[00:01:35] softball question: where can I edit the sections on the image page? For example, if I want to delete file history, where would I do this? Can I do this on an individual image page editing MediaWiki:Monobook.css? [00:05:02] alright now [00:05:06] alrighty now [00:05:15] softball question: where can I edit the sections on the image page? [00:06:26] 03(mod) api invalid title passed to langlinks causes internal error - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13442 (10aaron) [00:11:20] I moved a MediaWiki installation earlier and not every single image thumb copied over originally so the pages with said thumbs displayed errors within them. [00:11:31] I've since replace the thumbs with the proper JPEGs, but the errors are still on the affected pages. [00:11:52] The pages say they have no permission...but the JPEGs are definitely exactly where the error says they are to the character [00:12:00] and full world read perms as well [00:12:06] I tried restarting Apache to no avail [00:19:52] hi, does anyone know where i can get the AjaxLogin extension [00:20:37] did you try svn/mediawiki.org? [00:32:10] man this irc area is becoming worthless. no one ever responds anymore [00:32:51] Skizzerz: it's not there [00:33:00] hmm, do you know who wrote it? [00:33:26] Misza13 [00:35:29] ah, i think its still beta [00:35:38] I'll ask when I see him again [00:39:12] lol. 17:32 < odes2> man this irc area is becoming worthless. no one ever responds anymore [00:39:21] I have not found that to be my experience of #mediawiki, but.. [00:42:37] anybody familiar with mw's SVG implementation? pls contact me! [00:46:24] 03aaron * r32798 10/trunk/extensions/FlaggedRevs/FlaggedRevs.php: Add more heuristics [00:58:28] 03aaron * r32799 10/trunk/extensions/FlaggedRevs/FlaggedRevs.php: *ahem*, code works better when it can compile [01:15:11] 03(FIXED) Flag to mark accounts locked/hidden - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=6397 +comment (10JSchulz_4587) [01:26:24] chuck: still there? [01:26:45] Skizzerz: yep [01:27:00] k, I asked Misza about releasing the code for it [01:27:03] for you [01:27:13] he said "epic meh" [01:27:36] *Skizzerz will try again later [01:30:03] Skizzerz: :S [01:30:06] where did you ask him? [01:30:10] email [01:30:33] couldn't find him on IRC [01:32:36] 03(mod) renameuser exension doesn't change text on ipblocklist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13555 +comment (10JSchulz_4587) [01:38:01] 03aaron * r32800 10/trunk/extensions/Renameuser/SpecialRenameuser_body.php: Update ipblocks table [01:38:20] 03(FIXED) Blocking account creation on an IP address should apply to logged in users too - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13611 +comment (10JSchulz_4587) [01:39:07] 04(REOPENED) Blocking account creation on an IP address should apply to logged in users too - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13611 +comment (10JSchulz_4587) [01:39:53] 03(FIXED) renameuser exension doesn't change text on ipblocklist - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13555 +comment (10JSchulz_4587) [01:40:10] 03(mod) Add keyword __NOEDITTHISSECTION__ to disable edit link on a single section and it 's sub-sections instead of the whole page - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=8515 +comment (10waldir) [01:40:56] 14(WFM) Login error on temporary passwords - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13510 +comment (10JSchulz_4587) [01:52:00] 03(mod) Blocking account creation on an IP address should apply to logged in users too - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13611 +comment (10mike.lifeguard) [01:52:18] "ILIKEIT" :\ [02:16:44] has anyone gotten NiceCategoryList2/2.1 working in MW1.12? [02:23:01] 03aaron * r32801 10/trunk/extensions/FlaggedRevs/FlaggedRevs.php: Use estimates [02:56:14] 03(mod) Blocking account creation on an IP address should apply to logged in users too - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13611 (10brovvnout+wikibugzilla) [03:07:00] how do I add an image on bulletted lists? [03:07:18] I mean, custom "bullet" image... [03:08:28] nevermind... I found it... :) [03:56:20] !wg AutoPromote [03:56:20] --mwbot-- http://www.mediawiki.org/wiki/Manual:%24wgAutoPromote [04:01:59] !e ConfirmAccount [04:01:59] --mwbot-- http://www.mediawiki.org/wiki/Extension:ConfirmAccount [04:08:07] 03aaron * r32802 10/trunk/extensions/FlaggedRevs/FlaggedRevs.php: Thought this right was here already [04:23:14] i cant find a clear text explaining what to do to install a skin. abd btw, are they called skins ? or im searching wrong ? [04:24:02] broli: You get the .css file, copy it to skins/ [04:24:33] broli: I don't know if what you are searcing for is called skin or not. [04:24:48] Explain what you want to do, then I maybe know if it is skin. [04:25:23] i want the wiki to resemble the colors of the main website [04:26:04] Which main website? Anyway, you need a skin that is made for Mediawiki to change the look. [04:26:42] Or you can make your own skin, copy the Monobook -files to another name and modify the colors therein. [04:27:15] http://www.xero-clan.com <-- and i have -1 on knowledge about css, html, and java [04:28:19] What's the trick to get

working if I'm on shared hosting? (1&1)
[04:29:09] broli: If Monobook skin looks OK except for colors, you might get by with trial and error.
[04:30:18] tale  you mean that i can get some succes editing the ile and changing the f8f29f0 colors ?
[04:31:19] broli: Copy the Monobook* files to another name, change the colors there and use that modified skin.
[04:31:51] broli: That should get you a wiki with other colors.
[04:32:31] But do not modify the Monobook* files, so users that want that can get the original look.
[04:32:43] ok. then to find some guy at our website that i can force to work on it :P thanks for the help
[04:33:26] When all is OK, you can change the site to use as default skin the one you modified, so all users get your look except if they specifically set their account to use Monobook.
[04:36:46] thanks again
[04:41:17] 03aaron * r32803 10/trunk/extensions/FlaggedRevs/FlaggedRevs.php: Casing, whitespace
[04:42:05] I suppose if I compile it locally then upload it there it will probably not run... 
[05:33:18] Tried it... seems to almost work... 
[05:33:38] downloaded the math folder to my pc.. then compiled texvc...
[05:33:49] before compiling and setting it executable, I had
[05:33:54] Failed to parse (Missing texvc executable; please see math/README to configure.): \int_{-N}^{N} e^x\, dx
[05:33:57] now I have
[05:34:02] Failed to parse (unknown error): \int_{-N}^{N} e^x\, dx
[05:37:24] I just downloaded mediawiki, php, apache, and mysql. Which order do I need to install them in in order for it to work? (I'm doing this on my own computer for offline use)
[05:49:11] I think you can install MySQL, Apache, and PHP in any order. MediaWiki needs to be installed last.
[05:49:45] If they require something else, they should just tell you and quit the installation.
[05:53:14] Ok, thanks.
[06:05:04] solution : http://meta.wikimedia.org/wiki/Mimetex_alternative   thanks anyway
[06:28:10] Just edited MediaWiki:Sidebar and nothing appears new in the actual sidebar
[06:30:27] clear the cache
[06:30:34] did
[06:30:42] I'm not using the correct syntax I believe
[06:31:03] there a | thing with keywords that point I don't know where
[06:31:20] page | Title i think
[06:31:30] not even :S
[06:31:42] i dunno then
[06:31:50] oh
[06:32:05] i think they point to MediaWiki:whatever pages
[06:32:35] so whatever | Something links to the page which name is the only content in MediaWikiWhatever
[06:32:50] MediaWiki:Whatever'
[06:32:52] oh.. 
