[00:04:04] (03CR) 10Brian Wolff: "I don't know if this is a good idea. Well its true in the persian case, its hard to know if its true for all languages, including potentia" [core] - 10https://gerrit.wikimedia.org/r/89488 (owner: 10Bartosz Dziewoński) [02:02:16] (03PS3) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [02:05:17] (03CR) 10jenkins-bot: [V: 04-1] Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 (owner: 10Ori.livneh) [02:43:58] (03Abandoned) 10Legoktm: Add two new letters [extensions/AntiSpoof] - 10https://gerrit.wikimedia.org/r/38902 (owner: 10Legoktm) [04:24:49] (03PS4) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [04:26:00] (03PS5) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [04:29:19] (03CR) 10jenkins-bot: [V: 04-1] Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 (owner: 10Ori.livneh) [04:29:40] AHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHHH [04:30:33] Louder. [04:31:07] Krinkle|detached: https://integration.wikimedia.org/ci/job/mediawiki-core-jsduck/2677/console [04:31:58] Krinkle|detached: Neither complaint has to do with my change. I think you should hold off on making it voting. [06:05:20] (03PS6) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [06:08:23] (03CR) 10jenkins-bot: [V: 04-1] Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 (owner: 10Ori.livneh) [06:59:38] (03CR) 10Siebrand: [C: 031] "i18n/L10n reviewed." [extensions/OAuth] - 10https://gerrit.wikimedia.org/r/89236 (owner: 10Aaron Schulz) [07:01:01] (03CR) 10Siebrand: "(1 comment)" [extensions/PageDisqus] - 10https://gerrit.wikimedia.org/r/89222 (owner: 10Raimond Spekking) [07:13:34] ori-l: Nope, you're wrong :-) [07:13:47] ori-l: after @class, the default member context changes to that class [07:14:14] That explains one of two warnings [07:14:52] firstly, I didn't make anything voting recently. [07:15:18] It's voting for a small set of files you declare in the config file, right? [07:15:22] secondly, HTMLImageElement was introduced by someone and it slipt through review. That is supposed to be HTMLElement. [07:15:50] Jobs aren't voting based on content, they're either voting or not. [07:16:00] as for jsduck, it should run on all js files, not just when its config changes. [07:16:19] But there's a set of files that are explicitly mentioned; maybe they're listed for a different reason. [07:17:13] I think you're referring to the file name filter so that jobs don't run if nothing related to them was modified. e.g. jshint only runs if *.js or .jshintrc/.jshintignore was changed. and jsduck on *.js or jsduck.config etc. I think. [07:17:25] in those cases it might not run at all. [07:17:29] It's possible, yeah. [07:18:00] I've noticed the mw.Title error for a while in the jsduck-publish build in the last few days. [07:18:10] However I don't recall us having a jsduck test on patch set [07:18:11] Anyways, I don't like JSDuck. I want to frame my comments in the right way, so I'll say: the quality of JavaScript code has improved immeasurably in the last year, and it's especially evident in consistency of presentation and idiom, and it really allows you to focus on what the code is doing. [07:18:43] I know I added one in ve to avoid regressions, but I never enabled it for core (only in postmerge) because we weren't passing it yet. [07:19:06] I credit it to the judicious configuration of JSHint and Jenkins, which I think you've done a really great job with, and it goes beyond just setting up the tools -- I know the precise rules you chose etc. were determined by the state of the code, code conventions, etc. [07:19:23] But with JSDuck -- my problems are: [07:20:15] a) It's a tool for library documentation, but we're using it to document internal APIs; I'm not sure that the generated documentation is all that useful in such cases [07:20:49] b) It is full of entities that don't actually exist in JavaScript, borrowed from the world of classical inheritance languages [07:21:52] I find that the desire to avoid the -1 electro-shock is causing me to actually write code that panders to its style rather than what I think is best :/ [07:22:05] anyways, I'm rambling, hope this makes sense and that you take it in the right way. [07:24:10] a) I disagree. Its output doesn't seem oriented towards general library docs. Or rather, I don't think these APIs are that internal. Firstly, these are pubicly used by extensions, gadgets and ourselves. Not sure what else to say, how is it differnet from doxygen (yes, chapters to write about how awful doxygen is, but its purpose I mean - I don't defend doxygen in any way shape or form, but I do highly [07:24:10] recommend jsduck) [07:24:52] b) True. There are various concepts it supports documenting, that javascript doesn't have as a native language feature (private, static, mixin, ..) [07:25:19] However, javascript features are way more powerful than that. It can emulate all of it, and people do. [07:25:52] right, but these things exist as design patterns, and in certain cases they are not idiomatic to MediaWiki [07:25:55] there isn't really a way to properly document javacript in a generic way, because nothing is anything in javascript, it is what you pretend it is. [07:26:13] and JSDuck goes beyond merely 'supporting' them -- it nudged me into calling moduleStore a 'class' even though that is a pretty bizarre way to describe it [07:26:34] anyway, it comes down to: I dare you to find a better documenter for javascript, impress me. I tried really really hard to find anything even worth looking at, and all I found is garbage. [07:26:55] 'dare', challenge [07:27:04] bad choice of words. [07:27:29] I took it in the spirit it was intended :) Yeah, I'm not sure. I dabbled in YUIDoc and it didn't seem substantially better. [07:27:39] so far both in terms of actually understanding javascript and providing a good interface, jsduck has been the best I could find. [07:28:27] I'm a bit frustrated because it seems to be importing limitations from the world of languages like Java that don't exist in JavaScript [07:28:44] like, it's not a "static class" -- it's an object! That's all -- an object. [07:28:49] Although in some ways it seems like a generic port of antoher documenter made to work with js, it really does focus on javascript specific