[01:29:27] hi all, anyhow has any idea how to display the top 3 titles in the wiki main page ? [01:42:06] depends on what you mean by "top" [01:49:45] top meaning most number of likes [01:49:52] thanks for the response [01:50:34] and what is a "like"? is this some extension? [01:51:24] i used fblikebutton extension [01:52:21] well, if the fblikebutton extension doesn't provide such a feature, it probably doesn't exist [01:52:44] alright.. thanks for your response. Appreciated it. [04:01:51] Hey guys! I got a coding problem for ya'll. [04:03:19] I want to write a function which looks at the WP:GO page on the English Wikipedia (https://en.wikipedia.org/wiki/Wikipedia:Goings-on) and returns the freshest archived Goings-on page. [04:03:33] As of now that is (https://en.wikipedia.org/wiki/Wikipedia:Goings-on/May_24,_2015). [04:04:00] returns the URL*, and via API queries. [04:04:24] Because of the naming scheme, doing a listall request doesn't work too well: (https://en.wikipedia.org/w/api.php?action=query&list=allpages&apnamespace=4&aplimit=500&apprefix=Goings-on) [04:04:32] Does anyone have any alternative solutions that would work better? [04:04:59] can you guess the page title based off of the current date? [04:05:11] I'm thinking I'll just brute-force it and do a backwards check starting from the current date on the existence of the archive page, and keep going back until I get a page that exists. [04:05:23] ^ that's what I was thinking [04:05:34] you can batch multiple page existence checks in one request too [04:07:28] Right, ok, got it. [04:07:38] Time to dive the docs. [04:09:31] You wouldn't happen to know the specific API query I need would you? [04:09:46] * legoktm looks at docs [04:10:19] ResMar: https://en.wikipedia.org/w/api.php?action=query&titles=FooBar|Barack%20Obama|Nonexistent%20page [04:10:44] Ok, great, thanks. [04:10:52] ResMar: actually, use https://en.wikipedia.org/w/api.php?action=query&titles=FooBar|Barack%20Obama|Nonexistent%20page&formatversion=2 :) [04:10:55] I wanted to write a little importer script. [04:11:05] formatversion=2 makes the JSON output not suck [04:11:12] But now it's 450 lines long. [04:11:18] That includes comments but still. [04:11:28] Holy **** [04:11:36] WHY DID NO ONE TELL ME ABOUT THIS [04:11:50] I was on here earlier complaining about the shittiness of the API's JSON output. [04:12:07] :P [04:12:08] it's a new thing [04:12:18] are you on the mediawiki-api-announce mailing list? [04:12:20] You make a key out of an effectively random pageid? Really? [04:12:35] Why would anyone...ever... [04:12:52] No, I'm not I'm afraid. [04:12:55] Too many mailing lists. [04:12:56] the API was designed for XML, and then people realized it was terrible [04:13:12] the announce list is pretty low traffic, and has notifications of breaking changes and major milestones like formatversion=2 [04:13:24] Yeah but... [04:13:32] I have reason to be subscribed to every mailing list [04:13:39] I solve that problem by not subscribing to any of them. [04:13:49] Not even foundation-l, too spammy. [04:14:01] Not even? [04:14:11] foundation-l is one of the last lists you're likely to sign up to [04:14:58] Well, I have reason to [04:15:07] N&N at the Signpost has to be written somehow [04:16:27] Right, brute force... [05:38:09] hi all, can i know the difference between Auth_remoteuser extension and LDAP extension? [05:38:39] I am trying to implement an auto login feature . I am using PHP 5.3.3 (cgi-fgci), Apache 2.1.6, MySQL 5.1.73, mediawiki 1.22.0 [05:39:07] I've read both documents but could not know which one is better for my environment set up. Thanks to all advises. [06:30:20] loll mediawiki support desk [06:30:23] "I was trying to complete the quest Roving Elves and after i killed the 4 bosses i have to take the todo list to Arianwyn. I looked at the quest guide and went to where he is supposed to be located and he wasn't there. I cant find him anywhere." [06:30:53] send him to http://runescape.wikia.com thedj [06:31:48] not sure how he ended up on mediawiki.org, it's not like wikia links back anywhere obvious [06:42:19] onei: ppl just google for wiki help desk [06:42:33] we also get a lot of those ppl on en.wp help desk. [06:51:25] that's