[00:48:32] in visual editor, how can i put a footnote in? [00:55:15] k-man: Insert > More > Basic reference or type i don't have Basic reference or type in that menu [00:56:40] there's a keyboard shortcut too, but the mapping depends on your OS/browser. Hit the "?" menu item and select Keyboard shortcuts to see what it is (⌘+⇧+K on my macbook) [00:57:13] ok [00:57:32] so its ctrl+shift+k for me [00:58:29] oh. Basic reference, or just type the literal string it should notice that and pop up the correct UI form [00:59:21] just like typing {{ will open up the template insert dialog [00:59:25] oh [00:59:53] VE is kind of magic :) [01:00:22] do i need a footnote template to have a footnote? I want a little [1] next to some text, then as a footnote, i want "[1] some foot note words" [01:01:22] that should work without templates, but you will need to have the Cite extension installed I think [01:01:39] https://www.mediawiki.org/wiki/Extension:Cite [01:02:51] ok, its working [01:03:17] of course the footnote wasn't showing up until i saved the page. [01:03:28] :) [01:04:28] so the formatting seems a bit off - it doesn't have any proper deliniation from the page content to the footnote content. As I have a numbered list at the end of the page, it looks odd [01:04:41] is there a scratch mediawiki instance somewhere I can test it on? [01:04:53] or sandbox [01:05:14] https://test2.wikipedia.org/wiki/Main_Page [01:06:01] but yeah, you have to style around the tag you insert yourself [01:07:16] Usually I do "==References==" or something like that just before inserting [01:08:59] ah ok [01:13:23] nice tip, thanks bd808 [01:13:46] i didn't haev a so i guess it just puts it at the end if you don't specificaly have that? [01:28:16] is there a list of shortcuts for visual editor, like the is there a shortcut to insert ? [02:14:33] Hi, can anyone tell me what JSON files are? [05:22:37] Platonides: in terms of a MariaDB instance on Amazon RDS [07:09:59] Sometimes while I'm in UploadWizard, on the last step where it adds the file with the description and everything, while it uploads files I get 'You must be logged in to upload this file.' even though I just logged on or refresh another page on the wiki to make sure my session is still active... did this happened to anyone? [09:30:50] extarys_: not to me [09:31:04] extarys_: but please report a bug at phabricator in case someone else reproduces it waaaay later [10:37:20] Is there a system for installing extensions automatically? [10:39:02] spycrab0: https://www.mediawiki.org/wiki/Special:ExtensionDistributor ? [10:39:14] wait no sorry, forget it [10:40:19] spycrab0: i guess https://www.mediawiki.org/wiki/Extension:ManageWiki is maybe the best i know of [10:41:18] I'll check that out thanks. [10:44:06] Is there a way to get the latest wmf release over git? I mean I guess I could iterate over the tags but that seems cumbersome. So does manually updating every time. [10:54:18] if you are a developer, maybe mediawiki-vagrant is useful for you [10:55:04] I tried to upgrade wiki to 1.31 from 1.30 and decided to revert because extensions weren't working, now this doesn't even seem to recognize itself [10:55:08] https://wikizilla.org/wiki/Sanda [10:55:20] Every page is a 404 error [10:55:49] I restored the backup and then this happened. I don't know what to do. [11:07:00] The 404 image there is from the /wiki/images so I assume the website's still there... but everything shows the 404 page.... [15:25:29] Hi! I have specific question about the Wikimedia API for a project app [15:26:04] Is anyone here knowledgeable about the API? I've been messing around with the Sandbox and can't quite hone in on what I'm looking for [15:27:13] I'm specifically making a song search app that gives links to the searched for songs and then if clicked, will give every possible detail about that song [15:29:16] dsound: I think you might get more responses in #mediawiki-dev [15:29:40] thank you! [15:30:55] #mediawiki-dev is not a channel? [15:31:25] it is not, it forwards to #wikimedia-dev [15:31:36] #wikimedia-tech is probably the channel that spycrab0 meant [15:31:54] #wikimedia-dev would also be a poor channel for dsound's question [15:31:57] in any case, help with the API is likely best suited for this channel [15:32:34] Oh sorry, I misremembered the channel name :/ [15:33:29] dsound: your main issue is that the API can fetch page contents and do general searches, but data in wiki pages is largely unstructured. Wikidata is a project to expose that better, and you can see if your songs have any data items on wikidata that you could query. Otherwise you're mostly limited to parsing the infobox template or whatnot to extract the info from page text [15:33:45] dsound: This is page is definitely something you want to look at if you haven't already: https://www.mediawiki.org/wiki/API:Main_page [15:33:59] (which is prone to failure) [15:34:01] +1 for wikidata :) [15:34:02] dsound: checkout query.wikidata.org for being able to do SPARQL queries against wikidata [15:34:36] It really depends, do you just want a summary - in which case you could probably use the extracts api module to get like the first paragraph description from wikipedia [15:34:48] Otherwise, definitely checkout wikidata, for facts on an individual level [15:38:20] There are even quite a few music related examples in there. [15:39:16] Be warned, SPARQL has a tad bit of a learning curve, especially if you are not familar with the semantic web model [16:10:52] hello [18:07:57] just a vague hope: does anybody use the scribunto