[00:02:53] Dantman: https://www.mediawiki.org/wiki/Requests_for_comment/MediaWiki.org_Main_Page_tweaks/Sandbox/v10 - that better in the pop pages area? [00:02:53] sure [00:02:53] heatherw: any reaction to https://www.mediawiki.org/wiki/Requests_for_comment/MediaWiki.org_Main_Page_tweaks/Sandbox/v10 [00:02:53] I MWified the colors a bit [00:03:18] 03krinkle * 10/trunk/phase3/ (6 files in 3 dirs): (log message trimmed) [00:03:19] [mediawik.special.changeslist] Remove ugly arrow swapping hack [00:03:19] * Use CSS and change the background image, instead of displaying 2 tags and hiding one based on the parent element's class [00:03:19] * touches bug 17616 a bit, although it doesn't address the focus issue yet, but does improve the overal structure of this module [00:03:20] * Also speeds up by: [00:03:21] - Outputting less markup [00:03:22] - Having 1 embedded images in the CSS file instead of (N changes)* tags [00:03:49] AlisonW: Ryan_Lane [00:04:09] Ta Dantman ;-P [00:05:04] eh? [00:05:44] Hi Ryan ... Seems that according to new stuff I'm in the market for a gerrit signin ... can u help? [00:06:09] I can, though right now isn't the best time [00:06:26] okey ... will catch you another time ... [00:06:27] ^demon: ping [00:06:37] it doesn't have to be me [00:06:45] anyone that can give out svn accounts can give out gerrit accounts [00:06:49] Dantman: was, er, kind enough to point at you ;-P [00:06:54] * Ryan_Lane nods [00:07:18] <^demon> varnent: Pong [00:07:27] varnent, what's the though on my download button http://i.imgur.com/YeBB2.png [00:07:29] * ^demon is mildly distracted, the Oscars are starting. [00:07:38] varnent: same color comment :) maybe stick with cool or warm and not a mix of both. i still like the one i supported, which would be stronger with some text tweaks. [00:08:27] heatherw: any color examples you can send me? like a bag with a mix you'd suggest? :) [00:08:43] Dantman: I think that's an improved - see v10 - moved it around a bit.. [00:09:03] ^demon: what are your thoughts on wikicode based extensions? leave them on wiki or house them in a repo? [00:09:06] i could, but i don't have time right now, i'm sorry. if you can wait i will. [00:09:23] ^demon: let me word that better :) extensions only displayed via wikicode and not in repo [00:09:25] varnent: No, I'm talking about the Web 0.9 style outset border [00:09:28] <^demon> varnent: I think they should go in a repo or be deleted. [00:09:31] heatherw: no rush :) [00:09:38] <^demon> On-wiki code is evil, and I've been saying that for about 2 years. [00:09:41] varnent: cool :) [00:09:48] Dantman: yeah - I like that fade effect [00:10:03] 03krinkle * 10/trunk/extensions/CodeReview/backend/DiffHighlighter.php: (log message trimmed) [00:10:03] [CodeReview/DiffHighlighter] Restore inline wrapper for diff text [00:10:03] * Restore inline wrapper for diff text to allow styling of the diff text with the correct inline content width (style for td.ins/td.del styles the entire line rather than just the text) [00:10:03] * Use Html::element() consistently in both the top and bottom of formatLine() [00:10:04] for the line-numbers. No need to allow raw html in td.linenumbers, and if it [00:10:05] does it should be documented and also allowed in other parts of the code. [00:10:06] * main change: [00:10:43] ^demon: okay - so here's a thought..I think within a reasonable time after git is live, we can start cleaning those up a bit, (might as well start at some point) - curious how you'd like to handle that from a git management perspective.. [00:10:47] New code comment: Krinkle; This makes it possible to fix bug 27375.; [00:10:55] ^demon: Can you review r112458 ? [00:11:08] <^demon> varnent: They'd get new repos in mediawiki/extensions/ [00:11:22] <^demon> I'm perfectly fine with moving things over if people want to maintain them. [00:11:51] ^demon: since they would likely be imported by a small group of folks - any problems with setting up new repos each time? [00:12:13] 03(mod) Whitespace changes should be identifiable in CR diffs - 10https://bugzilla.wikimedia.org/27375 +comment (10krinklemail) [00:12:28] <^demon> Well as long as somebody's planning on maintaining them. The problem with on-wiki extensions is that they're typically abandoned. [00:12:38] <^demon> Moving it to git to remain abandoned is better, but not by much :p [00:12:45] ^demon: my thought, and I'm curious if you agree, is that we should assess if the extensions function (compatible with current version), if not is it valuable enough to salvage..if no..archive..if yes..put on repo in hopes of someone adopting it - but either way - get rid of all wikicode based code [00:13:03] <^demon> Sounds good. We can look at this after the migration :) [00:13:20] okay - good - that gives some direction to some other conversations I'm having with folks - ty :) [00:13:53] might be a hackathon type thing has been one thought for example [00:14:22] <^demon> Krinkle: Looks good, want to go ahead and merge + deploy? [00:14:35] ^demon: sure, I can't do that myself though [00:14:40] <^demon> I'll do it. [00:14:43] Thx [00:14:49] Dantman: how hard would it be to make the buttons look like that example image you sent? :) [00:15:01] box-shadow [00:15:19] I can write the css when we're actually writing css [00:15:38] Putting vendor prefixes in inline styles is annoying [00:15:51] fair enough :) [00:16:31] any thoughts on moving it all the way to the right? per v10 right now.. [00:16:41] 03demon * 10/branches/wmf/1.19wmf1/extensions/CodeReview/backend/DiffHighlighter.php: MFT r112458 [00:18:13] <^demon> Ugh, getting some password prompts on sync-file. [00:18:20] <^demon> Ryan_Lane: ^ ? :( [00:18:31] which systems? [00:18:36] are they pooled? [00:18:40] <^demon> Dunno [00:18:44] <^demon> Lemme pastebin [00:18:45] and why are you deploying on a sunday? [00:19:00] <^demon> http://p.defau.lt/?RMjokwJeglja_tz4pvoQ9A [00:19:09] <^demon> Ryan_Lane: Because code review diffs are ugly :) [00:19:15] -_- [00:19:26] Working on it.. [00:19:34] <^demon> And the oscars are still doing red carpet crap :p [00:19:39] (function ( $, mw, undefined ) { [00:19:39] "use strict"; [00:19:48] (all diffs had strike-through after ^demon 's deploy) [00:19:49] like some sort of perl/JS mashup? [00:20:08] <^demon> use strict exists in JS now. [00:20:12] Yep [00:20:12] <^demon> I remember reading about it. [00:20:16] TimStarling: ECMAScript5 [00:20:23] cute syntax [00:20:25] Be careful with that though [00:20:38] TimStarling: The way to trigger strict mode, very useful for making JS more throwy and less silent ignorant [00:20:38] TimStarling: That's so it still works in ES3 [00:20:49] it's the best they could do without breaking < ES5 [00:20:59] As far as I know no browser has yet implemented strict mode. [00:21:09] It's actually not such a good idea to start using it right now. [00:21:16] Will hopefully become default some day, but opt-in for now [00:21:27] At least not until you can actually test that your code works in strict mode. [00:21:54] it's in SignupAPI if you want to remove it [00:22:19] Am I allowed to add myself to the editor group on mediawikiwiki? [00:22:48] hi guys, i cannot get the questy extension to work --extension:confirmedit. it is not showing up at all in special:version. I got the simple captcha to work, and the ReCaptcha to work, but cant get questy to work. and suggestions to troubleshoot? [00:23:08] ^demon: okay, diff styles are working good now on mw.org [00:23:12] clear cache to confirm [00:23:27] <^demon> Krinkle: Well, other than the — and [00:23:36] sorry it's actually in Translate [00:23:38] hey dant [00:23:42] I was looking at the wrong window title [00:23:52] <^demon> Krinkle: https://www.mediawiki.org/wiki/Special:Code/MediaWiki/112458, left column. [00:23:58] ^demon: https://imgur.com/SqVJj [00:23:59] Translate/resources/ext.translate.quickedit.js [00:24:15] ^demon: Ow, crap [00:24:41] here is the extension. http://www.mediawiki.org/wiki/Extension:ConfirmEdit#QuestyCaptcha we really need it. if you ever have gotten spam, we get so much now [00:24:45] <^demon> The diffs looked fine how they were. The only reason we changed it was for hashar's backed-out "per-line commenting" [00:24:54] <^demon> I'd just as soon revert it all back [00:25:14] well, the line numbering is nice [00:25:30] <^demon> I think the table is horribly ugly. [00:25:31] although it does make it harder to quote a part of the diff now [00:25:38] yeah, that too [00:25:39] krenar what page? what edit group? seems like everyone is busy with something else [00:26:52] anno1404, I just wanted to mark https://www.mediawiki.org/w/index.php?title=Help:Extension:ParserFunctions&diff=504178&oldid=502977 as