[06:33:46] i think, but im not sure
[06:33:56] there should be a way to point directly to a page to
[06:33:57] I think you're right
[06:34:07] that's about what they explain in http://www.wikia.com/wiki/Help:Sidebar
[06:34:51] yes.. what I was seeing was of the kind portal-url|portal .. so I didn't understand, but it can be forum:index|Forum too
[06:35:01] the first points to MediaWiki:, like you said
[06:35:46] http://www.mediawiki.org/wiki/Manual:Interface/Sidebar
[06:36:20] yup, thanks it's mediawiki's equivalent page
[06:36:44] more detailed though
[06:37:10] if MediaWiki:Something doesn't exist, Something is treated as a [[Something]] or [Something] (if valid URI)
[06:37:47] so Foo|Bar would look up MediaWiki:Foo, and if it doesn't exist, it would link to [[Foo]], if MediaWiki:Bar/xx (userlang) exists, or MediaWiki:Bar exists, that is used, else [[Bar]]
[06:37:57] er "Bar" (as link text)
[06:38:56] thanks
[06:39:10] what a great piece of software... I like all this :)
[06:39:46] reason being of course, the link text will be translated if you need it to be
[06:40:53] http://www.google.com|google-text     where MediaWiki:Google-text contains "Google search", and MediaWiki:Google-text/de contains "Google Suche"
[06:41:09]

) [11:18:49] k trying [11:18:55] By default, I have anonymous user registration turned off. If I, as a sysop, want to add users manually, how would I go about this? I don't see a likely candidate in the maintenance directory. [11:19:21] k4jcw: You can create accounts at Special:Userlogin [11:19:30] ah, cool. [11:21:03] MONK3000: I think I could maybe do Advanced Search some time next week [11:21:38] what kind of advanced search? [11:21:50] http://www.mediawiki.org/wiki/User:Subfader#Advanced_Search_.5B150.E2.82.AC.5D [11:21:56] whoo. be sure it works as requested with the special needs fro my site [11:22:43] Sure. I'm thinking I might wanna make it a bit more extendable, you know, make it possible to add more textboxes on the fly with JavaScript [11:22:56] And the namespace thing is definitely possible [11:23:12] http://p.defau.lt/?tF8eRQFGMi4u6kY_E9CtRQ floats to the right better (might break in certain IE) [11:24:14] Roan: ok most important is that you do 1 search action combining all search aspects. e.g. find article with this word in article title but not in this category but with complete tracklist etc. [11:24:25] Yeah [11:24:39] Splarka: thanks works, can do the rest of the "hacking" [11:25:01] That (combination) is probably the nastiest part [11:25:10] enjoy, keep money, buy porn or candy with it [11:25:56] Splarka sneeks to bed [11:35:40] RoanKattouw, thanks (special:userlogin) [11:35:50] No problem [11:36:27] !hide [11:36:27] --mwbot-- I don't know anything about "hide". You might try: !csshideheader !dismissablesitenotice !hidetoc [11:36:58] How to hide som text in a page? [11:37:20] BjornH: What do you mean by hide? Do you want it to be showable? [11:37:28] Or to stay hidden at all times? [11:37:48] Probably wants [11:37:52] Maybe [11:38:02] Yes, only show in edit page, as a memo. [11:38:11] [11:38:24] Aha, tank you! [11:39:24] Just remember, commenting things and job security are mutually exclusive. [11:41:24] I only want to show examle for edit people. [11:45:31] The FAQ's suggestion to try and remove (the pages, ability to edit them and the tab) discussion/talk pages did not work for me, has anyone a good suggestion how to? [11:46:00] hi Jack_Phoenix [11:46:14] heya ialex, how goes? :-) [11:46:20] Note that my wiki uses only a modified monobook theme, all other themes are removed so changes to the css files are fine to remove the tab [11:46:30] Jack_Phoenix: fine, and you ? [11:46:42] I'm fine too, thanks. :) [11:46:47] (however I tried this, it did not appear to work, even with a cleared cache and the css files reloaded by the browser client) [12:01:44] 03(NEW) Bad searches, irritating - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13616 minor; normal; Wikimedia: General/Unknown; (stuartlocke) [12:17:21] got a little js problem: i have var url='http://www.google.com/custom?q=' + encodeURIComponent(wgTitle); now i want to add sth after Title) but it doesn't work ? [12:18:59] var url='http://www.google.com/custom?q=' + encodeURIComponent(wgTitle) + '&foo=bar'; [12:20:10] ah thanks. the ' is irritating in the ?q=' + encodeU [12:20:35] Why's that? [12:21:30] cos of url='.. i thought i isn't closed at the end before; [12:21:45] anyway. this doesn't work :( var url='http://www.google.com/custom?q=' + encodeURIComponent(wgTitle) + '&client=pub-123&forid=1&channel=123'; [12:22:15] wait it does [12:22:18] great [12:34:31] woot [12:38:40] ^o) [12:46:04] Is there any sort of index extension available? If you compare the Category to the table of contents in a book, I want to build an index. Something like [[Index:word1 word2 word3 word4]]. and the Index page would contain links to the various pages under each word1, word2, etc. [13:10:59] hey all, I've downloaded the enwiki dump and I'd like to run it offline. I've installed mediawiki etc fine, and I'm trying to import it, but its incredibly slow - 3.5 pages a second. Is there any method that's faster than using importDump.php? [13:12:02] I've seen some mentions of MWDumper being able to generated SQL from the XML, but it seems not to be maintained... [13:12:41] eAi2000: Maybe download an SQL dump [13:12:46] Dunno if they're available [13:12:52] As far as I'm aware they aren't [13:13:29] at 3.5 pages a second it's going to take over a week to do [13:13:34] lol [13:14:11] Did you download a dump with full history or something like that? [13:14:26] nope [13:14:31] ^^ [13:14:43] Then 3.5 pages per second is kinda slow IMO [13:15:03] enwiki-latest-pages-articles.xml [13:15:08] 15.03GB [13:15:44] I agree it's slow, it could be just because it's PHP or it could be a disk issue [13:17:47] anyone know what mediawiki's current database schema version is? [13:23:23] eAi2000, MWDumper is actually working perfectly [13:23:30] ah great [13:23:38] was worried that it hadn't been updated for 2 years [13:23:50] any idea if it's faster than the php method? [13:23:53] it should be [13:23:54] MWDumper is officially mantained [13:24:10] it's faster [13:24:13] using the php import is extremely slow [13:24:13] definitely [13:24:20] because it renders each page [13:24:25] Platonides, NicDumZ: He found out already ;) [13:24:47] He actually calculated that at its current rate of 3.5 pages/s it'll take him a week [13:25:10] :) [13:25:21] anyway it will probably take a day or so [13:25:39] I'll have a go [13:25:45] thanks [13:25:53] didn't want to waste my time generating a broken database :) [13:26:09] eAi2000, http://www.mediawiki.org/wiki/MWDumper is* helpful. [13:26:16] yep, I'm reading it [13:27:20] oh, and you'll probably need to import the other side dumps, such as pagelinks categorylinks, templatelinks, and logging [13:27:37] ok [13:27:41] they can all be generated by maintenance scripts [13:28:16] do you know which scripts in particular MinuteElectron? [13:28:21] Yeah except logging but you don't really need that [13:28:39] isn't faster to pipe sql in, instead of generating them ? :) [13:28:53] NicDumZ: Maybe, but downloading also takes time [13:29:16] BTW, those tables would be built gradually as pages are viewed (and parsed) [13:29:51] But if you really need it all at once you can download the SQL or run rebuildall.php (which will probably take very long) [13:30:34] actually, I think that downloading the sidedumps during the main import is the fastest way ;) [13:30:50] but that's not this important ^. [13:30:51] That's probably true [13:34:07] is the sql actually available for wikipedia? [13:34:16] I thought it had been replaced with xml only [13:35:43] they provide the sql [13:36:02] they only replaced the revision and page dumps with xml so they could filter out deleted revisions [13:42:09] well, I tried mwdumper and it did one step - 1000 pages, then an exception occurred - Duplicate entry '6' for key 1 [13:42:37] MinuteElectron: and because sql dumps would have been useless [13:42:39] I defined sth for all main namespace pages. if(wgNamespaceNumber == 0) addOnloadHook(function() {... but how can i exclude it from the main page? [13:42:55] you couldn't even have per-wiki dumps of page text, you'd