[01:47:39] are there any articles on wikipedia that load a different set of scripts than all the others? [01:48:24] Emw: I'd imagine so [01:49:05] do you know of any examples? [01:49:52] Emw: some mediawiki extensions could trigger script inclusion on a per page basis. I'm not sure if any on wikipedia do. Additionally the wikipedians themselves do all sorts of weird js stuff themselves, that could potentially be per page [01:50:05] i'm looking for an example of an example that loads a different set of scripts (e.g., an additional JS module) based on the article's identity [01:50:23] an example of an article* [01:51:59] bawolff: ah, ok. can you think of an example on any mediawiki deployment where a page loads a module that any other pages in the same namespace don't? [01:52:22] (specifically, a javascript module) [01:52:32] umm, I think the quiz extension does load a js file ( but i don't think that extension has been moved to resource loader yet) [01:53:49] oh found one [01:53:55] i'm not familiar with the quiz extension. do you know of a page that would have that different JS module (and any other page in that namespace that doesn't have the module)? [01:54:18] Emw: The categorytree will include the ext.categorytree module on a page only if it has the parser tag on that page [01:54:49] For example: https://en.wikipedia.org/w/index.php?title=Wikipedia:Sandbox&oldid=482127617 [01:55:33] has the ext.categoryTree module loaded, while almost no other page does [01:55:37] bawolff: fantastic. i'll look into that. is that parser tag just an XML tag in the HTML file? [01:55:53] no, not quite [01:56:45] xml style wiki tags aren't really xml at all. Whenever mediawiki sees a in a wikitext source, it calls a php function which tells mediawiki what html to put instead of the tag [01:59:31] Emw: If i may ask, what exactly are you curious about/trying to accomplish? Maybe I could help point you to something in the right direction [02:04:08] bawolff: i'm trying to figure out how to load certain scripts on a subset of pages, based on the page's identity, such that the scripts would be loaded if the user-editable page contained some special tag or other indicator that those scripts should be loaded. ideally i'd like to load the script via resource loader (e.g., with the script names URL-parameterized in a request to load.php) [02:06:04] the indicator/tag that triggers the script loading should be user-editable -- e.g., users should be able to add or remove the tag into the page, maybe something akin to a template [02:06:47] (if more information would help here, please let me know) [02:07:18] Emw: yeah, that could easily be done using a parser hook or function [02:07:42] aka, something like {{#loadmagic:}} [02:09:39] Emw: what you do is call $parser->getOutput()->addModules( array( 'my.module.name' ) ); [02:09:47] from either a parser hook or tag [02:09:55] bawolff: very neat. i'm new to the mediawiki codebase, and just stumbled onto /mediawiki/docs/hooks.txt. is there any existing hooks that i could use to do that, or would i have to define a unique hooks? [02:10:14] s/hooks/hook [02:10:35] This actually doesn't really use the hook interface, but more some of the parsers methods [02:10:51] see https://www.mediawiki.org/wiki/Manual:Tag_extensions and https://www.mediawiki.org/wiki/Manual:Parser_functions [02:11:12] (be warned, some of the documentation on mw.org is of variable quality unfortunately :s ) [02:14:24] bawolff: great, i think i might have the gist of it [02:14:47] Feel free to drop by anytime if you have any questions [02:15:01] will do. this is a huge help, thanks [02:15:08] ( #mediawiki is the slightly more appropriate channel, but either channel works) [02:18:57] ok. do wikipedia articles make much/any use of parser functions? would there be any reason a wikipedia article couldn't use a user-include-able parser function to load a particular set of scripts (if those scripts happened to already exist in the mediawiki deployment)? [02:19:35] There's lots of parser functions in use, just most of them hidden away in the template namespace [02:20:11] generally parser functions do something on the page, and script