things. That's not to say its perfect. There's a lot of things missing: [07:29:05] like documenting modules / groups or provided functionality [07:29:10] promise return values [07:29:15] callback signatures [07:29:16] etc. [07:29:45] well, [07:29:50] I'll give it another go, what can I say. [07:30:05] ori-l: anyway, in case of moduleStore, I'd say, 1) yes, it shouldn't limit you to documenting it that way, it should be more flexible., 2) I wouldn't document it as a class, but as an object that is a propery of mw.loader (which it is) [07:30:48] meh, no, @class is better here indeed. the problem is classes shouldn't be nested inside others. [07:30:57] a level should either be a class or a namespace, not both. [07:31:17] no scrap that [07:31:26] we do that with 'mw' as well, that should work fine. [07:32:06] and is a good example where jsduck supports something that's quite unique to javascript (dynamic objects and freedom of having a class inside a random object essentially) [07:32:41] so how do I return the default member context? is it sufficient to annotate one member to set a new implicit default? [07:33:10] ori-l: rambling, also. But it looks like in this case the 2 errors from jsduck test seem justified. The one was because of context and the other because of an undefined implied global. [07:33:55] moduleStore is exactly a singleton class, that's just jsduck terminology. I don't think this is nudging anything, they could've called it anything, doesn't change what they and we mean by it. [07:34:08] ori-l: nope, unfortunately it keeps no stack. That's by design they say. [07:34:37] implied context is linear only. So in general, separate the code from it after it finished (if possible without uglyfing the code) [07:34:40] I've been looking ways to document the JS code we write, but I never got as far as understanding how to do it with jsduck... it's certainly not obvious how to use it [07:34:49] it's only undefined because https://github.com/senchalabs/jsduck/blob/master/lib/jsduck/options.rb contains a small subset of objects, not even everything in the ecmascript specs [07:35:57] ori-l: Yeah, but that list is trivially extended from our config. I don't think it is its job to mention it all. It has more than enough by default. [07:36:59] Nikerabbit: I don't suppose it is incredibly easy, but what isn't obvious about it? Most of it works intuitively with what you'd do already even if it isn't parsedby jsduck (e.g. most of VE and MW files had comments that could transparently be interpreted as jsduck, standard old @method, @param, @returns etc. [07:37:29] Nikerabbit: it's 1 command (jsduck . --config) and then you got html output like https://doc.wikimedia.org/mediawiki-core/master/js/ [07:38:00] ofcourse big projects most likely need to override certain options but that's going to be that way no matter what. Certainly easier than doxygen :P but that's not a difficult one to beat. [07:38:51] ori-l: If you can't move the code to be linear, you'll have to add @member tags to the ones below. [07:38:52] well, the difficulty (at least for me) is figuring out what JSDuck concept to map the code that I have just written to [07:39:29] ori-l: in this case, if you move this singleton down to be the last key, it should work. [07:39:45] for example, an earlier patchset had moduleStore a self-standing object in top-level closure in mediawiki.js (never mind for the moment if that's the best place for it or not) [07:40:00] what is it? a property? a class? [07:40:07] not saying its pretty, but that's how you work around the context issue. Define the classes own properties first and then any child classes. Seems sensible from a structure point of view. [07:40:47] yeah, I don't mind that so much now that I understand it, though it still seems like a pretty trifling reason to -1 a patch [07:41:13] ori-l: yeah, patch set 2 looks like an @private @singleton @class mw.moduleStore [07:41:29] name 'mw.moduleStore' will be implied, no need to specify [07:41:55] well, not in that case, but with a small adjustment it will. [07:42:13] (not in this case because mw itself isn't specified until further down, btu otherwise it would) [07:43:18] but that's a little convoluted and tricky, no? [07:43:48] I feel like it's easier to write correct JavaScript than it is to write correct JSDuck, certainly that's been my track record with Jenkins recently :P [07:45:10] anyways, like I said, I'll get over feeling annoyed and I'll give it another shot and see how it goes [07:45:10] I've never felt limited by jsduck. When first getting to know it, there's a few things you need to learn. But at this point I don't ever feel like jsduck is limiting, and don't let it limit you. So far I always found a way to document something without any convoluted tricks. [07:45:30] what do you find convoluted, it needing @private, @singleton? Or what? [07:45:45] it only guessing the context 'mw' if it is after @class mw? [07:46:11] well, why is mw the context? [07:46:27] namespace [07:46:28] it happens to share a closure with the declaration of mw, but they're not the same [07:46:52] nobody says they're the same. It's not a property, it's a class [07:46:55] what's a namespace? (honest question) [07:48:34] like everything in javascript, nothing means anything. javascript code has by design no meaning. there is no class, method, namespace. there's functions and more functions. a class in this case is an interface consisting of properties and methods. A namespace would be an identifier that hosts other such groupings of data. in javascript terms, a namespace is an object. [07:49:09] so is 'mw' the namespace and 'mw' the object the same thing? [07:49:21] but since functions are objects, 'mw' is a namespace in mw.Title, but it is a singleton as well [07:49:45] containig mw.config = new mw.Map (property of mw singleton), and mw.log etc. [07:49:58] mw.Title and mw.Map are their own classes, but in the same namespace. [07:50:26] well, those are two cases which are class-y, so it's a better fit [07:50:33] that's not a jsduck-ism but what we did in our javascript, but this is roughly how it'd translate to jsduck syntax. [07:50:49] but mw.loader is also a class (singleton in this case) [07:51:02] right, but presumably documentation is a way to express