have to download all of it [13:42:56] MONK3000: if(wgNamespaceNumber == 0 && wgTitle != 'Main Page') [13:43:22] eAi2000: Start with a clean database [13:43:27] I have [13:43:33] That's very weird [13:43:41] I remove LocalSettings and set it up again with a new database name [13:43:46] Did you clean your database between the manual import and MWdumper? [13:44:01] I did that step just before [13:44:12] Main_Page was probably the only item in the db [13:47:33] deleted that page too, still get the error [13:48:47] flyingparchment: I don't understand, surely the e.g. enwiki_p database have been dumped? [13:49:04] MinuteElectron: enwiki_p is a toolserver view, the database is enwiki. and that database doesn not contain any text [13:49:04] flyingparchment: I know there is a reason within the architecture of Wikimedia, but I don't understand it. [13:49:07] eAi2000: Don't just clear the page table, clear ALL TABLES, preferably just DROP all of them [13:49:16] MinuteElectron: the text is batched onto external storage clusters, which are apaches with large, cheap SATA disks [13:49:28] I dropped the whole database and let mediawiki recreate it [13:49:37] because it's wasteful and expensive to put bulk data (text) on expensive SCSI disks [13:49:45] flyingparchment: oh, right; now I understand, this is what the "old ext.ts storage" messages are for? [13:49:48] Hmm then you shouldn't get any duplicate key errors [13:49:57] those are for the toolserver external storage replicas [13:50:01] (ts = toolserver) [13:50:15] yeah, but they are relicas of the large cheap SATA disks? [13:50:21] yes [13:50:27] right, thanks [13:50:30] actually, i'm slightly wrong here; you _could_ have per-wikis dumps [13:50:31] I udnerstand now. [13:50:35] what you can't have is current revisions only [13:50:52] (because the ES storage has nothing to indicate which text blobs are current revisions) [13:51:48] thanks roan [13:51:56] MONK3000: Did it work? [13:57:54] is image revert broken? [13:58:02] I don't know [13:58:16] I'd say ask Bryan, but he's not here [13:58:29] Why do you think so? [13:59:05] Bryan? [13:59:12] Bryan Tongh Minh [13:59:29] He does a lot of stuff at Commons and works on image handling code [13:59:34] Especially the API part of it [13:59:46] He implemented image deletion through the API yesterday [14:00:03] Oh wait no the day before [14:00:26] Anyway, if image revert is broken, he very likely knows about it [14:07:42] 03(NEW) Pseudo-namespace on Korean Wikipedia - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=13617 15enhancement; normal; Wikimedia: General/Unknown; (camway) [14:08:35] hi there [14:09:13] i need to know if i am right about the database structure [14:09:52] In small wiki installations there is only one DB on which all writes and reads are performed. [14:10:17] That's right [14:10:18] Bigger installations like wikipedia use multiple databases, one master db and some slave db servers [14:10:24] Yes [14:10:38] Reads are performed on the slave wherever possible [14:10:43] writes can or should only be performed on the master? [14:10:46] Writes always go to the master [14:10:56] The slaves regularly poll the master to ask about the latest writes [14:11:02] okay [14:11:11] Which is why they're always a few seconds behind (so-called 'lag' or 'replag') [14:11:24] ah okay [14:11:38] and the slaves then copy new content? [14:11:39] asga: Hold on a minute, I've got a good link for you [14:11:45] okay [14:11:48] thank you very much [14:12:07] http://en.wikipedia.org/w/api.php?action=query&meta=siteinfo&siprop=dbrepllag&sishowalldb [14:12:17] As you can see things are quiet on Wikipedia rigt now [14:12:28] All of the slaves are <1 second behind [14:12:38] And yes, the slaves copy the new content from the master [14:13:09] oh that link is cool [14:13:13] :) [14:13:26] now i can continue writing my thesis xD [14:13:32] asga: We once had a problem with a guy's watchlits [14:13:42] It contained 220,000+ entires which caused some crashes [14:13:53] So Brion deleted all those rows at once manually (on the master) [14:13:58] Slave lags were in the 10-minute range [14:14:11] The delete itself took 7 minutes [14:14:26] So if it's very busy, slave lag will be higher [14:14:35] okay [14:14:45] And if all slaves are >30 s behind, the wiki will be locked from editing automatically [14:14:50] To let the slaves catch up [14:14:55] oh oO [14:14:58] that is good to know [14:15:18] Because effectively, a 20s lag means any changes you make will only be visible after 20s [14:15:33] yes [14:16:20] And apart from the master/slave system we do load balancing too, obviously [14:16:55] if some extension or code uses wfGetDB( DB_SLAVE ) it is internally redirected to one of the slave servers? [14:17:08] Yep, the load balancer decides which one [14:17:12] lol [14:17:17] it must have been broken for months [14:17:22] LB? [14:17:25] i wanted to say that this must be the loadbalancer thing [14:17:25] Or image revert? [14:17:34] hi TimStarling [14:18:07] image revert [14:18:15] Ah [14:18:35] which sql dumps from http://download.wikimedia.org/enwiki/latest/ do I need to download to get a reasonable working copy? [14:20:03] okay ill write some text about mw db now... [14:20:07] thanks again RoanKattouw [14:20:27] asga: http://www.mediawiki.org/wiki/Manual:Database_layout [14:20:31] Not 100% up-to-date [14:20:42] But a very good picture [14:21:21] The most recent DB scheme is at http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/maintenance/tables.sql?view=markup [14:21:32] hehe [14:21:53] i already used both documents [14:22:00] Oh good [14:22:17] is it right that is intended to add user specific article restrictions? [14:22:28] No, not as far as I'm aware [14:23:03] If you want that, don't use MW ;) [14:23:14] no i don't want that [14:23:21] but in tables.sql it states: [14:23:22] -- Field for future support of per-user restriction. [14:23:22] pr_user int NULL, [14:23:35] page_restrictions table [14:23:50] Hmm I never knew it was there [14:23:54] Let's see who added it [14:24:08] TimStarling xD [14:24:17] xD [14:24:25] werdna [14:24:28] Oh wait no he didn't [14:25:08] Indeed, Werdna in r19095 [14:25:16] okay [14:25:21] asga: That was in Jan 2007 [14:25:33] oh... [14:25:38] RoanKattouw: hmm, maybe I should play with API img revert [14:25:44] i didn't look at the date it was added xD [14:25:48] Werdna created the page_restrictions table, not sure whether he asked anyone about pr_user [14:26:09] AaronSchulz: Well maybe fix regular image revert first? TimStarling says it's broken [14:26:17] maybe he will tell us something about it when hes here [14:26:25] When you've done that you'll probably know the code well enough to write an API module for ti [14:26:52] oh, I thought it had one already [14:27:00] Hmm the spurious thing is that he didn't add an index for pr_uesr [14:27:09] so regular img revert is broken? How so? [14:27:18] no, not months [14:27:19] AaronSchulz: Dunno, ask TimStarling, his words [14:27:23] just since r32181 [14:27:46] AaronSchulz: That's your rev [14:27:47] TimStarling: btw, can you comment on bug 7011 [14:27:52] !r 32181 | AaronSchulz [14:27:52] --mwbot-- AaronSchulz: http://svn.wikimedia.org/viewvc/mediawiki?view=rev&revision=32181 [14:27:54] yep Aaron's rev [14:28:39] AaronSchulz: obviously you didn't bother testing it [14:28:57] is there a limit on how much slave servers can be used? [14:29:09] Ok.... obviously [NL]Thoth doesn't like Guido van Rossum [14:29:13] asga: Not that I know of [14:29:37] But even enwiki gets by with a relatively low amount of slave servers [14:29:43] (less than I'd expected anyway) [14:30:09] ok [14:30:19] 13 DB servers in total (as you can see at the api.php link I gave you) [14:30:48] mind you, I shouldn't be laying blame yet, I haven't quite got it narrowed down [14:35:07] $file = wfFindFile( $this->title, $this->oldimage ); [14:35:12] yeah [14:35:18] function wfFindFile( $title, $time = false ) [14:35:22] oldimage is a filename, that implies it's a timestamp [14:35:35] but that's been there since the file was created [14:35:52] it's hard to believe Rob wouldn't have tested file reversion when he made a file called FileRevertForm.php [14:35:54] ah one more question xD [14:36:17] maybe it's been broken since July, or maybe there was some