[00:02:02] Nervosa: true wysiwyg is not really possible, and even comming close is hard. the best option (but far from perfect) is probably the FCKEditor integration [00:03:00] Duesentrieb: The editor they're using looks pretty good. [00:03:14] Looks like JavaScript. [00:03:51] [00:03:55] Im not really concerned about the wysiwyg more or less lookin for an easy way for users to make mages and add them to categories [00:04:42] Nervosa: there's a number of extensions for page templates and category selection. havn't tried any [00:05:08] I also also have seen kinda the samething on wikia sites like http://halo.wikia.com/wiki/Special:Createpage [00:06:20] Duesentrieb: ah, guess Ill have to test them out there. Was woundering if there was a commen one since I have seen it on a few other sites. [00:08:14] no idea, maybe it is :) [00:15:00] Wikia's looks like a custom extension; it's listed on Special:Version (with no documentation link), while armchairgm's isn't. [00:16:03] hey wiki fans... I'm back with more obscure questions [00:16:45] Is there a way to let anonymous have read access from only a speciffic IP range/address? and make everyone else login to read? [00:33:48] 03(mod) Enable Gadgets extension in cs.wikisource - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11994 +comment (10danny_b) [00:43:32] Did I stump the experts? [00:46:45] Klemo: that is not a feature built into MediaWiki [00:46:58] however, you could write an extension to meet that need [00:54:59] lol [00:55:21] I dont think i'm savey enough to write an extension... [00:56:02] well maybe... but i definatly dont have the time ti write one [00:58:48] sigh... [00:58:54] any who... laters wiki freaks... [00:58:55] love ya [01:30:34] Hi [01:32:02] I'm having problems with the svg mime type, when I upload an image of that type it is showed with a text saying "image/svg+xml" I have no idea about how to remove it. Please, could you help me? [02:40:46] hello everybody [02:41:09] how do i configure mediawiki so that it allows user's own javascript? [02:41:41] like it happens on wikipedia: [[User:MySelf/monobook.js]] ? [02:42:54] is it some extension? [02:56:22] php nuke platinum [02:56:25] _ [02:56:27] ? [02:59:38] Xerol [02:59:45] hoşgeidn kardeş [03:06:32] index applications are up to ... adler [03:06:36] at least that's in s3 :) [03:08:01] yay [03:08:26] brion: log_id too? ;) [03:08:36] nope :) [03:08:43] :'( [03:26:14] :P [03:26:17] Pathoschild [03:26:41] No, bi. [03:26:58] masterpersona gay 2 [03:27:20] Oh no I am so deeply insulted. [03:27:38] :Pp [03:28:00] *AaronSchulz is surrounded by more gays than he though [03:28:18] brion: promise to to start trying to kiss me one day will you? [03:28:30] *to not [03:28:34] opps ;) [03:28:49] The usual Freudian slip. [03:28:57] *brion is straight and taken :) [03:29:07] *AaronSchulz wants a girl too [03:29:12] Pathoschild love you :D [03:29:18] *AaronSchulz thinks about stealing brion's [03:29:25] *Pathoschild-gay :D [03:29:29] No stealing. [03:29:30] *Pathoschild doubts brion would approve. [03:29:40] :P [03:29:43] I approve of Pathoschild stealing. [03:29:44] :DD [03:29:48] :o [03:29:50] no noo noo [03:29:51] :DDD [03:29:52] *Pathoschild steals masterpersona. [03:29:58] ;DD [03:30:00] :DD [03:30:11] ok, enough irc kiddie spam [03:30:11] ok they need to make mac os x work in parallels / vmware... [03:30:23] or else be easy to keep old copies of safari around actually using old copies of webkit [03:31:15] G_SabinoMullate: ! [03:31:45] G_SabinoMullane: http://en.citizendium.org/wiki/Special:ConfirmAccounts [03:31:47] wtf? [03:35:53] Checking... [03:37:20] btw, I should make a separate list for "held" accounts [03:37:50] AaronSchulz: Line 994 of _body : $tbl_arch is not defined anywhere [03:39:36] ah, I see [03:41:43] ok, the first run-through of applications seems done [03:41:51] i'll have to do master switches then the next application [03:45:07] AaronSchulz: Rolled into place [03:45:26] G_SabinoMullane: gah, I need shell ;) [03:51:14] ok gonna swap on s3 [03:52:15] locking [03:53:48] bah [03:53:50] something borked [03:55:14] show master status\G [03:55:14] Empty set (0.00 sec) [03:55:16] the fuck? [03:55:59] ERROR 1186 (00000): Binlog closed, cannot RESET MASTER [03:56:28] weeee [03:57:10] ok i guess we won't be using db1 as master. [03:59:50] samuel looks happier [04:00:18] How do I download a database dump? [04:00:43] whenever I try as is I get 26% of the file (about a gigabyte and a bit) and then the download stops [04:00:47] at the exact same place each time [04:01:04] this is for the all pages, current versions only, 20071018 dump [04:02:01] the articles, templates, image descriptions dump stops after a couple kb [04:02:10] yelyos: double-check that your download tool handles large files [04:02:18] It's firefox - should it? [04:02:27] ought to [04:02:27] It shows me the expected file size properly [04:02:52] Just a quick question on mediawiki please [04:02:58] It's fine until it hits that magic 26% and then the download rate crashes to 0 [04:03:10] and refuses to restart [04:03:33] yelyos: link to file? [04:03:49] Opps sorry to interrupt i have a quick question on MediaWiki Please ? [04:03:50] http://download.wikimedia.org/enwiki/20071018/ [04:04:00] sakthi: shoot [04:04:03] Thanks [04:04:10] there's not really such a thing as interrupting in irc [04:04:27] imo anyway. [04:04:30] which file? [04:04:37] I have some how got media wiki installed on win 2003 server [04:04:48] pages-meta-current.xml.bz2 [04:05:01] yelyos: all i can say is, if one tool fails, try another :) [04:05:05] I tried pages-articles.xml.bz2 too [04:05:11] and it downloaded some 9kb file onto my computer [04:05:48] But when some one tries to create a new uses then it come up with an error can some one tell which file do i need to tweek to talk to proper mail server ? [04:06:08] all right, i'll try and see if it works for me... [04:06:27] starting on s2... [04:06:57] MZMcBride: well, it worked fine for more than a gigabyte [04:07:06] so it'd be a while before you'd know whether you were getting the same problem [04:07:24] yeah, i realize; i'll see if it goes past 26% [04:07:40] Guys whats happening to my question please [04:08:04] sakthi: what's the error? [04:08:35] is the e-mail address that you set-up the wiki with valid? [04:08:58] Could not send confirmation mail. Check address for invalid characters. Mailer returned: SMTP server response: 550 5.7.1 Unable to relay for hsukumar@bond.edu.au [04:09:49] i know it is some thing to do with tweaking some file to talk to proper mail server [04:10:12] can some one tell me what the file is please? [04:11:12] all settings are in LocalSettings.php [04:11:30] all available options for LocalSettings.php are found in includes/DefaultSettings.php [04:12:34] MZMcBride: so do you think if i change it to proper mail server then i will not be getting this error? [04:12:46] the only way to find out is to try [04:13:01] ok i will try it thanks mate [04:13:17] I'll be back in a bit. [04:13:23] ok [04:13:25] I'll try wget [04:13:28] see if it works better [04:21:38] Can I check the values of 2 variable in an #if statement? [04:21:46] like, if either are given [04:22:12] I tried: #if: {{{var1|{{{var2|}}}}}} [04:22:27] but that seems to never display [04:22:51] What is the full {{#if:}} statement? [04:23:27] {{ #if: {{{mother|{{{father|}}}}}} | [04:23:38] then there's a couple of lines of .. oh wait .. wikitable stuff [04:23:49] I've had to fix this before [04:23:51] meh [04:24:28] Yeah, wikiML table syntax will mess with the ParserFunction syntax. [04:24:54] yeah .. it works with normal tr/td tags [04:25:14] i use {{ #if: {{{mother|}}}{{{father|}}} | [04:26:43] No difference, beyond style. :) [05:15:15] I changed the wgHashedUploadDirectory to false... how can I move all the images out of the hashed dirs without breaking them?? [05:40:44] 03(mod) add "refresh this page" link - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11811 (10jidanni) [05:44:32] what is the simple way to add an inline image? [05:49:54] it seems like uploads are not enables on my wiki [05:50:07] is there a setting I need to set? [05:52:34] gongoputch: see LocalSettings.php and DefaultSettings.php [05:52:53] ok [06:00:51] it wants 'convert' ... is that from Imagemagick ? [06:02:27] nvr mnd I am stupid [06:06:19] !stupid | gongoputch [06:06:19] gongoputch : The wise man knows he is stupid. [06:12:42] :) [06:23:06] Things dont work in media wiki at all its hard and there are so many easy ones in the market and why do people still use mediawiki [06:31:47] sakthi: I started on it because it has such a userbase, and wikipedia looks good to me [06:32:10] I wish postgres support was more solid for it tho [06:34:17] TimStarling: just a short question to ParserFunctions. In expr.php are ~ 10 errormessages which are not localizable. do you have any objections to move these messages to ParserFunctions.i18n.php to make them localizable? [07:12:34] Raymond: you mean the expr_* messages? sure, you can move them [07:14:29] *AaronSchulz hugs TimStarling [07:15:33] TimStarling: yes. these one. ok, I will move them :) [07:17:27] 03(NEW) A way to select which parts of articles to show - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11996 15enhancement; normal; Wiktionary tools: General; (hippytrail) [07:22:07] 03(NEW) Convert the WT: PREFS javascript to a PHP MediaWiki extension - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11997 15enhancement; normal; Wiktionary tools: General; (hippytrail) [07:33:14] http://en.wikipedia.org/w/index.php?title=Template:Loop&action=edit [07:34:36] 03(NEW) A way to show only the few languages a user is interested in - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11998 15enhancement; normal; Wiktionary tools: General; (hippytrail) [07:35:17] mysql> select count(*) from templatelinks where tl_namespace=10 and tl_title='Loop'; [07:35:21] | 58375 | [07:38:38] I hate users [07:38:50] hand up if you're a user, I want to kick you [07:42:43] no users here? how odd [07:44:25] never... /me is a writer, photographer, programmer, but never a user [07:44:50] what programming languages? [07:45:01] do you program in wikitext? [07:45:07] say 'no' [07:45:13] TimStarling: http://en.wikipedia.org/wiki/Talk:Chemical_imbalance#Response_to_.27Voice_of_All.27 [07:45:17] noooo. a little bit php [07:45:24] grr...I should just leave this guy [07:45:32] I really don't give a shit for this anymore [07:45:51] never written an email client in mediawiki templates? [07:47:27] you mean http://en.wikipedia.org/wiki/Template:Loop ? looks very evil [07:47:51] 03(mod) A way to select which parts of articles to show - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=11996 +comment (10dmcdevit) [07:50:12] *Raymond_ offers user:Polonium to kick [08:03:28] 03(FIXED) ParserFunctions error message localisation - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=8773 +comment (10raimond.spekking) [08:03:41] <_wooz> llo [08:03:56] *Raymond_ kicks CIA-36 again [08:04:34] bah, someone is prolly DoSing msn hotmail now :( [08:05:46] Raymond_: dead for days [08:06:09] Nikerabbit: I know but sometimes kicking helps :) at least to me :) [08:07:25] or is that not what you meant? [08:08:01] better kicked bot than a silent bot :) [08:09:02] is there easy way to get diff on one revision in svn? [08:16:30] + $messages = array_merge( $messages_base, $messages_base ); ?? [08:18:53] ?? [08:19:11] Duesentrieb_: ? [08:20:00] does anyone know how to set the message that is displayed when a MediaWiki installation is turend to readOnly? [08:20:34] The message says "THe administrator who locked it offered this explanation: $1"... but I don't know how to set teh $1 [08:21:15] sean: add $wgReadOnly='your read only error message' to localSettings.php [08:21:43] thank