[02:09:37] Does anyone know what the API query for getting a list of subpages is? [02:09:54] I feel like it's in query somewhere but I can't find it and I feel like I'm going blind or something [02:09:58] https://en.wikipedia.org/w/api.php?action=help&modules=query [02:12:02] All subpages? [02:12:06] Or subpage of a specific page? [02:12:45] If it's for a specific subpage... https://en.wikipedia.org/w/api.php?action=help&modules=query%2Bprefixsearch should work [02:20:33] prefixsearch does fuzzy searches [02:20:58] you’ll want allpages with apprefix parameter [02:22:39] mmm [02:22:43] Huh, really? [02:23:20] I guess that sort of makes sense, but I find it surprising, for some reason. [02:23:47] really. you might not see a difference on your wiki because the built-in search backend is lame. [02:24:17] An enormous amount of institutional resources at the WMF are being siphoned into search [02:24:20] wikimedia wikis are using cirrussearch [02:24:59] CirrusSearch isn't the backend? [02:25:40] The old search algorithm is a crusty dinosaur, you're telling me we haven't nuked it from orbit yet? [02:26:03] lucene is long dead [02:26:16] But if you setup your own wiki, it'll just use mysql search by default [02:26:18] Which will suck [02:26:48] Okay. [02:26:54] We really ought to be paying down our technical debt [02:27:08] I'm confused [02:27:13] Why do you think we're not? [02:27:29] We can't expect people to use ElasticSearch on every wiki [02:27:30] Oh no, I just mean that I think it should be more of a priority, and done more aggressively. [02:27:40] I'm still confused [02:28:00] I'm confused that you're confused [02:28:07] What tech debt are you talking about [02:28:24] You should see the amount of work a few of us have done fixing up old code usage in the last few weeks... Having to fix code in many apparently abandoned extensions [02:28:31] Stuff that has been deprecated since 1.16 and such [02:28:58] Good to hear! [02:29:24] https://github.com/wikimedia/mediawiki/blob/master/RELEASE-NOTES-1.27 [02:29:27] Lots of removed stuff [02:29:43] Yeah it looks like it. [02:29:49] That's fraking great [02:30:22] I think I made/reviewed between 300 and 500 commits a few weekends ago [03:39:47] Hi,I am not able to add comment in https://www.mediawiki.org/wiki/Git/New_repositories/Requests for my extension PipVideoJS.The comments are addded to bottom row of some other requests [03:41:09] I wanted to add the comment that my gerrit account name is pip [08:21:33] Anyone have suggestions for this problem: What ever i try to do reagaring login in to my wiki i get "invalid username". If i manually add a user the user-table in the database gets populated by a user without username. Tried to debug the wiki but it shows nothing... My own account is logged on looks like because of cookies [12:31:19] Hey Guys.. Someone familiar with linking or embeding files in mediawiki form another server that is only reachable in the local network (but the file should be reachable over the internet). Like a FileProxy for MW [12:31:58] Or could it work via a network share? [12:38:22] if you really need that, NFS is probably best [12:39:50] or you could do an http reverse proxy set up for those requests, though that is likely more work than using NFS [12:47:01] Are there any tools to convert a Mediawiki database into a Git repository? [12:56:43] Does MW can use and "proxy" those files via nfs without problem?.. but i would like to maybe use a samba instead of an nfs [13:19:12] Dziuba: mediawiki just sees files [13:20:09] To answer myself: git comes with a remote mediawiki plugin: git clone mediawiki::http://where.your.wiki.is/ [13:20:09] and CIFS works just as well if that's more convenient for you [13:21:50] and im able to access files via the wiki that are local network only if i use cifs? [13:23:13] you mount the network share where MW expects to see the files and it accesses them as normal [13:23:42] that's the point of a network share [13:24:36] applications don't know or care that the files are backed by a network share instead of a local filesystem [13:25:55] they dont need to be hashed or need metadata for MW? [13:26:20] they need to be exactly what was there before [13:29:17] but its a file like a pdf thats constantly changing that should be available via MW but its not hashed or