okay [00:27:28] oh krenair [00:27:39] you have the ability to check it as okay? [00:27:51] you need permissions i think [00:27:53] let me check [00:28:00] I can add editor to myself. [00:28:24] I just wanted to know if I was allowed to do so without going to requests for permissions [00:28:44] i dont see how you could, i think it is a special permission [00:28:59] here is the extension: http://www.mediawiki.org/wiki/Extension:FlaggedRevs let me look into how you can become a member of the elite group [00:29:09] 03krinkle * 10/trunk/extensions/CodeReview/backend/DiffHighlighter.php: Follow-up r112458: html entities were showing up literally on the page. the numbers can be treated as to-be-escaped html but these were not, fixing them to be literal to avoid breakage through Html::element [00:29:15] Admins can add the editor right to themselves and others, anno1404 [00:29:34] <^demon> Krenair: Go ahead and give yourself +editor, that's fine [00:29:45] http://www.mediawiki.org/wiki/Extension:FlaggedRevs#How_to_assign_user_rights [00:29:48] thanks [00:29:58] oh so you assigned him the rights [00:30:02] thats good [00:30:09] Dantman: Chrome has had strict mode for a while iirc [00:30:34] Atleast a lot of it, I remember using it and catching an annoying little bug with it that I didn't catch other wise [00:30:51] demon, would you have a second about the captcha questycaptcha? I have tried for motnhs to get it to work -- and it won't [00:31:01] any suggestion would be most appreciated [00:31:24] <^demon> I've never used the captcha extensions. [00:31:28] ^demon: Do you mind reviewing r112458 on CR ? since it was merged+deployed already [00:31:39] <^demon> Oh, I never pressed save. [00:31:41] <^demon> One sec [00:31:42] lol [00:32:25] oops tyop: http://paste.tstarling.com/p/nnGydC.html [00:32:31] *typo [00:33:04] <^demon> TimStarling: Those all seem to be from the same author ;-) [00:33:28] TimStarling: haha, hilarious [00:33:40] I was about to fix it like I did in r111989, but I thought hang on, I'd better check if there are any more [00:33:55] okay thanks demon [00:33:56] I might be guilty of one of those Sripts though, I think I created one of those files and copy/pasted [00:34:30] they're mostly from Jeroen [00:34:55] I'll fix them with sed once I'm done with my current commit [00:35:06] ^demon|food: when you get back, can you merge+deploy r112460 as well ? The entity fix you asked for [00:35:07] here is my pastebin on what i have for questy captcha: localsettings.php: http://www.pasteall.org/29517 it does not appear in special:version page though [00:35:29] <^demon|food> Krinkle: I'll poke at it, yes. [00:35:33] k [00:39:52] hi guys, i cannot get the questy extension to work --extension:confirmedit. it is not showing up at all in special:version. I got the simple captcha to work, and the ReCaptcha to work, but cant get questy to work. and suggestions to troubleshoot? [00:51:57] 03robin * 10/trunk/phase3/languages/messages/MessagesTt.php: Remove empty messages introduced in r108445. Translation is present in fallback tt-cyrl. [00:56:43] 03tstarling * 10/trunk/extensions/ (5 files in 5 dirs): (bug 34538) Fixed compatibility with $wgResourceLoaderExperimentalAsyncLoading in extensions N-Z. Followup from r111989 where I did A-M. Untested. [01:00:05] hi, has sometime time to discuss a small problem in HttpFunctions factory ? [01:00:39] when called with options['userAgent'] , this is not correctly set [01:02:58] 03tstarling * 10/trunk/extensions/ (70 files in 26 dirs): Fixed a typo in a boilerplate comment that Jeroen copied to many files, JavasSript -> JavaScript [01:03:45] hello, figured it out, but the captcha only shows up when i am registering a page, never anywhere else. [01:04:00] any suggestions on how i can fix this? [01:06:09] Wikinaut: I have time to review a patch if you have one [01:06:35] hm not yet a patch. I will commit something for E:RSS, then I will explain here, ok ? [01:06:46] ok [01:07:06] 03demon * 10/branches/wmf/1.19wmf1/extensions/CodeReview/backend/DiffHighlighter.php: MFT r112460 [01:07:50] i figured it out. this irc group was no hel today :( [01:07:54] help* [01:11:01] 03wikinaut * 10/trunk/extensions/RSS/ (6 files): (log message trimmed) [01:11:01] fix for ultra bug 30028 . The RSS extension can parse RSS and ATOM feeds of [01:11:01] different flavours. The php xml