[00:08:20] 03(mod) Automated smallest-CIDR calculation from within Checkuser UI - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17018 (10JSchulz_4587) [00:11:34] hi [00:12:56] is it true there is no way to get double-strikethrough in html or whatever? [00:13:45] Hi I have one question, which are the issues why mediawiki uses by default binary fields instead utf8 ones? [00:14:56] Mike_lifeguard, right. [00:15:01] At least not reasonably. [00:16:38] http://www.w3.org/TR/css3-text/#decoration [00:16:43] CSS3 may allow it. [00:16:57] antiPoP- I believe MySQL 4.0 doesn't support utf8 at all, and newer versions don't support characters outside the BMP [00:16:59] But it's totally up in the air. [00:17:14] anomie, yep. [00:18:01] "text-decoration: double strike-through" might work in the future at some point, but not now. [00:18:15] anomie what do you mean for BMP? [00:18:28] *Mike_lifeguard keels over and DIES [00:18:48] (but thanks, Simetrical... I guess I'll stop suggesting we use double-strikeout then, eh? :D) [00:19:02] Good thought. [00:19:07] antiPoP, the basic multilingual plane. [00:19:17] http://en.wikipedia.org/wiki/Basic_multilingual_plane [00:19:58] Wikipedia, for instance, does contain characters outside of the BMP, so utf8 wouldn't be acceptable for it even if it worked in MySQL 4.0 (which Wikipedia uses). [00:20:30] 03(mod) Only ask gender by default if it will be in fact used for the current language - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18841 (10jidanni) [00:28:04] Is there a way to set a maximum character limit for names? [00:29:16] Hm. [00:29:18] !defaultsettings [00:29:18] --mwbot-- For the current version of DefaultSettings.php, see: . [00:29:31] Off-hand, the easiest way is probably to install Titleblacklist. [00:29:44] there is already a maximum length [00:29:50] Marybelle, doubt that will work. [00:29:52] Oh, maximum. [00:30:00] *Marybelle can't read. [00:30:07] carl-m: How can I change it because it's way too high. [00:30:25] $wgMaxNameChars = 255; [00:30:39] Put that in LocalSettings.php, set it to whatever you want, and then run... [00:30:46] ./maintenancecheckUsernames.php [00:31:24] Simetrical: The Titleblacklist has a option or something. You could do .{44} [00:31:31] I think I have that set on en.wiki, actually. [00:31:52] .{40,200} [00:32:00] *Marybelle stares. [00:32:09] Okay thanks [00:32:20] o_O [00:32:34] I wonder who added the 200... [00:33:00] Also is it possible to prevent a user from completely blanking a page? I've been getting quite a lot of vandalism and that seems to be what the users have been doing. [00:33:17] try abusefilter? [00:33:31] I'd also like to possibly have someone blanking a page an auto ban [00:33:35] Marybelle: so people can create 201-255chr names? [00:33:46] Splarka: I guess so. Seems rather stupid... [00:33:57] I'm fairly sure someone else is to blame. [00:34:00] stupid in MY en.wp? it's more likely than you think [00:34:31] eep, thunderstorms on the way, bbl [00:34:32] .{40,} [00:34:36] That's what I wrote... [00:36:48] 03(mod) Edit summaries are rendered differently to admins when contents have been concealed by RevisionDelete - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17060 (10mikelifeguard) [00:38:40] Check history? [00:39:41] 04(REOPENED) Automated smallest-CIDR calculation from within Checkuser UI - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17018 +comment (10mikelifeguard) [00:39:43] Yeah. [00:39:46] Found it and fixed it. [00:39:50] People are idiots. [00:39:59] {40,} was "broken." [00:40:03] Rawr. [00:40:12] I hope it wasn't me who fixed it :D [00:40:32] Which project was the one that banned all usernames over ten characters ... [00:40:55] enwikisource or something? [00:41:12] I think that was partially my fault too :( [00:42:09] [A-Z ]{10,} [00:42:13] Mike_lifeguard: they hated you so much they banned all usernames over 10 chars? [00:42:27] And .{30,} [00:42:29] just for an excuse to ban you? [00:42:31] From, shockingly, Wiktionary. [00:42:32] :( [00:42:38] nah, it was when I was switching everything from userblacklist to titleblacklist [00:42:59] Marybelle: .{30,} isn't a bad idea per se [00:43:11] there were some peculiarities in the regexes and I had to make educated guesses as to what they wanted to do, and sometimes I made mistakes [00:43:13] there does have to be some limit, and if a site wants it to be shorter, so be it [00:43:53] I just create the account on a lenient project and use SUL. :-) [00:43:58] there was another one where punctuation got disallowed, but that wasn't my fault, honest :P [00:46:10] hi all. anyone can help with a question? in twiki i use !+ to show/hide a text, how can i do with mediawiki? [00:47:29] What does "show/hide a text" mean? [00:47:49] I think the answer is that there's no such built-in functionality. [00:53:43] Simetrical, Like the "hide/show" link in the TOC. [00:53:57] There's no built-in function to do this for arbitrary text. [00:55:44] with twiki i use !+ [00:57:34] and colapse text [00:59:59] There's no such thing in MediaWiki. You'd have to write the JavaScript yourself, make a template or whatever. [01:00:58] Simetrical thanks for the reply... [01:01:32] I'm looking at the mediawiki schema, and seems that although articles are stored as binary, the page titles are in utf8.... [01:01:54] so I should not expect extended characters in the titles, rigth? [01:02:55] Simetrical, just this? heheheheh [01:28:11] What can I do to drive down the amount of time it takes to serve a page? [01:28:20] Currently it's taking .681 s [01:30:01] Aaron has a subpage about making MW faster. [01:30:06] On MW.org [01:31:04] chuck: not installing 80 zillion extentions helps too [01:31:41] Prodego: Oh hush :P [01:31:46] Maybe I'll prune some out [01:45:39] could i pick somebody's brain about makefiles and such? [01:47:33] i have some stuff which uses index files for wiki dumps and i want to automate building those indexes [02:09:50] 03(mod) Consolidate special pages with namespace selector - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=4204 (10jidanni) [02:09:51] 03(mod) Query pages should limit to content namespaces, not just main namespace - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=8130 (10jidanni) [02:24:11] 03(mod) WAI-ARIA landmark roles to improve accessibility in the Monobook skin - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18338 +accessibility (10jidanni) [02:30:29] hello [02:32:55] 03(mod) PNG cannot display thumbnails if the original over 12.5 megapixels - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18826 (10cuerden) [02:35:49] !r 50714 [02:35:49] --mwbot-- http://www.mediawiki.org/wiki/Special:Code/MediaWiki/50714 [02:35:53] Hello. [02:40:36] 03(mod) Thumbnailing produces much worse results for PNGs than JPG, but PNG is the only real option for serious restoration - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18014 (10tstarling) [02:41:59] was that to me? [02:42:11] The hello was. [02:42:24] The link was for me. [02:42:33] right, hello [02:42:52] Did that lock problem get resolved? [02:43:11] I haven't heard too many complaints today. [02:43:24] whoever it was probably stopped updating the template [02:43:33] hey TimStarling. [02:43:39] my days at an end ... seeya [02:43:44] Bah. [02:43:49] I had a question for him. Oh well. [02:43:56] *Marybelle notices Tim is Bugzilla trolling. [02:44:13] it's a serious question [02:44:29] he submits a lot of bug reports but never explains why he thinks we should work on them [02:45:20] "..why it has any relevance to Wikipedia?" is bear-poking ( http://enwp.org/WP:BEAR ) [02:45:29] apart from mentioning some relevance to serious restoration [02:45:43] Well, Commons is a media repository. [02:45:53] So presumably they do image work. [02:45:57] The Commoners. [02:46:00] serious restoration? [02:46:21] Well, little in Wikimedia is serious. Most of it is probably amateurish. But I'm sure they take it seriously. [02:46:27] he has to make his case either way, I'm not working on it as is [02:46:45] Fair enough. [02:47:23] speaking of bugzilla trolling, I'm just reading a WAI document defining accessibility and listing web-relevant disabilities [02:47:35] haven't seen any mention of text browser compatibility yet [02:47:40] This is Jidanni-related? [02:47:58] yes [02:48:07] Marybelle: who else would it be :P [02:48:47] I got WONTFIXED a few days ago. [02:49:14] I should go bug-war or something. [03:10:22] what is wrong with this syntax: [03:10:23] {{#ifeq:{{{1}}}||an unspecified matter|}} [03:10:52] harej: I'm not particularly experienced, but you might want {{{1|}}} [03:10:53] if a parameter is not provided, then it yields "an unspecified matter", while doing nothing if at least something is provided [03:11:53] See http://www.mediawiki.org/wiki/Help:Extension:ParserFunctions [03:13:58] harej: What's the problem? [03:14:05] Marybelle: I think i have it figuredo ut [03:14:11] Excellent. [03:14:15] moring [03:15:01] any idea about how to make a subpage for each index item [03:19:20] make a link for each index item to a non-existing subpage, click on each link and edit it [03:23:36] but my page is a full book [03:23:40] 04(REOPENED) Enable AbuseFilter on Cantonese Wikipedia - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18589 +comment (10shinjiman) [03:24:25] it import from msword.should i mannually split 500 sub page?. [03:25:22] no, you should split it by chapter [03:25:30] not by page [03:25:48] yeah. [03:26:46] isn't there any choice to auto finish it? [03:27:33] I doubt it [03:27:37] you know where google is though [03:28:24] i tried ,but get nothing [03:30:09] well, that's probably what there is then [03:37:37] ashley: That vandal you keep reverting on usability wiki has replaced the header with "Good site, admin" again, you might want to leave a message on his/her talk page for block for a few days or w/e [03:44:48] 03(mod) Enable AbuseFilter on Cantonese Wikipedia - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18589 (10shinjiman) [03:48:30] I just tried to use ini_set in my LocalSettings.php to increase the maximum file upload size to 5MB... and I'm still getting an error "The file you uploaded seems to be empty." I can't find anything else obviously wrong, as the default values of post_max_size and the $wgXXX vars seem to be ok. What could be wrong? [03:56:57] file permissions? [03:57:28] Where? on the server? I've been able to upload other (smaller) things. [03:57:41] I mean, I can just shrink it, but I'd rather know why it's failing. [04:02:48] jtbandes: do you know how to use phpinfo()? [04:03:26] TimStarling: yeah [04:03:41] Unless there's some aspect of it you're referring to that I haven't seen. [04:04:27] check that the post_max_size and upload_max_filesize in phpinfo() match what you set them to in php.ini [04:05:10] I didn't set them in php.ini, as I'm on a shared host and I can't access it (though I may be able to make a custom one... haven't tried). I used ini_set(), but should that not still work? [04:05:45] I'd doubt it, if it's a shared host [04:05:59] Doubt which part? [04:06:08] that you'd be able to override it with ini_set [04:06:13] no, ini_set does not work [04:07:29] it says PHP_INI_PERDIR in the manual, so you should be able to set it through .htaccess [04:07:52] Aha, I can use a custom php.ini. [04:19:47] Ok, I tried php.ini, but it still seems to be broken. [04:21:10] I'll just upload a smaller version for now. [04:58:53] jtbandes: did you check it against phpinfo() this time? [04:59:14] because there's a lot of things that can go wrong in between php.ini and phpinfo() [05:01:16] TimStarling: looks to be working [05:01:21] But then MW pays no attention. [05:10:22] has anyone here installed plexcel for their mediawiki install? [05:11:03] jtbandes: there's not much MW can do about it, besides informative error