intent in a clear and consistent way [07:51:18] and i was not thinking about it as being a singleton class, in fact that still strikes me as a bizarre and indirect way to describe it [07:51:34] Yes, and especially in the case of javascript (where the code itself makes it impossible to derive intend because almost everything is a function and an object) [07:51:40] so that's my benchmark for a documentation tool -- if it gives me a standardized way of saying what i want to say [07:51:52] and in this case it feels like it's making me say things that i don't want to say [07:52:40] ori-l: Think of it as a different language, not a different workflow. It's the exact same what you want to say, it just calls it something else, that's all. [07:53:26] class is a group of functionality and properties that can be instantiated/constructed. singleton is a single instance of such interface. [07:53:41] but the code is already a language, but it's a constrained language that has to make sense to the interpreter / compiler, so in-line documentation exists to provide a means of describing something directly [07:53:53] like "in case it's not obvious from the code, here's what i'm trying to do here" [07:53:53] which can be an iffy returning 'this', or a closure returning an object, it can be a private instance of a class (like mw.user used to be), or just a plain object with no closure at all. [07:54:05] those are implementation details, but what comes out is the same, a singleton. [07:54:20] in this case i feel like the language allows me to be more direct and expressive than the documentation, which is a strange situation to be in [07:54:21] Exactly [07:55:27] It forces a small level of abstraction that I think is good in documentation. Seeing how mw.user and mw.loader have evolved for a example is a good example of how implementation has radically changed, and yet documentation and interface to the outside stayed the same [07:56:16] mw.user used to be function User, prototype, mw.user = new User(); mw.loader used to be new (function() { return this; }()) [07:56:23] now they're both plainobjects [07:57:31] but remember that the default option for documentation is to just not have an enforced set of constraints [07:57:51] so it's not like you couldn't say that using plain prose [07:58:11] which makes me think that the test of the aptness of the documentation system is not the cases where it shines, but the cases where it is limiting or insufficiently expressive [07:58:29] Sure, but that's a tradeoff when using autogenerated documentation that is structured. [07:59:00] right, but i think the value of autogenerated documentation is specific to a subset of the code [07:59:44] Which I think is worth the theoratical 'it can't express this tricky bit of js', though so far I've been able to document even the trickier js libs with jsduck exactly because it doesn't require implementation details. It's on a more simplified level. [07:59:58] So it shoudl be able to document anything without being restrictive on implementation. [08:00:18] You seem to believe the opposite about the same software, which I find interesting. Maybe we havent' used it fully yet? [08:00:55] maybe. I'll keep an open mind give it a bit more time and see if it settles, how does that sound? [08:01:36] Sure :) And if you find anything that doesn't seem to fit in jsduck's model I'd love to hear it. And of course, alternatives are always welcome if you find one. [08:01:51] (03PS1) 10PleaseStand: Revert installer changes from e9e24cbfe78 [core] - 10https://gerrit.wikimedia.org/r/89498 [08:02:25] but I feel pretty strongly that at the very least it should be OK (not just jenkins-OK but socially-OK) to say "i think in this case i'm better off having concise free prose comments instead of formal doc blocks" [08:03:24] Also note that it's open source and seems open to suggestions and contributions from the community to improve it. The few things I found annoying about jsduck were mostly about subtle details that were incorrectly throwing something out or defaulting to error. Usually an easy thing to fix. The general base of its parse model and especially the user interface it produces is quite stable and unlike other [08:03:24] generators I found. [08:04:12] well, for one i think we could drop '@private' and replace it with '@public' [08:04:30] ori-l: prose is fine, but adding a few @'s in addition to that should be trivial. Why not? Keep the descriptions empty (just list the context, members, parameters etc. and be done) [08:04:41] ori-l: No, that makes no sense. [08:04:52] @private is used where something is in fact private, e.g. closure private. [08:05:14] but our pattern is to have almost all js code live in a closure [08:05:17] or in some rare cases were something is exposed but should not be used as an API. Though that should be rare. [08:05:32] often even a double closure, one imposed by mw.implement and another in the file source [08:05:49] so it's rather the exporting of stuff that is explicit and usual and deserves to be highlighted [08:05:55] *unusual [08:06:21] it having a closure doesn't make it private, it's private if it remains unexposed. [08:06:34] right, most stuff isn't [08:06:49] if it remains unexposed, it not having a doc block is fine. [08:06:55] sigh [08:06:57] $ ./mwjsduck-gen [08:06:57] /usr/local/Cellar/ruby/2.0.0-p247/lib/ruby/gems/2.0.0/gems/jsduck-5.3.3/lib/jsduck/options.rb:872:in `parse_options': invalid option: --config=/Users/ori/projects/vagrant/mediawiki/maintenance/jsduck/config.json (OptionParser::InvalidOption) [08:07:09] i guess i'm running a wrong version? [08:07:17] if an entire class is unexposed, but you want to document it, document it, and add @ignore or @private to the top level member (no need to repeat it for each method) [08:07:42] ori-l: indeed, they broke stuff in 5.x, we use 4.x [08:08:05] what version does jenkins use? [08:08:31] or rather, refactored. This why I like npm, because it allows specifying 4.x or ~4.10. It's unrealistic to work with global versions and expect everything to work with major version upgrades. [08:08:44] ori-l: see the generated pages, 4.10. something [08:09:06] (03CR) 10PleaseStand: "Daniel, as the one who asked (on Patch Set 49) for the installer change in the