[00:00:07] Romaine: no the rendered banner itself [00:00:11] Romaine: not text in banners [00:00:17] the banner seems subject to caching, I can't see the banner on my homewiki but on other Wikipedias I see it [00:01:30] Romaine: I'm going to nlwiki but still seeing no banner [00:01:31] I don't understand you [00:01:43] I asked other Wikipedians and they see it [00:01:52] liangent: which country you access internet from? [00:02:50] (it is GEO located) [00:03:09] Romaine: ouch [00:03:36] we only show the banner in countries that participate in Wiki Loves Monuments [00:04:59] Romaine: https://bits.wikimedia.org/geoiplookup gives me "Geo = {}" [00:05:12] ah nice :S [00:07:43] Romaine: do you have access to CN management? [00:07:58] and can you enable this banner to everyone on testwiki [00:08:10] I created the current CentralNotice if that is what you meant: yes [00:08:28] I think I can create a test banner for testwiki [00:09:08] Romaine: ok please create one [00:09:20] I use a test one [00:10:52] Romaine: I still see nothing on https://test.wikipedia.org/wiki/Main_Page ... [00:11:04] I actived now [00:11:12] their can be a delay [00:11:26] is a shift-reload needed? [00:12:33] I don't know [00:12:53] it is activated on test, in all languages, everywhere [00:15:06] mmm, I still do not see it [00:15:09] on testwiki [00:15:16] Romaine: and https://meta.wikimedia.org/wiki/Special:CentralNoticeBanners/edit/wlm_2013 has a "preview" link, but it's still not working for me [00:15:56] I see no preview link there [00:16:16] the whole CN management is counter intuitive for me [00:18:09] Romaine: it's in subtitle area [00:18:16] ie just below firstHeading [00:20:01] ah there, when I click it I get him [00:23:01] Romaine: anyway do you get different banners for https://meta.wikimedia.org/w/index.php?title=Special:Random&banner=wlm_2013&uselang=zh-cn&force=1 and https://meta.wikimedia.org/w/index.php?title=Special:Random&banner=wlm_2013&uselang=zh-tw&force=1 [00:23:40] I see some difference in it [00:24:37] here are zh-hant and zh-tw [00:25:04] can you copy the text you see in the banner? [00:25:05] compared with hk two characters different [00:25:20] in case you can't recognize chinese characters [00:25:29] zh, zh-hans and zh-cn are the same [00:25:54] yeah that's expected [00:26:06] zh-hant / zh-tw: ?????:????,???????????! [00:26:18] my client gives a problem I think [00:26:37] Romaine: ok anyway that seems working [00:27:31] zh-hk: ?????:????,???????????! [00:27:49] Romaine: still a bunch of question marks [00:27:53] zh, zh-hans, zh-cn: ?????:????,???????????! [00:28:00] ok, than you see the same [00:28:15] can I put them somwhere on a wiki page? [00:28:37] Romaine: yeah [00:31:07] https://meta.wikimedia.org/wiki/User:Romaine/Sandbox [00:31:42] Romaine: so I see nothing wrong here [00:35:19] [06:40] Hi Romaine, the centralnotice is still not showing the right terms in variant Chinese. [00:36:10] (03PS1) 10Yuvipanda: Use redis instead of memcached for primary cache [vagrant] - 10https://gerrit.wikimedia.org/r/82204 [00:36:11] (03PS1) 10Yuvipanda: Kill the memcached module [vagrant] - 10https://gerrit.wikimedia.org/r/82205 [00:36:18] (03Abandoned) 10Yuvipanda: Use redis instead of memcached for primary cache [vagrant] - 10https://gerrit.wikimedia.org/r/81222 (owner: 10Yuvipanda) [00:36:32] Romaine: No such nick or channel / Supaplextw [00:37:00] * YuviPanda pokes ori-l with https://gerrit.wikimedia.org/r/#/c/82204/ and https://gerrit.wikimedia.org/r/#/c/82205/ [00:38:09] was earlier yesterday [00:40:04] Romaine: I'm marking