[00:46:39] Nikerabbit, yo - are you in SF already? [01:38:00] leaving just now for airport [02:02:24] oh [02:02:44] have a nice flight [12:08:36] who was bz guy who asked me about integration of bugzilla to wm-bot? [12:08:47] hashar I guess heh [15:25:32] Anyone know Daniel Werner's username? [15:27:15] Or rather, could I get some input on some code real quick? [15:27:22] nullspoon: Probably Daniel_WMDE? [15:27:28] that was my guess [15:27:38] DanielW_WMDE [15:27:41] didn't wanna bother the wrong person though. ;) [15:28:31] nullspoon: mh? [15:29:05] DanielW_WMDE: mh? [15:29:13] Dunno that acronym [15:29:16] Daniel Werner, that's me^^ [15:29:32] ah ha. Good good. Thanks for getting back to me. [15:29:43] I wanted to run a few things by you before recomitting my code. [15:29:49] Iif you have a few seconds. [15:29:55] sure [15:29:57] Thanks [15:29:59] Change 11137 [15:30:03] Title.php [15:30:37] !g 11137 [15:30:37] https://gerrit.wikimedia.org/r/#q,11137,n,z [15:30:49] Geeze. I always forget about that. :) [15:31:05] DanielW_WMDE: So, first comment. You suggested I use empty instead of count to speed things up. [15:31:54] I agree. One thought on that though. I used count because the earlier code that checks wgWhitelistRead also uses isset and count. Should I update that to use empty as well? [15:33:05] if it is related to your commit, you can change it of course. In general, count() is slower than empty() for array checks [15:33:46] It's not really related to my commit. I'll make another one to update the wgWhitelistRead check. [15:34:04] sounds reasonable [15:34:12] Good good. [15:34:21] now my last one [15:34:24] second comment [15:35:00] Brian Wolf already rejected my commit on for checking dbName and name but then agreed I should check for both once we discussed it. [15:35:12] er, rejected my commit for checking* [15:35:39] I'm really open to either option here. [15:35:42] why check for both? [15:35:59] Well my reasoning for that was because dbName is used in the URI and name is used as the actual page title. [15:36:17] Which means the end admin sees both and might try to use the one seen in the URI. [15:36:50] (dbName is name with underscores instead of spaces) [15:38:13] DanielW_WMDE: I'm really not trying to be difficult. You guys run this project, not me, so I'm happy with whatever. [15:38:27] nullspoon: One moment please, I have to check for a few possible implications here [15:38:32] In the end I'm just super stoked that we have such amazing open source projects. [15:38:38] Sure thing. [15:44:48] nullspoon: seems like there are no problems, just use both then. [15:46:03] DanielW_WMDE: You sure? This is the second time it's come up with the same comments so I'm wondering if I shouldn't just get right of one of the checks. [15:47:00] DanielW_WMDE: I mean, I could just put the comments into DefaultSettings saying it only checks on the page title with spaces. [15:47:08] It'd speed things up a bit too. [15:47:22] nullspoon: honestly, I would still go for the non-DB form because it is a configuration thing which has to be documented properly and then it should be clear for the admin what to use. Having a lot of tollerance here makes the thing just more error-prone and slower. But it doesn't really matter in this case I suppose [15:48:46] DanielW_WMDE: I'll go with you suggestion of just checking $name then. You're likely a much better dev than I. After all, my change has been rejected a whopping 29 times. ;) [15:50:27] DanielW_WMDE: Thanks so much for your input. I'll go ahead and get this pushed for review. [15:51:16] nullspoon: don't worry, had a review sitting there for over two months as well ;) By the way, if you are doing this for some customer, you could simply put it into an extension, delivering it way earlier and then going through all the trouble of bringing it from the extension into core. [15:52:44] DanielW_WMDE: Thanks for the suggestion. I actually implemented this at work for my team's wiki and figured it'd be handy since