first place, if you could give me a good argument for fixin" [core] - 10https://gerrit.wikimedia.org/r/89498 (owner: 10PleaseStand) [08:09:13] 4.10.4 [08:09:28] gem install --user-install jsduck --version '< 5' [08:09:46] they won't break stuff in minor updates, so any 4.x is fine [08:10:00] we adopted it at 3.9 [08:11:01] i'll add it to mw-vagrant's default install [08:11:06] The current option model was deprecatedin 4.7 I think and removed in 5.0 [08:11:23] hence why 5.3.3. fails :) [08:13:29] pff, glad cvn-api rewrite is finished, finally working on labs from toolserver after moving 5 C# projects for CVN. [08:13:43] hrm, adding @class mw.loader to addStyleTag seems to have mollified it [08:13:53] ori-l: @member, not @class [08:14:04] heh, oops [08:14:04] it's not a class, but a member [08:14:33] anyhow, that might break the next one in line. Is it feasible to mode moduleStore down in the mw.loader object to be the last key? [08:14:37] that way it won't clash with anything [08:14:55] i guess so, sure [08:15:42] C# is a bitch to debug under xbuild, mono and MacOSX. [08:16:01] + it relying on sqlite, irc.wikimedia.org and freenode, and 2 other bots. [08:16:47] ori-l: You any good in C#? I fixed it mostly, but it seems to be suffering from a memory leak for over 2 years. We've been restarting it every 2 weeks to keep the labs instance from crashing OOM. [08:17:00] at least it's working now, but he, oh well. [08:17:04] never used it [08:17:10] hear good things :) people like LINQ [08:17:51] [[m:CVN]], github.com/countervandalism/CVNBot, #cvn-commons, cvn.wmflabs [08:18:19] http://ganglia.wmflabs.org/latest/?r=week&cs=&ce=&m=load_one&s=by+name&c=cvn&h=cvn-app1 [08:18:23] ori-l: ^ not healthy :P [08:18:33] someone wrote this in 2007, AlexZ and I maintained it since 2010. [08:18:52] Warning: /Users/ori/projects/vagrant/mediawiki/resources/mediawiki/mediawiki.js:1456: Unclosed HTML tag: [08:18:58] yeah, you showed it to me at wikimania 2012 [08:19:35] this one is my favorite: http://ganglia.wmflabs.org/latest/graph.php?r=month&z=xlarge&h=cvn-app1&m=load_one&s=by+name&mc=2&g=mem_report&c=cvn [08:19:36] I want to get rid of it sooner than later and turn it into a special page workflow based on websockets and stuff. [08:19:55] (03CR) 10Nemo bis: "(Discussion for that continues at http://lists.wikimedia.org/pipermail/wikitech-l/2013-October/072422.html .)" [core] - 10https://gerrit.wikimedia.org/r/82116 (owner: 10MaxSem) [08:19:59] irc is keeping good users out :) [08:20:03] and integration it terrible [08:20:08] UX wise [08:20:20] ori-l: yeah, sometimes we forget to restart it [08:20:55] and then it goes OOM, even the ganglia notifier is unable to phone home. some bots will stay working for a few hours so we don't always notice [08:22:28] ori-l: when I got sick of one of the bots crashing for other reasons (but unmaintained code even more unmaintained than CVNBot), I wrote up stillalive to keep them alive. https://github.com/countervandalism/stillalive [08:22:48] in hinsight, there's probably existing packages I could've used, but it seemed a fun experiment to write my own [08:23:51] it could all fit in crontab, but I wanted to keep it in a simple json file and have it autodeploy on multiiple instances and be able to move it accross [08:24:13] https://github.com/countervandalism/stillalive/blob/master/localSettings-cvn.json [08:24:15] seems neat. php is a strange choice for systems programming :P [08:24:26] yeah, I wanted to use nodejs and did for a bit [08:24:30] ah, that's pretty nice [08:24:31] but it became a mess. [08:24:38] the json config format i mean [08:24:44] yeah, no doubt :) [08:24:50] the php is not pretty nie [08:27:05] unrelated thought i had today -- we should ask nik and chad if it'd be possible to have some search interface for the mediawiki NS of all wikimedia wikis [08:27:51] ori-l: https://toolserver.org/~krinkle/tmp/PageScanner/ [08:27:57] I had the same wish in 2010 [08:28:04] never finished it though [08:28:18] the idea was to let it run on cron nightly and find deprecated patterns [08:28:29] 2011* [08:28:36] woudl be great to have Elastic search do this [08:28:38] oh, cool, yeah this is it exactly [08:28:56] https://toolserver.org/~krinkle/tmp/PageScanner//index.php/ShowCollections [08:29:42] hm.. doesn't seem to work anymore [08:29:49] either that or nothing is matching anywhere [08:29:57] I did work for 1 namespace 1 wiki at some point [08:30:13] (03PS7) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [08:30:53] oh well. [08:31:31] ori-l: btw, I hope to make some progress next week on phantomjs runner from jenkins to look for http errors and exceptions on wikis. [08:32:11] ori-l: when I tried to implement some of it, I got into a long tangent to get urls to all wikis. Afterwards I realised sitematrix API provides this. [08:32:27] https://gist.github.com/Krinkle/6878261 [08:32:29] :D [08:32:51] Though $ php tools/getwmfconfig.php --setting wgCanonicalServer --dblist=large [08:32:52] hah! [08:32:53] is nice [08:33:09] yeah, that'll be useful i bet [08:33:16] (second file for output) [08:33:50] \o/ [08:33:57] jenkins approved [08:34:12] i nearly had an aneurysm earlier [08:35:30] another way would be to regex out wgCanonicalServer from wgConf.php and correlate it directly to the dblist keys [08:36:01] but I previously used this to get all the wg udp prefixes, most of which are $lang.