you! [08:22:43] Has any of you ever had problems with replication in MediaWiki before? [08:22:56] I had replication going for more than a week... then this morning the site goes haywire [08:23:13] and randomly reads and diffs are chosing between current content and content that seems to be from before replication started [08:24:27] how many requests per second are you serving? [08:24:39] probably about 8 to 10 [08:24:57] I have a squid set up and memcached... so those are other possible points I could be screwing up [08:25:38] 8 to 10 at the frontend? so presumably not much more than that figure at the DB in terms of queries per second? [08:26:05] around 8 at the backend [08:26:06] but yeah [08:26:13] not much more at the back [08:26:41] so surely your DB servers couldn't be heavily loaded at that rate [08:27:00] they're actually fairly heavily loaded [08:27:18] master was running around 30% cpu yesterday and slave was 80-90% CPU [08:27:32] I'm on relatively cheap hardware I guess [08:27:42] the master has 4GB of RAM and the slave has 2GB [08:28:30] i ran "SHOW SLAVE STATUS" and there was no entry in seconds_behind_master... so I assume everything was caught up [08:28:57] our db9 at the moment is doing 3300 queries per second [08:29:08] ... but the weird thing is that the data isn't just slightly stale... on random requests it's like a week old (my gues is from before replication) [08:29:17] hmmm... I'm not sure how many queries per second I'm doing [08:29:22] how can I check that? [08:29:43] $ echo 'show status like "Questions"' | mysql -h db9 ; sleep 10 ; echo 'show status like "Questions"' | mysql -h db9 [08:29:58] $ echo $((3437651457-3437618449)) [08:29:58] 33008 [08:30:22] or 3300.8 qps [08:30:50] running that... [08:31:23] my point is that replication does require a lot of sysadmin work, especially if you're balancing read load [08:31:29] it's not particularly simple to maintain [08:32:22] hmm [08:32:23] so unless you've got a really big site, it might be easier to serve the whole load from the master, and keep the slave as a hot spare [08:32:46] I'm doing 100 queries per second right now [08:33:05] I think it's fairly big... [08:33:18] I couldn't handle the traffic on just the one server before [08:33:28] db9 is a quad-core Xeon with 16GB of memory [08:33:29] Over a million requests get PAST the squid per day [08:33:41] dang [08:34:05] your data set size is important though [08:34:21] if you're serving entirely from RAM, MySQL can serve a *lot* of queries [08:34:57] about 500,000 good pages [08:35:13] ok, so your site is bigger than you thought it was :) [08:35:32] I guess :) [08:35:41] ... that's not all going to fit into RAM I take it [08:35:43] lol [08:35:52] well... not how little RAM i have [08:36:20] try "show table status", that shows you the data set sizes [08:36:43] it's particularly important that you have enough memory to hold the indexes [08:36:47] oh... I'm runnin 100qps on my slave too apparently (so 200 qps)? [08:37:08] hey [08:37:16] hello domas [08:37:27] TimStarling: I realized it is very hot in North Pole! :) [08:37:33] yeah? [08:37:34] is it a common joke in Australia? [08:37:38] http://www.accuweather.com/world-index-forecast.asp?partner=accuweather&traveler=0&locCode=OCN|AU|WA|NORTH%20POLE&metric=1 [08:37:39] :) [08:38:10] re [08:38:10] some of these index sizes are... realllly big [08:38:10] mmm, should I bump up heating or dress more [08:38:25] Sean: sometimes by making the index bigger you make it faster. [08:38:46] and if anyone has single app server, running memcached is bad. [08:39:03] why is that? [08:39:03] anyone tried Sphinx on Windows? [08:39:06] domas: that's amusing, I wonder if that's an actual place [08:39:30] we had -12C here this night [08:39:37] now it warmed up to -5C [08:39:46] why does winter strike that fast. [08:39:53] *domas blames global warming [08:39:58] I only have one APP server... but 4 servers total... is that still bad to run