I've seen several people asking around for it. [15:53:20] DanielW_WMDE: Now i'm just working on it in my spare time because it's fun to work for an open source project that's contributing so much to the world and has such a great community. [15:53:33] DanielW_WMDE: (not to mention I use it for my personal blog) [15:54:11] great and thangs for contributing! keep up the good work ;) [15:54:58] Thanks for putting together an awesome project! :) I've learned a tremendous amount form attempting to contribute code to this project. [15:55:11] you can ask marktraceur. I'm becoming a git master. [15:55:19] :) [16:09:55] nullspoon: Nice pun. [16:10:24] marktraceur: oh my gosh I didn't even see that. [16:10:26] haha [16:10:52] My goodness I hope no one pulls code from me. It's gonna be buggy as ever. [16:10:59] they'll regret it for sure. haha [16:13:15] nullspoon: That's why we use code review and version control, friend. Don't worry about it :) [16:15:05] marktraceur: Which by the way, I'm thinking I might implement gerrit at work. [16:15:29] marktraceur: The wikimedia projects have been of great interest of me for several months now. [16:16:05] You guys have figured out ways to do the entire development to production workflow with 100% open source software, a feat no company would take on it seems. [16:16:26] nullspoon: /me disagrees with at least part of that statement [16:16:32] sorry - few [16:16:32] :) [16:16:39] But most of the infrastructure is free software, it's true [16:16:51] I mean, we've got rackspace using open stack and xen [16:16:58] my company for instance [16:17:04] we're on sharepoint for everything [16:17:13] Everyone hates it [16:17:24] *nod* sounds about right [16:17:31] I got fed up with it an set up a mediawiki instance and ported all of our docs over. [16:17:36] Just for my team [16:17:58] Now numerous other teams are wanting to piggy back off of our wiki becasue it works so well. [16:19:07] nullspoon: I used to work for Socialtext, they sell an Enterprise wiki (with other features bolted on). Sharepoint->wiki migrations were a source of income. No one likes Sharepoint. [16:19:07] Either way though, I'm just intruiged by how you guys do everything so well with *nearly* 100% open source [16:19:21] chrismcmahon: Seems like there's a pretty big market for it. I have only met one person who likes sharepoint and that person has never used anythign else. [16:20:21] I don't wanna bash Sharepoint too much. I mean, Microsoft did a pretty heft job there and it works for a lot of people. [16:20:46] For documentation only though, mediawiki beats the pants off of sharepoint. [16:21:30] It's funny to watch the more team members as they learn wikitext markup and templates. "Oh my gosh...look at this outofdate template I discovered!!!" [16:21:36] lol [16:21:46] more senior* [16:22:51] hey TrevorParscal [16:22:57] hello [16:23:00] I think you were quite a CSS guy? [16:23:12] nullspoon: Sharepoint is a glorified file server :) [16:23:18] sure, that's a way to describe me I think [16:24:36] TomDaley: Agreed. One that takes many a resource. [16:24:36] trying to get the sidebar to behave better when reducing the viewport size [16:24:43] but without placing the actual sidebar html before the content [16:25:14] I'm out of ideas, but maybe you have some brilliant one at some point [16:25:39] Platonides: right, so you need to consider 2 things [16:25:59] 1, if the sidebar is longer than the page, i needs to push the footer down [16:26:11] 2. if it's not the content should push the footer down [16:26:33] 2.a in this case, do you want the right side of the page to continue the background color of the sidebar? [16:27:04] I don't think the footer is the problem, if I remember right it has a clear: both [16:27:30] I don't really care about the sidebar bgcolor to expand or not [16:27:34] clear:both only works with floating though, if I give you a solution that uses absolute positioning for instance, you will be in hot water [16:27:53] also, is this supposed