https://bugzilla.wikimedia.org/show_bug.cgi?id=53641 as unconfirmed, but anyone actually facing this issue is welcome to confirm it with their detailed configurations provided [00:40:55] YuviPanda: have you tested them? [00:41:05] [20:21] The CentralNotice of WLM for zh-tw need some correction [00:41:06] [20:21] Romaine: Supaplextw is from taiwan and has a question about translations [00:41:06] [20:34] Supaplextw: your changes to the CentralNotice should have effect already, can you check? [00:41:06] [20:38] Romaine Nope. I have change the location for each verison of Chinese [00:41:06] [20:39] Supaplextw: you can put ?uselang=zh-hant or another zh-... at the end of the url [00:41:06] [20:39] then you test another variant [00:41:06] [20:45] Romaine: I test another zh variant, it's still the same. [00:41:07] [06:40] Hi Romaine, the centralnotice is still not showing the right terms in variant Chinese. [00:41:07] [06:53] Supaplextw: which language you have set in your preferences? [00:41:08] [06:53] which zh-...? [00:41:08] [06:54] zh-tw [00:41:59] (while I saw yesterday the same as now) [00:42:26] you got correct variants, right? [00:42:35] while Supaplextw doesn't? [00:42:49] I can't say my variants are correct, I can't read it [00:42:58] ori-l: yeah, tested it. puts things on redis [00:43:11] ori-l: but it needs a restart, though. by default it just keeps using the memcached instance. [00:43:22] some sort of cache, I'd suppose. [00:43:42] but yeah, i tested by doing a 'keys *' in redis-cli [00:43:49] Romaine: https://meta.wikimedia.org/w/index.php?title=User:Romaine/Sandbox&oldid=5768385 this is correct [00:44:26] Romaine: and for zh, using either variant (zh-cn or zh-tw) is ok theoretically [00:44:37] even mixed variant [00:44:39] that user was https://commons.wikimedia.org/wiki/User:Supaplex [00:46:21] Romaine: it seems that user's language settings are good [00:46:40] and it's expected to show zh-tw text in banners [00:46:52] YuviPanda: redis patch looks good, i'll do a vagrant destroy / vagrant up just in case tho [00:47:04] memcached i'd like to keep simply because we do still use it in prod (redis is only used for session cache) [00:47:09] keep the module, i mean [00:47:19] ori-l: in definitive evidence that memcached as a project is now dead, a patch as simple as https://github.com/memcached/memcached/pull/43 is now dead in the water for 2 months. Nothing has been merged for about 8 months now. [00:47:28] and nothing non trivial for a year or more [00:47:41] ori-l: pfft. fine, fine. [00:48:08] YuviPanda: i'd take that as definitive evidence that brad fitzpatrick gets things right [00:48:29] Romaine: but I can really do nothing now [00:48:40] it doesn't even compile out of the box on recent systems, and my trivial 1 line patch to fix that has been rotting for 3 months now. [00:49:22] I understand [00:49:33] ori-l: either way, feel free to abandon the memcached removal patch for now :) [00:50:38] Romaine: I guess I can ask some other chinese-speaking people to check it later [00:51:14] ow wait [00:51:27] Romaine: ? [00:52:00] he uses now this user name on iRC: Supaplextw1 [00:52:06] YuviPanda: k, thanks [00:52:16] I just pinged, but he hasn't responded yet [00:52:44] and his status is "Away: 我現在不在位子上" meaning afk in chinese [00:53:14] ok [00:53:15] gtg [00:53:18] see you [00:53:22] (03Abandoned) 10Ori.livneh: Kill the memcached module [vagrant] - 10https://gerrit.wikimedia.org/r/82205 (owner: 10Yuvipanda) [00:53:28] ok, pity [01:12:56] mwemcached [01:14:13] hi Reedy [01:14:19] ohai [01:23:44] (03PS5) 10Mwjames: [DIC]: \SMW\JobBase implements \SMW\DependencyRequestor [extensions/SemanticMediaWiki] - 10https://gerrit.wikimedia.org/r/81449 [01:52:59] (03PD2) 10AndyRussG: Add animateFromPrev step prop and animation [extensions/GuidedTour] - 10https://gerrit.wikimedia.org/r/82209 [01:53:53] (03PD2) 10AndyRussG: TourController w/ modifyStep( step, properties ) [extensions/GuidedTour] - 10https://gerrit.wikimedia.org/r/82208 [01:58:15] I am wondering how I could get a table of comment [02:00:33] On a talk page : a user post a comment then an another one answer to it. How to get a table of two dimensions where each first dimension content a reply or the message posted at creation time? it would do: list[0][] first message of the section list[1] first answer list[2][] second answer etc... [02:00:54] This is for writing a bot in C [02:04:05] Or something like a struct or enum. [02:10:26] If it is impossible, how to do something similar with Apibot in php. [02:14:09] (03PS1) 10Gnosygnu: v0.9.0.0 [xowa] - 10https://gerrit.wikimedia.org/r/82210 [02:17:10] (03CR) 10Gnosygnu: [C: 032 V: 032] "reviewing own commit" [xowa] - 10https://gerrit.wikimedia.org/r/82210 (owner: 10Gnosygnu) [02:20:47] Anyone? [02:28:24] (03CR) 10Mwjames: [C: 032] [DIC]: \SMW\JobBase implements \SMW\DependencyRequestor [extensions/SemanticMediaWiki] - 10https://gerrit.wikimedia.org/r/81449 (owner: 10Mwjames) [02:30:07] (03Merged) 10jenkins-bot: [DIC]: \SMW\JobBase implements \SMW\DependencyRequestor [extensions/SemanticMediaWiki] - 10https://gerrit.wikimedia.org/r/81449 (owner: 10Mwjames) [02:46:34] (03PS1) 10Gnosygnu: v0.9.0.0b [xowa] - 10https://gerrit.wikimedia.org/r/82211 [02:47:16] (03CR) 10Gnosygnu: [C: 032 V: 032] "self-review" [xowa] - 10https://gerrit.wikimedia.org/r/82211 (owner: 10Gnosygnu) [02:57:36] (03PS1) 10Gnosygnu: Revert "initial commit:2013-08-26" [xowa] - 10https://gerrit.wikimedia.org/r/82212 [02:57:37] (03PS1) 10Gnosygnu: resetting [xowa] - 10https://gerrit.wikimedia.org/r/82213 [02:57:58] (03Abandoned) 10Gnosygnu: resetting [xowa] - 10https://gerrit.wikimedia.org/r/82213 (owner: 10Gnosygnu) [02:58:12] (03CR) 10Gnosygnu: [C: 032 V: 032] Revert "initial commit:2013-08-26" [xowa] - 10https://gerrit.wikimedia.org/r/82212 (owner: 10Gnosygnu) [03:45:23] Yippie, build fixed! [03:45:23] Project browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_9 build #258: FIXED in 17 min: https://wmf.ci.cloudbees.com/job/browsertests-commons.wikimedia.beta.wmflabs.org-windows-internet_explorer_9/258/ [03:46:08] (03PS1) 10Rfaulk: add - more logging to check_lock method. [sartoris] - 10https://gerrit.wikimedia.org/r/82215 [03:46:09] (03PS1) 10Rfaulk: add - write deploy user name to lock file. [sartoris] - 10https://gerrit.wikimedia.org/r/82216 [03:46:10] (03PS1) 10Rfaulk: add - check lock file user, _get_current_lock_user. [sartoris] - 10https://gerrit.wikimedia.org/r/82217 [03:46:11] (03PS1) 10Rfaulk: add - implement _get_current_lock_user. [sartoris] - 10https://gerrit.wikimedia.org/r/82218 [03:47:48] (03CR) 10Rfaulk: [C: 032 V: 032] add - more logging to check_lock method. [sartoris] - 10https://gerrit.wikimedia.org/r/82215 (owner: 10Rfaulk) [03:48:00] (03CR) 10Rfaulk: [C: 032 V: 032] add - write deploy user name to lock file. [sartoris] - 10https://gerrit.wikimedia.org/r/82216 (owner: 10Rfaulk) [03:48:14] (03CR) 10Rfaulk: [C: 032 V: 032] add - check lock file user, _get_current_lock_user. [sartoris] - 10https://gerrit.wikimedia.org/r/82217 (owner: 10Rfaulk) [03:48:28] (03CR) 10Rfaulk: [C: 032 V: 032] add - implement _get_current_lock_user. [sartoris] - 