$site [08:36:24] so that way I could determine which are not matching their server name [08:36:32] e.g. #testwikidata.wikipedia [08:38:18] then fixed the default and listed all exceptions so that bots keep working who accounted for the mismatches, but prevent new exceptions from being created. [08:38:45] k, I'm finally tired. seeya [08:39:20] good night, thanks for helping with the doc stuff [08:39:39] err good morning i guess [09:07:39] (03PS8) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [09:10:18] (03PS9) 10Ori.livneh: Cache ResourceLoader modules in localStorage [core] - 10https://gerrit.wikimedia.org/r/86867 [09:18:15] (03PS1) 10Rillke: Parse pages and text in context of user language [extensions/GuidedTour] - 10https://gerrit.wikimedia.org/r/89499 [11:08:22] (03PS3) 10Ori.livneh: Have EL depend on Geshi, related minor refactoring [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [11:08:23] (03PS1) 10Ori.livneh: Lint. [vagrant] - 10https://gerrit.wikimedia.org/r/89507 [11:08:52] hmm, I wonder if that is a 'add trailing comma' thing :P [11:09:59] hmm [11:10:05] (03PS2) 10Ori.livneh: Lint. [vagrant] - 10https://gerrit.wikimedia.org/r/89507 [11:10:11] (03CR) 10Ori.livneh: [C: 032] Lint. [vagrant] - 10https://gerrit.wikimedia.org/r/89507 (owner: 10Ori.livneh) [11:10:14] (03Merged) 10jenkins-bot: Lint. [vagrant] - 10https://gerrit.wikimedia.org/r/89507 (owner: 10Ori.livneh) [11:10:33] (03PS4) 10Ori.livneh: Have EL depend on Geshi, related minor refactoring [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [11:12:23] (03PS5) 10Ori.livneh: Have EL depend on Geshi, related minor refactoring [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [11:12:29] (03CR) 10Ori.livneh: [C: 032] Add 'geshi' role for SyntaxHighlight_Geshi [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [11:12:32] (03Merged) 10jenkins-bot: Add 'geshi' role for SyntaxHighlight_Geshi [vagrant] - 10https://gerrit.wikimedia.org/r/84488 (owner: 10Mattflaschen) [11:17:27] (03CR) 10Physikerwelt: [C: 032] Insert defs into svg element, not before it [extensions/Math] - 10https://gerrit.wikimedia.org/r/89485 (owner: 10GWicke) [11:18:16] (03PS1) 10Ori.livneh: Make role names supplied as CLI args case-insensitive [vagrant] - 10https://gerrit.wikimedia.org/r/89508 [11:19:00] (03CR) 10Ori.livneh: [C: 032] Make role names supplied as CLI args case-insensitive [vagrant] - 10https://gerrit.wikimedia.org/r/89508 (owner: 10Ori.livneh) [11:19:03] (03Merged) 10jenkins-bot: Make role names supplied as CLI args case-insensitive [vagrant] - 10https://gerrit.wikimedia.org/r/89508 (owner: 10Ori.livneh) [11:20:01] (03PS5) 10Physikerwelt: Mark Math-specific functions in core as deprecated [core] - 10https://gerrit.wikimedia.org/r/87479 [12:28:17] (03PS1) 10Foxtrott: Enable caching [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89511 [12:28:53] (03PS2) 10Foxtrott: Enable caching [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89511 [12:30:33] (03CR) 10Foxtrott: [C: 032 V: 032] Enable caching [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89511 (owner: 10Foxtrott) [12:31:41] did I mention I hate xml [12:31:47] Hi all [12:31:54] how do I validate an xml file with custom namespaces? [12:32:32] Is it known that Special:UncategorizedPages isn't being updated since Sep 9/10 for most/all projects? [12:34:21] (03PS1) 10PleaseStand: Hide message for DBUnexpectedError exceptions [core] - 10https://gerrit.wikimedia.org/r/89512 [12:35:36] (03CR) 10jenkins-bot: [V: 04-1] Hide message for DBUnexpectedError exceptions [core] - 10https://gerrit.wikimedia.org/r/89512 (owner: 10PleaseStand) [12:37:41] (03PS2) 10PleaseStand: Hide message for DBUnexpectedError exceptions [core] - 10https://gerrit.wikimedia.org/r/89512 [12:42:13] jem-: yes [12:42:19] and that's a question for -tech [12:42:52] Nikerabbit: hmm, I remember this was already discussed in the past, was it still you? :) [12:43:14] Thanks, Nemo_bis [12:43:32] And sorry, sometimes it's difficult to know :) [12:43:39] I'll ask there [12:44:25] There is bug 15434, but that special page did get updated until septembre [12:44:28] ber* [12:44:46] yes, yes (I already replied in the other channel) [12:44:53] Ooops [12:44:59] Getting there [12:53:11] Nemo_bis: where [13:03:23] @seen valhallasw [13:03:23] MatmaRex: Last time I saw valhallasw they were joining the channel, they are still in the channel #wikimedia-labs at 10/13/2013 12:24:14 PM (39m9s ago) [13:32:23] (03PS1) 10Gerrit Patch Uploader: FormOptions: Document getValueReal() [core] - 10https://gerrit.wikimedia.org/r/89514 [13:38:38] (03PS1) 10coren: Make FOR UPDATE specify table in fetchFromConds() [core] - 10https://gerrit.wikimedia.org/r/89515 [13:39:51] (03Abandoned) 10coren: Make FOR UPDATE specify table in fetchFromConds() [core] - 10https://gerrit.wikimedia.org/r/89515 (owner: 10coren) [13:43:48] (03PS1) 10coren: Make FOR UPDATE specify table in fetchFromConds() [core] - 10https://gerrit.wikimedia.org/r/89516 [13:52:30] (03PS1) 10Gerrit Patch Uploader: Another test commit. [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89517 [13:52:32] (03CR) 10coren: [C: 031] "My original test case was deleting any image with a clone of trunk, which fails on Postgres with the expected error:" [core] - 10https://gerrit.wikimedia.org/r/89516 (owner: 10coren) [13:58:56] (03PS2) 10Gerrit Patch Uploader: FormOptions: Document getValueReal(), consistency fixes [core] - 10https://gerrit.wikimedia.org/r/89514 [14:01:48] (03PS1) 10Foxtrott: Do some basic error handling [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89518 [14:04:58] (03PS1) 10Foxtrott: fix bug 54931 (Legends of the Special:Preferences should appear as Nav Tabs) [skins/chameleon] - 10https://gerrit.wikimedia.org/r/89520 [14:05:19] (03PS2) 10Foxtrott: fix bug 54931 (Legends of the Special:Preferences should appear as Nav Tabs) [skins/chameleon] - 10https://gerrit.wikimedia.org/r/89520 [14:06:19] (03CR) 10Foxtrott: [C: 032 V: 032] fix bug 54931 (Legends of the Special:Preferences should appear as Nav Tabs) [skins/chameleon] - 10https://gerrit.wikimedia.org/r/89520 (owner: 10Foxtrott) [14:07:39] (03PS2) 10Foxtrott: Do some basic error handling [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89518 [14:08:28] (03PS3) 10Foxtrott: Do some basic error handling [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89518 [14:08:53] (03CR) 10Foxtrott: [C: 032 V: 032] Do some basic error handling [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89518 (owner: 10Foxtrott) [14:18:09] (03CR) 10Bartosz Dziewoński: "(This was submitted and amended using http://tools.wmflabs.org/gerrit-patch-uploader/ )" [core] - 10https://gerrit.wikimedia.org/r/89514 (owner: 10Gerrit Patch Uploader) [14:28:52] (03CR) 10Alex Monk: [C: 04-2] "I don't really think what we want is the ability to store notifications in another wiki's database, rather a central database table to hol" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/88668 (owner: 10Legoktm) [14:32:26] (03PS1) 10PleaseStand: Ensure that $call['args'] is set before using it [core] - 10https://gerrit.wikimedia.org/r/89521 [14:43:52] (03PS1) 10Physikerwelt: Fix: Add missing xmlns [extensions/Math] - 10https://gerrit.wikimedia.org/r/89522 [14:55:48] can anybody explain why this is not working as I would expect http://pastebin.com/Tb9my2HR ? [14:59:30] query( '//*' ) [14:59:34] Is that supposed to match something? [15:00:08] Hmm, never mind. [15:03:12] (03CR) 10GWicke: [C: 032] Fix: Add missing xmlns [extensions/Math] - 10https://gerrit.wikimedia.org/r/89522 (owner: 10Physikerwelt) [15:03:22] (03Merged) 10jenkins-bot: Fix: Add missing xmlns [extensions/Math] - 10https://gerrit.wikimedia.org/r/89522 (owner: 10Physikerwelt) [16:01:06] (03PS1) 10Gerrit Patch Uploader: Another test... [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89532 [16:03:40] (03PS2) 10Gerrit Patch Uploader: Another test... (new commit message) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89532 [16:04:00] (03PS1) 10Gerrit Patch Uploader: Another test... (new commit message) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89533 [16:04:33] (03PS1) 10Gerrit Patch Uploader: Another test... (new commit message 2) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89534 [16:08:08] (03PS1) 10Gerrit Patch Uploader: Another test... (new commit message 2) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89535 [16:08:37] (03Abandoned) 10Gerrit Patch Uploader: Something Something Patchy [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89414 (owner: 10Gerrit Patch Uploader) [16:08:44] (03Abandoned) 10Gerrit Patch Uploader: Another test... (new commit message 2) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89535 (owner: 10Gerrit Patch Uploader) [16:08:53] (03Abandoned) 10Gerrit Patch Uploader: Another test... (new commit message 2) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89534 (owner: 10Gerrit Patch Uploader) [16:09:00] (03Abandoned) 10Gerrit Patch Uploader: Another test... (new commit message) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89533 (owner: 10Gerrit Patch Uploader) [16:09:08] (03Abandoned) 10Gerrit Patch Uploader: Another test... (new commit message) [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89532 (owner: 10Gerrit Patch Uploader) [16:09:17] (03Abandoned) 10Gerrit Patch Uploader: Another test commit. [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89517 (owner: 10Gerrit Patch Uploader) [16:10:57] (03Abandoned) 10Merlijn van Deen: Another test commit, now in an integrated setting. [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89408 (owner: 10Merlijn van Deen) [16:11:06] (03Abandoned) 10Merlijn van Deen: Another test commit, now in an integrated setting. [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89409 (owner: 10Merlijn van Deen) [16:11:14] (03Abandoned) 10Merlijn van Deen: Another test commit, now in an integrated setting. [test/mediawiki/extensions/examples] - 10https://gerrit.wikimedia.org/r/89410 (owner: 10Merlijn van Deen) [16:16:16] (03CR) 10Alex Monk: [C: 032] Ensure that $call['args'] is set before using it [core] - 10https://gerrit.wikimedia.org/r/89521 (owner: 10PleaseStand) [16:19:08] (03Merged) 10jenkins-bot: Ensure that $call['args'] is set before using it [core] - 10https://gerrit.wikimedia.org/r/89521 (owner: 10PleaseStand) [16:34:48] (03CR) 10Alex Monk: [C: 04-1] "(1 comment)" [extensions/AbuseFilter] - 10https://gerrit.wikimedia.org/r/78491 (owner: 10Legoktm) [16:54:09] (03PS2) 10Umherirrender: Clean up old title on move before reset article id [core] - 10https://gerrit.wikimedia.org/r/89226 [16:59:32] (03CR) 10Alex Monk: [C: 04-1] "(1 comment)" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/88045 (owner: 10Ebrahim) [17:00:19] (03CR) 10Alex Monk: "(1 comment)" [extensions/Echo] - 10https://gerrit.wikimedia.org/r/88045 (owner: 10Ebrahim) [17:13:23] (03PS1) 10Foxtrott: Preparing version 0.1 [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89540 [17:14:06] (03CR) 10Foxtrott: [C: 032 V: 032] Preparing version 0.1 [extensions/Bootstrap] - 10https://gerrit.wikimedia.org/r/89540 (owner: 10Foxtrott) [18:27:16] Anyone know about the OAuth extension on Wikimedia wikis? [18:27:26] I'm curious if the intention is to eventually deploy it to all Wikimedia wikis. [18:27:31] It's currently on phase0 wikis, I believe. [18:30:01] Elsie: seems like it, since OAuth integration is a WMF engineering activity [18:30:22] Well, [18:30:30] I guess I'll file a tracking ticket, then. [18:30:45] tracking what? [18:30:57] Review and deploy OAuth extension to Wikimedia wikis [18:30:59] Maybe. [18:31:05] I dunno. [18:31:14] I want to file a bunch of bugs against OAuth, but I want to be able to track them. [18:31:21] Though I'm not sure they'll all be blockers to deployment. [18:32:00] I asked the same question about OpenID, there doesn't seem to be any tracking bug [18:32:20] There's a "become an OpenID provider" bug. [18:32:22] I know that. [18:32:23] https://www.mediawiki.org/wiki/Talk:Auth_systems [18:32:46] Bug 9604 has a bunch of blockers, but I doubt that bug is really used [18:33:05] There is no deployment blockers tracking bug which has ever told the truth, IIRC. [18:33:19] Perhaps that's why they didn't make one! [18:35:04] * Elsie NP: "Semi-Charmed Kind Of Life" by Third Eye Blind from "Third Eye Blind" [18:35:22] Elsie, and yeah, oauth is enabled on testwiki, test2wiki, mediawikiwiki and testwikidatawiki [18:38:58] (03PS1) 10Yaron Koren: Fix for param parsing changes from r3bd2ec0 [extensions/ExternalData] - 10https://gerrit.wikimedia.org/r/89541 [18:40:55] Krenair: Yes, phase0 wikis. :-) [19:01:30] (03PS1) 10Foxtrott: Preparing version 0.1 [skins/chameleon] - 10https://gerrit.wikimedia.org/r/89542 [19:01:48] (03CR) 10Parent5446: [C: 04-2] "This is not how options work in the MediaWiki database class. Adding the table name won't just add it onto the query. Instead the Database" [core] - 10https://gerrit.wikimedia.org/r/89516 (owner: 10coren) [19:02:17] (03CR) 10Foxtrott: [C: 032 V: 032] Preparing version 0.1 [skins/chameleon] - 10https://gerrit.wikimedia.org/r/89542 (owner: 10Foxtrott) [19:03:11] (03CR) 10Yaron Koren: [C: 032 V: 031] Fix for param parsing changes from r3bd2ec0 [extensions/ExternalData] - 10https://gerrit.wikimedia.org/r/89541 (owner: 10Yaron Koren) [19:04:08] https://meta.wikimedia.org/wiki/Tech_debt [19:04:13] I started expanding this page. [19:05:33] (03CR) 10Yaron Koren: [V: 