loaded by them would be to support whatever action the parser function actually does [02:21:27] btw, for personal user scripts, people generally use the gadgets interface, which associates certain scripts with options in special:preferences [02:21:40] good to know [02:22:25] btw, if you're poking around the code base, http://svn.wikimedia.org/doc/ is generally a big help imo [02:22:48] oh very cool. i didn't know about that. [02:23:53] Especially when you're just getting started, its helpful when you need to lookup what methods of commonly used classes (like Title for example) do [02:34:23] Emw: btw, what type of scripts would you want to include on a per page basis? [02:35:42] so i'm thinking about trying to develop a way to integrate some WebGL-enabled interactivity features into mediawiki. my idea is to allow users to add edit some tag (maybe a parser function embedded in a template), and have that trigger a request for scripts (three.js and another script) and data that would put a manipulable 3D structure into the lead image area in an infobox. initially,... [02:35:44] ...this would make sense for chemical structures (e.g. simple molecules and proteins), but it could have applications to other three-dimensional subjects [02:36:16] Hmm, 3d models of chemicals have been on some people's wish list for quite a while [02:36:51] I know Jmol has been discussed in the past [02:36:55] ( https://www.mediawiki.org/wiki/Extension:Jmol ) [02:37:03] right, ideas about integrating jmol have been brewing for a few years [02:37:38] the big disadvantage to that is that it requires installed java, launching an applet, etc. [02:38:48] i think a webgl version of something like jmol -- GLmol (http://webglmol.sourceforge.jp/index-en.html) -- is a pretty promising alternative [02:40:02] the main advantage being that webgl is baked into the browser. the main disadvantage i see is that it's not supported in all browsers. IE9 and below don't support it, and it looks like IE10 might not either [02:41:56] thanks for pointing out that Jmol extension. i didn't realize one existed, and i can probably get a better idea of the mediawiki integration code i'd need by looking into that [02:42:35] Well given that the jmol extension was rejected as not being up to Wikimedia's standards (if i recall) it may not be the best example of integrating with mediawiki [02:43:19] any recollection on why it wasn't up to wikimedia's standards? [02:43:23] although i think the extension has been improved since then [02:43:31] It had XSS issues at one point [02:43:38] see also https://bugzilla.wikimedia.org/show_bug.cgi?id=16491 [02:44:49] * bawolff is giving a quick look through of the extension to see if anything sticks out [02:45:05] it does not use resource loader, but that's probably because it was written before RL existed [02:46:19] There's several things in that extension that kind of are just weird from a more php prespective [02:46:36] require_once("Title.php"); , eventhough Title.php is autoloaded [02:46:54] doing things like global $wgJmolForceNameSpace; from a global context [02:49:19] The way it sets up its hooks is kind of weird... [02:55:10] bawolff: one question i see in that bugzilla ticket that would apply to a potential GLmol extension is that of how the 3D data would be incorporated. one way would presumably be to allow a few new types of media files to be uploaded in Commons (CML, PDB, etc.); another way suggested in the ticket seems to be including the data within a special tag like (special user-editable data) [02:55:53] do you have any impression on which approach would be better? [02:55:59] I personally thing the upload a file method would be better [02:56:29] Especially if the CML files are more than say 10 lines long, and its unlikely your average editor would be able to meaningfully edit them [02:58:05] great point. so i'm thinking about doing this with PDB files, which are usually 100+ lines and require very specialized knowledge to meaningfully edit [02:58:45] yeah, in that case i definitly agree that image files are preferable [02:58:57] however, the number of PDB files that would be uploaded would be pretty small (probably < 10,000) [02:59:18] I still think thats ok [02:59:26] ah, ok [03:00:27] alright, i think i'm going to get back to work on this. thanks a ton for all the help and information [03:00:28] Unfortunatly the docs for making custom media handlers is kind of scarce. You may want to look at the oggHandler extension as an example though (Its an extension which inserts a video player into articles for video files, which i imagine would probably be somewhat similar to inserting a custom thing for rendering chemical files) [03:00:50] hmm [03:02:34] so the custom rendering is really controlled only by two or three javascript files and a PDB file, e.g. the three js files and one pdb file here: http://webglmol.sourceforge.jp/glmol/viewer.html [03:04:09] i know next to nothing about custom media handlers. given that glmol example being controlled by a two special JS files and chemical data file, do you think something like a glmol extension for mediawiki would need a custom media handler? [03:06:00] If you want it to be integrated with file handling (So writing [[file:Foo.pdb]] makes 3d thing appear) then it needs to be a custom media handler [03:07:30] it would probably be simpler then some of the other custom media handlers though [03:09:55] oh ok. thinking about it a little more, that seems like it would probably be a good idea [03:10:29] Well anyways i have to go. Good luck [16:48:41] hexmode: chrismcmahon after bug 35266, I'm wondering if we should start making a list of common mistakes that should be machine identifiable, and look at making a lint esk test for it [16:48:59] ie if there's '' with {$ in it, chances are it wants to be "" [16:49:19] yes [16:49:25] yes yes yes :) [16:55:01] * sumanah agrees [16:56:11] I suppose it might aswell be a wiki page, rather than a load of bug comments [16:56:22] Shall I start a [[Git/Lint check]] page [17:00:11] Reedy: you could add to [[Pre-commit checklist]], too [17:01:19] Reedy: I do see a lot of strange/inconsistent issues on labs beta instances. Frustrating when it's hard to tell if it's a code problem or an env problem. [17:03:02] chrismcmahon: report any you find so that we can address them. Bugzilla project "Wikimedia Labs", component "WMF Beta project" [17:03:18] I just started https://www.mediawiki.org/wiki/Common_commit_errors [17:03:44] Not sure if needs to be pre commit, certainly on git, post commit and having the bot review it [17:04:23] Reedy: the pre-commit checklist is stuff people are supposed to do -- sanity checks. [17:04:30] Yeah [17:04:41] Expecting people to do stuff is expecting issues ;) [17:04:47] heh [17:05:06] Things like the quote issue most people won't notice with looking over it [17:05:19] When you reported the bug, it was quite obvious to guess what the problem was! :p [17:05:26] fwiw, I didn't write it for other people, but as a reminder for myself. [17:06:01] Yeah [17:06:08] chrismcmahon: I think hexmode is right, either way, if it's a code error, or a config/setup difference, it needs dealing with to some extent, so it can at least be triaged [17:09:17] Reedy: it wasn't obvious to me :-) (at least not yet) [17:52:18] hi ChristianWikia [17:54:01] sumanah: Christian is AFK for a sec [17:54:21] rmoen: I wanted to ask whether the Wikia folks would be interested in mentoring for Google Summer of Code [17:55:05] sumanah: Inez said yes [17:55:39] Inez_: you should sign up at https://www.mediawiki.org/wiki/Summer_of_Code_2012 [17:55:46] Inez_: https://www.mediawiki.org/wiki/Summer_of_Code_2012#Mentor_signup [17:55:59] Inez_: you should also read https://www.mediawiki.org/wiki/Summer_of_Code_2012/management#GSoC_management_philosophy [17:56:02] thanks rmoen [17:56:25] sumanah: np[ [17:56:34] -[ [18:05:51] ? zhwiki shows a diff for edit preview? [18:06:21] hrm... maybe I just haven't done this on "add topic" before [18:07:54] yes, zhwiki has some weird settings [18:08:09] can't remember all of them :D [18:08:22] oh, hexmode, and thanks for your cute barnstar :) [18:08:29] Nemo_bis: :) [18:13:09] Nemo_bis: remember, today's the last day for Open Source Bridge talk submissions [18:13:33] thanks sumanah, but I'm already overwhelmed with things to do in