uploaded.. MW should access the original file thats on the other Server and Provide it even for people that access the MW over the internet [13:30:54] if it's not uploaded via MW, then it's not available via MW [13:31:36] but you can still link to whatever on your pages and use a reverse proxy to make that work [13:32:21] it would likely only be some more configuration in your webserver [13:35:36] Maybe ill try a reverse proxy.. never worked with it before but i test it out.. if it doesnt work ill just write a script that uploads the file via the api [13:36:53] There's a maintenance script to import files [13:37:12] you can also mount a network share somewhere outside MW and have your webserver serve the files in it [13:39:35] first decide whether you want these files to be tracked by MW or not and go from there [13:41:25] my apache already serve the files but only local.. so people accessing mediawiki from the "outside" cant access the files linked in MW.. i was thinking that maybe MW could proxy them for me [13:42:56] but seems like i need to upload them or proxy them by myself (but never worked with proxy before so we will see) :) [13:51:39] how to create forms using mediawiki? [13:52:26] I have got mediawiki installed and setup on my localhost, how can I create a form page with save and submit buttons on it? [15:57:51] Hi I am Aditya Gulati. I am participating in GCI 2015. [16:16:41] adityagulati: hello, and welcome. [17:50:08] Anyone that can assist for a bit, I cant really get this problem. All of a sudden when my users want to login they get "you have not specified a valid username" but they have, only reason i can login is because of cookies. Also its not possible to change usernames and account creation just creates an account but the username is not forwarded to the database.. Only thing was done before this [17:50:08] was to install UploadWizard extension disabled it and all other unnecessary extensions but no difference. [17:50:16] anyone that have a clue or suggestion? [17:50:46] What MW version? [17:50:54] Are you using any auth plugins etc? [17:53:39] 1.26.2 , only running network auth to get VisualEditor to work [17:54:02] Why do you need networkauth for ve? [17:55:05] Its a private wiki with ACL and it was the only way i got it to work [17:56:04] Because Parsoid needs to be able to view pages [17:56:36] I'd imagine oauth would maybe be an option, but whatever [17:57:11] ill check it out, but disabled networkauth and HaloACL at the moment, and still get problems [17:57:19] an other thing i just noticed [17:57:36] tried to klick my username to create an article, but then gives me catchable fatala error [17:57:42] jober: What kinds of usernames are they putting into the login box? [17:58:07] Ooh, what's the error? [17:58:14] Just normal onese, my own "Joel".. [17:58:16] the error is : [17:58:18] Catchable fatal error: Method User::__toString() must return a string value in /var/lib/mediawiki/includes/Block.php on line 249 [17:58:37] Hmm, have you updated MediaWiki lately? [17:58:47] nope, running 1.26.2 since start [17:58:51] Hm. [17:58:51] just got it up last week [17:59:00] NetworkAuth code is awful [17:59:27] jober: How did you do the installation? [17:59:51] apt-get, its running on a Ubunto port for ARM, running it on a Raspberry pi 2 [17:59:55] Oh god [18:00:07] OK so, disclaimer [18:00:14] The Debian package of MediaWiki is bad [18:00:17] It always has been [18:01:01] but its worked quite normal, i run a private wiki at the office, with the normal install (not apt-get) and its bascially the same on the surface [18:01:14] OK, but [18:01:27] jober: You got it working a week ago, and everyone registered, thus logging in. Right? [18:01:34] I think the cookies by default expire in one week [18:01:58] sounds about right [18:02:36] That's the end of my ability to sleuth, though. I don't know NetworkAuth. [18:03:06] I run cognito mode in Chrome to try to login to my own account, but just gives me same info, oh and one more error similar to the other one [18:03:07] Catchable fatal error: Method User::__toString() must return a string value in /var/lib/mediawiki/includes/Block.php on line 266 [18:03:39] Right, so that's the problem [18:03:49] But more likely, something is being subclassed and causing the problem [18:04:49] And NA was updated