10https://gerrit.wikimedia.org/r/82218 (owner: 10Rfaulk) [03:53:41] GitBlit's code source click handler is seriously lame... It won't let you highlight and select code. [03:55:55] Dantman: what do you mean? it does for me [03:56:46] Ok, looks like you "can" but not if you start from text. [03:56:53] https://git.wikimedia.org/blob/mediawiki%2Fextensions%2FVector/57546a76b73623d7df6fcc2517494227da1a3221/modules%2Fext.vector.expandableSearch.js#L49 [03:57:11] Click and drag from the outside of text and it'll select. [03:57:27] huh, that is quite annoying [03:57:50] I ran into it while trying to copy #a0d8ff [03:58:15] Naturally I started inside of text cause that chunk of code is inside a JS string and I Just wanted the css value. [03:58:30] Similar case for double-click. [03:58:52] Traditionally double-click will select a word. But in this case it just selects the line twice. [03:59:19] the javascript code it's loading seems to consist primarily of a syntax highlighter & bootstrap [03:59:26] so i think this behavior probably comes from bootstrap [03:59:33] though it may be that bootstrap is being misused somehow [04:00:52] I'd say it's the syntax highlighter. Bootstrap itself doesn't have any handling for line selections. [04:02:25] The implementor probably decided that letting you select a line by clicking on a line rather than just clicking on the line number was a "good" idea. Without considering that there are existing behaviours that are expected of plain text code you output onto the page. [04:02:26] it's because the text is surrounded by link tags [04:03:57] Yeah actually it's probably not bootstrap or the syntax highligher. It's probably GitBlit's fault. Those tags show up in the unmodified HTML source. [04:03:59] which doesn't have to do with JS [04:04:02] right [04:04:33] https://github.com/gitblit/gitblit/blob/0096569bfdb8eb3cd7483a0d40f42c99b5be93bb/src/main/java/com/gitblit/wicket/pages/BlobPage.java#L166 [04:05:01] ;) which leads me back to a only slightly modified version of my original comment "GitBlit's code source line behaviour is seriously lame" [04:05:11] the code seems pretty straightforward & hackable, and the one patch i submitted upstream was merged quickly [04:05:41] Gahhh... Java [04:06:00] you don't even really need to touch the java, just change string literals [04:06:50] ^_^ Oh no, more of an "my eyes! I wasn't expecting to look at java right now." [04:06:56] kind of comment [04:07:00] String linePattern = "{1}\r"; [04:07:48] Though also a little bit of an "blech, ANOTHER java web application we're using" [04:12:09] the line numbers are just s and the line content are s, i wonder if you could get the desired behavior by just swapping the tags [04:14:34] Dantman or ori-l: Is there a MediaWiki maintenance script to normalize page titles? [04:14:53] ori-l: Ideally yes, but unfortunately it doesn't look that simple. [04:14:54] (03PS3) 10AndyRussG: Add animateFromPrev step prop and animation [extensions/GuidedTour] - 10https://gerrit.wikimedia.org/r/82209 [04:14:58] Apparently we currently have inaccessible titles. [04:15:17] Pages with titles that pre-date newer restrictive rules. [04:15:38] cleanupTitles.php? [04:15:40] do you have an example? [04:15:44] I do! [04:16:12] and yes, it does sound like that's what you're after [04:16:15] http://p.defau.lt/?dfWS7L3SrvgEUjnteUp87w [04:17:10] ori-l: The line numbers have some weird screwed up css going on with a border top and margin top that makes it so that the actual interactive part of the number (ie: where the mouse would have to click for say line 81 