messages [05:11:08] im mostly through it and have stuck on an issue at the last step, which is getting the sso going when using a browser (works fine using wfetch like in the documentation) [05:11:12] file uploads are handled by PHP before MW starts [05:11:41] devians-work: never heard of it [05:11:50] Well, it shows the limit as 2MB on the upload page, and it's not working to upload something that big. [05:12:29] jtbandes: is the wiki public? [05:12:51] TimStarling, its a php extension that allows for integration with active directory [05:14:19] and there's a MW auth extension for it? [05:14:57] TimStarling: not exactly, by which I mean yes, but anonymous users can't upload, and the wiki isn't very heavily used yet. what do you want to see? [05:21:01] jtbandes: the phpinfo() output, and whether or not it's generated for the same directory as the wiki [05:25:12] TimStarling: you win, it isn't loading the custom php.ini that changes the value from the wiki's directory, only the one in the docroot. Running php -i via SSH while in the wiki's directory works, though. [05:30:59] 03(NEW) Customizing the dimension of pictures in categories - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18843 enhancement; Normal; MediaWiki: Categories; (public.wiki) [05:38:52] urgh, active directory sso for mediawiki seems to be a bit of a tall order. i have this plexcel thing installed but it is being a royal pain in the butt. [05:40:21] it's not like there's a massive user community [05:40:41] pardon? [05:41:01] probably just some guy like you who spent a few hours writing the auth plugin and then forgot about it [05:41:10] you shouldn't expect great things [05:41:22] plexcel? or ad sso for mediawiki? [05:41:39] the second one [05:43:15] yeah there seems to be a pile of 'i hacked this together' solutions but its nothing solid [05:44:09] no big deal if you lower your expectations and learn PHP [05:48:50] or you could ask for your money back... [05:49:06] heh [05:51:09] been using php for 6 or so years now. and who is seriously stupid enough to pay for something that expensive before seeing if it even works? [05:51:17] the language isnt the problem, its time. [06:06:48] 14(DUP) LoginForm reports wrong password when CAPTCHA is wrong - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18798 +comment (10tstarling) [06:06:53] 03(mod) Vague error on captcha mismatch during login - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=12206 +comment (10tstarling) [06:24:03] 03(mod) Add setting to hide a list of preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18761 +comment (10tstarling) [06:32:23] Hi, anyone here who could help me setting up user rights with the Lockdown Extension? [07:10:31] grrrrrr [07:11:46] https://bugzilla.wikimedia.org/show_bug.cgi?id=18828#c4 -> #5 -> http://www.mediawiki.org/w/index.php?title=Help:Magic_words&curid=1748&action=history [07:12:52] 03(mod) Can't link to [[{{FULLPAGENAME}}]] any more - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18828 (10herd) [07:19:52] 04(REOPENED) Have a permalink type URL as a user preference - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18796 +comment (10Gerard.meijssen) [07:34:15] *Splarka wonders if https://bugzilla.wikimedia.org/show_bug.cgi?id=11281 should dupe to https://bugzilla.wikimedia.org/show_bug.cgi?id=4970 [07:34:20] *Splarka guesses not [07:38:35] i don't think 'duplicate' in that sense means what you think it does [07:38:46] (you want to make a copy of the bug at another url?) [07:39:55] no [07:40:11] dupe, as in DUP, not duplicate [07:40:19] but that isn't a verb [07:40:26] neither is graceful [07:40:36] your point? [07:40:55] so it's illegal to make up new words [07:41:01] you shouldn't do it [07:41:10] it is a civil offense, it isn't criminal [07:41:15] *Splarka pays the fine and goes on [07:42:16] anyway, thanks for the waste of time, in actual constructivenessiveness: I guess what 11281 wants is a way to control the limit in the UI, not necessarily exceeding the limit set in Default/LocalSettings [07:43:59] a synergenic solutionitude might be ampleish duplifaction of the paginationables in other listovules [07:47:46] 03(mod) Page history lines are rendered differently to admins when contents have been concealed by RevisionDelete - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17060 summary; +comment (10happy-melon) [08:01:50] 03tstarling * r50763 10/trunk/phase3/includes/LinksUpdate.php: Reverted r46706, misses essential updates, the feature can easily be implemented another way (see CodeReview). For 1.15. [08:05:51] 03tstarling * r50764 10/branches/REL1_15/phase3/ (RELEASE-NOTES includes/LinksUpdate.php): Reverting r46706 (category redirects), wrong design, buggy [08:10:49] 04(REOPENED) Automatic category redirects - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3311 +comment (10tstarling) [09:07:20] 03(mod) Free up "Project" as a namespace (for Wikipedia:Projects) by using an alternate token - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18793 +comment (10roan.kattouw) [09:46:25] hello guys [09:46:56] $title_cat = $title->getParentCategories() returns an array, how can I retrieve the name of the categorie(s) in it ? [09:50:38] 03(mod) Automatic category redirects - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3311 (10catlow) [09:51:13] I don't suppose anybody here is possibly using mediawiki on a BSD machine w/MySQL on the back end? Oh and hearty greetings to everyone out there. :) [09:52:44] After much fighting with familiarizing myself with OpenBSD vs. Linux, I've got mediawiki up and happy with apache & php, but I'm having problems getting it to understand that the MySQL database server is, indeed, there, and I am not turning up anything too helpful in google... My own exploits for configuring the database/account/password correctly for the wiki & database are not seeming to work [09:56:13] !man Running_MediaWiki_on_FreeBSD | dgetsman [09:56:13] --mwbot-- dgetsman: http://www.mediawiki.org/wiki/Manual:Running_MediaWiki_on_FreeBSD [09:56:41] p858snake|: was typing out that macro less work than just copypaste linking them to it? [09:57:16] well i knew what the article was actually called so yes [10:03:54] msg mwbot2 that doesn't seem to have exactly what I need; I've followed all of those steps. are you aware of any bsd specific or even more general MySQL<-->Mediawiki troubleshooting pages? I haven't turned any up [10:04:33] p858snake|: case and point^ [10:05:19] mwbot is often a false-authority abstraction layer [10:08:49] hah [10:09:04] teach me to type before coffee [10:35:02] 03(mod) PNG cannot display thumbnails if the original over 12.5 megapixels - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18826 (10mikelifeguard) [10:50:58] wouldn't want to dirty the hands with inaccurate advice [11:02:24] 03(mod) Only ask gender by default if it will be in fact used for the current language - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18841 (10niklas.laxstrom) [11:05:30] 03(mod) Free up "Project" as a namespace (for Wikipedia:Projects) by using an alternate token - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18793 (10niklas.laxstrom) [11:11:09] anyone with IE 8 ? [11:12:26] *Splarka opens up his Angry Mob Emporium franchise, polishes the pitchforks and oils up the torches, and looks at thedj [11:12:31] would have to reboot [11:12:52] http://en.wikipedia.org/wiki/Wikipedia:Village_pump_(technical)#Weird_behavior_when_editing [11:13:15] no bug to free up "Image" as a namespace yet? [11:13:18] we need a solution/work-around for that... [11:13:39] workaround - use FF/chrome/!IE [11:14:02] yes everyone knows.... [11:14:11] :P [11:14:23] don't use IE. unfortunately, i'm not employed by every single company in the world to make that happen. [11:14:35] :) [11:14:37] You should be ;) [11:14:57] "don't use MediaWiki" [11:15:10] *Reedy glares at Splarka [11:15:32] seriously, some people just complain too much, the returns they are diminishing [11:15:39] *cough*jidanni*cough* [11:17:05] here is my problem with IE... [11:17:43] the people that report problems with IE and expect you to fix them, are of the intelligence and capability level consistent with those who choose to use IE, or are forced into a situation whereby they have no choice but to use IE [11:18:51] so when trying to get them to help you help them, they are typical IE users... X_X [11:19:14] I for some reason use IE for some sites, FF for others and chrome for more.. [11:19:42] I used to use IE for flash as it was more reliable, that was IE 5.0 for those keeping score [11:19:46] but also have safari and opera installed for teh lulz [11:19:55] i use safari for browsing and the others for testing. (except IE8) [11:40:29] Hi there, I have a problem with installation of MediaWiki 1.14.0. Here a the response during installation process: [11:40:31] PHP 5.2.5-tuxtools installed [11:40:32] Found database drivers for: MySQL [11:40:34] PHP server API is cgi; using ugly URLs (index.php5?title=Page_Title) [11:40:35] Have XML / Latin1-UTF-8 conversion support. [11:40:37] Warning: A value for session.save_path has not been set in PHP.ini. If the default value causes problems with saving session data, set it to a valid path which is read/write/execute for the user your web server is running under. [11:40:38] PHP's memory_limit is 32M. [11:40:40] Couldn't find Turck MMCache, eAccelerator, APC or XCache; cannot use these for object caching. [11:40:42] GNU diff3 not found. [11:40:43] Found GD graphics library built-in, image thumbnailing will be enabled if you enable uploads. [11:40:45] Installation directory: /www/htdocs/tanja/wiki [11:40:46] Script URI path: /wiki [11:40:48] Installing MediaWiki with php5 file extensions [11:40:48] please stop pasting [11:40:49] Environment checked. You can install MediaWiki. [11:40:51] Warning: $wgSecretKey key is insecure, generated with mt_rand(). Consider changing it manually. [11:40:53] Generating configuration file... [11:40:55] Database type: MySQL [11:40:57] Warning: Unterminated comment starting line 334 in /www/htdocs/tanja/wiki/includes/WebRequest.php on line 334 Parse error: syntax error, unexpected $end, expecting T_FUNCTION in /www/htdocs/tanja/wiki/includes/WebRequest.php on line 334 [11:41:01] I really don't know what to do (after checking FAQ, and the other regular support-stuff) and would be very very thankful if someone gives me instruction about solving this problem. best regards... [11:47:37] 03(mod) Automatic category redirects - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3311 (10tstarling) [11:48:26] how to check that mediawiki is in "edit mode" via php? [11:49:08] Splarka: don't temp me.... i will file one about it [11:57:09] found it. :) ($mediaWiki->getVal('Action');) [11:58:12] ugga [11:58:15] note that 'submit' is also an edit window generally (usually a preview, or other edit that doesn't 302 redirect to the article page after submission as a save does) [11:58:29] *Splarka pokles Rabbit [12:00:10] auxelius: did you upload with ftp? [12:00:40] yes [12:01:10] looks like all the files were not uploaded properly [12:01:59] do you mean that they are burst while ftp-process? [12:03:16] I'm just saying that there should be more than 334 lines in that file [12:03:34] and be warned that other actions may also be harder if you do not have shell access [12:11:41] in fact there are more than 334 lines in that file (733)! [12:14:07] well line 334 should be empty, so there can hardly be a parser error there [12:14:39] I have a comment at that line... [12:14:49] auxelius: can you paste the line? [12:15:00] only that one [12:15:11] /** [12:15:13] * Fetch a boolean value from the input or return $default if not set. [12:15:15] * Guaranteed to return true or false, with normal PHP semantics for [12:15:16] * boolean interpretation of strings. [12:15:18] * @param $name string [12:15:20] * @param $default bool [12:15:22] * @return bool [12:15:23] */ [12:15:24] auxelius: please do NOT paste [12:15:42] I'm sorry [12:15:59] If you need to show us multiple lines, please use an online pastebin [12:17:51] Sorry again - It's my first time [12:18:22] auxelius: and that error happens again if you try again? [12:18:41] again and again!!! [12:20:02] I even tried out with index.php and index.php5 [12:24:23] auxelius: are you uploading from windows? [12:24:38] yes [12:28:07] auxelius: with text or binary mode? [12:29:25] I configured filezilla to decide automatic! [12:34:49] 03(mod) Automatic category redirects - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3311 (10catlow) [12:40:35] nice to see that demon has taken up the text browser cause, committing lots of jidanni's patches [12:41:49] someone's got to help all those people surfing the web from 1970's era dumb terminals [12:42:11] :O [12:45:24] auxelius: no, either the files are corrupted during upload or your php is broken [12:45:30] 03ashley * r50765 10/trunk/extensions/SocialProfile/UserProfile/SpecialUpdateProfile.php: [12:45:30] SocialProfile: tweaks to UpdateProfile special page - only spam unconfirmed [12:45:30] users with that confirm email message (it's broken anyway, the links aren't [12:45:30] working) and add ids to some elements to that they can be hidden with css [12:46:05] http://www.freedomscientific.com/fs_support/BulletinView.asp?QC=1165 [12:47:12] so blind people can read documents with CSS, just not Jidanni with w3m [12:47:39] Nikerabbit: should I upload the files again in binary mode? [12:48:41] TimStarling: but JAWS apparently ignores aural and speech media types, it tries to be "internet explorer" [12:48:45] auxelius: you can try [12:49:08] TimStarling: where are those jidanni patches? [12:49:11] saper: I think you can configure that [12:49:18] in bugzilla [12:49:19] though it's been a while since I've used JAWS [12:49:20] Mike_lifeguard: I don't think so... [12:49:39] I miss non-CSS "?" for redlinks very much, is that included there? [12:50:44] saper: what sort of browser do you use? [12:50:54] TimStarling: w3m [12:51:03] withing emacs (seldom) and standalone [12:51:11] it's great [12:51:13] with images or without? [12:51:16] for mediawiki [12:51:19] usually without [12:51:41] in emacs I switch them on sometimes or per-image (like somebody's foto) [12:51:53] do you use it because your only internet access is via a text-only dumb terminal? [12:52:02] no [12:52:22] I have a fully featured laptop with 3D acceleraated graphics capable to run compiz [12:52:52] it's much quicker to start w3m with a link from irc recent changes or just like that than wait for seamonkey/firefox whatever [12:53:24] I also like "buffer" feature very much - like history <- -> but easier to use and you do not "lose" forward pacges [12:53:27] pages [12:53:52] so I can have preview and edit box in two "buffers" if I want [12:54:32] so kind of like tabs? [12:54:52] except wrapped in an antiquated user interface? [12:54:55] w3m has tabs in the firefox sense [12:55:05] but I don't use them [12:55:21] buffer always full screen [12:55:35] it can contain page info or page source, if you need [12:56:10] *TimStarling is not buying [12:56:36] well, I am much faster with this [12:56:58] last time I was fixing some stupid links that were too complicated to do them with bot [12:57:11] (or actually doing python logic for them would take too much time) [12:57:14] and it was great [12:57:22] the only think I don't have are redlinks [12:57:40] due to this (pardon me) stupid "after-content:" CSS gimmick [12:57:57] you should try an already-running firefox with CSS and images disabled [12:58:06] no [12:58:07] should be pretty quick [12:58:18] I don't want "already running" up [12:58:25] I need to start up and tear down quickly [12:58:32] low RAM? [12:58:34] w3m is like less [12:59:14] it's like telling me "why don't you page all your manpages and textfiles via the standalone less monster" :) [12:59:37] I read my manpages in firefox [12:59:38] TimStarling: do you know "dwm" window manager? [12:59:44] I have a bookmarklet and man2html [13:00:02] so for you firefox is your emacs :) [13:00:07] for me it's screen and xterm [13:00:43] just say it, you are stuck in your old habits :o [13:00:44] http://www.cactusflower.org/why-screen-readers-dont-support-aural-css [13:01:12] Nikerabbit: nope, it's postindustrial, I used to run gnome/kde and those kinds of things :) [13:01:22] is dwm one of those many lightweight window managers designed for machines with a meg or two of RAM? [13:01:57] it's ca 1000 lines of source code [13:02:33] Screen readers don't use aural CSS for the same reason Opera's small-screen mode totally mangles the display semantics of CSS and HTML. [13:02:46] the xerox desktop paradigm is gone awol, nobody uses a window system like it was meant in the 70'ies, with menus and many windows etc. [13:03:14] Because people write web pages for normal browsers, and it turns out it's a lot more reliable and efficient to make a "normal" -> "weird" translator than to get authors to mark stuff up for weird uses properly to begin with. [13:03:17] almost everyone I know uses maximised windows [13:03:43] Simetrical: for all, people have difficulty to imagine how it's like [13:03:44] saper: hilarious: You must enable scripting in your browser to post comments. [13:03:57] Nikerabbit: hehe lemme check [13:03:58] Why would anyone not want to use AJAX?! [13:04:09] since it's a gimmick :) [13:04:26] It's not a gimmick, it's actually quite handy for speeding up some things. [13:04:30] Although it's sometimes overused. [13:04:56] It allows you to do things in the natural way rather than shoehorning everything into HTML/HTTP's "totally separate pages" model. [13:05:57] Nikerabbit: hehe preview works fine, submit gives me 404 [13:06:13] Anyway, one nice thing about some parts of the W3C today is that they focus on stuff people will actually implement and use, and try to require use-cases for new features, rather than making up what they think a priori must be a good idea. [13:06:27] Like aural stylesheets. [13:06:31] Or longdesc. [13:06:43] well what used to be called hypertext evolves into the big permanent pipe between client and server [13:06:55] saper: I don't even get a form [13:07:04] Nikerabbit: I do and I can submit that [13:07:09] Which is a more natural way to do applications. The original web was intended for documents only, nothing very interactive. [13:07:16] Did the original HTML even have
? [13:07:55] 2.01 I am sure [13:08:00] 1.0 no idea [13:08:11] the web was not meant for "applications" that's the thing [13:08:21] the original HTML was thrown together by TBL in an afternoon [13:08:24] Yes, but now it's used for them, so . . . [13:08:37] I remember big discussions ca. 1994-1995 whether browser is the right user interface to do things like that [13:08:44] I don't see any good replacements for HTML for web applications. [13:08:46] the answer at the time was, mostly, "no" [13:08:47] what are we even discussing right now [13:08:55] Simetrical: well "web applications" :) [13:08:55] Unless we're going to all go to XUL or something. [13:09:19] Heh, okay. Thin-client untrusted Internet applications. [13:09:20] the web as defined is a request-response model [13:09:36] which makes sense, since the number #1 rule of IPC over net is to reduce number of calls [13:09:54] I think everybody trying to tune their web server for the ajax application quick finds that out [13:09:59] quickly [13:10:34] But AJAX can also reduce the number of requests, or at least make them smaller. [13:10:51] saper: if the answer in 1994-95 was "yes", maybe we wouldn't have three partially incompatible solutions for doing it [13:10:52] New page load to load reply form -> only load an HTML snippet instead. [13:10:59] from what I've experienced it can increase number of calls 5-10 fold ;) [13:11:07] Depends how you use it. [13:11:12] a standard at the right time would have saved a lot of pain [13:11:17] Look at AJAX watch in MediaWiki. [13:11:24] That strictly reduces network traffic, I think. [13:11:33] TimStarling: it was time when you know started to appear [13:11:37] (I should know, since I committed it, but I didn't write it . . .) [13:11:51] and java applets to display what is today an animated gif [13:12:09] If you're dealing with idiots like whoever designed the new Python documentation search, now, AJAX can increase the number of requests. [13:12:26] So can using too many images, scripts, or styles. [13:12:45] Simetrical: it always will, every HTTP request is the HTTP request, that's all. [13:12:45] My forum has like 50 hits/page view for uncached views. All the little icons, avatars, . . . [13:12:56] What always will what? [13:13:22] every XMLHTTPRequest is _the_ request [13:13:43] . . . what? [13:13:44] sure, if one needed to reload the whole page with a form 20 times [13:13:52] to get all validations right [13:13:52] the goal of the modern web isn't to minimise network traffic or remote calls [13:13:56] ajax helps [13:14:01] it's to improve user experience [13:14:08] TimStarling: that's the change of paradigm that's all [13:14:12] If you use AJAX for form validation, yes, that will increase calls. [13:14:13] if that requires a bit of extra network traffic, so be it [13:14:16] infrastructure follows that [13:14:26] infrastructure follows bittorrent [13:14:26] Or displaying search suggestions. [13:14:35] Yeah, the web isn't the big bandwidth hog these days. [13:14:37] the web just follows along [13:14:52] tries to compete with Bittorrent to get "fair share" :-) [13:15:37] obviously there are bad uses for AJAX, there are bad uses for anything [13:15:51] *saper imagines a nice new net neutrality argument "we will make the web a bandwidth hog if you won't let us use p2p" :) [13:15:55] but there are lots of cases where it is provably a good thing for the experience of the vast majority of users [13:16:26] sure, as there are thousands of use cases of web [13:16:34] the web [13:18:20] as there are many client side use cases, i.e. w3m :) [13:18:38] if you use w3m on a perfectly good computer then that's your own silly fault [13:19:05] that said, MediaWiki uses ajax in a way that is compatible with w3m [13:19:17] Some of the stuff is worth fixing for other reasons than terminal users, though. [13:19:23] you can watch and unwatch pages in w3m, you wouldn't know that you could do it faster in firefox unless you tried [13:19:27] . . . although maybe not most. [13:19:29] mediawiki works very fine for me, expcept for the red links [13:19:45] I wish standard skin could have the real "?" mark [13:19:52] you know why it doesn't, right? [13:19:56] Kills parser cache. [13:20:05] I was talking to the w3m user [13:20:07] and that "https://bugzilla.wikimedia.org/show_bug.cgi?id=2557" could finally be reviewed :) [13:20:07] Thus making everything slower for you. ;) [13:20:23] hi all. Can mediawiki username use only numbers? Ex: 1234567890 [13:20:32] Well, I was talking to him too, who says I was answering your question? [13:20:37] Simetrical: really? it's not output after parsing? [13:20:43] fernando: yes, probably [13:20:48] fernando: regular article name rules apply [13:20:49] fernando, yes, unless you have AntiSpoof. [13:21:23] saper, well, I guess you could write up a whole new postprocessor. [13:21:54] That could be run quickly on the output of the parser cache, and wouldn't be cached itself. [13:22:05] We could save a whole bunch of existing cache fragmentation that way. [13:22:16] Simetrical: so adding css class for the redlink does not break things and "?" does? [13:22:24] The CSS class is there for everyone. [13:22:34] The "?" would have to be there for only some users. [13:22:39] Thus, cache fragmentation, lower hit rate. [13:23:11] Your