the next couple of months [18:13:39] a deadline every day [18:13:40] ok [18:13:43] yeah [19:04:50] hexmode, what can be done to move this bug towards a conclusion? https://bugzilla.wikimedia.org/show_bug.cgi?id=27860 [19:05:24] * hexmode clicks since he doesn't know bugs from the number alone [19:05:29] (usually) [19:26:13] hexmode, yes, links are to be clicked :p [19:52:01] Nemo_bis: updated [19:53:13] hexmode, thank you! I hope thoughts come :) [19:54:12] anyway I doubt that the new behaviour has created any new customs which a revert would break, annoying people (while keeping it does) [19:55:08] Nemo_bis: I dunno.... I'm used to people saying "wah!" no matter what ;) [19:55:16] hehe [19:57:17] hexmode, if your comment means that a new config would be created, with the old behaviour as default, and en.wiki keeping its new toy to force people do what they want, I'm super ok with it [19:57:36] :) [22:05:46] sumanah: how does http://wikimania2012.wikimedia.org/wiki/Submissions/The_Wikipedia_Mobile_Experience look to you? [22:05:58] preilly: I will look! just a moment. [22:06:38] sumanah: okay, thanks! I really appreciate it. [22:06:43] :) [22:06:54] preilly: it needs a copyedit -- I can do that in a moment [22:07:14] preilly: but let me first talk about more substantive things [22:08:14] sumanah: okay [22:08:30] preilly: like: will this session also include a live demo or 3? :D [22:09:02] sumanah: now, that's a good question [22:09:18] preilly: because that would be cool of course :D [22:10:38] sumanah: I just added "This presentation will include demos of what things might come to look like in the future (e.g., visual editing on an iPhone) for example." to the end [22:10:48] * sumanah reloads [22:10:49] sumanah: what is the copy edit that it needs? [22:10:58] preilly: I'll do it, give me a moment [22:12:47] sumanah: thanks, for the edit [22:12:51] preilly: done. [22:12:53] you're welcome! [22:13:13] preilly: "and help shape the future of Wikipedia on mobile devices" if you have specific questions/structure to how you want more info or feedback from the attendees, that might be welcome [22:13:24] sumanah: good point [22:13:25] make it clear how hands-on this will be (or whether it will not be hands-on) [22:14:46] hi sumanah: question about git migration stuff here: https://bugzilla.wikimedia.org/showdependencytree.cgi?id=22596&hide_resolved=1 How are you distinguishing between Mar 21 blockers and stuff we should get to really soon after Mar 21? [22:15:40] sumanah: okay, thanks [22:15:46] robla: "the most recent comment from me" is not a good answer, but it is the one I have right now :-/ [22:15:47] sumanah: I greatly appreciate the feedback [22:16:08] preilly: I am grateful to you for writing a proposal and I can give you more feedback once you iterate again :D [22:16:20] robla: (the most recent comment from me on each bug) + whether it's high/highest priority [22:16:40] sumanah: all highest/high priority are blockers? [22:16:48] robla: lemme check to see whether I am lying [22:16:54] k. [22:18:10] * robla thinks Ryan_Lane may have actually completed https://bugzilla.wikimedia.org/show_bug.cgi?id=35209 [22:20:09] robla: pretty sure Antoine is on top of https://bugzilla.wikimedia.org/show_bug.cgi?id=34141. not sure if it's 100% but it's close. [22:20:43] inded. I just forgot to mark it fixed [22:20:51] robla: yes, all the High or Highest ones are blockers, and the Normal ones are not. [22:20:59] Ryan_Lane: s/inded/indeed [22:21:11] s/preilly/troll/ [22:21:13] :) [22:21:16] Ryan_Lane: FTFY [22:21:17] Ryan_Lane: oh! the final step on that is to actually inform the community via an email to wikitech-l I think [22:21:29] Ryan_Lane: can you do that, as I'm trying to do 5 things at once? there's a talk deadline today :( [22:21:37] Ryan_Lane: FTFY again [22:21:40] :D [22:21:52] sumanah: it's not going to be as informative as you'd like [22:22:13] Ryan_Lane: I'll take that risk and can add more info tomorrow :-/ [22:27:05] robla: status on all the open blockers, then: Antoine is still working on the l10n scripts; on the 21st we will pull the switch re fenari; I haven't talked with Roan and Krinkle yet re the updated