[06:10:54] hey all, i would like to implement the auto login feature on my mediawiki. I've found the ldap authentication. [06:10:59] is that the right extension to use? [06:11:28] my environment is apache 2.1.6, mediawiki 1.22.0, php 5.3.3 , is the environment suitable to use? thanks all [06:12:54] 1.22.0 is no longer supported, you should consider upgrading [06:13:34] thanks for the response. [06:14:17] assuming i upgraded, how should i go about implementing the ldap extension? i have read all the document and not fully understand it [06:15:01] Do I need to have AD directory? or the Kerberos ?? what are the difference between these 2? thanks [06:18:18] I don't know anything about the LDAP extension I'm afraid, so can't really help [06:19:01] alright thanks thought :) [07:04:25] i just installed mediawiki on debian 8, edited /etc/mediawiki/apache.conf to enabled alias, and restarted apache. when i went to localhost/mediawiki i get a 404 error. anyone know how to fix this? [07:07:26] jaypro: Look at error logs? [07:07:42] The first step is to get http://localhost working with the appropriate Web root. [07:07:55] So that http://localhost maps to some directory on your file system. [07:08:05] Configuring MediaWiki comes after that. [07:10:33] hey all, i would like to implement the auto login feature on my mediawiki. I've found the ldap authentication [07:10:38] my environment is apache 2.1.6, mediawiki 1.22.0, php 5.3.3 , is the environment suitable to use? [07:10:49] I have read all the document regarding ldap and not fully understand it. Do I need to have AD directory? or the Kerberos ?? what are the difference between these 2? thanks [07:29:06] Katie everything seems to be pointing to the correct place [07:29:56] jaypro: You should be able to access main entry points like index.php and api.php. [07:30:11] There are also text files in the MediaWiki root directory that should load in your browser, if everything is working correctly. [07:30:46] You can also create a "test.php" script with ' And that your paths are working correctly. [07:31:37] well... php works... i have php files in /var/www/html/ ...so php works [07:32:38] mediawiki is not in /var/www/html... so i figure the alias should work [07:34:23] /etc/mediawiki/apache.conf writes Alias /mediawiki /var/lib/mediawiki [07:34:33] so i checked /var/lib/mediawiki... and the files are there [07:35:41] the only difference i see between debian 7 and 8 was that now the webroot files are in /var/www/html/ instead of /var/www/ ... which i would imagine would not be an issue [07:38:02] Instead of an Apache alias, you could try a filesystem symlink maybe. [07:38:23] Can you pastebin the apache.conf file? [07:48:49] actually getting tired.. i'll try again tomorrow. thank you, Katie! [09:44:42] howdy- is there an extension, to anyone's knowledge, that would let a page be published at a certain time/date but be entered beforehand? [09:45:04] (presumably hidden or as a draft or somesuch) [09:45:43] alternatively, if there'd be a way to enter a draft and then publish it via a mysql query, that'd work for me too (atd<3) [09:47:40] r00t^2: https://www.mediawiki.org/wiki/Extension:Drafts [09:49:33] yeah, i knew you could generate drafts. was hoping there'd be a way i could programmaticlly publish a saved draft from the system shell [09:51:42] that extension provides an api, maybe you can use it [09:52:44] hmmm, no, the api is for saving a draft, but not to make the draft an actual article [10:48:11] hi, can I export all pages to xml without putting them into a category? [10:52:43] linux-amiga: yes, go to [[Special:Export]] and add the pages you want to export [10:53:07] Yes, but if I have a lot of pages? [10:55:29] i wish there was an "all" option [10:59:47] linux-amiga: if you have shell access, there's a maintenance script to dump all pages, or select by namespace [11:00:01] ah, good to know [11:14:25] linux-amiga you can export a whole wiki via the python script by https://github.com/wikiteam/wikiteam [11:16:36] i see there are solutions ... [11:17:17] however not for people who cannot grasp them [13:40:20] Maintenance scripts: What are the chunks good for? Will a script run faster if I raise 50 to 100? [13:42:05] what maintenance script are we talking about? [13:44:17] Subfader, it's so the script doesn't run huge write queries that take ages for the databases to handle and slow down real client requests iirc [13:45:39] Hmh, that would mean that a chunk of 50 runs 50 simultaneous queries [13:48:15] Vulpix: I wrote me an own ages ago. I copied some existing script. It grabs URLs from the externallinks table, runs curl with it and writes in new stuff afterwards. The script is very slow (cos of curl) so I wonder if I can speed it up [13:50:00] more likely 50 rows affected by a single query [16:56:23] Hi, I've got a bit of a problem with the Echo extension - I'm creating a new skin, and for some reason the preferences button in echo doesn't work [16:56:31] It instead points to undefined and has no name [16:56:43] I have no clue what to look at to fix it [17:33:42] soeb: i think Echo is looking for some existing link to Special:Preferences on the page and copies that [17:34:12] soeb: probably you need to ensure that link in your skin's menu has id="pt-preferences", or something like this [17:34:59] Ooh. Thanks, that's probably it then. We've made some changes to the personalbar and used different ids [17:36:16] That's one less issue then, cheers! Would have been great to be able to find that info elsewhere, but this works for me :) [17:39:56] I still have some other issues with it (for some reason the height of the overlay is the maximum of the page), but this one was the biggest [17:41:35] soeb: file bugs :) can't promise a timely fix, but at the very least the info will be out there [17:42:09] I'm convinced it's our issue, works on Vector, breaks on the new skin [17:42:37] But I know, I'm waiting on some Hovercards and Flow bugs for a while now, wanting to use fun new features but can't :( [17:44:31] MatmaRex: oh wait, seems this one isn't one of mine, just a change of behaviour in Echo - I can live with that, since it's not a bug [18:38:57] In the limit report, what does "Post‐expand include size" refer to? [18:43:49] harej: I'd say it's the size of transcluded pages/templates on the current page once all of them have been parsed recursively (that is, counting HTML and not wikitext) [18:44:11] so the post-expansion size of the page. [18:44:59] I'd say not the entire page, just what's transcluded on it (because of the "includes") [18:45:40] Though if the entire page consists of transclusions... [18:46:09] yes, if the entire page is a transclusion, it won't matter [18:46:21] *it won't matter the difference [18:54:10] I wonder why MediaWiki 1.25 installs setting $wgScriptExtension = ".php5"; by default :( [18:57:41] aargh, apache has DirectoryIndex index.php5 before index.php :S [18:57:48] Vulpix: where do you see that? [18:58:32] /etc/apache2/conf.d/php5.conf [18:58:50] on my local openSuSE installation [18:59:02] no, I mean mw setting wgScriptExtension to php5 by default [18:59:24] well, in the generated LocalSettings.php after installation [19:00:17] ah [19:00:19] I accessed /mw-config/ and since apache had DirectoryIndex php5 before plain php, it went to index.php5, and the script got that extension instead of the default php [19:45:26] hi, how do i change the "create account or login" text from skin header? [19:45:44] !qqx | Tuju [19:45:44] Tuju: If you want to find a particular system message being used on a page, you can append the uselang=qqx parameter to the URL. That will replace all system messages used on the page by their message names. See https://www.mediawiki.org/wiki/Qqx [19:46:11] find that message using the trick above and edit it [19:46:34] is that in skin source files or in database? [19:46:52] !message [19:46:52] There is no such key, you probably want to try: !allmessages, !anonnotice, !charinsert, !doesntwork, !gerriterror, !gsod, !interface, !memoserv, !qqx, !repeat, !screenshot, !sitenotice, [19:46:59] !interface [19:46:59] You can edit the text which appears in most parts of the interface by editing the right pages in the MediaWiki: namespace. See [19:47:11] ^ that [19:48:05] that should be in index.php?Special:AllMessages ? [19:49:55] or just [[MediaWiki:whateve-rmessage-name]] [19:51:05] actually i already changed couple of such messages but that didn't change that header. [19:51:19] could it be "Createaccount-title" [19:51:50] i've translated skind in finnish, but that createaccount-title doesn't match that appears in page header. [19:56:21] with uselang=qqx you should see the message name. If editing it doesn't change, check if you have $wgUseDatabaseMessages set to true (which is the default value) [19:56:32] yep, that helped a lot! [19:56:49] userloginnocreate is at least one [19:58:06] Nav-login-createaccount is what I'm looking for. [20:00:48] it works great! How can i clean up a bit that username + password page? [20:10:35] found it, in templates [21:22:16] how do i prevent firefox error "Blocked loading mixed content" while using https? [21:22:37] Load your scripts over HTTPS. :-) [21:22:41] Or use protocol-relative paths. [21:22:51] i tried latter, no success [21:22:58] this is stock mediawiki installation. [21:23:36] Which script is coming over HTTP? [21:23:46] Can you locate it in the HTML page source? [21:25:02] those are jquery etc [21:27:24] mmmm...maybe it got fixed by removing http from $wgServer [21:27:48] $wgServer is pretty important. [21:36:50] yep, but it breaks wgSecureLogin [21:36:55] crap. [21:37:39] and once I put that http: back into wgServer, SecureLogin works, but it breaks those mixed links in firefox. [21:38:40] duh, it's late here already. i'm gonna give up for today. [23:08:29] are edits associated with user_id or user_name? [23:13:22] onei, both [23:13:39] rev_user -> user_id [23:13:45] rev_user_text -> user_text, usually [23:14:37] is there any reason for using both? Seems simpler to associate by one (which could make renames easier too) [23:15:51] anonymous edits [23:16:48] I'm not particularly sure why we both with user_name for logged in users though, maybe it's just to keep queries simple [23:17:51] could they not be signed an arbitrary id? It seems rather odd to treat IPs differently in that context [23:17:59] or are there legal ramifications [23:21:19] essentially, I'm just wondering why user_name isn't treated as a display name [23:29:16] onei, there are other cases where user names appear anonymously [23:29:20] imports [23:29:27] certain system users iirc