page views would be noticeably slower because every page would probably have to be reparsed for you, unless a lot of users use the "?" thing relative to the rate of changes. [13:23:27] Of course, if you use other weird settings they might already require a reparse most of the time. [13:23:44] Simetrical: well, but let me think.. the class is there only for non-existing links, right? [13:23:49] Yes. [13:24:17] maybe he's got heading autonumbering and stub links on already [13:24:32] I'm tryingo to use ldap here. My usernames are numbers only. It stop in: Munged username: 1234567890. then Entering modifyUITemplate and stop [13:24:33] then it wouldn't make any difference [13:24:37] I should turn off stub links, it's probably slowing down my Wikipedia viewing. [13:24:47] Although it's let me spot some vandalism sometimes. [13:24:58] fernando, LDAP might have its own rules. Dunno about that. [13:27:15] Simetrical: and I don't think there are that many standard/simple skin users out there to slow everyone down, if at all [13:27:28] Skin doesn't affect parser cache. [13:28:33] Hey Reedy [13:28:38] :) [13:29:08] so the classes are just cached in the parser output as read HTML (output from Linker::link() method)? [13:31:00] The parser cache is the raw HTML corresponding to the page contents. So yes, it will just contain stuff like . [13:31:33] Parsing large pages (converting from wikitext -> HTML) can take 10 seconds or more, so a high parser cache hit rate is essential for a good user experience. [13:31:44] And the parser uses most of the cluster's CPU time, also, so it's good to avoid that. [13:32:22] so stuff like "languageLinks" for normal links would help, wouldn't that? [13:32:31] What? [13:33:03] hm there is Parser-mLinks table already... [13:33:16] looking at "ParserOutput.php" [13:33:48] saper: the issue is that if you change the content based on a preference, you have to re-render the entire page for each value of that preference [13:34:01] it doesn't really matter how efficient the feature in question is [13:42:02] it'd be possible to do a post-processor like what Simetrical said [13:42:25] as long as it doesn't attempt to implement a full XHTML parser in PHP or something [13:43:19] sure [13:43:42] but don't things like toc or editsections depend on preferences, too? [13:44:09] Some parser-cache-fragmenting preferences already exist, but we try to minimize them. [13:44:14] The TOC one is CSS-based too. [13:44:21] Editsections probably is as well. [13:44:46] mm [13:44:59] post-processor would be quite awesome [13:45:18] couldn't the parser leave magic markers to the cache to make it easier to post-process? [13:45:25] of course [13:45:34] TimStarling: so we need something that instead of storing mText and mLinks of ParserOutput goes a bit up? [13:45:57] you could probably set it up so that the post-process was just a single strtr/fss_exec_replace [13:45:58] there is NOPARSE stuff in there currently, what is this? [13:46:40] $text = str_replace($this->mUniqPrefix.'NOPARSE', '', $text); [13:46:53] Parser.php:899 [13:47:32] fine, so the question is what the immediate output format of the parser should be not to make it slow and overly complicated [13:49:35] 03(mod) Automatic category redirects - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3311 (10roan.kattouw) [13:50:33] HTML with comments for markers [13:51:06] like the markers we use for RIL->RLH [13:52:30] ril rlh? [13:53:16] those acronyms are in the source now, I'm allowed to use them [13:53:39] does that mean I should know what they are? [13:54:02] replaceInternalLinks and replaceLinkHolders [13:54:20] parser/Parser.php: * Process [[ ]] wikilinks (RIL) [13:54:50] Q: if this is so easy and nice, why hasn't it done before? [13:55:34] it's easy and nice for w3m users [13:55:57] not sure how much benefit there is for everyone else [13:57:01] we all have lots of projects on our plates [13:59:10] We could implement {{USERNAME}} if we had a postprocessor, at least in a primitive form. [13:59:18] It wouldn't work for conditionals and such, of course. [13:59:19] Simetrical: but not for anonymous? [13:59:30] :( [13:59:31] Well, that would fragment *Squid* cache. :) [13:59:55] So only for logged-in users. [14:00:14] maybe we should just turn off parser cache entirely [14:00:18] when people put things in double braces, they expect them to work a certain way [14:00:25] you could implement though [14:00:30] it is not like we can't afford 50x more servers [14:00:46] We're talking about fragmenting the parser cache *less*, domas. [14:00:48] Using a postprocessor. [14:01:00] Which would just be an extra strtr() per request. [14:01:29] well, there should be no {{}}'s in content that are not cacheable [14:01:42] easy :) [14:03:16] you would have to be careful what you can substitute, and where, to avoid messing up the link tables and opening holes in the spam filters [14:03:40] but question marks after red links? yes that's easy [14:04:37] numbered headings? [14:05:44] yes, that too [14:06:23] it's a bit tricky without getting into regexes [14:06:42] but you could uncomment things with strtr alone, I think [14:06:45] would be nice to know how much fragmentation there is now [14:06:55] hi brion [14:07:10] mornin dudes [14:07:30] well in terms of averages, most parsing is with the default options [14:07:33] somehow my guess is that language code is the biggest offender [14:07:52] so the case is not really there if you want to save hardware [14:07:59] What would language code affect? [14:08:17] toc, section edit links, things done using {{int:}} [14:08:25] The benefit would be latency and the ability to add new features, not hardware savings. [14:08:28] Hmm, right. [14:08:31] I think you'd have to look at features more hardware, and at latency for users who use those features [14:08:32] in theory things like toc and edit links could be post-replaced [14:08:43] actual use of {{int:}} for localized templates is harder [14:08:43] Yeah, that's what we're talking about. [14:08:48] [this is more common on for instance commons i think] [14:09:05] *more than hardware [14:09:11] do we want to get some stats on parse cache fragmentation? [14:09:18] 03(mod) Automatic category redirects - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=3311 (10Simetrical+wikibugs) [14:09:18] that would give us a nice way to let {{int:}} work but break it in links etc [14:09:28] I've done it before, last time the answer was not much [14:09:47] I think our cache lifetime in memcached has got even shorter since then [14:10:03] so you'd have even more parsing of default options [14:11:19] we should just disable those features [14:11:21] if we could reduce all parse-time options to post-parse replacements, would it be beneficial to have a 'permanent cache', eg keep the rendering made at edit time until something pushes it to expiration such as a template change? [14:11:25] question marks after red links? [14:11:35] domas: that's where this discussion started [14:11:37] domas: that's all CSS :) [14:11:47] doesn't affect parsing since like 2003 [14:11:54] damn! [14:12:17] This all started when someone complained that question mark redlinks don't work in w3m. :) [14:12:18] so, hold on, someone said that we should support people who don't have CSS? [14:12:27] yes, specifically w3m [14:12:34] who? [14:12:40] which is a browser used by a small but crazy minority [14:12:47] saper? [14:12:53] and jidanni [14:13:00] yep [14:13:26] saper; get a browser with CSS support, then [14:13:29] if jidanni used the same software as me i think i'd stop using it [14:13:30] what is the problem? [14:13:42] saper: should we support gopher? [14:13:43] domas: sorry, it's the web, it's not firefox-only world... [14:13:56] saper: I don't use firefox [14:14:00] and CSS works just fine for me [14:14:03] saper: it's the web, it has well known standards. if you want to use software that doesn't support them, that's your problem [14:14:32] I am currently translating some WMF stuff using two w3m instances [14:15:22] saper: so? [14:15:29] saper: should I give you list of browsers? [14:15:42] I do some testing using 'telnet rr.pmtpa.wikimedia.org 80' [14:15:44] domas: that redlink is the only thing that does not work for me [14:16:06] translating? [14:16:26] it doesn't mean that we should start making better support for people who use telnet to browse HTTP web sites [14:16:32] yep, there are translation requests on meta [14:18:07] ok quick background on the link formatting [14:18:26] the question-mark formatting is how our old old old old old software in 2001-02 formatted links initially [14:18:36] yes, before 1.6 I think [14:18:39] the preference exists for people to get the old style back if they want it [14:18:47] 1.6? more like before 1.0 [14:19:29] the question mark may have been a rational choice, you know [14:19:29] there's no functional requirement for that formatting, so its implementation being optional isn't terribly problematic [14:19:34] despite the fact that everyone hated it [14:19:38] hehe [14:19:50] colourblind people can use it, and people on monchrome displays [14:20:54] although the most common types of colourblindness are probably ok with distinguishing red and blue [14:21:07] colourblind people can click the link [14:21:10] to find out if page exists or not [14:21:20] *domas ducks [14:21:26] mmm [14:21:34] Or hover to get the URL. [14:21:38] yup! [14:22:03] Even if you're red-blue colorblind, you might be able to tell them apart by their brightness. [14:22:25] there's a website which filters HTML to simulate various kinds of colourblindness [14:22:43] there's a couple of them where the difference is very subtle [14:25:50] Do those actually work, or did some dude make them up based on something he read in a textbook somewhere? [14:25:52] I am not a big fan for question mark, either (except for the WikiWikiWebStyleLinks? maybe) [14:26:02] I mean, do they actually simulate colorblindness accurately. [14:26:16] I didn't look at it that closely [14:26:22] obviously they can't do it perfectly [14:26:25] I daresay CSS question mark in preferences is plain useless, nobody wants it except those, who really have to. [14:26:54] we've got a few colorblind folks around we can ask for sanity checks :) [14:27:26] 10% of men are supposed to be red-green colorblind, aren't they? [14:27:36] something like that [14:27:39] one of the interesting things I found out last time I was reading this is that some "colourblind" people (in fact the most common type) actually have perfectly working three-colour vision [14:27:41] i think i know more colourblind women though [14:27:52] i.e. R-G colourblindness in men [14:28:14] it's just that one of the cones has a peak sensitivity that is shifted [14:28:22] saper: so why are you using w3m instead of proper browser? :) [14:28:23] and also it's less sensitive overall [14:28:58] so that may mean that for them, a typical monitor can't cover the full colour range of their vision [14:29:31] it's hard to simulate that with a filter [14:30:04] Our whole display technology is remarkably primitive. If you looked at it in a spectroscope it would look like a joke, nothing at all like reality. [14:30:12] domas: because the way of editing wiki is different. It's like telling python programmer using vi to switch to eclipse, because it's so much superior. [14:30:48] how different? [14:31:19] Mm. Sometimes I think our UI is terrible, then I get reassured: http://img.thedailywtf.com/images/200905/errord/Picture%201.png [14:31:21] is it worth sacrificing standards then? [14:31:58] bwaha [14:34:34] Simetrical: yeah, I think this is a lot better: http://spexi.ircvisa.com/stuff/ok.jpg [14:36:08] Simetrical: yes, it is terrible. But its simplicity still keeps us afloat :) [14:37:41] domas: "standards" are not the goal in itself, look at how many browsers support "content" and "after-content" stuff. Not to mention things like CSS3 multi-column layout. Would you push everyone to forget about their -moz- and -webkit- quirks in the name of standards? [14:38:44] We should never be using -moz- and -webkit- stuff without the equivalent standard property being used too. [14:39:44] We try to degrade gracefully if you don't support CSS, but "have to look at link URL to find if the page exists" is graceful enough for such egregious refusal to use modern technology. [14:40:04] Given the costs of putting in a literal question mark. [14:42:25] Man, we need a policy on class spam. [14:43:42] CSS classes? [14:44:27] Yes. [14:44:46] Removing all classes and id's from the Wikipedia main page results in a 9% decrease in HTML size, unless I screwed up my regexes. [14:45:01] I tend to argue that redlink is the single basic feature of the wiki. it should be available. I am not pushing for it, I 'd like to help find a sound technical solution to provide this possibilty, doing some work myself if required. [14:45:15] Do something that doesn't break parser cache. [14:45:21] Then fine. [14:45:25] Otherwise no. [14:45:50] Oh, haha. [14:46:03] My "main page" is actually an error page because I was using GET without specifying a user agent. [14:46:04] Silly me. [14:47:39] Okay, only 8%. [14:47:59] I have to admit I prefer classes over descendant selectors [14:48:23] I had some issues with descendant selectors screwing up my layout in OggHandler and ImageMap [14:48:40] but there is a bandwidth overhead [14:49:34] There's no limit to the number of classes you can add. [14:49:56] Someone just added a class for every row of watchlists containing the page name, apparently. [14:50:03] That's the point at which I have to say you should just be using JS. [14:50:17] well yeah [14:50:27] . . . of course, I added the page-* class with every page's name to each page, but that's just one class per page. [14:51:01] working out what precedent you're setting is always difficult [14:51:31] I wrote wfMsg() and wfMsgReal() [14:51:42] thought that would be enough wfMsg*() functions [14:51:55] 03demon * r50766 10/trunk/phase3/includes/templates/PHP4.php: Partial revert of r49700: per Tim, this should probably stay, since it is likely to still work in a PHP4 environment. [14:52:24] committing but not on IRC? [14:53:54] Aaron does that, it annoys me. [14:54:33] 03demon * r50767 10/trunk/phase3/includes/specials/SpecialExport.php: Code style tweak, use strval() to make this all much shorter. [14:55:02] 03shinjiman * r50768 10/trunk/phase3/includes/zhtable/ (Makefile Makefile.py): changing the Unihan.zip source from ftp to http in order to getting more stable download. [14:55:03] I've just been reviewing his stuff, not finished yet [14:55:12] :o [14:55:50] *Nikerabbit sits in a train [14:56:50] I reset all the status fields in CodeReview from r48811 back to new (except fixme/reverted) so that I could review them all [14:57:02] there wasn't any other way to do it, that I could think of [14:57:11] CodeReview is too nice to just do it the old fashioned way [14:58:26] ok this is getting silly ... since svn update yesterday fcgi segfaults if i use Oracle [14:58:53] I've had fcgi segfault on random things too. [14:58:55] svn update of mediawiki? [14:58:58] PHP is fun! [14:58:58] yes [14:59:05] you wouldn't think it would do that [14:59:09] it IS fun ... if it works [14:59:22] unless it's infinite recursion [14:59:31] there are lots of thing you can do with the oci8 extension that will crash php [14:59:34] it's very badly written [14:59:39] Infinite recursion should throw an extension, not segfault. [14:59:41] you know there was an infinite loop bug in SQLite, not infinite recursion though [14:59:43] Exception. [14:59:48] flyingparchment i TOTALLY agree on that one [14:59:58] (the api sucks as well) [15:00:02] there's a patch you can apply [15:00:11] or you can use xdebug [15:00:43] been doing that for the last 24hrs excluding sleep [15:02:25] is there a linux command to tell you whether a file has DOS line endings? [15:03:09] hmmm ... dunno [15:03:25] I usually just open it in gvim and see what it says, but that's a bit cumbersome if there's more than one file [15:03:42] TimStarling, http://stackoverflow.com/questions/73833/how-do-you-search-for-files-containing-dos-line-endings-crlf-with-grep-under-li [15:05:06] dos2unix < infile | diff -q - infile [15:05:09] I think the short answer there is "grep -l" [15:05:26] would probably work [15:05:51] How do you specify \0xd so grep accepts it? [15:06:02] ctrl-v ctrl-m [15:06:03] sed would proly own the problem [15:06:45] ctrl-v in the shell makes the next character literal [15:07:56] I only knew that worked in vim, didn't know it worked in the shell. Cool. [15:08:15] no wonder demon sets eol-style on files so often [15:08:25] 03(NEW) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 normal; Normal; MediaWiki extensions: Spam Blacklist; (mikelifeguard) [15:08:31] he has an editor which automatically converts unix files to dos, when he saves them [15:08:37] doesn't even warn him [15:08:39] what about 'file *'? [15:08:53] file didn't appear to work [15:09:04] $ file RT_body.php~ [15:09:04] RT_body.php~: PHP script text [15:09:11] doesn't mention line endings [15:09:18] grep -l does work though [15:10:09] git has the best way to handle line endings. Everything is Unix, and if your system is broken and uses DOS, your git client can convert it and not bother anyone else. ;) [15:10:22] that's how many VCSs work apart from svn [15:10:28] s/many/most [15:11:03] svn tried to be clever and '8-bit clean', ignoring the fact that 98% of files stored will be text files, which now require special handling [15:12:19] 03aboostani * r50769 10/trunk/phase3/includes/ (DefaultSettings.php parser/Parser.php): Reverting the code that encloses article sections in their own div for the time being, because there are certain scenarios that break the document structure. [15:12:25] it happily manages conversions, but you have to enable them as non-default. it's danged annoying :) [15:12:52] right... the default should be text files, and binary should require special handling [15:12:57] svn got it backwards [15:13:01] i think i've got some helper scripts around somewhere to do things like check if all your files are set w/ the right line end style [15:13:06] i'll have to dig them up and chekc them in [15:13:22] we could set up a pre-commit hook to verify and bitch at people too ;) [15:13:22] 14(DUP) Customizing the dimension of pictures in categories - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18843 +comment (10innocentkiller) [15:13:27] 03(mod) Size thumbnails in galleries according to user preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=1340 +comment (10innocentkiller) [15:13:59] ok, heading into the office. see y'all [15:15:49] and while we're on the subject, what's with editors that leave off the newline at the end of the file? [15:16:00] that's been a convention in text editors since the beginning [15:16:13] we have all kinds of problem with that on the toolserver [15:16:26] vixie cron ignores line that don't end in a newline, and someone set the default editor to one that forgets it [15:17:00] Files like that are annoying if you cat them. [15:17:04] yes [15:17:10] presumably that's how the convention started [15:17:31] no, i think it was because otherwise you can't tell if the file is truncated in the middle of a line [15:18:10] if so, it's not exactly a bulletproof checksum method [15:19:29] 03yaron * r50770 10/trunk/extensions/ReplaceText/SpecialReplaceText.php: Added namespace selector to page, fixed bad-page-ID bug [15:21:26] hello [15:21:57] hi [15:22:08] how are you [15:22:42] 03yaron * r50771 10/trunk/extensions/ReplaceText/ReplaceTextJob.php: Minor formatting change [15:22:47] 03(mod) Changes to DSN for more width and finer css/js control - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18249 +need-review +patch (10innocentkiller) [15:22:56] were are you from [15:24:02] 03yaron * r50772 10/trunk/extensions/ReplaceText/ReplaceText.php: Added Admin Links hook [15:24:06] hi, I'm from the internet [15:24:13] omg, rly? [15:24:26] 03yaron * r50773 10/trunk/extensions/ReplaceText/README: New version: 0.6 [15:26:33] ok need pointers ... on initial load ... after bagostuff _doquery ... what is setup next ... [15:27:30] look at includes/Setup.php [15:30:00] hi Roan [15:30:07] Hi Tim [15:37:55] <^demon> Oops, one of the tags is broken: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/tag [15:38:16] Hey, that's the empty tag [15:38:24] There was that revision that displayed Tags: , 1.15 [15:38:45] hi ^demon [15:38:55] <^demon> Hi Tim. [15:38:59] <^demon> Roan: I wish unused tags would disappear :) [15:39:16] Yeah, but then this tag isn't unused :P [15:39:23] we were saying just before you came in, it'd be cool if your editor didn't automatically convert files to DOS line endings [15:40:03] <^demon> I'm sending off this laptop for repair today anyway, so you'll be free of my windows machine for the time being. Problem solved. [15:40:35] which editor do you use? [15:40:59] <^demon> For larger stuff, Aptana (based on Eclipse). For quicker edits, a project called Intype. [15:41:20] what is {{#pos:}} and {{#rpos}}? [15:41:32] Shouldn't svn:eol-style=native take care of all the line-ending stuff? [15:41:38] !parserfunctions | Ose [15:41:38] --mwbot-- Ose: "Parser functions" are a way to extend the wiki syntax. ParserFunctions is an extension that provides the basic set of parser functions (you have to install it separately!). For help using parser functions, please see . For details about the extension, see . [15:41:42] some people don't set it [15:42:11] then demon edits their files without noticing, and you get this: http://www.mediawiki.org/wiki/Special:Code/MediaWiki/50035 [15:42:20] I know about parserfunctions, but these two are new for me [15:42:21] <^demon> You have to explicitly set it on new files. I've forgotten on a few recent commits, apologies. [15:42:45] Oh wait, #pos: might be a StringFunction [15:43:08] it's certainly not on http://www.mediawiki.org/wiki/Help:Extension:ParserFunctions [15:43:17] !e StringFunctions | Ose [15:43:17] --mwbot-- Ose: http://www.mediawiki.org/wiki/Extension:StringFunctions [15:43:17] like explode? [15:43:17] ^demon: http://www.mediawiki.org/wiki/Subversion/auto-props [15:43:22] There maybe? [15:43:30] aha [15:43:34] I don't believe TortoiseSVN supports auto-props [15:43:35] thanks [15:43:49] RoanKattouw: it has a screenshot from TortoiseSVN on that page [15:43:52] so I think it probably does [15:44:07] Hmm [15:44:10] That's new for me [15:44:15] <^demon> Yes, it does. I thought I had that set. Might've gotten lost in an upgrade. [15:44:33] anyway, I wasn't complaining about your new files [15:44:39] so {{#pos:b|abc}} will return "2"? [15:44:47] just about the fact that your editor converts all files to DOS when you click save [15:45:03] unlike, say, vim, which leaves the files with the same line ending as they had [15:45:26] Ose, try it and see? [15:45:28] There are worse editors out there, like the one that converts tabs to spaces [15:45:40] I think ultraedit leaves the same line endings as well [15:45:44] At least it's friendly enough to tell you everytime it does it, but it's very annoying [15:45:54] 03(FIXED) Enable AbuseFilter on Cantonese Wikipedia - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18589 +comment (10rhalsell) [15:45:56] <^demon> Roan: Ew. jEdit does, but doesn't tell you. [15:46:08] <^demon> tabs->spaces, that is. [15:47:20] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 +comment (10happy-melon) [15:48:11] ^demon: didn't the block form give warnings before changing block settings if you tried to block someone who was already blocked? [15:48:16] because it isn't now [15:48:21] ^demon: I'm still reviewing your code, do you want comments here on in CodeReview? [15:49:55] ^demon / TimStarling the block conflict detecting part of Special:Block seems to be broken [15:49:59] if one of you could check that out [15:50:18] Prodego: on what wiki? [15:50:21] en [15:50:45] totally broken? or partially? [15:51:04] FlyingToaster, Mentifisto, and I had a bunch of blocks where we conflicted, but didn't get any messages [15:51:30] http://en.wikipedia.org/w/index.php?title=Special:Log&type=block&user=&page=User:63.78.151.178 [15:51:50] TimStarling: What're you reviewing? [15:52:04] r48811 - HEAD [15:52:21] ~1.8k revs [15:52:29] Does that mean you guys are finally gonna scap? :D [15:52:40] yes [15:52:44] Awesome [15:52:46] just me though, brion's off getting married [15:52:57] really? [15:52:59] Yeah, heard he was gonna, is that today? [15:53:07] thursday [15:53:28] Right [15:53:31] he's at work today [15:54:11] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 +comment (10mikelifeguard) [15:55:14] ok 'nother noob question ... how do i see all revisions where a certain file was changed and diff if possible [15:55:27] svn log? [15:55:44] svn blame I think [15:55:48] svn log is slow trans-pacific [15:55:51] you should use viewvc [15:56:04] well, unless you're in the US I guess [15:56:04] Or git-svn. [15:56:05] freakolowsky: http://svn.wikimedia.org/viewvc/mediawiki/trunk/phase3/includes/Article.php?view=log [15:56:09] Or try view=annotate [15:56:20] (for a per-line who-changed-this-last-and-when) [15:56:24] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 +comment (10Simetrical+wikibugs) [15:56:39] yes, or git-svn or svk or whatever [16:08:53] 03(mod) Strange spacing before [[irc:...]] links - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18207 +comment (10tstarling) [16:09:07] 03demon * r50774 10/trunk/phase3/includes/specials/SpecialBlockip.php: Followup to 50635: Consolidate all of these hacky global checks into a clean method. Also move a global and document it. [16:10:49] 03demon * r50775 10/trunk/phase3/includes/specials/SpecialBlockip.php: Remove the globals. [16:15:52] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 (10mikelifeguard) [16:19:50] 03(mod) Show estimate of number of edits to be reverted in rollback link - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18834 (10jan) [16:24:11] 03demon * r50776 10/trunk/phase3/includes/ (AuthPlugin.php Preferences.php): [16:24:11] Per code review: Redo these allowBlah() functions (except allowPasswordChange) [16:24:11] into a general allowPropChange(), update Preferences to use this. Checked [16:24:11] extensions in SVN using AuthPlugin and none have implemented any of the allow*() [16:24:11] functions, so nothing should break. [16:34:02] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 (10Simetrical+wikibugs) [16:34:33] YAAAAAAAAAAAAAAAAAAAAY ... monstah ... found the problem ... [16:35:03] 03demon * r50777 10/trunk/phase3/includes/ (diff/DifferenceEngine.php filerepo/FSRepo.php): Cleanup to r50310 & 50313: Don't use @ on chmod() and dl(), use wfSuppressWarnings()/wfRestoreWarnings() [16:35:11] 03(NEW) Allow Scribus (.sla / .sla.gz) file uploads on Wikimedia - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18845 enhancement; Normal; Wikimedia: Site requests; (brion) [16:35:37] 03(mod) PostgreSQL wrong protocol version - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=16937 (10etienne.massip) [16:38:34] 03shinjiman * r50778 10/trunk/phase3/languages/ (Language.php messages/MessagesZh_tw.php): added Minguo calendar support for the Taiwan Chinese language [16:38:50] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 +comment (10mikelifeguard) [16:39:37] i have pretome(torrent tracker) invites to giveaway [16:39:39] 03(mod) Only ask gender by default if it will be in fact used for the current language - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18841 +comment (10brion) [16:41:07] @NE1 ... how long after commit are changes visible on the test server? ... tested it localy, but i have little data in the DB ... so just to be on the safe side [16:41:31] "Note also that the optional gender setting could be useful for non-grammatical [16:41:31] features such as selection of default avatar images, should we end up with such [16:41:31] a thing. [16:41:31] " [16:41:37] avatars.... [16:41:40] freakolowsky, you mean test.wikipedia.org? Only at scap. [16:41:42] seriously? [16:41:48] There hasn't been a scap for like two months. [16:42:02] It goes live on test very shortly before it goes live on the whole site. [16:42:15] 03(mod) Add setting to hide a list of preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18761 +comment (10brion) [16:42:23] oh ... hmmm ... any other public sites that use svn build [16:42:34] BetaWiki is more cutting-edge. [16:43:53] betawiki? [16:44:34] freakolowsky: translatewiki.net formerly known as betawiki :) [16:44:53] :D .... just found it in the exsact moment [16:45:05] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 +comment (10brion) [16:46:29] heh [16:46:40] well, I'm searching if the actual numbers are in my talk page archives :D [16:47:06] Nike is like ... Nikerabbit ?? [16:47:20] yes [16:48:25] oh from Finland ... and a student ... if he's anything like the Finish ppl i know he's pro'ly dead drunk by now :D [16:53:37] i'm writing a work and include some facts about MediaWiki. Is it somewhere stated that features are only added to core code when the feature is essential for wikipedia or useful otherwise and stable and working in all browsers? I wrote it down but can't find a page to give as reference. [16:54:59] hello brion [16:55:40] Subfader, no, that's not true. [16:55:55] Features are added to the code when some developer wants to add them and they pass review. [16:56:06] There's no central planning of features. [16:56:12] 03yaron * r50779 10/trunk/extensions/ReplaceText/SpecialReplaceText.php: Added sorting of result, better formatting of strings in user messages (both from Capmo) [16:56:31] Wikipedia's interest overrides other interests if there's a conflict, but it's rare that there's a real conflict, since you can just tack on some more config settings. [16:56:41] There are plenty of completely broken features that are disabled by default. [16:57:11] And plenty of features that don't work in all browsers, e.g., external link icons not working right in IE6. [16:57:22] So no, no hard and fast rules here. [16:57:24] ok right. i use that statement while I try to explain why Commons misses some basic features which are available as extension though [16:57:56] As much as MW is written for wikipedia and such, it isn't written only for it [16:58:02] like multicategory search and multi-upload, while that last is not stable as extension [16:58:19] There are minimum standards for inclusion, yes. [16:58:24] There's also the problem of reviewing things. [16:58:34] but if it's essentially needed for WP some dev(s) will try to add it to core [16:58:43] It looks like multiupload is going to be solved in the nearish future, if the new-upload branch doesn't die (isn't that one thing it does?). [16:58:52] Most devs come from Wikipedia, and are more interested in it than in Commons. [16:58:57] Because Wikipedia is bigger. [16:59:07] If more Commonists became developers, there'd be more features for Commons. [16:59:08] Easy. [16:59:11] 03freakolowsky * r50780 10/trunk/phase3/includes/ (3 files in 2 dirs): [16:59:11] Preparations for Oracle database abstraction update. Take II. [16:59:11] Replaced hardcoded LIMIT clauses with database function (except in maintenance). [16:59:11] Created unionQueries in Database and overloaded it in DatabaseOracle (not commited yet). [16:59:11] Replaced all UNION clauses with function calls (except in maintenance). [16:59:36] well both features i mentioned are essential for lots of WM projects. but tthat's not the point :) [16:59:51] 03(NEW) Password failure after upgrade to MW 1.13 or 1.14 - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18846 major; Normal; MediaWiki: Installation; (tstarling) [17:00:00] oukiedoukie ... commit done ... now for some relaxation ... dunno why but the whole day i have the feeling it's friday :D ?!??! [17:00:07] 03(mod) Release MediaWiki 1.15 (tracking) - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18629 (10tstarling) [17:00:35] The point is that developers mostly write whatever features they feel like. [17:00:45] ok so i will remove my "requirements for core feature additions". [17:00:52] There are requirements for core feature additions. [17:01:07] Simetrical actually most devs write whatever features their bosses feel like :D [17:01:12] But one of them is sufficient attention from someone with commit access, to correct any deficiencies. [17:01:17] freakolowsky, MediaWiki devs, I'm talking about. [17:01:29] Most MediaWiki devs are volunteering, or otherwise get to mostly decide what they want to do. [17:01:55] erm ... yes ... my JOB is to develop MW (Oracle support and some custom exts) [17:02:06] can you point me to a page where it describes the requiremtents for core additions? [17:02:29] Subfader, there is none. [17:02:31] It's all ad hoc. [17:02:41] People review the code and point out any problems they have with it. [17:02:50] If all the problems get fixed, then it might get committed. Otherwise, it probably won't. [17:02:50] can't quote Simetrical (MW ev) on #mediawiki chan via IRC :D [17:02:59] Sure you can. [17:03:08] 03shinjiman * r50781 10/trunk/phase3/RELEASE-NOTES: follow up r50778, updating the release notes. [17:03:14] "Personal correspondence with Aryeh Gregor (MediaWiki developer since 2006)." [17:03:41] brion: avatars, please never impliment those :) [17:03:41] alright that sounds more scientific ;) [17:03:52] you may commit code to the core, providing you pledge fealty to Brion [17:03:59] could link to the chan archive though [17:04:41] Yeah, a link would work. [17:05:22] mwahaha [17:05:34] actually avatars would be great for liquidthreads ;) [17:06:45] but is liquidthreads good? [17:06:54] how many active devs are there? 