to use media queries to expand/collapse this automatically or what? [16:28:12] I tried with absolute positioning, but with a small width it overlapped the content [16:28:22] the page has to be far to wide for it to show up [16:28:27] and left the margin as blank space at the right :( [16:29:02] you mean the archive section? [16:29:20] it is probably done with AJAX [16:29:28] blogger code [16:29:54] ah, we're not trying to give several resolution sizes with media queries [16:30:27] (or a resized window) [16:30:54] you currently need 1585px width for it to show [16:30:57] well, at full screen on my 13" mac book air, the side bar was still on the bottom [16:31:11] I had to make the browser 200px or so wider to get it to come up to the top right [16:31:15] this page is simply too wide [16:31:16] yes, that's the kind of thing it would be good to avoid :) [16:31:40] you should design for a 1024px wide monitor, which means about 960px after scrollbars and borders and such [16:31:45] maybe 1000px [16:33:37] I didn't make the original design [16:33:42] just tried to adapt it [16:33:48] you need to put the content and sidebar in a div that's horizontal margins are set to auto (so it's centered) and it's width is set to like 960 - 100px [16:34:54] then set the content and sidebar widths to some split of the remaining content (don't use any margins or borders and leave some gap for the distance between them, at least a few pixels for rounding errors) [16:35:13] float the content left, float the sidebar right, use a clear both after them, and call it a day [21:05:36] Krinkle, RoanKattouw, anyone : the .js in resources/ has Doxygen-style comments, but Doxygen doesn't handle JavaScript without hacks (I filed bug 40143). What do people use to generate JS docs? [21:07:23] spagewmf: personally i never generate the docs; i read them in the source or let IDEs like NetBeans read them for me [21:09:47] spagewmf: I believe there was something about hacks? Hacks sounds like fun. [21:10:15] Brion I too read the source, I'm looking for something to encourage me to use the conventions properly. Maybe you can show me an IDE handling JS comments at Tech Days. [21:12:30] marktraceur I linked in the bug. The hacks rewrite JS to C or Java so that Doxygen will process it. Seems crazy, but "Add support to parse/document other languages" has highest degree of difficulty at http://www.stack.nl/~dimitri/doxygen/todo.html [21:12:49] Hm [21:12:51] Very odd [21:14:10] BTW, if you install doxygen, use --no-install-recommends (`sudo apt-get --no-install-recommends install doxygen`) to avoid pulling in 430MB of LaTex support. [21:14:59] Haha, awesome [21:15:31] !b 40143 [21:15:31] https://bugzilla.wikimedia.org/show_bug.cgi?id=40143 [21:16:29] Something tells me it might be easier to just use a different system for the JavaScript [21:17:21] jsdoc? Seems like a reasonable name for a javascript docs generator [21:19:17] HA [21:19:22] This is a silly page [21:19:26] https://en.wikipedia.org/wiki/Comparison_of_documentation_generators#Language_support [21:19:36] JavaScript isn't one of the languages in the table [21:19:42] But JSDoc is one of the list items [21:20:34] Oh I see [21:20:36] Two tables [21:21:18] Could use Sphinx [21:21:48] marktraceur yeah I used a different generator off github for a JS project, some "bear" name? There's some benefit to using the same tooling; OTOH doxygen might get confused and cross-reference PHP functions and JS functions. [21:22:26] Eugh, sounds narsty. [21:23:21] marktraceur: there's a Sphinx expert whom I met at OSCON and who emailed me asking if anyone at WMF wanted to meet & talk Sphinx when he's in SF next [21:23:43] Oooh [21:23:56] Shall I fwd? [21:24:10] It looks like Sphinx (sphinx!) is not a code comments -> HTML generator, though [21:25:25] spagewmf: Looks like we'd need to document in separate files, and given the dogged longevity of bug #1.... [21:25:53] Putting up barriers like "opening a new file" and "learning a new markup language" might