contributing change [14:36:36] there obviously can only be one master server... or am i wrong? [14:36:43] * @param mixed $time The 14-char timestamp the file should have [14:36:55] function findFile( $title, $time = false ) [14:37:18] well, an archive name is going in where a timestamp should [14:37:22] yeah, it's a time, it's right from there down, it's always been a time [14:37:56] asga: I don't know, but I think so [14:38:02] the error is passing a filename to wfFindFile, the other error is using it at all since last time I checked it, it was rubbish [14:38:16] OldLocalFile, that is [14:40:18] LocalRepo has newFromArchiveName(), it could use that in theory [14:40:39] yeah, filedeleteform uses that [14:41:20] you're not seriously suggesting that this thing has been broken for 9 months with nobody noticing are you? [14:42:12] actually, they both should just use a timestamp though, it is indexed better [14:45:21] is the actual text for enwiki articles available as an sql dump? [14:45:59] I can't see a dump of the text table... [14:46:44] the text table doesn't have any text in it [14:46:53] so it wouldn't help you much [14:47:21] where is the text then? [14:47:35] archive? [14:47:37] RoanKattouw: allo [14:47:39] in external storage [14:47:46] AzaTht: Hi [14:47:56] ah [14:47:58] useful [14:48:22] RoanKattouw: have hear some people wanting to be able to get non-redirect new pages via the api [14:48:39] so theres not much point downloading the sql then really if you want a local copy of the pages [14:48:56] no [14:49:00] that's what the XML is for [14:49:09] can't get that to work in any reasonable fashion :( [14:49:22] the php is too slow and the java gives an exception [14:49:41] RoanKattouw: could such a filter be implemented? [14:49:46] can't get the java to run in eclipse properly either [14:54:45] TimStarling, I readded articledelete hooks for file deletions, maybe that's inteferring with something? i haven't had problem on my local conf, but I don't really understand the oldimage magic [14:55:21] you haven't had a problem with file revert? [14:55:48] you mean file undelete? no, it works fine [14:55:57] no, I mean revert [14:56:04] http://test.wikipedia.org/wiki/Image:Test.jpg [14:56:12] log in, go there, click the link that says "revert" [14:56:16] RoanKattouw: uh? [14:57:02] AzaTht: Lemme check [14:58:42] AzaTht: Correct me if I'm wrong, but isn't that exactly what Special:Allpages does? [14:58:44] 03(FIXED) Bad searches, irritating - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13616 +comment (10Bryan.TongMinh) [14:59:28] TimStarling: how does wikipedia save page text in external storage? [14:59:59] see ExternalStoreDB.php [15:00:03] RoanKattouw: the request was to be able to filter out redirects from list=recentchanges [15:00:14] unless you have a more specific question? [15:00:15] i.e. only get "real" new pages [15:00:58] There's a problem with that [15:01:18] Any implementation would also drop page creations that weren't created as redirects, but turned into redirects later on [15:01:57] And the other way around, page creations that were created as redirects but became real pages later would _not_ be filtered out [15:02:34] This is because the recentchanges table has no rc_is_redirect field, so you'd have to check the page table's page_is_redirect field, which contains the /current/ redirect/nonredirect status of the apge [15:02:36] apge [15:02:38] page [15:03:43] hmm, I think r32806 is not needed [15:03:54] !r 32806 [15:03:54] --mwbot-- http://svn.wikimedia.org/viewvc/mediawiki?view=rev&revision=32806 [15:04:25] AzaTht: Would that be a problem? [15:04:57] because r32733 doesn't affect anything when php parses it [15:09:15] AaronSchulz: could you point me to in which revision and in what part of code you made log entries in changes list to group under the pages they affect? [15:10:17] look at recentchange.php [15:11:01] the log type used to be the title, now, the relevant page's title is the title, and the log type is stored in the newer rc columns [15:11:17] ugh [15:11:27] I was wondering why there is two kinds of log entries in the table [15:11:45] rebuildrecentchange probably adds wrong too? [15:16:02] no, it's fine [15:16:07] AaronSchulz ran it several times this week [15:16:19] ok, back to fixing reverts [15:16:19] hmm [15:16:29] oh well [15:16:56] what's going on with CIA? [15:17:24] ... is it malfunctioning AGAIN? [15:17:29] Hell it's not even here :O [15:17:37] I'll go whine about this in #cia [15:18:46] Oh hey wait ALL CIAs are gone, probably due to the weird splits that happened a few hours ago [15:19:05] No that can't be, CIA-44 left just now [15:19:23] But BearPerson is away, so it may take a while before they're restarted [15:19:24] CIA has been broken since yesterday [15:19:37] That'll have looked nice in #commits xD [15:19:42] CIA-1 has left Freenode [15:19:47] CIA-2 has left Freenode [15:19:49] Etc [15:20:15] food [15:23:34] RoanKattouw: it's what RC says that's relevant, not the eventaul future [15:23:53] AzaTht: The point is that the RC doesn't say* anything about redirect status [15:24:01] that's true [15:24:15] but does the rediect table gets updated on creatre? [15:24:41] AzaTht: Yes, but also on change (i.e. a page becomes a redirect, is no longer a redirect or changes targets) [15:24:51] The redirect table only reflects the current status of redirect, not past [15:25:06] RoanKattouw: true, but that's irrelevant, as RC doesn't reflect the past anyway [15:25:15] So if you create a redirect and then turn it into a stub, there'll be no flag anywhere saying it once was a redirect [15:25:48] or perhaps it's relevant [15:25:53] AzaTht: It might; pages coulld be created as redirects, then become stubs, but with rctype=new the latter wouldn't be shown [15:25:57] as the same page can be multple times in RCX [15:26:08] Exactly [15:27:08] though, we are only intrested in NEW pages [15:27:20] rctype=new will show all creations [15:27:33] New pages are pages that haven't been edited since their creation [15:27:35] There's a difference [15:27:43] hehe [15:28:19] perhaps someone could add a field to flag if the new page was created as an redirect or not [15:28:33] perhaps brion can do it ツ [15:28:36] Unlikely, hardly anyone is interested in that information [15:28:50] there is [15:29:07] file revert worked in 1.12, I just tested it [15:29:18] Who? And for what purposes? [15:30:01] RoanKattouw: patrollers [15:30:09] TimStarling! [15:30:17] hi brion [15:30:35] RoanKattouw: many of them are getting tired of patrolling redirects [15:30:41] sorry about not being around, they wouldn't let me use the internet while I was on the plane [15:30:45] heh [15:31:03] what's the status on configuration for the new load balancer? i'd like to get code updates going again RSN [15:31:13] TimStarling: you didn't sat in first class? [15:31:45] Well even redirects should be patrolled IMO, you never know if someone redirects [[George w. Bush]] to [[Baboon]] or something like that [15:32:26] well, the external storage part is broken, I knew it was untested, so I tested it and fixed it on the plane [15:32:36] great :D [15:32:42] but my current working copy also touches file reversion [15:32:52] so I tried to test that too, and discovered it's already broken [15:32:56] in HEAD and live on the site [15:33:02] d'oh [15:33:32] the current puzzle is how it was managing to work in 1.12, because the code was plainly broken [15:33:39] heh [15:33:49] TimStarling: I did change time <= to == [15:33:53] this is the bug: https://bugzilla.wikimedia.org/show_bug.cgi?id=13532 [15:34:03] so it may have kept matching shit [15:34:11] AaronSchulz: yes, that would do it [15:34:11] (previously) [15:34:21] it was meant to be <= [15:34:39] well... its me again ;) [15:34:42] the bug is reported as being postgresql-specific, but actually it's just completely broken [15:34:56] asga: Fire [15:35:00] is it possible to use different article languages in a basic media wiki installation? [15:35:14] the function should get a file with said timestamp [15:35:18] asga: The best way is probably to have a separate wiki for each language [15:35:25] that's not what was documented [15:35:33] Like Wikipedia does [15:35:54] where? [15:35:54] it was <= to support stable versions [15:35:55] Alternatively, you