Gerrit interface since late Wednesday so I need to check on that; I need to add people to the Gerrit project owner groups [22:28:22] preilly: one more thing - sometimes it makes a good impression if you specifically mention particular non-Wikipedia projects (Commons, et al.) that will benefit from your work :) [22:30:09] sumanah: okay, cool [22:30:21] preilly: also, now that you've written the abstract, you can probably find a more specifically engaging title than "The Wikipedia Mobile Experience" -- maybe something about mobile CONTRIBUTION. "uploading, editing, and curating from your phone: the magical mobile future" [22:30:30] ok, maybe that is not good [22:31:46] preilly: I remember tfinc mentioning how tablets make certain kinds of gnome/curation/antivandalism work possible from mobile devices? I bet that's something Wikimania people will love to hear [22:33:46] sumanah: ah, also a good point [22:34:10] * robla wanders off for a little bit, but will want to follow up on https://bugzilla.wikimedia.org/show_bug.cgi?id=35148 (Set up Gerrit project owner group for MediaWiki core + WMF-deployed extensions) [22:40:16] preilly: quick check: your Gerrit username is preilly, right? [22:40:26] sumanah: yes [22:40:37] good [22:40:44] permissions thing, you know how it is [22:41:45] tfinc: hey, do you have a Gerrit account? [22:42:05] sumanah: he is currently in a meeting [22:42:11] sumanah: not sure [22:42:18] and yes .. i'm on a call [22:42:19] interview [22:42:49] tfinc: can you log in at https://labsconsole.wikimedia.org ? I need to create a Labsconsole/Gerrit account for you because you will be a project owner for MediaWiki core [22:43:20] i do have a labs console account [22:43:44] tfinc: what's the username? [22:43:50] sumanah: tfinc [22:43:56] Tfinc to be exact [22:44:59] tfinc: neither tfinc nor Tfinc is a registered user in Gerrit. Argh! Ryan_Lane can you help solve this mystery? I want to add Tomasz as a project owner but when I type in things like "tfinc" and "Tfinc" and his email into the Members field and hit Add, I get a "not a registered user" error [22:45:29] preilly: hey, could you add an email address for yourself in your labsconsole/Gerrit preferences? you'll need that so you get notified of things [22:46:01] sumanah: my labs console preference has preilly@wikimedia.org [22:46:45] preilly: you may need to add it separately in Gerrit or something, because when I look at the list of project owners for the MediaWiki project, your username has no email associated.... [22:46:59] preilly: see https://gerrit.wikimedia.org/r/#admin,group,11 [22:47:33] sumanah: https://gerrit.wikimedia.org/r/#settings,contact contains the correct email address [22:48:14] preilly: thanks. something just reset or something; now your email shows up appropriately in the list [22:48:33] sumanah: no worries [22:49:10] he already has an account [22:49:34] sumanah: if he's never logged into gerrit, then he doesn't have an account there [22:49:38] it works like mediawiki [22:49:47] when you log in, your account is created [22:50:06] tfinc: ok, please log into https://gerrit.wikimedia.org with your Labsconsole.wikimedia.org credentials and then I can add you to the relevant group. [22:50:10] and it syncs the info from ldap [22:50:12] Thank you Ryan_Lane - I had forgotten that [22:50:18] no worries [22:51:07] Ryan_Lane: per our conversation earlier this week I am manually adding project owners to the mediawiki Gerrit project owners group so we can remove the "wmf" LDAP group, so we can use "wmf" to include non-owner people [22:51:22] * Ryan_Lane nods [22:52:15] Ryan_Lane: under what circumstances do we remove someone from the "wmf" LDAP group? like if they stop working for WMF? [22:52:27] yes [22:52:53] Ryan_Lane: then there's a little cleaning to do, IIUC [22:53:12] * Ryan_Lane shrugs [22:53:13] :) [22:53:20] yeah, we need some process for this [22:53:28] Ryan_Lane: would I file an rt ticket for the one-time cleanup? [22:53:34] sure [22:53:45] I don't know who should necessarily manage that group [22:53:48] probably the devs [22:53:51] since it's for wmf devs [22:54:12] you have the ability to clean it up, using modify-ldap-group [22:54:16] on formey [22:54:28] Ryan_Lane: remind me, how do I just see everyone who's in a specific LDAP group? [22:54:34] authz? [22:54:42] ldaplist -l group [22:55:06] or getent group [22:56:48] ok, and currently does that list ("wmf" LDAP group) drive anything in particular? I say that because I notice there are some incongruencies between it and https://gerrit.wikimedia.org/r/gitweb?p=operations/puppet.git;a=blob;f=manifests/admins.pp;h=2080ad4588963dc512543978936ac5367c8d1efd;hb=HEAD (admins::mortals) and I just want to get the Gerrit list right [22:57:04] it allows people to do group administration in gerrit [22:57:08] and create repos [22:57:19] (that first one explains why I am in it :D) [22:57:31] well, that and you are wmf ;) [22:57:47] right :) [22:59:15] so, to trace the causality here: if a person in the "wmf" LDAP group can do group administration in Gerrit, doesn't that mean that she could add herself as a project owner for anything? I presume we're just going to trust people not to use that loophole when they shouldn't :) [22:59:39] we should likely remove wmf from that privilege [22:59:50] in fact... [22:59:52] * Ryan_Lane does that [23:00:01] I was thinking about that the other day [23:00:25] I'm going to make another group [23:00:28] who should be in it? [23:00:30] Ryan_Lane: ok, good thing I got under the wire and added myself as a project owner! [23:00:44] Ryan_Lane: admins::mortals & admins::roots ? [23:00:50] yeah, we should really be added groups as project owners [23:00:55] hm [23:01:10] for now I'll leave in ops, and add a few other people manually [23:01:22] Ryan_Lane: take a look at https://gerrit.wikimedia.org/r/#admin,group,11 [23:01:23] in the future we'll have a user who can create repos and assign project owners [23:01:39] and we'll control it via labsconsole [23:02:23] some of those users are redundant, since they are in the groups :) [23:02:30] yeah [23:02:36] I can do cleanup once it's more settled [23:03:15] I'm adding chad, hashar, and you as admins for now [23:03:16] and removing wmf [23:03:37] Ryan_Lane: ok, you're removing "wmf" from what specifically? I was about to remove it from the mediawiki core Gerrit project group [23:03:45] Administrators of gerrit [23:03:50] ok [23:04:02] but yeah, remove it from there as well [23:04:04] * sumanah removes [23:04:15] at some point I'm only going to have ops as gerrit administrators [23:04:17] just didn't want a crazy edit conflict making Gerrit go total [23:04:26] once we can control it via labsconsole [23:05:17] yeah [23:06:21] nimish_g_: hey, have you ever logged in to Gerrit? [23:06:30] also poke Ryan Faulkner to do so please [23:07:24] sumanah: I haven't [23:08:58] nimish_g_: 'twould be good if you would [23:09:45] sumanah!: hi Sumana, nimish tells me you werte looking for me [23:10:09] rfaulk: yeah! could you please log in to Gerrit? that way I can give you various permissions [23:13:22] sumanah: what's the login for gerrit? is it the same as for labs? [23:13:28] nimish_g_: yes [23:13:29] sumanah!: hmm not sure that i have an account there yet [23:13:39] rfaulk: do you have a labsconsole.wikimedia.org account? [23:13:42] it's the same credentials [23:15:02] i don't think i do [23:15:17] ok, logged in [23:17:21] rfaulk: here is what I need from you [23:17:22] Preferred wiki username. This will also be the user's git username, so legal name would be reasonable [23:17:22] Preferred email address [23:17:22] Their svn account name [23:17:33] rfaulk: join #wikimedia-labs [23:17:33] thanks nimish_g_ [23:17:42] he *must* have one [23:17:53] he does [23:17:58] hm [23:18:00] maybe he doesn't [23:18:03] he = me or faulkner? [23:18:09] nimish_g_: you're all set :) [23:18:12] ah. he didn't have a USERINFO file [23:18:14] ok cool [23:18:18] rfaulk: tsk tsk!!! [23:18:23] rfaulk: !! [23:18:25] SCORN [23:18:27] :D [23:18:28] feel my scorn, Ryan [23:18:35] from across a continent [23:18:50] heh [23:18:58] ok, dinner [23:18:59] it is felt deeply [23:19:19] sumanah: saw your note on #mw....I think the backlog answers any questions I have already [23:19:27] see ya....have a great weekend! [23:20:04] ok! thanks robla and talk to you Monday.