is actually around line 71) [04:17:42] ori-l: I'll just file a bug. [04:18:00] So if we just made that small change, all of a sudden when you'd try to click the line number for line 70 you'd jump to #L80 [04:18:45] 70 GOTO 80 [04:19:00] Elsie: what's wrong with that title, btw? [04:19:37] ori-l: The first letter of a title gets auto-capitalized. [04:19:54] Normalization, old page, etc. [04:21:59] https://bugzilla.wikimedia.org/show_bug.cgi?id=53670 [04:24:48] MediaWiki timestamps apparently double as Gerrit identifiers... [04:25:02] wtf [04:28:57] I'm not sure cleanupTitles.php checks a normalized page title object against the stored value. [04:29:27] Hmmm. [04:30:16] so, a couple of clues [04:30:49] first, i set the title of a page on my mediawiki instance to that title [04:30:55] and i ran cleanupTitles.php, which worked [04:31:08] Processing page... [04:31:09] page 11 (User:ɑʀʇʉʀɵ/SmallCaps.charset) doesn't match self. [04:31:09] renaming 11 (2,'ɑʀʇʉʀɵ/SmallCaps.charset') to (2,'Ɑʀʇʉʀɵ/SmallCaps.charset') [04:31:10] Finished page... 1 of 9 rows updated [04:31:39] Hmmm. [04:31:55] but, oddly, when i look at the updated table contents, the new capital is represented as a unicode escape [04:31:55] So perhaps all we need is a script run. [04:31:56] | 11 | 2 | ʀʇʉʀɵ/SmallCaps.charset | | 1 | 0 | 1 | 0.56617282533 | 201 [04:32:20] Eep. [04:32:29] Is that valid? [04:32:31] and if you look up that glyph [04:32:57] you'll see that it was part of unicode 5.1.0, which came out in 2008 [04:33:06] that page was last edited in 2006, but i'm not sure that's relevant [04:33:34] https://en.wikipedia.org/wiki/Latin_alpha#Encoding_and_forms [04:33:35] Hi I am trying join wikimedia-tech [04:33:38] Does the renamed title function in your test instance? [04:33:51] With the Unicode escape. [04:33:52] good question, let me see [04:34:04] Can database fields contain Unicode escapes? [04:34:06] Sorry it was an error of my client [04:34:06] I've never seen that. [04:34:13] Hi ytrezq. [04:34:55] the title functions properly. the database field isn't special; that's just how my terminal rendered it [04:35:05] Oh. [04:35:12] It write this phrase when /join wikimedia-tech [04:35:15] Sorry [04:35:19] Why did your terminal only escape one character? [04:35:59] maybe because the terminal or font definitions aren't up-to-date with respect to the unicode standard [04:36:10] Weird. [04:36:13] which could mean either mediawiki or mysql aren't either [04:36:14] All right. [04:36:37] I think re-running cleanupTitles.php should resolve the issue. [04:36:41] though cleanupTitles.php worked correctly, so we're back to "maybe we just need to run the script" [04:36:41] right [04:36:47] Agh... I try to do an actual test... and find that for some reason my copy of MW will not use simplesearch at all. [04:36:54] Of course now the renamed titles exist. [04:37:19] when was that last time cleanupTitles.php was run? [04:37:21] Though those four pages... no idea why they're on en.wiki to begin with. [04:37:28] ori-l: 2012, allegedly. [04:37:33] I linked the bug in my bug. [04:37:34] Also, [04:37:47] I think there may be an associated user with an inaccessible user name. [04:38:03] I feel like this has happened to Mu or something too... [04:38:56] user_name: ɑʀʇʉʀɵ [04:38:58] There you go. [04:39:12] https://en.wikipedia.org/wiki/User:%E2%B1%AD%CA%80%CA%87%CA%89%CA%80%C9%B5 [04:39:23] And the software pretends he or she doesn't exist. [04:39:29] Fun. [04:53:41] so, I think I chased this down [04:54:08] I was reading about supporting Harry Potter characters. [04:54:20] Gellert