could do what www.mediawiki.org does and use [[Foo]] [[Foo/nl]] [[Foo/de]] etc. [15:36:02] RoanKattouw: it was MrZ's request [15:36:12] And who's MrZ? [15:36:16] TimStarling: oh, yeah, that was when I was crazy and wanted it more like "wayback machine" [15:36:28] then I realized that was totally impractical ;) [15:36:33] okay so a basic installation only supports button labels and menu items in a internationalized version? [15:36:47] RoanKattouw: MrZ-man [15:36:48] asga: Yeah, articles would have to be translated manually of course [15:36:54] yes :) [15:36:58] it's documented in FileRepo.php [15:37:04] AzaTht: Who is that? xD [15:37:10] Still doesn't ring a bell [15:37:11] RoanKattouw: en-admins [15:37:13] I thought I updated the docs, maybe I missed some [15:37:18] * Find an instance of the named file that existed at the specified time [15:37:19] AzaTht: Ah. [15:37:44] just wondered if there is a technique distinguishing such different language articles automatically (not just by adding a language prefix... or if, then automatically add it) [15:37:46] not "find an instance of the named file with the specified timestamp", that's what you changed it to [15:37:48] or something like that [15:37:58] but that answered my question :) [15:37:59] thanks [15:38:01] asga: Not really [15:38:13] (its just that i don't want to write something that isn't true) [15:38:21] No problem [15:38:56] The interface has been translated into a lot of languages, and every user can select their own interface language, but this doesn't go for articles [15:39:01] RoanKattouw kicks CIA-44 [15:39:06] RoanKattouw slaps CIA-44 [15:39:09] RoanKattouw kicks CIA-44 [15:39:16] Why the hell doesn't that work? [15:39:17] RoanKattouw: apr 01 23:28:33 I considered the API, but it doesn't distinguish redirects [15:39:43] Hmm [15:39:53] ow [15:39:53] ow [15:40:00] O.o [15:40:02] Well I could* do it, but only with current redirect status, past redirect status is nearly impossible to get [15:40:03] lol? [15:40:15] brion: when the new LBFactory stuff is brought in, the wikimedia configuration needs to be updated simultaneously [15:40:16] Hmm it took some time for it to feel that [15:40:23] because I've removed $wgAlternateMaster [15:40:30] TimStarling: right. [15:40:38] RoanKattouw: better than nothing at least [15:40:40] I have a new db.php here that can basically be swapped in [15:40:43] i poked a bit at redoing the config file (offline) but i'm not really sure what i'm doing ;) [15:40:46] if you've got one ready, then great [15:40:52] yeah, best to use my one :) [15:40:56] yeah :D [15:41:10] RoanKattouw: and the probabillity of an redirect going to non-redirect without anyone checking the redirect first is pretty low [15:41:22] I've been doing all my testing with the wikimedia configuration file, and that hack to make all database servers ultimately go to localhost [15:41:44] brion: I wonder if this is the longest something has ever been broken [15:41:50] brion: don't say " but i'm not really sure what i'm doing" loud ツ [15:42:04] :D [15:42:31] AaronSchulz: did you understand what was borked with the revert? [15:42:46] I couldn't find where those came from there [15:42:52] so AaronSchulz, can you explain why you needed to change the interface? [15:43:15] 03aaron * r32813 10/trunk/phase3/includes/filerepo/FileRepo.php: Update doc [15:43:28] AaronSchulz: did you get my email, sir? [15:43:41] will check later [15:44:19] TimStarling: hmm? [15:44:25] the diff http://en.wikipedia.org/w/index.php?title=Carnivorous_plant&diff=203521746&oldid=202790305 doesn't mach reality [15:44:47] AaronSchulz: the FileRepo::newFromTitle() interface [15:44:49] AzaTht: the mystery repeated lines? [15:44:58] yeah, known issue which we haven't been able to track down yet [15:45:01] AaronSchulz: just in case, I mean the email I sent you about a week ago.. about Oversight, etc [15:45:02] brion: hmm, perhaps [15:45:06] or findFile or whatever it was [15:45:11] AzaTht: That's true [15:45:15] TimStarling: all it needs is to match the timestamp. In fact, I don't want older ones than requested [15:45:23] which can happen if, say a revision was deleted [15:45:24] Please file a bug at BugZilla, I'm not at home now so I can't commit anything [15:45:29] 03(mod) Bad searches, irritating - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13616 (10N/A) [15:45:40] all what needs? what is the application? [15:45:50] hmm [15:45:54] brion: ok, so there is an bug allready then [15:45:58] either it matches the ones at review time or it is blue linked (unless current version behavior is set on) [15:46:03] FlaggedRevs [15:46:14] http://www.mediawiki.org/wiki/Summer_of_Code_2008 is looking like an early draft instead of a list of projects to choose from [15:46:15] has anyone gotten NiceCategoryList2/2.1 working in MW1.12? [15:46:39] TimStarling: also, by having it match, it makes an easy way to get old versions of images using a nice index [15:47:18] Hi! What's the trick so that I can add HTML in a page... like a form? [15:47:30] 03(NEW) filter recent changes by redirect/nonredirect - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13618 15enhancement; normal; MediaWiki: API; (azatoth) [15:47:33] did you review the other uses of that interface besides FlaggedRevs? [15:47:39] Max_-: The htmlets extension [15:47:44] !e htmlets | Max_- [15:47:44] --mwbot-- Max_-: http://www.mediawiki.org/wiki/Extension:htmlets [15:48:07] Oops http://www.mediawiki.org/wiki/Extension:HTMLets [15:48:44] RoanKattouw: Thanks, looking at this [15:49:16] !htmlets is an extension that provides a way to include (inline) static HTML snippets into wiki pages. (http://www.mediawiki.org/wiki/Extension:HTMLets) [15:49:16] --mwbot-- Keyword "htmlets" already exists. Use "!htmlets del" to remove it. [15:49:24] !htmlets [15:49:24] --mwbot-- For allowing any and all HTML, see . This is of course VERY DANGEROUS. Safer options include , and . [15:49:34] !:-P [15:49:34] --mwbot-- I don't know anything about ":-p". [15:49:48] TimStarling: I couldn't find any other uses [15:50:10] 03(ASSIGNED) filter recent changes by redirect/nonredirect - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13618 +comment (10roan.kattouw) [15:50:28] ohh rawhtml will be simpler for me [15:50:53] 03(mod) Bad searches, irritating - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13616 +comment (10roan.kattouw) [15:51:11] Hmm it's actually quite busy here considering it's Saturday, even Tim and Brion are here [15:52:00] RoanKattouw: think something if fnucked upon all borkage [15:52:14] Probably xD [15:52:50] AaronSchulz: looks like you didn't update the documentation on GlobalFunctions.php either [15:54:01] do you want me to make these changes for you? [15:54:17] 03aaron * r32814 10/trunk/phase3/includes/GlobalFunctions.php: And here... [15:54:54] I guess not [15:59:42] return wfTimestamp( filemtime( $path ) ); [15:59:49] TimStarling: does that make any sense? [15:59:57] Err I don't think so [16:00:00] function wfTimestamp($outputtype=TS_UNIX,$ts=0) [16:00:06] wfTimestamp($format, $ts) right? [16:00:12] that's what I thought [16:01:29] AaronSchulz: just leave FileRevertForm.php, I'll fix that one [16:01:38] I'm testing my fix now [16:01:54] can you make it use the img,timestamp index ? [16:02:36] why? [16:03:13] performance? [16:03:27] what's the difference? [16:03:45] I don't see an archivename index, just a file name one [16:03:58] so that could be linear for the number of file revisions [16:04:09] KEY `oi_name_archive_name` (`oi_name`,`oi_archive_name`(14)), [16:04:21] looks like an oi_archive_name index to me [16:04:25] I must be missing an index [16:04:40] odd, I'll add that one manually [16:04:51] it's on enwiki [16:06:10] well well well, I'm fighting to find an interesting but big enough task for GSoC... :) [16:09:38] 03aaron * r32815 10/trunk/phase3/includes/filerepo/File.php: Use TS_MW [16:10:29] brion: ok, I am still thinking about block logs and renames [16:11:18] AaronSchulz: the sensible thing at this point may be just to do the renames :P [16:11:21] for the sake of utility, I am leaning towards doing it [16:11:26] it'd be ideal if the log can show both the original and new name, if possible [16:11:41] duplicating? hrm [16:13:28] brion, would I sound annoying poking you again about svn accounts ? :þ [16:14:18] anyway, the reason OldLocalFile is crap is because it's heavily performance-optimised for the history display case [16:14:24] and it's not even being used for that [16:14:41] when you do a file revert, it doesn't use the index [16:14:50] it loads 20 rows from the image history and saves them in a big blob [16:19:20] how about I just remove that caching code? [16:19:32] AaronSchulz was about to suggest that [16:19:39] brion: http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/SpecialUndelete.php?r1=32675&r2=32726&sortby=date [16:19:43] ahhh :) [16:22:43] brion: anyway, copying log stuff gives me the willies, I think I'll just do the renaming [16:23:36] AaronSchulz: any reason to pass around string titles there? would feel easier to pass a title obj :) [16:23:48] but yes, it's shorter than it used to be ;) [16:23:57] NicDumZ: did you send me a username & ssh key? [16:24:24] brion, valhallasw did, on the 8th [16:24:45] brion: if you want to make it titles, then knock yourself out :) [16:24:53] NicDumZ: then i'll get to it when i reach it :) [16:25:16] ah, nice. [16:25:49] AaronSchulz, are you planning on adding the "invert selection" I wrote on the previous version of S:Log ? :) [16:26:13] not really, the rewrite was just me getting sidetracked [16:27:20] perhaps I could give a look [16:31:46] brion: you there? [16:32:07] busy [16:32:09] what's up [16:32:22] ah [16:32:38] was just looking for a green light in terms of writing some image redirect handler [16:32:48] (for core) [16:33:58] like if you had #REDIRECT [[Image:Y.png]] on the Image:X.png page, calling [[Image:X.png]] will display [[Image:Y.png]] on the page (provided that X.png does not exist) [16:34:56] brion: what about the userrights log? [16:35:02] brion: so good idea or no? [16:35:13] AaronSchulz: what about it? [16:35:21] rename it to? [16:35:23] too [16:35:28] brion: have you ever noticed that your coding style tends towards inconsistent error reporting? [16:35:40] Skizzerz: er, you mean exactly like it does right now? [16:35:50] ... [16:35:57] I'm pretty sure it doesn't do that atm [16:35:59] you often give a nice detailed error message for an unreachable default case in a switch [16:36:01] TimStarling: in what way? [16:36:14] and then forget to implement any kind of error reporting for common user errors [16:36:15] Skizzerz tests [16:36:27] shrug* [16:36:38] or just use a boolean return [16:36:45] and tell the user that something bad happened :) [16:37:27] so how about that rights log? [16:39:59] brion: should I just be doing the block log? [16:40:08] AaronSchulz: you mean, update the targets when users are renamed? [16:40:16] i suppose that would be sensible; be consistent [16:41:17] ok [16:41:41] just those 2 then [16:44:49] 03(FIXED) Update block log when renaming an account - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=7011 +comment (10JSchulz_4587) [16:49:00] 03(FIXED) Reverting Images broken - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13532 summary; +comment (10tstarling) [17:14:13] TimStarling: r13532 ? [17:16:00] aha, you mixed up rev and bug number at https://bugzilla.wikimedia.org/show_bug.cgi?id=13532 [17:16:30] assume you meant r32817 [17:16:41] no, it was decided to name r19285 as the 0th revision [17:16:57] ;) [17:16:59] hehe [17:17:19] "Indonesian (id) translation for ImageMap extension" [17:17:25] !r 19285 [17:17:25] --mwbot-- http://svn.wikimedia.org/viewvc/mediawiki?view=rev&revision=19285 [17:17:43] an really important revision to use as base rev ツ [17:19:01] is there any way to get the source of an article directly, or do i have to get the ?action=edit page and parse it? [17:19:26] Qhestion: use the api [17:19:42] ur use action=raw [17:20:33] i probably should not try that, but that action is blocked on the english wikipedia [17:20:47] not really [17:20:49] which means, it doesnt help me. forgot to mention that, but this is exactly what i want to do. [17:21:08] just it's blocked from /wiki/... [17:21:23] ? [17:22:27] i really dont understand what you mean... [17:24:12] i want to use encodeURIComponent(wgTitle) for categories but giving out the name without "Category:" in front. any idea? [17:27:28] MONK3000: wgTitle will not have Category: and wgPageName will [17:42:28] hmmm when did RoanKattouw leave? [18:34:02] i use encodeURIComponent(wgTitle) to call a page title. now i want to use it for category names, but need to call it without "Category:" in front. any idea how to? [18:35:58] 03(FIXED) Unquoted value in JSON response - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=11633 +comment (10Bryan.TongMinh) [18:36:07] what are you doing? [18:36:09] MONK3000: you want something like substr [18:36:29] mediawiki has good escaping functions :o [18:36:39] oh javascript? [18:39:27] MONK3000: by default wgTitle doesn't have the namespace [18:41:02] hi ialex [18:41:15] hey Nikerabbit [18:42:37] ialex hates when CIA-44 doesn't report the commits :( [18:43:01] ialex: I think r32806 is not needed [18:44:29] Nikerabbit: otherwise linkprefix would be like before next time MessagesEn.php would be rebuilt [18:45:19] ialex: '\\blaa' === '\blaa' in php [18:45:47] and both php and the regex library understand \x00 syntax, so it isn't actually a problem [18:46:02] Nikerabbit: say this to Siebrand :) [18:46:14] ialex: I've said it, I think [18:46:41] ialex: I was also confused myself before I realised that the regex understand \x00 too [18:47:09] Nikerabbit: ok then [18:48:30] it is very weird that php parsers \\ to \ in '-quoted strings [18:52:39] anyone working with semantic media wiki? [18:54:23] 03(FIXED) Please, may we use square brackets in JSON callbacks? - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=12136 +comment (10Bryan.TongMinh) [18:57:51] sorry was making dinner. yes javascript, and the namespace is defined somewhere else [19:01:15] ok i was mistaken it works without the prefix [19:01:47] hmm [19:01:55] isn't count($errors) > 0 redundant [19:16:34] i have probs creating an array for if(wgNamespaceNumber == 0 && wgTitle != 'Main Page') it shoudl also include NS 14 [19:17:30] if(wgNamespaceNumber =array(0,14) && wgTitle != 'Main Page') will make it work on all namespaces [19:18:17] that code looks very wrong [19:18:26] if(wgNamespaceNumber =array(0,14) && wgTitle != 'Main Page') [19:18:29] oops [19:18:35] oh, is it javascript? [19:18:53] if( (wgNamespaceNumber == 0 || wgNamespaceNumber == 14) && wgTitle != 'Main Page') [19:19:10] flyingparchment: yes [19:19:20] people should mention little details like that :) [19:19:31] hi flyingparchment and ialex :-) [19:19:46] hello Jack_Phoenix ;) [19:20:04] Nikerabbit: not if the array has a negative number of elements! [19:20:09] rar [19:20:23] flyingparchment names this concept 'imaginary arrays' [19:20:30] great! thought i can use array= whenever i need "more" [19:20:53] flyingparchment: is it related to imaginary wifes? :o [19:23:25] are imaginary wifes imaginary girlfriends that you proposed to and they accepted? [19:23:53] sure? [19:26:56] yep. [19:27:03] ups [19:28:12] maybe someone can help me out with this: how can i make the link open in a new window? the link is created with this js: http://rafb.net/p/fObOtk10.html [19:29:07] would love to include it this way as it 100% works "javascript:void(0)" onclick="window.open('http://...') but tried everything in var url= [19:32:00] MONK3000: try #mediawiki-scripts [19:32:15] google://"javascript open in new window" [19:33:01] 03(FIXED) Unknown_action error should return an error object - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=11673 +comment (10Bryan.TongMinh) [19:34:21] productive evening [19:35:16] Mike: thanks for the hint [20:02:10] 03nikerabbit * r32826 10/trunk/phase3/ (RELEASE-NOTES includes/SpecialUserlogin.php): * (bug 13463) Login successful page doesn't use user's preferred interface language [20:02:23] \o/ CIA-48 [20:02:36] YAY [20:03:15] yaaay, no need to call the FBI :) [20:05:16] yay [20:05:23] 03(FIXED) zh-tw user still sees zh-cn upon login first page - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=13463 +comment (10niklas.laxstrom) [20:05:24] ialex slaps Jack_Phoenix [20:05:27] :) [20:05:32] hehe [20:06:08] 03(NEW) Add non-breaking spaces in appropriate places automatically - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13619 15enhancement; normal; MediaWiki: Page rendering; (l8qjvx002) [20:07:17] 03(mod) Syntax extensions: special character for non-breaking space ( ) - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3461 +comment (10l8qjvx002) [20:24:18] i use encodeURIComponent(wgTitle) to create a link to google the article name. now my articles all have a date in front, e.g. "2008-03-21 - Text". is there a way to only include the text but not the date? maybe exclude numbers and special characters from wgTitle? [20:24:58] cos google result are way better without the date [20:25:48] Ah... Godspeed goodpeople. [20:28:50] MONK3000, what are you talking about? [20:31:18] lch: http://rafb.net/p/uHa8do30.html this produces an image which links to a google search, searching the exact article title. but i want to exclude a part or numbers from the search string (wgTitle) [20:35:59] MONK3000, what's in wgTitle? [20:36:29] it calls the page title [20:37:30] have a look at function getLemma () in http://de.wikipedia.org/wiki/Benutzer:Dapete/Temp [20:37:34] maybe that does what you want [20:37:51] thanks. checkin [20:40:10] Hello, I recently came across something which looks like a wikimedia extension, but I can't figure what it is. I'd like to implement it on my website, could anyone help me identify it? [20:40:27] lch: doesn't seem so. it is checking / calidating something. [20:40:34] kaos_nk; wikimedia is the foundation, the software is MediaWiki [20:40:45] ah, sorry, mediawiki then [20:41:07] link would help to id [20:41:20] kaos_nk: what does the extension do ? [20:41:52] well, the actual wiki is: http://getfansub.com/wiki/Default but the author implemented something on http://getfansub.com/ which I'd like to reproduce [20:42:20] updates of pages, mostly [20:42:45] 03(NEW) #iferror should catch imagemap errors - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=13620 15enhancement; normal; MediaWiki extensions: ImageMap; (rene.kijewski) [20:43:33] I've looked into the source code on the actual index page (the 'non-wiki' one), but it looks like the same code as on every other wiki page, expect it's it's editable or viewable, the wiki source [20:43:46] it's not editable [20:44:37] kaos_nk: http://getfansub.com/ is not a page generated by MediaWiki, it only takes the look and feel from MediaWiki [20:44:52] yeah, that's what I figured out [20:45:08] I was just wondering if it was some sort of extension or tool, since it doesn't look like it was hand-written [20:46:06] kaos_nk: you can just copy a html source from a page generated by MediaWIki and replace some parts of it [20:46:15] yeah, I suppose [20:48:08] well, thanks [20:50:14] hmm, can I limit the number of versions stored in the history, to save space? [20:53:44] I have a few wiki pages with an ampersand on their titles, and they aren't showing up [20:54:04] "Screen Saver & Graphics Display - Einstein@Home" takes you to the "Screen Saver" page instead [20:54:18] only way to reach it is using index.php?action=view&title= [20:54:40] I tried to rename the page, but Special:Movepage isn't working correctly with the ampersand either (tries to move Screen Saver page instead) [20:55:06] Where is the image file stored that has the logo on the upper left corner of the wiki? [20:55:19] !logo | lymeca [20:55:19] --mwbot-- lymeca: The logo that appears in the top left of each page is determined by the $wgLogo configuration setting in the LocalSettings.php file. To change this you simply need to change the value of $wgLogo to point to the URL of your own logo image. See for more information. [20:55:31] PovAddict: Rewrite problem, most likely [20:55:40] Are you using pretty urls? [20:55:41] lymeca: the default logo has the instructions written on it :) [20:55:50] Dashiva: no idea, I'm not the admin [20:56:12] (and the admin is asleep by now...) [20:56:20] Link to the wiki then? [20:56:56] well, I managed to get to the correct movepage using /index.php?title=Special:Movepage/the_%26_page_name&action=view [20:57:08] http://www.boinc-wiki.info/index.php?title=Screen_Saver_%26_Graphics_Display_-_Einstein%40Home&action=view [20:57:30] that is the working link [20:57:49] go to [edit] and back to [article], and it will stop working [20:59:01] Thank you Skizzerz [20:59:34] PovAddict: Yeah, it's broken redirection [21:00:54] The URL is probably rewritten to index.php?title=Screen_Saver_&_Graphics_Display... which means it looks for a page called "Screen Saver " [21:01:15] yeah, then mediawiki trims the space and I end up at "Screen Saver" [21:01:19] which is an existing page [21:01:35] So all you can do is tell the admin to fix the redirection [21:01:51] ampersands in titles seem to be calling for trouble anyway... [21:02:43] MinuteElectron: how do you get your Mibbit extension to show channel joins and parts (and quits)? [21:03:02] i know this bug by creating pages via & in the url then the rest of the article name cutted [21:03:11] well it's no bug in my case :) [21:04:20] PovAddict: It works fine as long as you do it properly (pathinfo or escaping) [21:04:55] so, is there a way to limit the number of revisions? [21:08:38] kaos: why would you want to do that? [21:09:50] since I don't need more anything more than the last 10 revisions, and I want to save up on space [21:09:54] -more [21:10:28] so you don't want to limit revisions (that sounds like blocking edit when you reach a limit!), you want to delete old ones [21:10:49] yup [21:10:57] sorry, maybe I misformulated my query [21:11:36] Skizzerz_Away: It shoudl by default, I only wrote the interface to the chat client; someone else developed the actual chat bit. See http://www.mibbit.com/ [21:13:33] irc.wikimedia.org is down [21:13:37] is it not [21:14:37] well, just found $wgCompressRevisions. I suppose that is a start [21:17:20] MinuteElectron: hmm... lemme try something then [21:18:50] White_Cat: Wrong channel. [21:19:14] MinuteElectron and where do you expect me to post it? [21:19:22] #wikipedia-tech ? [21:19:27] I am quieted there [21:19:38] Yes, I know you are; nevertheless this is the wrong channel. [21:20:00] I was told to make these posts here [21:20:06] by whom? [21:20:16] White_Cat: mediawiki != wikimedia [21:20:16] by the person banning me from #wikipedia-tech [21:20:25] you are welcome to talk to him about it [21:20:52] Whats his face, my mind went blank [21:21:00] mark? [21:21:03] domas [21:21:08] Domas yea [21:21:09] ...I think [21:21:10] MinuteElectron: do joins/parts show on the betawiki one? [21:21:10] I think so [21:21:16] Skizzerz_Away: yeah [21:21:44] So unless Domas decides to unban me from #wikimedia-tech I will be forced to post requests here [21:21:53] mind pasting the $wgMibbitExtraParameters varaiable? [21:21:59] as he old me to use this channel [21:22:03] told [21:22:12] so please take it with him, not me [21:22:20] Skizzerz_Away: AFAIK they don't use any extra ones, just the defaults. [21:22:22] the irc channel still needs attention [21:22:27] hmm... [21:22:33] err I mean server [21:22:56] White_Cat: And I'm sure other people will also notice the problem and post there. [21:23:06] as has happened several times today already :) [21:23:06] MinuteElectron right [21:23:07] except irc.wikimedia.org works just fine [21:23:31] I get Connection timed out's [21:23:45] you manage to waste our time even here and now [21:24:04] why are you all so hostile towards me [21:24:10] whay have I freaking done to you? [21:24:48] I suppose I'd need to go into the sql database and run some query which deletes old revisions [21:25:04] hello [21:25:21] kaos_nk: there is a script for this, but it only keep current revisions [21:25:34] ialex: link? [21:26:06] kaos_nk: maintenance/deleteOldRevisions.php [21:26:07] if the only database identifier revisions have is their id, then there's no workaround for it [21:26:19] thanks [21:28:54] right now my wiki's URLs are http://server/index.php/Main_Page and I want to get rid of the index.php but none of the solutions are working for me. [21:29:07] I've tried putting many different things in my .htaccess file as per mediawiki.org [21:30:03] and in my LocalSettings.php "$wgArticlePath = "/$1";" and $wgUsePathInfo = false; [21:33:24] lymeca: don't put it into domain root [21:34:51] Nikerabbit: I've heard conflicting