last 3 years ago, which means it's my first suspect [18:05:41] Ok, so basically make sure everything with NA is disabled and run php update.php [18:05:43] ? [18:06:26] I'd say try that first [18:07:04] doing it now :) [18:12:04] so, tested logging in, it fails in incognito mode, tried to clik my username, same err as before and tried to add a new useraccount it populates the user table with a user with no (empty) username... :/ [18:25:52] trying to remove all the packages that came with imagemagick and build-essential [18:26:00] hope its that crap that didnt work :) [18:40:08] hi [18:40:16] i am trying to configure short URLs for mediawiki on nginx [18:40:31] http://www.website.com/wiki/Page_title [18:40:33] is how i want it [18:40:42] and i have mediawiki installed in /w/, as per the standard [18:41:11] i have tried the short URL tool from redwerks [18:41:34] which may have worked in the past for me, but for some reason does not now. i have the fastcgi_params location correct as well as the password [18:41:41] and the error i'm getting [18:41:46] is that all of the pages have no styles [18:42:02] Hello! This page has a table with anchors in it (I've included one as an example). The page renders the table before other parts, however, so the anchor tag doesn't work correctly--the table position is adjusted after the browser jumps to the anchor. Is there any way to fix this? https://wiki.waze.com/wiki/Glossary#JAI [18:42:18] and when i try to access a page that uses the old format (/w/index.php?title=Page_title) it just displays a blank page [18:42:32] are there any accurate guides for nginx out there [18:42:36] or is there something i'm missing? [18:48:06] hey [18:48:29] i can't really help, but try this if you haven't already: [18:48:33] !shorturls [18:48:33] To create simple URLs (such as the /wiki/PAGENAME style URLs on Wikimedia sites), follow the instructions at or try the new beta tool at . There are instructions for most different webserver setups. If you have problems getting the rewrite rules to work, see !rewriteproblem [18:48:48] !rewriteproblem [18:48:48] 1) Try as a fail-safe method; 2) Do not put the files into the document root; 3) Do not map the pages into the document root (see !nullpath); 4) Use different, non-overlapping paths for real files and virtual pages; 5) Do not set a RewriteBase; 6) Put all rules into the .htaccess file in the document root. See also !shorturls [18:49:15] thanks, MatmaRex_mobile [19:29:09] Hello! This page has a table with anchors in it (I've included one as an example). The page renders the table before other parts, however, so the anchor tag doesn't work correctly--the table position is adjusted after the browser jumps to the anchor. Is there any way to fix this? https://wiki.waze.com/wiki/Glossary#JAI [19:39:52] !debug [19:39:52] For information on debugging (including viewing errors), see http://www.mediawiki.org/wiki/Manual:How_to_debug . A list of related configuration variables is at https://www.mediawiki.org/wiki/Manual:Configuration_settings#Debug.2Flogging [21:03:18] heya. in the 3rd paragraph on https://www.mediawiki.org/wiki/VisualEditor , it links to "VisualEditor:Welcome" ( https://www.mediawiki.org/w/index.php?title=VisualEditor:Welcome&useskin=vector ), but that page has been deleted [21:03:34] somebody with write access to the visual editor page could possibly deal with that :) [21:04:13] oh, i actually might have write access. just not sure whether that link should be removed or replaced with something else [21:04:26] ah, brilliant... why an external link is used for VisualEditor:Welcome? [21:04:59] the admin deleting it was unable to see where it was linked to [21:06:36] ugh [21:06:54] also, the visual editor has progressed nicely since i last looked at it. congratulations and thanks to everybody involved :) [21:07:02] i like the table support a lot :) [21:07:06] because of the 'useskin=vector', it seems [21:07:14] should i undelete it? [21:07:21] it seems to be just a sandbox page [21:07:48] I'd just delete the link [21:07:52] so i'd rather remove the link [21:07:56] well, there's also https://www.mediawiki.org/wiki/VisualEditor:Test [21:08:05] and https://www.mediawiki.org/wiki/Project:Sandbox [21:08:09] lots of sandboxes :) [21:08:32] ping me if you need deleting/undeleting pages at