Grindelwald... [04:54:23] ultimately what capitalizes the first letter in MediaWiki is the function 'uc' in Language.php [04:54:36] which relies on mb_strtoupper [04:54:41] Yes. [04:55:15] until 2010, mb_strtoupper was using a unicode table from 2002: https://bugs.php.net/bug.php?id=52981 [04:55:42] Nice. [04:55:46] the fix in 2010 was included in php 5.3.3 [04:56:11] which we *probably* weren't running in 2012 [05:07:10] Yippie, build fixed! [05:07:10] Project browsertests-en.wikipedia.org-linux-chrome build #289: FIXED in 4 min 10 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-linux-chrome/289/ [05:07:12] yes: judging by http://apt.wikimedia.org/wikimedia/pool/main/p/php5/ we upgraded from 5.3.2 to 5.3.10 in august 2012 [05:08:03] the cleanupTitles.php run you linked to was from may [05:10:03] I'll update the bug [05:11:01] Yippie, build fixed! [05:11:02] Project browsertests-en.wikipedia.org-linux-firefox build #275: FIXED in 3 min 50 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-linux-firefox/275/ [05:21:16] Yippie, build fixed! [05:21:17] Project browsertests-en.wikipedia.org-windows-internet_explorer_8 build #248: FIXED in 5 min 17 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_8/248/ [05:24:52] Yippie, build fixed! [05:24:53] Project browsertests-en.wikipedia.org-windows-internet_explorer_9 build #236: FIXED in 6 min 11 sec: https://wmf.ci.cloudbees.com/job/browsertests-en.wikipedia.org-windows-internet_explorer_9/236/ [05:27:35] ori-l: helloooo from back in NYC! [05:27:42] we didn't get a chance to chat in person [05:27:49] but.. hellooooo still. [05:27:51] :) [05:30:08] oh, blah. bummer! [05:30:27] i hope it was a useful / fun visit. [05:33:31] hey ori-l [05:33:44] hey [05:33:49] funny you should pop up [05:33:54] i was about to merge your patch [05:33:57] hah! [05:34:02] i was about to ask you about that :) [05:34:16] also asking about using Redis for other things in Mediawiki. specifically as a generic cache [05:34:35] I realized that invalidating multiple language versions of something would be a lot easier if I just put them as keys in a hash [05:34:56] so the way our production redis is configured - can I just get a BagOStuff and use it for whatever? [05:35:44] but yes, patch merging first :) [05:36:12] (03CR) 10Ori.livneh: [C: 032] "Thanks for persisting. :D" [vagrant] - 10https://gerrit.wikimedia.org/r/82204 (owner: 10Yuvipanda) [05:36:13] The visit was great, very packed but awesome [05:36:16] (03Merged) 10jenkins-bot: Use redis instead of memcached for primary cache [vagrant] - 10https://gerrit.wikimedia.org/r/82204 (owner: 10Yuvipanda) [05:36:53] ty ori-l :) [05:38:08] so the way our production redis is configured - can I just get a BagOStuff and use it for whatever? [05:38:14] I think so, but I'd check with Aaron anyway [05:38:30] hmm [05:38:31] ok [05:39:00] the 'parsed' JSON needs to vary by language, and contains parsed wikitext. [05:39:09] I wonder if I could put that in the parsercache instead [05:39:24] maybe, I don't know very much about that. [05:39:39] hmm, okay [05:39:44] I'll check with Aaron at some point, I guess. [05:39:57] oooh, also, let me try using Redis for the l10n cache [05:40:00] too many queries for that [05:40:02] from mysql [05:40:20] I know you're going to tell me I'm in my 'use redis everywhere' phase :P [05:41:20] I think the only thing you need to be mindful of is the fact that the production redises are meant primarily for session cache, and whenever the eviction policy is discussed it's with that in mind [05:41:50] I think after some