things about my choice but this is what I want to do [21:35:08] getting rid of index.php should be the same either way...? [21:37:13] lymeca: no [21:37:26] in domain root it's bad and bad and thins like that [21:37:33] in subfolder it is pretty easy and ok [21:38:17] "it's bad and bad and thins like that" .......excuse me? [21:38:32] things [21:38:44] That's not quite what I didn't understand. [21:40:52] >> Warning: this method may create an unstable URL structure and leave some page names unusable on your wiki. See [[Manual:Wiki in site root directory]]. [21:42:24] to put it short: it is not recommended [21:44:54] Nikerabbit: I know it is not recommended, I've read. I still need help getting rid of index.php from the URLs regardless of whether or not I have a wiki/ directory [21:44:56] it's not working [21:45:28] !shorturl | lymeca [21:45:28] --mwbot-- lymeca: To create simple URLs (such as the /wiki/PAGENAME style URLs on Wikimedia sites), follow the instructions at . There are instructions for most different webserver setups. If you have problems getting the rewrite rules to work, see !rewriteproblem [21:47:27] lch: I've read those pages extensively and tried most if not all of the solutions for my category [21:47:32] I can't get it working [21:47:53] "I've tried putting many different things in my .htaccess file" doesn't sound like you know what you are doing [21:48:32] it's more or less trivial, you set up a rewriterule and you tell mediawiki through the LocalSettings.php that there is one [21:49:41] otherwise you need to be more verbose and give more info what problems you are encountering [21:50:17] turning on logging for the rewriteengine and watching what is reported would be a good start [21:53:15] lch: I mean I've tried following multiple different pages on mediawiki.org on how to get rid of index.php. THey all say to put different things in my /.htaccess file [21:53:46] Telling MediaWiki via LocalSettings.php is more or less trivial [21:54:03] sure, since the same thing can be achieved in lots of different ways [21:54:12] but every time I do, not matter what my rewriteengine rules are, Apache says it can't find the page [21:54:22] then you need to read the logs [21:54:25] index.php is removed from the URL for sure [21:54:28] Apache's logs? [21:54:28] and find out what the problem is [21:54:31] sure [21:55:06] the most obvious things why it would be failing: [21:55:39] 1) the rewriterules are wrong or not working, solution: turn on logging for the rewriteengine and read the logs [21:56:41] lch: I get a bunch of these: "File does not exist: /usr/share/mediawiki/Main_Page" [21:56:58] where, in the error log? [21:57:09] yes [21:57:24] 2) your apache doesn't care for the stuff in your .htaccess files because it's setup not to, solution: put some other stuff in there to check if that gets activated, and maybe move the .htaccess rules into the httpd.conf itself [21:57:28] there are also manymany lines File does not exist: /usr/share/mediawiki/favicon.ico [21:58:01] yeah, that's because clients always request a favicon [22:01:25] if apache isn't configured to use the .htaccess files in that directory through AllowOverride, it won't do that [22:06:19] lymeca, try adding the following to your .htaccess or httpd.conf file: [22:06:22] RewriteLog /var/log/httpd/my_rewrite_log [22:06:22] RewriteLogLevel 3 [22:06:42] and then look if that log gets written and what's inside there [22:07:05] use some other location for the log if your system is different [22:14:44] I would like to set up a system on our wiki in which various people maintain a small article that contains information about the status of their particular projects. Then this will be summarized in an overall status article. I'm trying to sort out if i should develop a template for this or create a plugin. It has to be very easy for the users. Using the statusitem1=blah|date1=blah|statusitem2=foo|date2=foo type template isn't going to work well. Any [22:21:59] 03siebrand * r32827 10/trunk/phase3/languages/messages/ (38 files): Localisation updates for core messages from Betawiki (2008-04-06 00:14 CEST) [22:26:39] is there no way to force lowercase word for the URL of a particular page? [22:32:02] andrew`: you can transclude normal pages, which would let them use whatever format they want [22:32:21] yep, but I want them all to be the same format [22:32:27] otherwise it will get messy very fast [22:38:35] 03siebrand * r32828 10/trunk/extensions/ (100 files in 97 dirs): Localisation updates for extension messages from Betawiki (2008-04-06 00:14 CEST) [22:39:28] apokryphos: sorry for the delayed response, but using {{DISPLAYTITLE:}} should do what you want [22:39:29] what about a spreadsheet style editor [22:40:19] apokryphos: like if a page was named "iPod", MediaWiki would display it as "IPod", but putting {{DISPLAYTITLE:iPod}} would force the first letter to be lowercase [22:40:31] (if you meant something else by your question please let me know) [22:41:47] Skizzerz: do I have to configure that template? Doesn't seem to work when I try [22:42:20] what version of mediawiki are you using? [22:45:53] Skizzerz: MediaWiki: 1.9.3 [22:46:34] apokryphos: that would be why. DISPLAYTITLE wasn't introduced until 1.11.0 [22:46:43] :( [22:46:51] Skizzerz: no way around it then? [22:47:10] you could use a javascript hack [22:47:33] Skizzerz: don't have access to the server :/ [22:47:48] you don't need it [22:47:55] just requires editing MediaWiki:Common.js [22:48:37] very hacky :( [22:49:21] it's what Wikipedia used before displaytitle came around [22:49:26] works the same way, but uses javascript [22:49:44] crazy that there was no proper solution for so long =) [23:05:42] Checking tsearch2 permissions...FAILED to access pg_ts_cfg. Make sure that the user "*" has SELECT access to all four tsearch2 tables [23:05:48] What are the "four tsearch2 tables"? [23:28:01] Shouldn't Apache display favicon.ico dropped into the DOcumentRoot? It's a PNG file 16x16 [23:29:52] the client picks it up eventually; you could add it with the $wgFavicon (or whatever it's called) [23:31:21] simonrvn: Also, I've got my wiki pages in my documentroot right now and it's working great except for the index.php/ in the URLs which I want to get rid of. I've tried basically all methods for getting rid of it on mediawiki.org but ti no avail. [23:31:44] i used the "root access/Alias" method, works fine [23:32:08] lymeca, favicon.ico must be, as the extension indicates, an Icon graphic [23:32:29] you can specify a PNG graphic as favicon through CSS, too [23:33:16] smart browser will then use that one, but you still can't stop that they'll aim for the favicon.ico in the root [23:34:14] lymeca, here's a freeware converter: http://www.winterdrache.de/freeware/png2ico/ [23:34:28] simonrvn: So you don't have a wiki/ in your URLs either? [23:37:01] i do, 'cept i don't have /index.php/ [23:38:57] I just attempted the 'root access/Alias' method, and I get You don't have permission to access /w/Main_Page on this server.' [23:39:14] lymeca, have you tried logging the RewriteEngine redirects? [23:42:28] k4jcw: That looks like the redirect is wrong [23:42:49] I'd agree, except that an 'ls' of the path I have on the Alias points write to index.php [23:43:15] The user should access /wiki/Pagename and it should rewrite to /w/index.php/Pagename [23:43:49] Perhaps what needs to be done for apache2 is different than for apache? [23:46:14] Here's the setup. I have a running installation in /wiki. cp -a wiki mediawiki so there's a duplication installation. Edit /mediawiki/LocalSettings.php, set $wgScriptPath = "/mediawiki"; Go to /mediawiki, all is well. Now I want to eliminate the index.php, so we follow the instructions, only the target will the /w, so it'll be site.com://w/Main_Page [23:47:14] $wgArticlePath = '/w/$1'; $wgUsePathInfo = true; and set Alias /w /var/www/localhost/htdocs/mediawiki/index.php. Restart apache2, go to site://w/Main_Page, get 403 error. [23:51:19] Hrm. site.com is a symlink to localhost in the /var/www directory. I wonder if apache has trouble with that when doing an alias? It's not broken anything else before, and I believe that was an apache manual suggestion for having multiple site names use the same directory. [23:59:35] Dashiva, got it fixed.