interesting [08:38:07] how do i put 2 images side-by-side in mediawiki? [08:40:55] morning all, got a fun one here - we've got an internal wiki we use, which is cool, but we've got a small issue with it. it seems that the category gremlin has gotten to it, and we're now seeing items appearing in categories they're not supposed to, and not appearing in categories that they are actually in - any suggestions? [08:41:56] case in point, a page in category "Customers", is appearing in the category "Suppliers" [08:42:45] I'm confused as to the reason why [08:45:33] I'll be back and forth in and out of meetings - but if anyone has any suggestions, please shout up and I'll review them when I get back [08:46:25] stangeland, put them into a table [08:47:08] MaxSem, i put them in a gallery [08:47:12] seems to work nicely [08:47:19] heh [09:30:15] hi [09:30:29] I'm installing mediawiki the first time. [09:30:56] Is it a good idea to create a separate mysql user for this? [09:31:38] This is on my server and I have no other use for mysql atm, but this may change. [09:32:09] Or ... can I migrate the wiki database to another user easily later? [09:32:52] I think I can. [09:34:19] it's not terribly hard to move it to another user, and you should always use a dedicated user (least privilege and all), [09:34:34] k [09:35:15] Which privileges does the wiki user need? [09:38:50] that I'm not 100% sure of, but if definitely doesn't need access to anything but its own database [09:39:08] k [09:39:31] global permissions are dangerous even for a dedicated mysql server because it enlarges the attack surface by quite a bit [09:39:42] and may give filesystem access indirectly [09:46:21] also, SUPER privileges can lock you out of your own server if you reach max_connections [09:50:59] Hi does anyone know of an extension that allows a user to overwrite an image with a different file type [09:51:16] currently the file mimechecker is making it not possible to do [09:55:13] If I enter the mysql root user at http://xxx/mediawiki/mw-config/index.php/?page=DBConnect will it use it as wiki user, or create a wiki user? [10:01:31] bbl [11:40:19] I'm trying to run unit tests on 1 extension, 1 extension only. Ok, so I can get phpunit.php to work for Just The One Test. But when I try to do Code Coverage it covers something like 10KLOC , while I'd really just like to stick to 'my' extension ... [11:41:02] (obv code coverage analysis over All Of Mediawiki kind of makes it hard to see what's going on with just the 1 extension) .... Any ideas how to constrain that? [12:33:38] Bonjour, quelqu'un parle français ici ? :) [12:41:41] oui mais c'est un chan anglophone [12:44:05] Je peux quand même m'exprimer en anglais ? [12:44:09] en français, pardon [13:25:22] CRH|Cloud: Oui! [15:12:55] Funnyn the "Traditio" wiki is a pioneer when it comes to MediaWiki releases :) https://wikiapiary.com/wiki/Generator:MediaWiki_1.25.1 [15:21:09] Anyone here do developer support? Is there a channel for that? [15:24:59] legalhazard: #wikimedia-dev [16:23:22] dear colleagues, i'm having trouble installing the code editor extension [16:24:05] i git cloned it from gerrit into extensions/CodeEditor and required it in LocalSettings, but it doesn't show up under Special:Version [16:28:00] re [16:28:03] If I enter the mysql root user at http://xxx/mediawiki/mw-config/index.php/?page=DBConnect will it use it as wiki user, or create a wiki user? [16:30:09] darsie: it should create the wiki user [16:30:15] ok [16:30:36] but do double check. chaning the user later isn't hard [16:30:44] yes [16:48:38] i just git-pulled the latest mediawiki, now getting this error msg: [16:48:41] Fatal error: Class 'UtfNormal\Validator' not found in /usr/share/nginx/sites/wiki/languages/Language.php on line 3097 [16:49:02] despens: run composer [16:49:23] * bawolff thought we had a better error message for that now [16:50:35] !composer [16:50:35] Composer is a dependency manager for PHP. MediaWiki uses it to include certain external libraries. See https://www.mediawiki.org/wiki/Composer for more details. [16:50:57] That's totally not the type of help page I was looking for [16:51:23] m) of course, forgot about composer [16:51:25] thanks! [16:52:25] bawolff: what should be on the help page? [16:52:56] instructions for what despens needs to do to make a git checkout work [16:54:15] that's on https://www.mediawiki.org/wiki/Download_from_Git#Fetch_external_libraries [16:56:31] !composer del [16:56:31] Successfully removed composer [16:57:12] !composer is Composer is a dependency manager for PHP. MediaWiki uses it to include certain external libraries. See https://www.mediawiki.org/wiki/Composer for more details. For information on how to install MediaWiki dependencies with composer see https://www.mediawiki.org/wiki/Download_from_Git#Fetch_external_libraries [16:57:13] Key was added [17:31:35] !composer [17:31:35] Composer is a dependency manager for PHP. MediaWiki uses it to include certain external libraries. See https://www.mediawiki.org/wiki/Composer for more details. For information on how to install MediaWiki dependencies with composer see https://www.mediawiki.org/wiki/Download_from_Git#Fetch_external_libraries [17:31:59] hey, i am having difficulties installing CodeEditor [17:32:29] it doesn't show up at Special:Version [17:33:34] i did wfLoadExtension( 'CodeEditor' ); $wgCodeEditorEnableCore = true; [17:33:46] but there is not even an error message [17:35:43] does code editor conflict with visual editor? [17:44:46] despens: the CodeEditor only loads on certain pages like JS ones or CSS (User:FooBar/common.js) [17:44:54] despens: and VisualEditor should not load on those pages [17:48:04] legoktm: why don't we have codeeditor for wikitext? [17:48:13] that can't be very hard, right?... [17:48:21] DanielK_WMDE_: there was a patch open for that... [17:48:48] DanielK_WMDE_: It's called VisualEditor. ;-) [17:49:02] James_F: not quite the same thing... [17:49:07] DanielK_WMDE_: oh, maybe not. Well, someone would have to write the ace implementation for it [17:49:19] DanielK_WMDE_: I know. [17:58:37] legoktm: the CodeEditor is not listed on Special:Version and is not invoked when i edit MediaWiki:Common.css ... :/ [17:59:14] despens: if it's not on special:version something is going wrong. What version of CodeEditor and MediaWiki are you using? [18:00:04] i cloned the codeeditor from gerrit and my mediawiki is 1.25alpha (9f01c8d) [18:01:02] ok, there's probably a version mismatch. [18:01:21] first, you should be using 1.25.1 which is stable now [18:01:39] Where could I find the code in mediawiki that does HTTP Basic Auth? [18:01:40] a lot of things have changed since february [18:02:13] legoktm: yeah if i git pull the latest mw, nothing works anymore ;) [18:02:55] what's breaking? you probably should be sticking to stable releases instead of alphas [18:03:51] debajit: HTTP basic auth? We don't do that (in default mediawiki. There's probably an extension that does) [18:04:56] debajit: Extension:HttpAuth unmaintained ? but links to https://www.mediawiki.org/wiki/Extension:Auth_remoteuser [18:07:47] bawolff: I'm looking at the wiki API authentication [18:08:48] debajit: The mediawiki api does not use HTTP basic auth [18:09:10] anomie: heya, are you available? I'm updating https://www.mediawiki.org/wiki/API:Query#Continuing_queries and https://www.mediawiki.org/wiki/API:Raw_query_continue and tackling batchcomplete. so many questions... [18:09:10] bawolff: Thanks for the clarification. It must be something we have added in our mediawiki deployment [18:09:55] OMG, he's trying to fix bug 1. Get Him!!! [18:10:17] bawolff: If I needed to see how API auth worked, where do I start looking? (I'm having a bit of trouble figuring out my way through the code) [18:10:30] !authplugin [18:10:30] http://www.mediawiki.org/wiki/AuthPlugin [18:10:35] has some information [18:10:50] bawolff: ;-) BTW T2001 is around for nostalgia, prefer the #documentation tag. [18:11:22] anomie: I'll just fire away here, maybe legoktm has an opinion [18:12:10] spagewmf: I'm here [18:12:47] 1) I'm adding a FAQ item about API changes, do you agree with Yuri's suggestion to specify formatversion=1 if you're change-averse? [18:12:52] On a serious note, I'm appreciative of everyone trying to make our docs better. So thanks :) [18:13:21] spagewmf: honestly, I never properly understood how old query continuation worked and just used whatever pywikibot