extension in termux? I can't get it to work [18:13:05] If you describe whatever issue your having, maybe someone will know the answer [18:35:48] I'm not sure myself. When running with nornal standalonelua I get 'Syntax Error "(" unexpected' on all templates using scribunto. Using a lua installation leads to 'libreadline.so.7 not found'. I have no clue how to go on, since I neither have an idea how to make scribunto log more than the errorlog nor what is wrong with my libreadline. [18:36:46] I was vaguely hoping for someone using scribunto and saying 'haha, there's a simple catch' [18:39:10] What OS? [18:40:30] android 7 [18:41:02] Uh [18:41:09] How are you even trying to use scribunto on android? [18:41:19] Termux busybox [18:42:55] It..works quite well, but not well enough yet. And I don't really want to take the step of installing linux yet [18:42:59] * Reedy scratches his head [18:43:30] Can you install libreadline in your package manager? [18:43:44] When you say "lua installation" what does that mean? [18:43:55] is that using lua from your package manager? [18:44:09] Using the lua binary from the package manager, yes [18:44:09] https://wiki.termux.com/wiki/Package_Management [18:44:19] Well actually, i guess you must, as i assume android is on arm and we don't bundle an arm version [18:44:53] https://github.com/termux/termux-packages/tree/master/packages/readline [18:44:57] Definitely seems to be packages [18:44:59] You're probably going to have to compile everything yourself, if I'm being honest [18:45:09] any os I install will be arm still, since arm is the hardware [18:45:16] This is kinda uncharted territory [18:45:35] So..learning curve ahead, thank you for the warning [18:46:05] Try installing readline 7 in the meantime, it looks like that's what they're packaging [18:46:17] Also, based on above... This is very much unsupported [18:46:44] I already have it installed :/ and thank you for bearing with me [18:46:46] heh [18:46:47] https://www.mediawiki.org/wiki/Topic:Sqeu40egi1wa49a9 [18:46:52] "This SO question seems relevant. Apparently, the lua binary included with Scribunto is not ok with Raspberry Pi, which may be a similar problem with a synology. [18:46:53] Basically, try to install lua from your distribution and set $wgScribuntoEngineConf['luastandalone']['luaPath'] to the path of the lua executable and see if that works." [18:46:58] http://stackoverflow.com/questions/29555459/mediawiki-scribunto-lua-error-internal-error-the-interpreter-exited-with-statu [18:48:29] did it, was source of the libreadline missing/not being linkable error. [18:49:39] Guess we should fix the error message. "syntax error: unexpected "("" is a terrible error for architecture not supported [18:50:11] Well, that's not our fault :P [18:50:21] That's just the underlying issue from trying to run the binary [18:50:39] I'm happy to have an explanation at least :) [18:51:00] computers suck? [18:51:44] Dependency management in particular sucks ;) [18:52:18] scribuntoProblem: Did you try installing readline package? [18:54:38] bawolff: I have the stable version installed [18:54:45] Didn't try dev yet [18:54:53] Do you know what version? [18:55:06] key thing is you need version 7 [18:55:24] well, dev doesn't mean it's not stable [18:55:32] I got 7.0.5-2 [18:55:45] Dev means it may or may be not stable [18:55:50] No [18:55:54] -dev usually means things like headers and stuff that are useful for debugging [18:56:13] I got a libreadline.so.7 in my lib folder, so all seams great [18:56:17] A dev package is different from a dev branch [18:56:21] scribuntoProblem: Well if you have that, its probably a path issue I would guess. For some reason its looking for libreadline in the wrong place [18:57:21] Thats kind of a good idea, since busybox paths are messed up [19:00:32] Reedy: I'm clueless concerning the ecosystem: a dev-package is nothing in development but something aimed at developers? [19:01:12] scribuntoProblem: Its for when you are writing your own programs [19:01:42] typically -dev packages contain header files, which are neccesary if you want to create new programs that are using the library [19:03:02] so the hacky way to fix your issue, would be to figure out where its looking for libraries, and symlink libreadline.so.7 there [19:04:07] So..you did the easy part of the solution :/ I'm yet to figure out where it is looking [19:04:26] Header files are like the *.h interfaces in c ? [19:04:43] yes [19:04:46] And please shout at me if its too offtopic [19:08:58] Its fine in this channel, but you may get better answers at a support forum for your specific OS, as its really a question about your OS package manager [19:09:56] You already made me atleast understand my problem [19:11:13] Currently I just want to use the foreach template, so if all fails I can fall back to php. Even if I wasted more characters on the problem than I would have to repeat [20:13:03] Just a shimmer of notunderstanding again. Mediawiki is a bunch of scripts for a httpd demon.. how does it bite down on my operating system that hard? [20:14:47] itsallmyfault: Because abstractions are fun! [20:15:19] it's also not exactly a lightweight application [20:15:20] I'm all for abstraction [20:15:54] I just hate myself for missing the php and linux train [20:16:04] rendering any given page involves issuing about a dozen db queries (or over a hundred if you're using db to cache i18n messages) [20:16:38] on top of that, there's a lot of string parsing and manipulation, both of page content as well as other stuff to get it to display [20:17:08] I