be problematic [21:25:54] marktraceur: bug 1 is descriptive, not prescriptive! let us fight towards making it ever less true over time :) [21:26:21] I agree [21:26:38] But I think we can do that with a tool that's *like* doxygen, except....not evil to JS [21:27:27] Robodoc seems mostly OK [21:27:32] Lots of caps lock [21:27:47] Haaaa, tiddlywiki [21:33:18] * paulproteus waves to marktraceur + all [21:33:18] * paulproteus is on the third floor, near some couches and some windows. Maybe that doesn't help for clarity? [21:33:18] Anyway, marktraceur, for notes for the all-staff, I was curious what you thought about using your extension for Etherpadded MediaWiki. Good idea, bad idea, great idea, terrifying idea? [21:35:00] paulproteus: Oh, pff, this is silly, I'll come find you [21:35:02] b [21:35:04] marktraceur thanks for your thoughts. Maybe Krinkle and RoanKattouw already use a JS doc tool or IDE for all the resource JS. [21:35:05] (thumbs up) [21:35:24] We don't have a JS doc tool AFAIK [22:14:09] Krinkle|detached: Forgot to add the link to [1] in that email :) [22:14:35] Did something change recently about the load-order or user JS? [22:14:40] or = of [22:15:33] It currently loads before RL modules, but I thought it used to load after [22:20:17] Krinkle: ^ [22:30:35] spagewmf: RoanKattouw_away: Indeed, no parsing of JS docs yet (other than by us humans, though consistency is important for humans as well). Though there's 2 links: [22:30:50] https://www.mediawiki.org/wiki/CC/JS#Documentation_comments [22:31:09] https://www.mediawiki.org/wiki/Requests_for_comment/Documentation_overhaul [22:32:32] jdlrobson: Wait, FIFTH floor kitchen? [22:32:41] * RoanKattouw just went on a fruitless hunt for chocolate [22:40:48] RoanKattouw: We are flooded with chocolate on the 6th floor [22:40:57] Swiss and Belgian [22:43:38] I've brought French candy, but no chocolate. [22:43:53] guillom: What sort of French candy? [22:44:16] marktraceur: nougat, pates de fruits, violettes [22:44:34] Interesting! [22:44:36] I'm going to give them to Mel to share them during the all-staff [22:44:42] Sounds good [22:45:02] The last time, I put them in the kitchen, but given that everything is off-site this year, the kitchen might not work too well. [22:45:24] *nod* [22:45:34] Unless you wanted to keep the candy, in which case that's a perfect plan [22:46:32] Ironically, I'm not a big fan of candy. But they had a lot of success among my dear colleagues :) [22:46:48] What do you want from us? Sugar is awesome. [22:46:51] :) [22:46:53] Then again, I'm sure wine would have a lot of success, and I don't drink wine either. [22:47:00] (Yes, I'm the Frenchman who doesn't drink wine.) [22:47:42] It's OK [22:51:58] kaldari: Where where [22:52:10] I looked in the kitchen and didn't see any [22:53:41] near the entrance [23:01:11] I wonder if that was 6th floor [23:01:14] it sure wasn't 5th though [23:01:27] * apergos goes chocolate hunting [23:01:40] scsi bug hunting isn't getting me very far [23:10:26] if you use loader.using with multiple dependancies, like... mw.loader.using( ['jquery.foobar', 'mediawiki.foobar']... the dependancies are loaded asynchronously, correct? [23:10:44] RoanKattouw: ^ [23:11:21] Ye [23:11:32] thanks [23:11:36] You don't know when or in which order they will load, unless they also have a dependency link between them [23:11:46] All you know is that your using callback will only run after they [23:11:51] 've both loaded, whenever that is [23:12:08] that's what I though, just wanted to confirm [23:26:19] Anyone else been getting this from git review -d ? [23:26:20] Could not parse json query response: 'generator' object has no attribute 'next' [23:38:23] Krenair: You need to update git-review [23:38:43] If you're using Mac OS, you will probably also need to upgrade pip before you can upgrade git-review [23:38:49] I did that a few days ago... Let's see if I can remember the command [23:38:56] ubuntu 12.04 [23:39:01] pip install -U git-review [23:39:04] (+sudo)