[00:15:38] hi jhobs. not sure if you're around. [02:13:19] T104094: do the screenshots match the intent of the Task? [12:37:31] T104094: do the screenshots match the intent of the Task? [15:46:21] maroloccio: i've followed up on the phabricator ticket :) [16:03:43] niedzielski: looks like we're seeing a possible crash on db upgrade -- https://phabricator.wikimedia.org/T129232 [16:04:33] dbrant: are you itching to do Q4 work breakdown, or would you be open to doing it next week? [16:05:01] dbrant: thanks, i'll check it out [16:05:16] mbinder: if it works better for you to move it, then it's fine with me [16:17:35] niedzielski: perfect, thank you. sorry about the duplicate message. [16:17:48] maroloccio: np, thank you! :) [16:36:58] bearND: hey, any idea what caused the deployment failure yesterday? i assume it had to do with the node upgrade. but the logs don't provide any clues. [16:42:26] bearND: ah, i see there's some discussion in #services in the backscroll [16:42:40] mdholloway: still no good idea. the tons of 404s are red herrings. those existed before the deploy as well. [16:43:37] mdholloway: i've created https://phabricator.wikimedia.org/T129235 to get rid of the log spam btw. [16:44:09] yeah i just (i think) fixed that: https://phabricator.wikimedia.org/T129229 (sorry for the duplicate phab task) [16:44:17] bearND: was just about to mention it on services [18:13:14] nzr: hey, did you say you already have the new language icon? [18:13:27] the icon is the same. the size and the color [18:13:33] is different [18:13:39] can we increase the size a little bit [18:13:40] nzr: will you upload it? [18:13:46] its there on phab? [18:14:04] yes, but it's dark and smaller [18:14:08] compared to the pencil icon [18:14:12] nzr: ^ [18:14:38] https://phabricator.wikimedia.org/T128350#2099200 nzr [18:14:48] yes. the size and color can be controlled from css? [18:15:14] nzr: i doubt it, it's an svg [18:15:19] okay just a sec [18:15:22] in a meeting [18:15:25] nzr: sure [18:26:37] niedzielski: up for a new suspected bug ? :) [18:26:57] matanya: sure, what'd you find? [18:27:55] niedzielski: new alpha: editing something, saving, old text appears. hard refresh, still old text, close the app, and reopen, still old text [18:29:11] matanya: i think this is a known issue. just a moment [18:29:41] matanya: T123938 [18:30:53] and more one minor thing niedzielski [18:31:19] OS X 10.8, Android Studio 1.5.1, Plug-in CheckStyle-IDEA 4.26 crashes (http://imgur.com/Wq0uhkW). Could I just CLI check? (e.g. http://checkstyle.sourceforge.net/cmdline.html) [18:31:22] the show in browser link takes you to the desktop version, i'd suggest using the mobile view [18:34:27] maroloccio: sorry, this is a bit of a problem area in our setup. there's two options here: 1) point AS to run on Java SDK 8 (you'll still want the Wikipedia project pointed at Java SDK 7), or 2) keep both on Java 7 and use an older version of the Checkstyle plugin [18:34:41] maroloccio: 1) is what I have (also on OSX) and what I'd recommend [18:35:29] matanya: that sounds reasonable. i think we used to have that behavior but we must have regressed when we started using non-mobile MediaWiki API URLs. i'll post a bug for that shortly if we don't already have one [18:35:50] matanya: thanks for the find! [18:35:56] sure [18:36:38] niedzielski: about that, is it still the case that we need to keep the project on Java 7 for Ubuntu? i'm not at home atm so can't double-check my setup on my laptop but I think the project works fine on JDK 8 on Arch [18:36:49] maroloccio: i also had problems (i don't remember what) using the GUI plugin but figured it was because i was on linux. i normally just run checkstyle before pushing a patch. if i forget, CI will check it too [18:36:57] niedzielski: er, about that (meaning my convo with marloccio above) [18:37:42] mdholloway: hm, i seem to recall that Java 8 wasn't officially supported but don't have any documentation at hand either way [18:38:12] mdholloway: i think we also had some trouble with one of the plugins for the java-mwapi library depending on it. i think we could work around that [18:38:14] niedzielski: could be. i think i got OpenJDK 8 from the (not-officially-supported) Arch User Repo [18:39:10] niedzielski: but tbh I haven't done too much Android development on that setup [18:40:45] mdholloway: i'm currently on 1.7.0_95. i think try with resources and maybe something else aren't available on older apis but otherwise java 7 is fine. i haven't tried 8 [18:43:00] mdholloway: In /Library/Java/JavaVirtualMachines/, I swapped jdk1.8.0_65.jdk