memcached? [08:39:58] Sean: ok, so let's assume that you do actually need replication [08:40:01] Sean: because there're IPC-less methods for same. [08:40:26] ok, thanks :) [08:40:42] hmm [08:40:51] the MediaWiki load balancer will distribute load between non-lagged slaves [08:41:00] non-loaded too [08:41:00] ok [08:41:26] hmmm, I have to remember what the defaults are and what it is that we use... [08:41:31] TimStarling: I had a question for you as load balancer expert. Should we remove master_pos_waits() from "Threads_running" or not? :) [08:42:42] it's a bit tricky, because if there are a lot of threads waiting, they have the potential to overload the server once they come out of wait [08:43:11] indeed [08:43:20] that was what made me hesitate [08:43:24] on the other hand, they may not [08:43:40] because the lag can be caused by some really long single write statement [08:43:51] now that makes "all servers busy" kind of thing [08:44:00] mmm [08:44:42] I don't understand how week-old data could crop up all of the sudden from replication [08:44:49] is there something simple I'm missing? [08:45:02] (like does it roll-back under some circumstances?) [08:45:24] domas: maybe you could take them out of threads_running, but put them in another variable [08:45:41] TimStarling: that was the idea [08:45:55] that way we can use threads_running+threads_lag_wait at the application level if need be [08:46:21] TimStarling: nowai I'd silently remove that without adding elsewhere [08:46:24] that would be fine by me [08:47:06] the term "threads_running" implies they're running, which waiting threads aren't [08:47:30] if you add another variable, then you give more flexibility to the app [08:47:44] then the app can choose what the definition should be [08:48:30] Sean: sounds like MediaWiki was not using your slave server at all [08:48:31] mhm [08:48:40] then something changed and it started using it, so you started getting old data [08:49:35] the default maximum lag is indeed infinite [08:49:46] does anyone know how to add the default mediawiki tables to a database with an arbitrary prefix, without using the normal webinstall? [08:50:00] hmm [08:50:29] you have to use something like $wgDBservers[1]['max lag'] = 30; [08:50:51] then mediawiki will stop using it if replication stops [08:50:57] hmm [08:51:00] I'll add that in [08:51:06] how can I tell if the replication is working? [08:51:13] show slave status; [08:51:17] I ran queries and got WEIRD results [08:51:18] k [08:51:24] and also show processlist [08:51:46] show processlist will show you the replication threads, and their status gives you some useful diagnostic information [08:52:00] it says it's waiting for master to send event [08:52:39] well, you can sanity-check it by looking for the most recent events in other ways [08:52:52] such as "select max(rc_id) from recentchanges" [08:53:28] on a working slave it'll be incremented at each edit [08:53:43] on a stopped slave, the value will be significantly behind the master [08:53:50] the slave has a HIGHER value [08:53:55] :[ [08:54:08] that would do it, yeah... [08:54:13] I wonder if I made MediaWiki do WRITES to the slave [08:54:25] you need to have read_only set on the slave [08:54:36] TimStarling: I'm not sure that parameter is used if slave thread is stopped [08:54:37] 03(mod) Install the StringFunctions extension - 10http://bugzilla.wikimedia.org/show_bug.cgi?id=6455 (10ross) [08:54:37] where... in wgDBservers? [08:54:47] in mysql, it's a mysql global variable [08:54:52] set @@read_only=1; [08:55:02] *domas giggles [08:55:33] k [08:55:35] because if mediawiki writes to a slave for whatever reason, you've got a bit of a disaster [08:55:42] which is what I suppose you've got [08:55:45] haha [08:55:48] i certainly have [08:55:58] the site is totally fubar [08:56:10] mediawiki doesn't write to slaves, usually [08:56:19] cause that would be trapped by our dberror.log :) [08:56:19] it's happened