did. [18:13:35] rawcontinue= will be mandatory for 1.26, AIUI formatversion=1 is future-proofing [18:16:58] the new section is https://www.mediawiki.org/wiki/API:FAQ#deal_with_2015.27s_API_changes.3F [18:17:18] 2) I went to update https://www.mediawiki.org/wiki/API:Raw_query_continue and found the XML results are quite different [18:18:13] bawolff: Thanks! [18:18:33] 2a) List of became List of, [18:20:12] * bawolff notes that a compliant xml client should probably handle either variation just fine [18:20:14] 2b) now has an accontinue attr, the example output has acfrom. Was that a copy-paste error, or did the output change [18:21:58] Quick question, how do I use superscript? [18:22:15] Pache: Umm, I think ^{Text in superscript} works [18:22:56] Ty [18:23:05] spagewmf: 1) Yeah, that's a good idea. I'd have mentioned it myself but I thought it might scare people ;) [18:23:46] anomie: when might we flip the default formatversion? [18:24:54] spagewmf: 2) First query is fine to change to match current output. Second and third we should tweak the query somehow to more-or-less match. [18:26:22] anomie: might the wiki page date from when query-continue returned acfrom not accontinue? Maybe yuri (no-autocomplete, not here) would know. [18:26:36] spagewmf: For 2b, I believe that changed in 2b3f4d821cd41. [18:26:54] which was in 2012 [18:26:59] hi [18:27:12] bawolff: nice sleuthing, thanks! [18:27:12] spagewmf: I have no intention at this time to ever change the default formatversion. If we were to do so, we'd announce the impending change during version V, have warnings in V+1, and not change until at least V+2. [18:27:18] how to enter break via mw edit button? [18:27:43] Juandev: you mean a line break or break into debugger or ? [18:28:09] spagewmf: Might be. Doesn't matter for that one. [18:28:20] Juandev: Which edit toolbar? [18:28:28] spagewmf: now I add tag foo, but Id like to start foo at the new line [18:28:48] Juandev: Just write
before [18:28:49] this is my code> https://cs.wikisource.org/wiki/U%C5%BEivatel:Juandev/vector.js [18:28:55] 2a) Yeah, that came along with the introduction of formatversion=2. Feel free to just update that, or just change the whole thing to json. [18:29:27] spagewmf: 2b) The output did change at some point. [18:29:55] bawolff: probably afterno? so: " tagOpen: "
","? [18:30:31] I don't actually know what does [18:31:07] yeah,
's can go in poem [18:31:20] or you can just add a newline to the start of your poem probably [18:33:16] lets see [18:33:58] anomie 3) the Query_raw_continue Continuation example [1] no longer returns tag [18:34:01] [1] https://www.mediawiki.org/w/api.php?action=query&titles=API:Main_page&generator=categories&prop=categories|links&gcllimit=2&cllimit=1&rawcontinue=&format=xml [18:37:51] ah, I think it's just that the first couple of pages it returns don't have any links. 3) solved, though it's a strange example [18:39:02] spagewmf: 3) The lack of links is probably from https://www.mediawiki.org/w/index.php?title=Category:Manual&diff=next&oldid=982782 and https://www.mediawiki.org/w/index.php?title=Category:Documentation&diff=760817&oldid=504785. [18:42:25] spagewmf: It *is* a contrived example, to show some of the complexity needed to use raw continuation correctly. The specific queries used don't really matter, as long as the pattern follows: first has , then we eventually get a , then finally just . [18:46:53] thanks. anomie 4), the dreaded batchcomplete. The fog is lifting on that one thanks to your https://phabricator.wikimedia.org/T84977#936343 . I think what might help readers is a sketch of how the API returns results for multiple props, each potentially with its own foolimit. [18:47:25] so not actually a question yet 8-) https://www.mediawiki.org/wiki/API:Query#batchcomplete so far. [18:47:57] anomie: and that's it. thanks good sir. [18:48:52] spagewmf: I don't know if the python code there is good or not, and it doesn't actually use batchcomplete, but the paragraph seems sane. [18:50:43] anomie: heh, why written English should use XML to denote nesting depth. I'll move it up. [18:55:03] I wish every sample code fragment linked to where it originated. https://phabricator.wikimedia.org/T91626 or something