032] Fix for param parsing changes from r3bd2ec0 [extensions/ExternalData] - 10https://gerrit.wikimedia.org/r/89541 (owner: 10Yaron Koren) [19:06:19] Elsie: your list seems to be more about rubbish to remove [19:06:38] That's what I consider "technical debt", mostly. [19:06:47] Though the term is both vague and expansive. [19:06:58] Technical debts is not mainly covered by your list [19:07:10] Perhaps the page needs a better name? :-) [19:07:14] I suggest a rename to "Technical trashbin" or something [19:07:20] Heh. [19:07:23] That's not very nice. [19:07:23] "Technical cellar debt" [19:07:33] Some of these extensions once served us well. [19:07:37] Tech attic [19:07:43] yeah [19:08:00] Anyway, I've been working on this kind of cleanup all year. [19:08:05] And I want to be able to track some of it. [19:08:12] My cellar is as big as half of my home [19:08:21] Nemo_bis is a vampire. [19:08:45] Frankly I don't see anything particularly relevant in that list [19:09:36] It would be like cleaning the windows of my cellar while leaving it covered with trash [19:10:04] Krenair, Nemo_bis: I copied you two on https://bugzilla.wikimedia.org/show_bug.cgi?id=55678 [19:10:25] Nemo_bis: I'm not sure what "trash" is in this context. [19:10:41] The idea is to reduce the amount of complexity and code in the ecosystem. [19:10:50] But eliminating unneeded extensions, database tables, etc. [19:10:56] Tech cleanup, maybe. [19:11:09] Though I'm not sure "tech debt" is a the wrong term. [19:11:10] -a [19:11:12] Tech purging [19:11:16] :-) [19:11:23] Tech pruning [19:11:30] Tech ruska [19:11:43] /wp ruska [19:12:03] /lastlog ruska [19:14:17] You and your Meta wiki pages I have to find a good title and category for... Commented on talk, that's officially a warning. [19:16:35] Meta-Wiki * [19:16:58] Replied. [19:19:08] * Elsie NP: "Amazing Grace" by Faith Hill [19:19:13] https://bugzilla.wikimedia.org/show_bug.cgi?id=55455 [19:20:16] Elsie: I meant Meta[-Wiki] wiki pages ;) [19:26:11] Of course. [19:28:24] https://bugzilla.wikimedia.org/show_bug.cgi?id=55679 [19:28:28] "Review and deploy OAuth extension to Wikimedia wikis" [19:28:47] I copied Brad J. and Chris S. [19:28:48] Elsie-snitch [19:29:22] :-) [19:29:36] Just trying to be as transparent as possible! [19:29:42] I replied to you on Meta-Wiki, BTW. [19:29:45] FDC report. [19:29:52] Good, trashcking bugs are very good for getting things our of my todos [19:30:14] Hmpf. Why remind me such horrible things [19:30:38] Heh. [19:31:01] Ther should be a throttling for how often it's allowed to mention FDC in a single day. [19:31:15] There [19:31:24] Grahhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhhh. [19:31:32] Fucking Bugzilla is still downloading attachments. [19:34:07] (03PS3) 10Ebrahim: Native digits on 'mw-echo-overlay-title-text' [extensions/Echo] - 10https://gerrit.wikimedia.org/r/88045 [19:35:11] (03PS1) 10Jeroen De Dauw: Add newline [extensions/SubPageList] - 10https://gerrit.wikimedia.org/r/89602 [19:35:21] (03CR) 10Jeroen De Dauw: [C: 032 V: 032] Add newline [extensions/SubPageList] - 10https://gerrit.wikimedia.org/r/89602 (owner: 10Jeroen De Dauw) [20:06:01] (03CR) 10Alex Monk: [C: 032] Native digits on 'mw-echo-overlay-title-text' [extensions/Echo] - 10https://gerrit.wikimedia.org/r/88045 (owner: 10Ebrahim) [20:06:10] (03Merged) 10jenkins-bot: Native digits on 'mw-echo-overlay-title-text' [extensions/Echo] - 10https://gerrit.wikimedia.org/r/88045 (owner: 10Ebrahim) [20:34:12] anybody bored? five really tirvial changes to review: https://gerrit.wikimedia.org/r/#/q/status:open+project:mediawiki/core+branch:master+topic:watchlist,n,z [20:37:33] (03PS3) 10Reedy: SpecialWatchlist: Remove dead cutoff code [core] - 10https://gerrit.wikimedia.org/r/89324 (owner: 10Bartosz Dziewoński) [20:37:38] (03CR) 10Reedy: [C: 032] SpecialWatchlist: Remove dead cutoff code [core] - 10https://gerrit.wikimedia.org/r/89324 (owner: 10Bartosz Dziewoński) [20:37:45] (03PS2) 10Reedy: SpecialWatchlist: Remove old irrelevant comment [core] - 10https://gerrit.wikimedia.org/r/89338 (owner: 10Bartosz Dziewoński) [20:37:51] (03CR) 10Reedy: [C: 032] SpecialWatchlist: Remove old irrelevant comment [core] - 10https://gerrit.wikimedia.org/r/89338 (owner: 10Bartosz Dziewoński) [20:38:08] (03PS2) 10Reedy: SpecialRecentChanges: Don't call getUser() repeatedly in getDefaultOptions() [core] - 10https://gerrit.wikimedia.org/r/89378 (owner: 10Bartosz Dziewoński) [20:38:13] (03CR) 10Reedy: [C: 032] SpecialRecentChanges: Don't call getUser() repeatedly in getDefaultOptions() [core] - 10https://gerrit.wikimedia.org/r/89378 (owner: 10Bartosz Dziewoński) [20:39:15] hey, thanks Reedy. i have some more stuff if you're still bored :P [20:39:36] Jenkins needs 15 minutes to catch up ;) [20:40:07] Reedy: https://gerrit.wikimedia.org/r/#/c/86319/ and its dependency badly need love (prefereably before the release) [20:42:28] (03Merged) 10jenkins-bot: SpecialWatchlist: Remove dead cutoff code [core] - 10https://gerrit.wikimedia.org/r/89324 (owner: 10Bartosz Dziewoński) [20:45:37] (03Merged) 10jenkins-bot: SpecialWatchlist: Remove old irrelevant comment [core] - 10https://gerrit.wikimedia.org/r/89338 (owner: 10Bartosz Dziewoński) [20:48:46] (03Merged) 10jenkins-bot: SpecialRecentChanges: Don't call getUser() repeatedly in getDefaultOptions() [core] - 10https://gerrit.wikimedia.org/r/89378 (owner: 10Bartosz Dziewoński) [20:50:12] (03CR) 10Reedy: [C: 032] Rename SpecialRecentChanges::addRecentChangesJS() to addModules() [core] - 10https://gerrit.wikimedia.org/r/89475 (owner: 10Bartosz Dziewoński) [20:53:17] (03PS1) 10Bartosz Dziewoński: Set 'watchcreations' and 'watchdefault' options to true [core] - 10https://gerrit.wikimedia.org/r/89604 [20:54:15] (03CR) 10Bartosz Dziewoński: [C: 04-1] "Needs https://gerrit.wikimedia.org/r/89603 merged first or Wikimedia wikis will allegedly explode." [core] - 10https://gerrit.wikimedia.org/r/89604 (owner: 10Bartosz Dziewoński) [20:55:04] Nemo_bis: ^ [20:55:30] (03Merged) 10jenkins-bot: Rename SpecialRecentChanges::addRecentChangesJS() to addModules() [core] - 10https://gerrit.wikimedia.org/r/89475 (owner: 10Bartosz Dziewoński) [20:56:53] (03CR) 10jenkins-bot: [V: 04-1] Set 'watchcreations' and 'watchdefault' options to true [core] - 10https://gerrit.wikimedia.org/r/89604 (owner: 10Bartosz Dziewoński) [21:01:51] (03CR) 10Bartosz Dziewoński: "Anybody is welcome to fix those tests, because they look nasty." [core] - 10https://gerrit.wikimedia.org/r/89604 (owner: 10Bartosz Dziewoński) [21:02:32] (03PS2) 10Bartosz Dziewoński: Set 'watchcreations' and 'watchdefault' options to true [core] - 10https://gerrit.wikimedia.org/r/89604 [21:03:42] Only 1 failure? I thought jenkins complained only above a hundred or so :) [21:04:17] Nemo_bis: it looks like it's a failure cause by other tests editing pages and accidentally adding entries to the watchlist [21:04:23] good luck tracking them down [21:04:51] (judging by the watchlist length, there are at least three) [21:05:34] (03CR) 10jenkins-bot: [V: 04-1] Set 'watchcreations' and 'watchdefault' options to true [core] - 10https://gerrit.wikimedia.org/r/89604 (owner: 10Bartosz Dziewoński) [21:05:54] what was the bug about tests making stupid assumptions about configs? [21:09:15] (03CR) 10Reedy: [C: 04-1] "(2 comments)" [core] - 10https://gerrit.wikimedia.org/r/86282 (owner: 10Bartosz Dziewoński) [21:10:19] (03CR) 10Bartosz Dziewoński: [C: 04-1] "This change is uncool now anyway." [core] - 10https://gerrit.wikimedia.org/r/86282 (owner: 10Bartosz Dziewoński) [21:15:03] MatmaRex: 86319 should almost be merged in first... [21:16:56] * Reedy moves it [21:17:47] (03PS3) 10Reedy: Remove calls to deprecated HTMLForm#setTitle [core] - 10https://gerrit.wikimedia.org/r/86319 (owner: 10Bartosz Dziewoński) [21:17:50] Reedy: okay [21:18:09] Reedy: you know you can do that via gerrit? cherry-pick to master :D [21:18:25] git fetch ssh://reedy@gerrit.wikimedia.org:29418/mediawiki/core refs/changes/19/86319/2 && git cherry-pick FETCH_HEAD && git review -R [21:18:29] Hard work, yo [21:18:49] (03CR) 10Reedy: [C: 032] Remove calls to deprecated HTMLForm#setTitle [core] - 10https://gerrit.wikimedia.org/r/86319 (owner: 10Bartosz Dziewoński) [21:18:52] (03CR) 10Bartosz Dziewoński: "(It's being hard-deprecated in https://gerrit.wikimedia.org/r/#/c/86304/ .)" [core] - 10https://gerrit.wikimedia.org/r/86319 (owner: 10Bartosz Dziewoński) [21:19:12] Reedy: trying to pull from or push to gerrit takes inordinate amounts of time for me [21:19:17] Easier to find and fix any other usages [21:19:27] like 40 seconds for a single 'git review' [21:19:43] git review is generally slow [21:19:57] I know git push origin HEAD:refs/for/master is quicker [21:20:47] (03PS1) 10Aklapper: Starting a bug summary with ">" would make it purple [wikimedia/bugzilla/modifications] - 10https://gerrit.wikimedia.org/r/89606 [21:20:48] gerrit is generally slow, too [21:20:56] the fact that git review does multiple ssh connections on each run is another thing [21:21:44] (03CR) 10Aklapper: [C: 031] "+1: Tested on http://boogs.wmflabs.org/show_bug.cgi?id=5 and decreases our diff to upstream a little bit" [wikimedia/bugzilla/modifications] - 10https://gerrit.wikimedia.org/r/89606 (owner: 10Aklapper) [21:21:55] afk [21:22:37] (03Merged) 10jenkins-bot: Remove calls to deprecated HTMLForm#setTitle [core] - 10https://gerrit.wikimedia.org/r/86319 (owner: 10Bartosz Dziewoński) [21:25:58] MatmaRex: 73 usages in the extensions in our repos [21:26:25] 45 in OSM [21:26:32] And numerous others in other extensions we care about [21:29:22] (03CR) 10Reedy: [C: 04-1] "73 usages in the extensions in our repos." [core] - 10https://gerrit.wikimedia.org/r/86304 (owner: 10Bartosz Dziewoński) [21:35:02] (03PS2) 10PleaseStand: LocalSettingsGenerator: add $wgGitBin [core] - 10https://gerrit.wikimedia.org/r/89498 [21:39:29] (03PS2) 10Aklapper: Italicize bug numbers of unconfirmed tickets in "Blocks/Depends on" field. [wikimedia/bugzilla/modifications] - 10https://gerrit.wikimedia.org/r/87053 [21:40:40] (03CR) 10Aklapper: [C: 031] "+1. Tested on http://boogs.wmflabs.org/show_bug.cgi?id=5 in the "Depends on" field (bug report number which is a dependency and which is U" [wikimedia/bugzilla/modifications] - 10https://gerrit.wikimedia.org/r/87053 (owner: 10Aklapper) [21:45:33] (03PS23) 10Reedy: Add archive, externallinks PK [core] - 10https://gerrit.wikimedia.org/r/51675 [22:22:35] MatmaRex, if EDIT_CLEAR is no longer used why don't we get rid of it? [22:25:49] Krenair: i dunno, ask whoever didn't remove it back in 2005 or so [22:25:58] Krenair: technically it's a public interface [22:28:38] MatmaRex: draft of dependency stuff: https://gerrit.wikimedia.org/r/#/c/89608/ [22:37:00] ori-l: meh, looks reasonable to me [22:39:00] it doesn't need to be shiny [22:39:15] unless you want to compile graphviz to javascript and render that to convas [22:39:19] canvas* :D [22:40:26] OK, I'll publish it [22:40:39] (03Draft1) 10Ori.livneh: List dependency relationships in mw.loader.inspect() [core] - 10https://gerrit.wikimedia.org/r/89608 [22:43:38] MatmaRex, well in this case I think any third party code depending on it SHOULD break - it's a feature removal [22:44:14] if I'm understanding it right that is [22:45:11] Krenair: gah, that was killed aeons ago and i only simplified some code [22:45:30] There's nothing in our extension repos [22:49:26] Krenair: if you want to help refactoring watchlist, i have some wip code i'm getting tired of :P [23:51:31] (03PS1) 10Legoktm: Add a tab with a link to edit on mediawiki.org [extensions/HelpPages] - 10https://gerrit.wikimedia.org/r/89609 [23:52:05] legoktm: Yo. [23:52:09] Hi [23:52:17] legoktm: Got a few minutes to poke at mw.loader.inspect()? [23:52:32] Probably [23:52:43] I need a small enhancement implemented. Have you tried it yet? [23:52:59] Basically I'd like to whitelist certain modules from inspect()'s output. [23:53:09] Because I've already determined that they're fine to load. [23:53:16] I haven't, but I saw the emails [23:53:18] For example, mediawiki.legacy.wikibits. [23:53:22] :-) [23:53:27] Try it out real quick in a console? [23:54:35] Just did [23:57:13] I suppose you could do something like mw.loader.inspect(['mediawiki.legacy.wikibits']) where the array is your whitelist [23:57:44] I'd like to be able to define a whitelist in a user subpage. [23:57:50] And then have inspect() account for it. [23:58:45] $wgInspectModulesToNotInspect = array( 'mediawiki.legacy.wikibits' ); [23:59:19] Is that a real global? [23:59:42] No :P [23:59:53] Shame.