a few times on wikipedia, due to misconfiguration [08:56:33] TimStarling: well, someone promoted slave to a master, in that misconfiguration [08:56:33] let me see if i'm doing this right [08:56:37] there might be some scripts in the maintenance directory to help you out [08:56:56] so I have wgDBuser, wgDB[whateever] all set to the master [08:57:08] first of all, set your site to read-only in MW [08:57:08] then I have wgDBservers as an array [08:57:10] and that array contains the master and the slave [08:57:21] so that the problem doesn't get any worse [08:57:22] yeah, it's read only at the moment [08:57:29] I set that right when I got in this room [08:57:35] http://lyricwiki.org [08:58:06] in wgDBservers both the master and the slave are there, but the slave is configured to take 70% of the load [08:58:25] next you need to look at MediaWiki's tables on each server and work out what has been written to the slave alone [08:58:43] ok [08:58:52] sometimes you can just promote the slave to be the new master, if the switch was clean [08:59:04] alrighty... I'll start digging [08:59:08] but if writes were mixed, some to the real master and some to the slave, then you'd have to merge them [08:59:08] I have a weird anomoly that might help me figure things out though... [08:59:20] so right before I turned the site to ReadOnly, I created a page [08:59:27] 'LyricWiki:Test' [08:59:36] then I went to each database to see which it was created on [08:59:58] and (maybe I'm just messing up the query here)... it looked like it wasn't in either database [09:00:03] even though the page shows up on the site [09:00:18] fun [09:00:43] does that make ANY sense? [09:00:47] here's the query I used: [09:01:01] SELECT * FROM wiki_page WHERE page_title='LyricWiki:Test' [09:01:16] oh okay... wtf [09:01:19] now the page isn't on the site [09:01:42] it's probably there, you're probably just not familiar with the schema [09:01:53] cache [09:02:03] try: SELECT * FROM wiki_page WHERE page_namespace=4 and page_title='Test'; [09:02:11] oh snap [09:02:14] that makes sense [09:02:18] I'll try it [09:02:57] sweet! [09:03:00] okay... it's on the master [09:03:05] unsweet... [09:03:15] lol... I think the writes were done to both database [09:03:16] s [09:03:34] I'm just surprised that it went for so long without doing anything weird on the site [09:03:35] until today [09:03:47] almost everything was messed up today [09:03:56] but yesterday everything was cool; [09:06:00] I did change the weights on the servers last night... that might have done it [09:06:07] I had it at 99% slave, 1% master [09:06:33] so my slave is probably in a better spot... and I just wasn't noticing the 1/100 requests that were getting messed up [09:07:37] how are you telling mediawiki which one is the master and which is the slave? [09:08:23] well, from the docs, I thought it just assumed that wgDBuser, wgDBpassword, etc. were for the master [09:08:29] and everything in wgDBservers were used as slaves [09:08:35] this is probably wrong, huh? [09:08:36] well, they're not [09:08:39] yep [09:08:40] OUCH [09:08:46] how am I supposed to tell it? [09:09:39] looks like it's not documented [09:10:03] hmm [09:10:06] how do you guys do it? [09:10:07] the first member in $wgDBservers is the master, the server with index 0 [09:10:17] WOW [09:10:21] OUCH [09:10:26] I had it the other way [09:10:34] if you set $wgDBservers, then the single-server variables are generally ignored [09:11:04] hmm [09:11:16] alright [09:11:27] looking at the data [09:11:37] it seems that my slave should now be the master [09:11:55] it's not entirely clean-cut since 30% of the writes were going to the master today [09:12:19] ...wait... why were any writes going to it if the first server is used as the master? Will MediaWiki send writes to all of the servers unless they are read_only? [09:12:38] writes will go just to first server in array [09:12:42] fun. :) [09:13:17] oh... okay [09:14:52]