simpler. [18:59:26] I think yurik wrote that python [19:16:11] legoktm: i needed to run an alpha because of wikibase, visual editor, and other bleeding edge extensions. i guess i will need to upgrade to a stable and stick with it [19:16:25] hi [19:16:37] I can see a git repository and git commit history in mediawiki? [19:18:08] The_Phototrapher: https://phabricator.wikimedia.org/diffusion/MW/ [19:18:52] phabricator? [19:21:35] legoktm: https://secure.phabricator.com/book/phabricator/article/diffusion/ ?? [19:24:58] The_Phototrapher: um, what are you trying to do? [19:25:31] legoktm: I want see simply a commit history on web [19:25:52] The_Phototrapher: https://phabricator.wikimedia.org/diffusion/MW/history/master/ [19:26:59] legoktm: exactly that, however, phabricator has another things [19:29:51] bawolff: I have installed in a server mediawiki and in another there are severals git repositories [19:30:35] If you're looking to just use mediawiki, I'd reccomend just download the tar file, its probably easier [19:32:29] * SamB_7 wonders if something can be done about lonely alphabetic headings on category pages [19:33:00] SamB_7: What sort of something did you have in mind? [19:34:09] (that is, to avoid column breaks immediately after the alphabetic headings) [19:34:52] Oh I thought we did that. Its certainly possible so long as we are still using table layouts [19:35:10] There was some patches to convert it to css layouts where its slightly harder to do that, but I think those got reverted [19:36:24] hmm, well, I forget where I saw this anyway [19:36:52] bawolff: they didn't [19:37:22] oh well, I don't know anything about css columns. I have no idea if that's tricky then [19:37:23] SamB_7: bawolff: there's a very simple CSS rule for this, but it's only supported by old Opera 12 and IE 11, if i remember correctly [19:38:05] I guess using table-based layouts has basically two drawbacks: can't adapt the number of columns to the browser window, and can't know if any given page name will fit on one line [19:38:34] and its tables. Everyone hates tables [19:39:03] MatmaRex: got a bugzilla.mozilla.org link? [19:40:08] bawolff: honestly, I don't hate tables for layout as much as everyone else seems to. I just think it should be possible to mark them as non-semantic, and provide a "reading order" [19:41:15] and, obviously, they aren't the most appropriate way to do homogenous columns [19:41:47] SamB_7: nope. but here's a jsfiddle you can test with: http://jsfiddle.net/9rb5wtj1/1/ [19:41:50] (from https://phabricator.wikimedia.org/T55130#1098438) [19:42:51] SamB_7: +1 to that. I'd love to have for non-semantic tables. [19:43:10] MediaWiki could just handle that and convert it to or

as desired [19:44:59] also, comparing with a category page on Wikia, which I imagine is using the table-based rendering, the CSS-based layout is (iirc) also missing " Cont." headings [19:45:15] (where the r in iirc can be either reason or recall) [19:45:54] http://www.wikia.com/Special:Version hmm [19:46:32] I mean seriously, how would you do that in CSS? [19:47:16] That might even be tricky in TeX [19:48:10] (because, you know, you'd have to avoid inserting one immediately before a new letter) [19:51:45] (though I don't expect any of you are even as familiar with the relevant arcane corners of TeX as I am, which is "not very" -- there just isn't that much call for such knowledge) [19:57:02] (the one thing I can tell you with confidence is that multi-column layouts are basically done by telling TeX itself that you want a much narrower box of text on the page than you do, hiding it away, and then asking it to please try again before bundling all the text up and actually shipping out a page) [20:04:25] lonely headers, in contrast, are easily avoided simply simply by assigning a sufficiently large penalty to the vertical glue (basically, whitespace) immediately following the heading, so that breaking the page there will result in a large enough "badness" that some other break will be chosen instead [20:07:55] (which, given that the relevant algorithms in TeX have no knowledge of any columnar layout that may be applied to the typeset "page" bodies, applies identically to columns as to pages)