mean, if you setup profiling, I could tell you in detail what's taking so much resources [20:17:09] plus the file IO to load its hundreds of php files [20:17:48] (since you probably aren't using opcode caching, or if you are, it may be way too constrained with available RAM to be useful) [20:18:25] Thats a result of the classic stateful vs stateless services debate.. and not an issue of technology [20:18:40] it's an issue of resources [20:19:10] a high-performance webserver or db server will have oodles of memory available to use for caching, so that things don't have to make it to PHP or the db server or whatever if they don't absolutely have to [20:19:39] So the webserver is like an oracle? [20:19:50] no? [20:20:16] just RAM is faster than hard disk, and fetching pre-made results is faster than asking the app to dynamically create it all over again [20:20:41] But it relies on 'same request->same output:' [20:20:56] yes, the application tells the server what aspects of the web page (if any) are cacheable [20:21:20] But it caches parts of things, and then does later things with it [20:21:44] If you want to compare it to a hash oracle, varnish would be the relavent layer [20:22:16] I missed a whole lot of stuff:/ [20:22:42] tl;dr don't expect to get good performance when running a webserver on your phone or table [20:22:45] *tablet [20:23:08] In particular, MW is generally optimized for a big powerful webserver environment [20:23:27] When there are tradeoffs to be made performance wise, it took the choice that makes the most sense for a big server deployment [20:24:22] Not to say you can't optimize it for small-form devices...but its definitely not the target audiance [20:24:42] I understand that path. But for me, mediawiki provides great support for personal notes. [20:25:47] That's good. Even if optimized for the big server case, we aim to make sure it still works decently on smaller things [20:26:36] Just to make sure: it does. I just have an issue with a small subset of templates [20:27:12] If your on a resource constrained device, there's probably a lot of stuff to do with optimizing caching layer that can help with performance. As probably suddenly you can't store it all, but you could probably make choices over which types of cache are more important then others [20:27:30] But that's definitely a bunch of work, and something you'd have to guide with profiling [20:28:21] As an aside, for personal notes, some people really like tiddlywiki https://tiddlywiki.com/ [20:29:34] I tried that too, but mediawiki has the structure of wikipedia, so every person is already familiar to it [20:29:45] Woo, world domination! [20:30:12] Tiddlywiki is great! But not great enough. [20:31:19] oh gosh. My wiki is aimed at other developers [20:33:08] btw, if you want to go down the rabbit hole of performance optimization, see https://www.mediawiki.org/wiki/Manual:Profiling [20:35:14] I'm a professional developer in mainframes, so even talking about using megabytes of ram blows my mind. Currently I don't experience any performance issues on the 1gb tablet running my wiki and the browser [20:36:01] Something something, kids these days and their ram hungry ways ;) [20:40:04] Its a fpm interpreting a couple of thousand lines of php [20:40:14] Cant be that bad [20:45:02] Hi, Anyone know any tool that can change the formatting of page wikicode like described here: https://pastebin.com/H2ZVRAbs (like prettyprint), ty [20:46:43] anything could regex something like that [20:47:45] I'm looking for something written that does that (Can write that with pywikibot, but checking before) [20:49:48] but, those aren't the same thing semantically; the extra line breaks can cause issues in certain cases [20:50:05] (notably, if you're using implicit/unnamed template parameters) [20:51:14] You mean that the param would be "first\n"? [20:53:09] perhaps it'd be easier to explain by way of example [20:54:54] https://dpaste.de/FfJk [20:55:08] so as long as you're using named params, you're safe [20:55:22] the moment you go into using {{{1}}}, {{{2}}} and the like, the newlines in the template call matter [20:58:36] tried that, and got you, why that? the param value may be parse at the same way no? [21:00:11] no, because the "pretty-print" version has newlines at the end of each param value [21:00:27] those newlines are NOT stripped when using implicit (unnamed) parameters [21:00:38] but they are stripped when using named parameters [21:03:15] niso: see https://www.thetestwiki.org/wiki/Sandbox for a demo [21:03:39] so, any script you end up using will need to ensure it only does the pretty-printing for named params [21:07:44] yea, ive tested that on my wiki also, ty. but why this is the case for unamed params? they cannot be treated (without showing the user) as named params with indexing? in your exmaple just add "|1=Test|2=cat"? [21:08:56] niso: I don't know the reasoning, it's just how it is. The moment you name the params (and you can always name a param, even if it's a number), it strips newlines [21:09:11] ty! [21:10:51] I think its a backcompat thing [21:11:30] The idea to kill whitespace came later, and named parameters were introduced later [21:11:35] maybe [21:12:23] that's likely the case [21:39:17] Heh, just ended up writing a mediawiki extension manager in python :P [22:10:03] Is there a way to delete files in batch? [22:10:11] Or UploadWizard to overwrite existing images with a new version? [23:42:18] How can I add bootstrap of foundation to help make nice pages? :P [23:51:16] extarys_: there is a bootstrap skin