back-and-forth we ended up with 'volatile-lru' (which is the default and probably what you expect anyway) and a 1GB size limit [05:42:08] hmm, yeah. I think they're also rdb [05:42:09] which is okay [05:42:26] it's not faster than memcached, btw [05:42:29] not by a long shot [05:42:46] i mean, operation-for-operation [05:42:57] if you can leverage redis's data structures to do less work, of course that can be faster [05:43:36] yes - putting data that is from the same source but varies by language into a hash should make invalidation easier/faster, no? [05:44:05] hmm, the message cache doesn't seem to work. I still see it hitting the db [05:48:07] TIL: Redis uses port 6379 because it spells out the last name of Italian showgirl Alessia Merz, which is an inside joke in Salvatore Sanfilipe's circle of friends: http://oldblog.antirez.com/post/redis-as-LRU-cache.html [05:50:12] how does 6379 spell Merz? [05:50:55] https://commons.wikimedia.org/wiki/File:Telephone-keypad.png [05:51:00] aaaah [05:51:02] right [05:51:04] I forgot those [05:55:17] I'm going to sleep now [05:55:18] hopefully [05:55:20] night, ori-l [05:57:30] good ight [05:57:32] night [06:38:32] (03PS1) 10Netbrain: Added a new parameter (activeicon)) to semantic maps. [extensions/SemanticMaps] - 10https://gerrit.wikimedia.org/r/82219 [07:17:19] grr, screwit. [07:29:28] Yippie, build fixed! [07:29:28] Project MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-firefox build #12: FIXED in 23 min: https://wmf.ci.cloudbees.com/job/MobileFrontend-en.m.wikipedia.beta.wmflabs.org-linux-firefox/12/ [07:43:55] Yippie, build fixed! [07:43:55] Project MobileFrontend-en.m.wikipedia.org-linux-firefox build #12: FIXED in 14 min: https://wmf.ci.cloudbees.com/job/MobileFrontend-en.m.wikipedia.org-linux-firefox/12/ [07:49:16] Yippie, build fixed! [07:49:17] Project UniversalLanguageSelector-sandbox.translatewiki.net-linux-chrome build #12: FIXED in 1 min 11 sec: https://wmf.ci.cloudbees.com/job/UniversalLanguageSelector-sandbox.translatewiki.net-linux-chrome/12/ [07:50:18] Yippie, build fixed! [07:50:19] Project UniversalLanguageSelector-sandbox.translatewiki.net-linux-firefox build #12: FIXED in 1 min 1 sec: https://wmf.ci.cloudbees.com/job/UniversalLanguageSelector-sandbox.translatewiki.net-linux-firefox/12/ [07:53:32] Hi all [07:54:51] Hello guys [07:56:02] got some questions and problems with $wgParser and parsers at all:) if someone good in it and have free time, please, write me in private. Need consultation. Thanks!:) [08:08:46] (03CR) 10Daniel Kinzler: [C: 031] "the approach seems effective and sane, don't know about the details" [core] - 10https://gerrit.wikimedia.org/r/82122 (owner: 10Hoo man) [08:10:22] Yippie, build fixed! [08:10:23] Project VisualEditor-test2.wikipedia.org-linux-firefox build #10: FIXED in 20 min: https://wmf.ci.cloudbees.com/job/VisualEditor-test2.wikipedia.org-linux-firefox/10/ [08:11:47] (03CR) 10Daniel Friesen: "(3 comments)" [core] - 10https://gerrit.wikimedia.org/r/82100 (owner: 10Matmarex) [08:43:27] (03CR) 10Daniel Friesen: "(1 comment)" [core] - 10https://gerrit.wikimedia.org/r/82100 (owner: 10Matmarex) [08:45:47] (03PS1) 10Hashar: more extensions [extensions] - 10https://gerrit.wikimedia.org/r/82222 [08:46:05] (03CR) 10Hashar: [C: 032 V: 032] more extensions [extensions] - 10https://gerrit.wikimedia.org/r/82222 (owner: 10Hashar) [08:58:52] in gerrit, i'm working on an extension. a branch patch set has been reviewed with a +1. i want to now merge that patch set with the master branch. does anyone know the steps needed to accomplish