40? just a guess [17:07:03] to be politically correct they should probably all be gender-neutral objects [17:07:58] like rocks and trees and stuff [17:07:59] 03demon * r50782 10/trunk/phase3/ (4 files in 2 dirs): (bug 18761) Add setting to hide a list of preferences. [17:08:02] Subfader: Maybe, but I'm pretty sure it's not more than that [17:08:35] alright, thanks :) [17:08:44] There are only ~150 people with commit access and most of those are inactive or hardly active [17:08:52] Subfader, I doubt there are 40. [17:09:10] Unless you include all the people who work on one specific extension and whatnot. [17:09:22] https://www.ohloh.net/p/mediawiki/contributors [17:09:23] General-purpose active core devs is more like 10 or 15, I'd think. [17:09:51] 03(FIXED) Add setting to hide a list of preferences - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18761 +comment (10innocentkiller) [17:10:13] oh, that's a nice source [17:10:30] Note, though, that the total number of commits might be totally irrelevant [17:11:04] Yeah. The highest commit number is probably siebrand committing localizations, which isn't technically development (not that it's not important). [17:11:10] <^demon> It would be neat to find out the most active per project in code-reivew. Like most active on an extension or core (could do by path?) [17:11:13] For instance, robchurch hasn't been active for years, but still ranks in the top 10 of commits in all time [17:11:21] And yes, Siebrand ranks like 3rd [17:13:56] yeah i'll add it like "about 30 active developers (out of 193 people with commit access)" [17:13:58] by commit count? of course aaron is #2 by commit count [17:14:11] he hits svn ci at the end of each line [17:14:22] 03shinjiman * r50783 10/trunk/extensions/cldr/ (Makefile README): follow up r50758, bump the cldr version to 1.7.0 [17:14:50] lol [17:15:11] you should look at svn annotate counts [17:16:39] one of the most annoying things is boss showing up with an urgent patch right in the moment you are about to pack up and go home :S [17:20:34] 03freakolowsky * r50784 10/trunk/phase3/RELEASE-NOTES: Forgot RELEASE-NOTES [17:21:13] <^demon> freakolowsky: Release-notes go at bottom of relevent section. [17:21:48] oh ... [17:24:01] 03freakolowsky * r50785 10/trunk/phase3/RELEASE-NOTES: order fixed [17:24:12] <^demon> :) [17:25:27] Actually, I'm not sure that belongs in RELEASE-NOTES at all [17:25:37] +* Database: unionQueries function to be used for UNION sql construction, so [17:25:39] + it can be overloaded on DB abstraction level for DB specific functionality [17:25:44] That has zero relevance to end users IMO [17:26:21] well then where should i document that kind of stuff? [17:26:32] Don't? :) [17:26:51] There's also the commit message [17:26:52] Internal stuff that only we care about doesn't have to be documented anywhere, it's documented in the SVN log. [17:27:13] funny ... and then i should go union hunting every once in a while because noone will know they should use a function instead ... like limitResults [17:27:41] Well people shouldn't be using raw SQL anyway [17:27:51] tell that to the ppl [17:27:55] New code should use Database::select() [17:28:14] I know there's a lot of raw SQL in specials/Special*.php , but I've fixed a lot of that in my branch already [17:30:10] but until all that code in trunk does like the wind does ... [17:30:20] brb ... need some nicotine enhanced fresh air [17:31:07] select() doesn't support everything. [17:31:15] Like, for instance, union. :) [17:31:49] Yeah; freakolowsky introduced a Database function for unions now, so people should use that instead [17:32:18] And if Brion/Tim/whoever else reviews code keeps that in mind, any usage of raw UNIONs will be caught and killed immediately [17:35:02] You mean like all the raw LIMITs have been killed? [17:35:08] It needs everyone to be diligent. [17:35:21] Simetrical except the ones in maintenance ... yes [17:35:24] i think [17:35:32] :D [17:35:32] Well, yes, because you just did them now. [17:35:41] Not because of our top-notch code review process. [17:35:42] Well if Oracle support takes off, people will notice raw LIMITs and stuff like that because they break on Oracle [17:35:47] is there a way to have mediawiki run shell scripts to grab information? [17:35:48] Yes, hopefully. [17:35:53] !api | kdub [17:35:53] --mwbot-- kdub: The MediaWiki API provides direct, high-level access to the data contained in the MediaWiki databases. Client programs should be able to use the API to login, get data, and post changes. Find out more at . For client libraries in various languages, see . [17:35:56] kdub, you mean like by writing an extension? [17:36:02] What are you trying to do? [17:36:23] basically, i want to have a text file on the server that the wiki can read and update the site accordingly [17:36:50] Then you probably want to write an extension [17:37:10] In that extension you'd use the PHP function system() to execute command-line stuff [17:37:14] That's not what you want to do, that's a whole class of problems that includes what you want to do. [17:37:20] What *specifically* do you want to do? [17:37:26] What's the purpose of this file, what type of update to the site? [17:38:28] anyways ... been @ work >13hrs ... starving ... cd /home/freak; cat fidge > /dev/mouth [17:38:31] I'm not going to test Oracle, MSSQL or DB2 pre-release [17:38:36] Why does everyone use $foo[] = 'bar' instead of $foo []= 'bar'? [17:38:40] if they work, they work, if not, file a bug [17:38:51] []= is logically a type of assignment operator. [17:39:12] Simetrical: I'm guessing that's how it appears in the PHP manual [17:39:24] Hmm, probably. [17:39:31] Also, $a+ = 1; doesn't work. [17:39:34] Simetrical: i'm looking for a way to auto-update the status of servers or accounts for an internal wiki [17:39:36] It only works for []=. [17:39:38] Simetrical: No, $foo[] is logically an array index meaning, $foo[n] where n is the first free index in $foo [17:39:40] Tim ... no need ... u just test MySQL so i don't break anything ... i have enough Ora testers busting my chops here :D [17:39:54] RoanKattouw, hmm. Interesting interpretation. [17:40:00] That's like C initialization, then. [17:40:07] int a[] = { 0, 1, 2 }; [17:40:14] That's probably how it originated, indeed. [17:40:27] That's my useless fact learned for the day, then. [17:41:24] TimStarling : PS: http://www.abakus.si/jk/jksvn/ ... my Oracle dev-test box ... and it's public [17:41:25] The fact that []= , [] = and += work but + = doesn't is probably due to the fact that you can't split a multi-char operator. += is an operator, but []= isn't (otoh [] and = are) [17:41:45] Apparently. [17:41:51] [] isn't an operator, though. [17:41:57] $a[]; is a syntax error. [17:42:04] Right? [17:42:19] *RoanKattouw tests [17:42:29] Fatal error: Cannot use [] for reading [17:42:33] <^demon> Yeah. [17:42:37] Makes sense [17:42:58] If it refers to the first nonexistent index, after all . . . [17:43:05] PHP's view of operators is a bit different to that in most languages [17:43:14] Interesting, though, that PHP has expressions that are lvalues but not rvalues [17:43:20] While in C all lvalues are also rvalues [17:43:22] a lot more is done in the parser than in C [17:44:31] for instance you can write $a[0] = 1; but not ($a)[0] = 1; [17:44:51] who has called me? [17:45:24] PHP loves special cases. [17:45:33] Did you know you can't use func_get_args() as a function argument? [17:45:35] because the PHP parser has a production rule like: var "[" expr "]" = expr [17:46:02] so the whole thing is an array index assignment expression [17:46:17] rather than indexing operators generating an lvalue which is assigned [17:46:17] Ah, so the bug that f()[0] doesn't work is also do to that rule (or a similar one)? [17:46:35] similar issue [17:46:49] no I am not an alcoholist! [17:47:03] Hmm, the func_get_args() thing seems to work now. [17:47:17] It'd make sense to forbid f()[0] from being used as an lvalue, though, but as an rvalue it should work [17:47:27] Yes [17:47:27] . [17:47:34] I've run into that, and it's pretty stupid. [17:48:04] my stock answer to that one is that it took them a whole major version to implement foo()->bar() [17:48:18] so maybe we can expect foo()[0] by PHP 7.0 [17:48:37] Why can't they just do things more sanely, like C has done for 30+ years? [17:48:50] Which language? [17:48:54] <^demon> PHP [17:49:36] because they just weren't as smart as people like Dennis Ritchie [17:49:59] Some stupid people seem to have managed to successfully copy smart people. [17:50:06] I guess it's potluck, though. [17:50:15] The real question is, why did they actually succeed if their language is such trash? [17:50:42] safe_mode/open_basedir [17:50:54] a feature which they hate despite all the glory it's brought them [17:51:23] That's what I gloomily suspected. [17:51:31] Why can't the hosts just use Unix permissions? [17:51:36] 1and1.com does. [17:51:45] You have to use CGI, that's all. [17:52:11] dunno [17:52:24] also it might be said that PHP, unlike perl, has a decent apache module [17:52:36] :O you mean once upon a time stuff like $parser->getTitle()->getPrefixedText() didn't work? Code back then must've been a mess [17:52:54] and IIRC PHP spent some time being incubated as an apache project [17:53:02] so they probably benefitted from apache's success [17:53:48] RoanKattouw: yes it didn't work, in PHP 4, you should read MW 1.6.x some time [17:54:23] I intend to avoid that as long as possible :) [17:54:40] :o [17:54:49] we didn't do $t = $parser->getTitle(); t->getPrefixedText(); though [17:54:50] <^demon> Roan: backport the API to 1.6.x :) [17:54:54] $ echo 'a();' | php [17:55:04] Okay, so why is that an instant segfault? [17:55:08] $parser->mTitle->getPrefixedText() was allowed so we mostly used that [17:55:10] Is there some infinite recursion I'm not spotting? [17:55:15] Simetrical: A::A() is the constructor [17:55:15] helped that PHP 4 didn't have visibility [17:55:20] Oops. [17:55:36] heh [17:55:37] I need to be more creative with my test method names. [17:55:44] lol [17:58:18] good night [17:58:25] Night [18:02:06] :o [18:02:51] Simetrical: I've run into that too few times [18:02:58] also that I cannot name function new() [18:47:58] 03thedevilonline * r50786 10/trunk/extensions/LocalisationUpdate/ (5 files): * Initial version of an extension to timely update localisation messages in Mediawiki, using the new 'MessageNotInMwNs'-hook [18:53:54] 03thedevilonline * r50787 10/trunk/extensions/LocalisationUpdate/KNOWN_ISSUS.txt: - Added 'Known Issues' for the LocalisationUpdate extension [19:06:08] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 (10vasilvv) [19:06:25] 03(mod) Automated smallest-CIDR calculation from within Checkuser UI - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17018 (10JSchulz_4587) [19:09:40] 03(mod) Page history lines are rendered differently to admins when contents have been concealed by RevisionDelete - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17060 (10JSchulz_4587) [19:19:54] 03aaron * r50788 10/trunk/phase3/includes/ (LogEventsList.php PageHistory.php): (bug 17060) Renders edit comments the same for all users wrt to revdelete [19:19:59] 03(FIXED) Page history lines are rendered differently to admins when contents have been concealed by RevisionDelete - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=17060 +comment (10JSchulz_4587) [19:29:14] ssh: connect to host svn.wikimedia.org port 22: Connection timed out [19:29:28] the timeout is absurdly slow.. can I fix it? [19:29:32] short* [19:30:38] Hello, we use mediawiki to manage technical content with a lot of abbreviations und we have a lot of redirects. The search result has a tick box at the bottom called "List redirects". How can I make that box ticked by default? [19:31:30] mw 1.14.0 we use [19:39:10] 03thedevilonline * r50789 10/trunk/extensions/LocalisationUpdate/ (3 files): [19:39:10] - Added initial localisation for the LocalisationUpdate extensions [19:39:10] - Removed excessive comments in install.sql [19:39:10] - Added known issues to the header of LocalisationUpdate.php as well [19:40:21] why is uselang affecting search results?!? [19:43:34] Nikerabbit: Example? [19:44:10] Nikerabbit: FWIW SSH status on srv201 is OK [19:44:36] 03nikerabbit * r50790 10/trunk/phase3/ (2 files in 2 dirs): * Fix plural issues [19:44:37] RoanKattouw: user error [19:44:41] Anyone know any plugins for WikiML in Textmate? [19:44:55] I want to include the number of images in a category, within a template... is there an easy way to do it? (e.g. "View Images (x)" where x is the number of images in a given category) [19:47:20] can one pass parameters via the url? http://wiki.com/Template:blah?param1=whatever (i thought this was possible but it doesn't seem to work) [19:47:28] Morder: No [19:47:37] well pfft :( [19:47:41] hehe, thanks [19:47:56] Morder: no, that's not how templates work [19:47:59] {{blah|param1=whatever}} works in wikitext, but you probably already knew that [19:48:04] yes [19:48:18] Morder: if you want that, use Special:ExpandTemplates [19:48:34] well... not quite what you want, probably [19:49:44] not quite - i just wanted to pass a parameter from one page to another using a link [19:50:18] no big deal - was hoping for something i didn't know :) [19:57:24] I will ask once more :-)Hello, we use mediawiki to manage technical content with a lot of abbreviations und we have a lot of redirects. The search result has a tick box at the bottom called "List redirects". How can I make that box ticked by default? [20:00:09] rakekniven: I don't