with jdk1.7.0_21.jdk for project "wikipedia" (/Contents/Home) alone but still, the IDE fails [18:43:13] mdholloway: I will consider running CS from the command line [18:54:20] jhobs: I'm done with my part. I've assigned the task to you for tests. [19:05:51] Thanks bmansurov [19:13:29] bmansurov: the icon thing [19:17:35] yep [19:25:39] bmansurov: i am giving you a new svg. i talked to volker [19:25:49] cool [19:58:06] bmansurov: do you wanna jump on a hangout. i wanna see how the patch looks [19:58:09] or just a screenshot. [20:03:43] nzr: sure, tracy island? [20:04:28] https://plus.google.com/hangouts/_/wikimedia.org/tracyisland [20:44:17] niedzielski: is the db patch all set? [20:45:19] dbrant: it is. i wanted to make a second patch on top of it so every database upgrade is performed the same regardless of the previous version. it's almost complete but has no effect on the fix [21:49:51] i just landed gerrit review #276000. what a time to be alive! [21:52:43] niedzielski: woot! [21:53:14] mdholloway: good thing i had my coding goggles on! [21:53:36] niedzielski: :) [22:13:43] niedzielski: okay if i call you on your cell for our 1:1? [22:13:56] mhurd: ^ okay if i call you on your cell for our 1:1? [22:14:20] dr0ptp4kt: yeah, that's cool :) i'm in a meeting right now but 20 minutes from now should be good [22:14:32] niedzielski: cool, thx [22:17:04] dr0ptp4kt: ya! [22:21:24] dr0ptp4kt: around? could you merge https://gerrit.wikimedia.org/r/276013 ? Can't find baha and it's urgent. [22:21:31] (assuming you have +2 on MobileFrontend :-)) [22:22:11] or bd808 :) [22:22:23] translatebot misbehaving [22:22:56] jdlrobson: those are removed qqq's? [22:23:36] bd808: yeh they got re-added by transltebot [22:23:58] thx! [22:23:58] *nod* bannana checker will catch if wrong I suppose [22:24:38] jdlrobson: that's totally a safe self-merge change too ;) [22:24:56] bd808: i can think of certain people that would have shouted at me ;-) [22:25:05] we really need to revisit our self merge policy [22:25:07] jdli see bd808 beat me to it jdlrobson [22:25:14] but i tried :) [22:25:53] self-merges cause fr-tech a bit of trouble, but for tiny things like that, meh [22:25:54] jdlrobson: what yuri and i used to do for simple i18n was a +1 from one person then a +2 self merge [22:26:40] complex i18n on the other hand...different story [22:26:51] I'm not pro self-merging a huge logic change in core, but lets of things are fine for self-merge IMO [22:27:06] *lots [22:31:43] dr0ptp4kt: yeh but at that point it's not really self merging :) [22:32:01] i've been shouted at for self merging patches that add @skip tags to browser tests hah [22:32:13] i think it's mostly common sense [22:33:30] should legoktm's RejectParserCacheValue hook be adapted / reintroduced to ensure the remaining pages that are bad are purged? [22:34:26] jdlrobson ^ opinion? [22:35:00] jdlrobson: @skip tags are actually a non-trivial change as they alter the test coverage. small number of chars !== small & non-controversial change [22:35:33] tgr: ^^^opinion? [22:35:49] ori: I'm not sure. I'm not clear on impact right now - whether we are talking 1% of pages or 50%. [22:36:24] it's probably less than 1%, but if you come for a specific page, and you hit one, it's 100% of the pages you see [22:36:35] in other words, I think it's worth fixing these cases even if they are few and far in between [22:36:38] if we have a way of banning them from varnish then sure [22:36:50] otherwise I don't think it would make much difference [22:37:00] ori: we should wait a few days to make sure it's actually gone (editors stop reporting new cases) and we're not just hiding it [22:37:19] legoktm: why not log in addition to rejecting the entry? [22:37:43] oh, that's probably even better [22:38:31] and yes, we can purge varnish from the hook too, I think I did that in wikihiero [22:38:56] you still need an uncached visit to the page to trigger this, but it means that a logged-in user visiting an affected page fixes it for anons too [22:39:46] legoktm: there's already a 'T124356' log channel that I have declared and enabled [22:40:39] sounds good, that's where the core debug log goes right now [22:42:46] https://gerrit.wikimedia.org/r/#/c/220362/1/SyntaxHighlight_GeSHi.class.php [22:43:03] so just call $page->getTitle()->purgeSquid(); whenever you hit a bad cached entry [22:43:21] but be careful, if you get the logic wrong you'll be wiping out varnish and the parser cache [22:43:30] no pressure :) [23:44:15] dbrant|brb: Do you remember why we decided to reveal TOC button only when scrolling?