; "Generate WmlContext on demand, make it load stuff from context itself" [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7743 [19:57:51] preilly, you might want to pick up https://gerrit.wikimedia.org/r/#/c/7616/ [20:00:13] MaxSem: I've got it already [20:00:43] MaxSem: that was so 5 minutes ago [20:01:39] * MaxSem sees nothing in deletion log. oh well [20:03:12] eh http://dpaste.com/748633/ [20:07:28] MaxSem: that is fixed [20:09:00] preilly, tfinc: is this a show stopping bug for deployment? https://bugzilla.wikimedia.org/show_bug.cgi?id=36030 [20:09:05] New review: preilly; "(no comment)" [mediawiki/extensions/MobileFrontend] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/7743 [20:09:07] Change merged: preilly; [mediawiki/extensions/MobileFrontend] (master) - https://gerrit.wikimedia.org/r/7743 [20:10:12] awjr: https://gerrit.wikimedia.org/r/7718 didn't fix it? [20:10:38] preilly does not appear to have [20:11:01] awjr: hmm [20:11:06] on android 2.3.6 the problem is less egregious [20:11:15] i presume due to that fix [20:11:40] awjr: I don't think it's a show stopper [20:12:13] word [20:12:26] MaxSem: how are you running spam cleanup? [20:12:43] now that i have lion im trying to install the new xcode so i can test with ios 5 - but otherwise all the other issues appear resolved, or resolved-enough-for-now on android [20:13:03] awjr: okay great [20:13:50] php maintenance/cleanupSpam.php --delete [20:14:03] e.g. *.viagra.com [20:14:11] MaxSem: so you are manually finding links? [20:14:39] yeah, spam cleanup sucks [20:14:46] MaxSem: I've been manually using cleanup.php from the SpamBlacklist extension [20:14:55]