see that in the new search ui [20:00:39] Are there any screenshots to view? [20:00:55] yes [20:01:09] could you point me to the right place? [20:01:58] example: http://translatewiki.net/sandwiki/index.php?title=Special%3ASearch&search=FreeCol&fulltext=Etsi [20:03:13] ooh, that's bad [20:03:29] =? [20:04:10] bad? [20:04:21] wait a second please :-) [20:05:02] Simetrical: Commoners, not Commonists. :P [20:05:36] Nikerabbit: Maybe I have to ask for another topic then. We have a page called "PPC5012". If I search for "5012" the page is not listed. So we have a redirect called "PPC 5012". Please note the space in the middle. This redirect is found when I search for "5012". How to get better search results? [20:06:37] 03(NEW) Implement gendered Chinese pronouns, for that personal touch - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18847 enhancement; Normal; MediaWiki: Internationalization; (jidanni) [20:06:40] 03(mod) Password failure after upgrade to MW 1.13 or 1.14 - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18846 (10brion) [20:06:43] 03(mod) Only ask gender by default if it will be in fact used for the current language - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18841 (10jidanni) [20:07:22] A new jidanni bug. :-) [20:07:55] rakekniven: do you use lucene? [20:08:10] No, built in search [20:08:13] Marybelle: is that trolling or not?? [20:09:07] 03(mod) Only ask gender by default if it will be in fact used for the current language - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18841 (10jidanni) [20:09:19] Maybe? [20:09:21] jidanni isn't trolling. I probably am. [20:11:37] Nikerabbit: Any idea? Help is really appreciated. [20:13:04] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 (10brion) [20:18:43] 03siebrand * r50791 10/trunk/phase3/languages/messages/ (48 files): Localisation updates for core messages from translatewiki.net (2009-05-19 19:49 UTC) [20:26:07] 03siebrand * r50792 10/trunk/extensions/ (136 files in 122 dirs): Localisation updates for extension messages from translatewiki.net (2009-05-19 19:49 UTC) [20:52:44] 03thedevilonline * r50793 10/trunk/extensions/LocalisationUpdate/LocalisationUpdate.i18n.php: - Fixed small typo bug [20:53:58] 03thedevilonline * r50794 10/trunk/extensions/LocalisationUpdate/LocalisationUpdate.i18n.php: And another bug fixed [20:55:04] 03greg * r50795 10/trunk/phase3/maintenance/postgres/compare_schemas.pl: Handle multi-col primary keys, cleanup new tables. [20:55:42] 03greg * r50796 10/trunk/phase3/maintenance/ (postgres/tables.sql tables.sql): Clean up whitespace, add log_search to Postgres schema. [21:24:33] 03(mod) Can't link to [[{{FULLPAGENAME}}]] any more - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18828 (10jidanni) [21:33:52] 14(INVALID) Implement gendered Chinese pronouns, for that personal touch - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18847 +comment (10siebrand) [21:34:14] *siebrand burps softly. [21:37:16] How many languages is MediaWiki translated into? [21:37:43] 'bout 12. [21:37:56] .-. [21:37:57] No, way more [21:38:11] *charitwo fires Marybelle [21:38:15] :O [21:38:22] there are message files for 338 languages [21:38:45] siebrand: Do you know? [21:39:03] or check the 'my preference' page on any wiki [21:39:10] and see how many languages you can pick from [21:39:49] http://translatewiki.net/wiki/Translating:Languages [21:42:06] hi! math rendering gives Failed to parse (Missing texvc executable), but I'm able to execute texvc. What might be wrong with it? [21:43:08] hey mmarc__, did you follow the steps on http://www.mediawiki.org/wiki/Texvc [21:43:08] ? [21:43:18] strange, when I execute texvc it gives "-", is it OK? [21:43:54] you could check http://www.mediawiki.org/wiki/Manual:Enable_TeX/problems#Error_:_Failed_to_parse_.28Missing_texvc_executable.29 [21:44:26] that last page contains the answer to most of your quesions [21:46:01] thanks, yes, i looked there [21:46:16] chuck: http://translatewiki.net/wiki/Translating:Group_statistics [21:46:24] but... not enough [21:49:18] please, tell: what texvc executable should show on run? Any options maybe, a --help option? [21:49:34] I have just nothing, that's strange! [21:51:19] and how do I inspect the $wgTexvc value? [21:55:06] Well, I asked about 12-14 hours ago and didn't really get a decent response other than a pointer to the manual page, so I guess it's aged enough for me to ask again. :) [21:56:43] I'm utilizing OpenBSD, PHP5, and MySQL under apache2. Took me a little while to get apache+php working on this system because I've been sysadmining for quite some time on a system that utilizes the separate configuration file version of apache instead of the monolith and a few other subtle differences [21:57:49] anyway, php and everything is working now, and I've got all of the php5-extensions and mysql packages installed that are needed. however I'm still getting the installation page message about "Could not find a suitable database driver!" and the notes to compile php using --with-mysql (my php was a package precompiled and has mysql support and the extensions are installed) [21:58:04] mmarc__: php maintenance/eval.php [21:58:11] any ideas or pointers to FAQs that might cover this? I haven't been able to uncover ANYTHING helpful for my situation via googling so far [21:58:12] Then, you can do this: var_dump($wgTexvc); [22:00:58] chuck: PHP Notice: Undefined variable: DIR in /var/www/html/wiki/maintenance/commandLine.inc on line 185 [22:01:10] 03purodha * r50797 10/trunk/extensions/GetFamily/GetFamily.i18n.php: [22:01:10] Consistent spelling, following a hint of EugeneZelenko at [22:01:10] http://translatewiki.net/w/i.php?title=Support&oldid=1225614#pywikipediabot_in_GetFamily_messages [22:02:07] dgetsman: I'd suggest making a simple test case and if that doesn't work, ask #OpenBSD [22:03:35] dgetsman, yes, try a testcase like: See if that produces any errors anywhere. [22:04:05] You might want this instead: Say what the output of that is. [22:04:55] Chinese has a different pronoun for gods? [22:05:05] That's . . . interesting. [22:06:46] doing a testcase [22:07:01] I need a regex which matches with {{Rfctag with an optional space between the two brackets and Rfctag and an optional | 1= with an optional space between | and 1= [22:07:26] and an optional space between the g and the | [22:07:49] please :-) [22:08:08] guys, just a simple question: what your math/texvc outputs when you just run it? Could you please check it (not too hard, if you all have mediawiki on your hand). Mine gives "-", even after i remaked it. [22:08:34] \{ ?\{ ?Rfctag ?\| ?1= [22:08:39] oh wow. that certainly does look like it's a bsd package error that doesn't have the correct links; my error was undefined function mysql_connect(); I'm heading over to #openbsd. thanks guys [22:08:44] That's PCRE, POSIX will have some backslashes different. [22:09:12] mmarc__, mine too. [22:09:48] Simetrical: Don't I need to put that between /'s? [22:09:58] Simetrical: thank you, i can hope it's ok now. Hm, what's wrong then... [22:10:04] harej, yeah, sure. [22:10:26] Simetrical: and is that case insensitive? [22:10:40] harej, it is if you use the i flag after the ending /. [22:10:45] which i do [22:10:47] thank you very much [22:11:08] 03purodha * r50798 10/trunk/extensions/Configure/Configure.settings.i18n.php: [22:11:08] Consistent spelling of "httpOnly", following hint at [22:11:08] http://translatewiki.net/w/i.php?title=Support&oldid=1225658#Group_Configure-settings [22:15:59] 03(mod) Implement gendered Chinese pronouns, for that personal touch - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18847 (10Simetrical+wikibugs) [22:20:12] what if I've got another texvc in /usr/bin? [22:21:34] I guess, if mediawiki picks texvc by path, for example /math/texvc, it should be that one, not form /usr/bin [22:22:27] Simetrical: the furries might like the 'animal' option, and anyone with the 'founder' user group could be referred to as 'god' [22:22:42] lol, furries [22:23:07] just looking for some potential problems, since it's absolutely unclear, why texvc is not found, and given steps do not help, because I don't know PHP. I remember math worked long ago, there are lots of compiled formulas in our wiki. [22:26:19] 03(mod) Free up "Project" as a namespace (for Wikipedia:Projects) by using an alternate token - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18793 (10stvrtg) [22:27:58] 03(mod) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 (10Simetrical+wikibugs) [22:28:58] OK, very nice [22:30:00] *Splarka notices Simetrical saper TimStarling and Nikerabbit talking about "?" redlinks and postprocessors... [22:30:16] got texvc working after defining $wgTexvc='/usr/bin/texvc/'; instead of ./math/texvc [22:30:16] what do you think? [22:30:37] so the problem is that non-CSS browsers have no way of styling redlinks, which is really the only time a "?" is handy, for non CSS browsers [22:30:51] the obvious solution is to /always/ put in the question mark, and then remove it for CSS browsers [22:31:14] ~~~? [22:31:55] no muss, no fuss, no cache, no bash, no fail, no wail, no cry, no sigh [22:32:13] Ugh. [22:32:26] Also massive extra bytes for presentation. [22:32:31] Screw text browser users. ;) [22:33:05] unf unf [22:33:14] well, it could be minimized [22:33:45] ~~~? and then a.new i {display:none;} [22:34:04] *Splarka didn't think Simetrical was a byte biter though [22:39:35] 03simetrical * r50799 10/trunk/extensions/LocalisationUpdate/ (KNOWN_ISSUES.txt KNOWN_ISSUS.txt): Spelling [22:52:40] 03(NEW) Incorrect link to Yahoo Search - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18848 enhancement; Normal; Wikimedia: General/Unknown; (jtatsu) [22:53:09] 03(mod) Implement gendered Chinese pronouns, for that personal touch - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18847 (10jidanni) [22:58:03] 04(REOPENED) Implement gendered Chinese pronouns, for that personal touch - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18847 +comment (10brion) [23:03:49] 03dale * r50800 10/branches/new-upload/phase3/ (13 files in 8 dirs): [23:03:49] * refactoring of firefogg code (still under way) [23:03:49] ** working simple Make_Ogg.html stand alone example [23:03:49] ** improved passthough mode support [23:03:49] ** cleaned up mwUploadPage.js implementation of firefogg [23:04:17] bah, brion just wants to be able to select 'god' in his chinese prefs [23:05:40] hehe [23:08:06] 03(mod) Incorrect link to Yahoo Search - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18848 (10brion) [23:18:59] 03(FIXED) Incorrect link to Yahoo Search - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18848 +comment (10brion) [23:19:35] 03(mod) Incorrect link to Yahoo Search - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18848 (10brion) [23:19:54] Splarka: we could do that for English too [23:19:58] requires capitalizing the pronouns [23:20:23] "Welcome to Wikipedia, Brion! Don't forget to update Your preferences." [23:21:01] :-P [23:22:18] xD [23:22:47] brion: We could also whip up some CSS magic to do small caps text like some bibles do for "LORD" and "HE" [23:24:39] 03(WONTFIX) Make all grouping non-capturing to save on processing - 10https://bugzilla.wikimedia.org/show_bug.cgi?id=18844 +comment (10brion) [23:24:55] And you shall know that my name is Brion VIBBER, when I wreak my vengeance upon thee! [23:28:05] 03siebrand * r50801 10/trunk/phase3/includes/OutputPage.php: Fix typo in comment [23:29:32] 03siebrand * r50802 10/trunk/extensions/Vote/ (Vote.i18n.php Vote.page.php Vote.php): Merge 'vote-login-link' into 'vote-login' for more transparant localisation [23:33:15] Surely "thou shalt". [23:36:54] eh whatever [23:36:59] grammatical agreement is for mortals [23:53:18] 03dale * r50803 10/branches/new-upload/phase3/js2/ (7 files in 4 dirs): update fixes