that? [09:01:53] dan-nl: gerrit merges everything to master by default, iirc [09:01:56] what's the patch? [09:02:12] https://gerrit.wikimedia.org/r/#/c/81256/ [09:02:43] ori-l: so i just need to wait for another +1 and then gerrit will merge it? [09:03:29] almost: our arithmetic is from outer space, so +1 + +1 != +2 [09:03:33] someone needs to +2 it [09:03:57] in other words, +1s are not cumulative; someone needs to be bold and merge your patch by marking it '+2' [09:04:16] ori-l: ah, okay, thanks. i guess i could do that, ja? [09:04:39] depends; is this deployed on Wikimedia's cluster? [09:05:22] if not (or not yet, because it's still in early development), are you are the primary author/maintainer, you can merge your own change [09:05:28] *and you are [09:05:56] AngularJS is pretty sweet [09:06:02] yes, i believe that is the case. i will try it out now that i know what to do [09:06:16] cool. [09:06:25] YuviPanda: what do you like about it? [09:06:29] (http://4dmv.localtunnel.com) [09:06:37] that entire thing is about 15 lines of my code [09:06:47] ori-l: it does some interesting stuff with dependencies. [09:06:51] and two way binding [09:07:05] so you update the model, and the view updates automatically. and vice versa too. [09:07:05] (03CR) 10Dan-nl: [C: 032] "approving with +2 so the patchset will merge with master." [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/81256 (owner: 10Dan-nl) [09:07:19] and it apparently does this in an effecient way [09:07:40] ori-l: and best part is that unlike, say, WPF/Silverlight, this databinding is done via a rather simple templating system that's very much just HTML + some templating [09:07:47] 'Pussylovers mobile Porn'? [09:07:55] err? [09:08:09] eewww [09:08:13] localtunnel has been hacked [09:08:25] grumble grumble [09:08:56] dan-nl: you'll actually have to merge it yourself by also adding a +2 / verified and choosing 'submit' [09:08:58] ori-l: http://4qd4.localtunnel.com/#/ is better [09:09:13] dan-nl: this is because you don't have continuous integration set up for that extension [09:09:39] (03CR) 10Dan-nl: [V: 032] "approving with +2 so the patchset will merge with master." [extensions/GWToolset] - 10https://gerrit.wikimedia.org/r/81256 (owner: 10Dan-nl) [09:09:43] the automatic-merge-on-+2 is contingent on having CI set up, and it's not automated yet [09:09:51] (but still slow, of course - it is serving off my local slow connection) [09:10:22] dan-nl: cool, it merged. [09:10:27] now to templatize that ugly JSON [09:10:44] YuviPanda: what should i search for (campaign name)? [09:10:58] 'uk' perhaps [09:11:02] wlm-uk maybe [09:11:03] wlm-in [09:11:10] anything fromt he Campaign: namespace [09:11:11] ori-l: thanks … is there a way for me to set-up the ci on the extension? [09:11:14] oh, neat [09:11:34] ori-l: the good part is just how little code all this is. Rather clean as well. [09:11:47] I didn't have to do any of the show/hide/render code myself. [09:12:09] yes, data binding ftw :) [09:12:14] yeah :) [09:12:17] coming to mediawiki circa 2024 [09:12:24] pffft, optimist [09:13:02] dan-nl: yeah, job definitions are YAML files, simple to edit & versioned in git [09:13:06] I need to keep my eyes out of mw as well more, I think. This entire bowser/grunt thing passed me by while I was busy doing android stuff [09:13:07] i'll dig up some examples, hang on [09:13:24] * YuviPanda goes to eat [09:13:25] also http://yeoman.io/ [09:13:41] ori-l: that's what i'm using :) [09:13:50] hipster! [09:13:56] also, it is good to be back using bootsrap. [09:13:59] grids! <3 [09:14:12]