mw.org, not touching this otherwise :) [21:08:40] https://www.mediawiki.org/wiki/Project:Sandbox says "To edit, click here or "edit" at the top of the page", but that opens visual editor. the same notice says "There's also a Sandbox for the upcoming VisualEditor" [21:09:14] i guess that infobox is slightly outdated by now [22:36:16] So now migrated the broken wiki to a VPS instead of a raspberry [22:36:54] Does anyone knows about UploadWizard, and PHP Fatal error: Call to undefined method UploadBase::getMaxPhpUploadSize() in /var/www/wiki/extensions/UploadWizard/UploadWizard.config.php on line 566?? [22:37:24] basically it simply says: "'maxPhpUploadSize' => UploadBase::getMaxPhpUploadSize()," on that line [22:39:40] Hmmm [22:40:03] jober: That's usually an issue with updating MW and not updating the dependencies [22:40:47] At least based on every time I've seen it [22:40:53] jober: What version of PHP are you running? [22:40:59] looks like i found the problem [22:41:03] OK [22:41:10] I was wrong to ask that anyway [22:41:11] it cloned a too new git version [22:41:32] Yeah, I forgot you were running 1.26.2 for a second [22:41:35] updated to rel_1.26 and it works now :) [22:42:03] but couldnt fix the issue in the raspberry anyhow, perhaps some extention broke a table or so [22:42:51] imported the sql into the vps, but put in the wiki without apt-get, so perhaps that makes a difference [22:43:05] Hm. [22:43:15] will try to run it ona Pinecone instead of a raspberry [22:43:20] but thats a few months away [22:43:29] but X64 not arm [22:43:38] So many fruits running MediaWiki these days [22:43:44] Fruits and nuts [22:43:50] yeah, indeed [22:44:02] hade some huge issues with the arm versions for a while [22:44:05] with Parsoid [22:44:14] but the dev version worked ok in thet end [22:44:20] until what ever broke, broke [22:46:33] personal comments… [22:47:49] <- looks vaguely like a coconut> [22:47:52] roundish. [22:49:02] Hah [22:52:16] thanks for the mental support and suggestions marktraceur, always nice to have a IRC group to cry out at :) [23:04:00] Hello! Can you help me? I am trying to login using Wikipedias API. It always gives me "NotExists", although I am entering my correct username. [23:08:01] Parckwart: Hi, one sec [23:08:23] Parckwart: What is your username? [23:08:30] Ventaro [23:08:33] German Wikipedia [23:09:33] I agree with you [23:09:36] https://en.wikipedia.org/wiki/Special:Log/Ventaro [23:09:40] But anyway [23:09:51] Parckwart: Can you pastebin the exact error? [23:11:10] is there a way to force mediawiki to do things? I'm trying to install centralauth but it just doesn't do some things and I don't know why [23:12:00] marktraceur: http://pastebin.com/jtSAPh4F [23:12:07] A lot of ugly html mess [23:13:02] Parckwart: Are you requesting an HTML page on the API? [23:13:40] Yes [23:13:57] I could also use ?format=php [23:14:18] Or XML or something [23:15:01] json would be best [23:15:33] Okay… [23:15:46] Chiyo: What do you want to force it to do? [23:15:56] {"login":{"result":"NotExists"}} [23:16:37] Parckwart: Okay, weird [23:16:40] What are you trying to login with? [23:16:45] curl [23:17:16] curl -b cookies -c cookies --data-urlencode 'lgname=Ventaro&lgpassworood=password' 'https://de.wikipedia.org/w/api.php?action=login&lgtoken=token&format=json' [23:18:29] Why is lgpassword apparently localised? [23:19:16] Oh, its just a paste error [23:19:25] curl -b cookies -c cookies --data-urlencode 'lgname=Ventaro&lgpassword=password' 'https://de.wikipedia.org/w/api.php?action=login&lgtoken=token&format=json' [23:19:52] marktraceur, errm, never mind for a minute as I think I missed an important variable [23:19:53] You know you need to do one request, get the token, then do a second request, with the same details AND the token? [23:20:18] Yes [23:21:15] So what are the two requests you're doing? [23:21:32] The first one without the lgtoken [23:21:44] The second is what i just pasted [23:22:05] You are passing back the token you're being given? [23:22:19] The token is allright, I think, because otherwise it tells be "WrongToken" [23:22:27] Yes, sure [23:37:32] is there a way to block spiders from a specific page? [23:38:59] that is, may I add rewritten urls in robots.txt? [23:40:43] https://www.mediawiki.org/wiki/Manual:Robots.txt [23:41:43] thanks