dom xpath query uses now a namespace-safe method [01:11:01] to find all elements like item (RSS, RDF) or entry (ATOM). Further fixed a [01:11:02] hidden problem when the feed url was redirecting, this threw the Cannot parse [01:11:04] RSS for XML error, which is now history. Introduced a new parameter [01:11:05] wgRSSUrlNumberOfAllowedRedirects which defaults to zero, i.e. no redirects are [01:11:40] 03(FIXED) "Error parsing XML and HTML for RSS" - improve and harden Extension:RSS when parsing differently flavoured RSS feeds - 10https://bugzilla.wikimedia.org/30028 +comment (10T. Gries) [01:11:47] New code comment: Tim Starling; Fixed in r112451.; [01:12:45] Tim: [01:12:55] $client = HttpRequest::factory( $url, [01:12:57] array( [01:12:58] - 'timeout' => $wgRSSFetchTimeout, [01:13:00] - 'proxy' => $wgRSSProxy, [01:13:01] - 'noProxy' => $noProxy, [01:13:03] + 'timeout' => $wgRSSFetchTimeout, [01:13:05] + 'followRedirects' => true, [01:13:06] + 'maxRedirects' => $maxRedirects, [01:13:08] + 'proxy' => $wgRSSProxy, [01:13:09] + 'noProxy' => $noProxy, [01:13:11] + 'userAgent' => $wgRSSUserAgent, [01:13:15] I added the option setting of "userAgent" [01:13:18] but this had no effect [01:14:05] Kept the following line $client->setUserAgent( $wgRSSUserAgent ); in the code [01:14:07] which works [01:14:33] perhaps you know, why it fails in HttpFunctions:factory [01:14:40] I have no idea [01:18:32] New code comment: Tim Starling; Yes, I think rethrowing the exception is the best option, at least in the short term. In theory, it'; [01:18:39] grrr [01:19:39] missed a key and hit tab instead of something else, then space a few ms later, which caused the "save changes" button to be clicked [01:23:18] when i try to put in the text feild how do i change it to HTML [01:25:55] how do i put HTML into a wiki page [01:26:32] New code comment: Tim Starling; ...is set to "error". Support for requests for modules with dependencies that have errors does not s; [01:27:51] A) ETag support. B) Built in 404 handler support. [01:28:10] how do i do that? [01:28:42] because on my wiki i want to put into a page but it is HTML not wiki markup [01:30:54] what type of wiki software is good for per/partial page restriction [01:32:07] sfdvgqasdfd: This isn't really the place you'd ask that [01:32:25] then where would i ask [01:32:38] Dunno... maybe wikimatrix [01:32:50] Or whatever it was called... that site that compares wiki engines [01:32:56] /join #wikimatrix [01:34:54] The Nginx and Lighttpd setups on Ubuntu have gotten interestingly smooth to setup [01:35:27] Lighttpd comes bundled with a fastcgi-php mod, you just enable it and install php5-cgi and php is up and running. [01:36:16] Nginx's example configs include a .php fastcgi line for localhost:9000 and if you install php5-fpm a fpm fastcgi server will automatically be running at that address [01:36:52] ^_^ So it's practically as simple as installing apache2 and php5 was [01:37:07] 03tstarling * 10/trunk/phase3/includes/HttpFunctions.php: Make userAgent also work as an option to MWHttpRequest::factory() and MWHttpRequest::__construct(), not just Http::request(). [01:37:23] Wikinaut: done [01:37:42] I'm sure you could have done a patch yourself if you tried [01:38:01] was just writing a bugzilla... (not yet sent) [01:38:46] Tim, I was not sure where these lines must go to... [01:40:03] Ohhhhhhh.... I should use that as my default fastcgi_param [01:41:50] Tim: negative [01:41:54] (sorry [01:41:57] sorry [01:42:02] need to svn up core [01:43:21] Tim: ok ! [01:44:32] 03wikinaut * 10/trunk/extensions/RSS/RSSParser.php: removed superfluous code for setting userAgent since r112466 [01:46:02] New code comment: Wikinaut; ok. Working for example in r112467; [01:47:49] good night [01:49:02] http://spyinginc.com/wiki/password_space.html [01:49:24] finaly found out how to put html into a site [01:53:14] (.) (.) [01:53:14] <> [01:53:14] (________) [01:53:14] [02:01:17] Krinkle-away: I'm around now, but now you're -away :) [02:01:33] don't worry, my mind is still here [02:02:08] trying to remember what it was I was pinging you for [02:02:36] hexmode: Ah, right. the 'nodeploy' tag. Is that anywhere used by automated processes or list generators? [02:02:47] Or just to manually filter Special:Code lists ? [02:03:03] Krinkle: it is used by the revision report [02:03:16] maybe crstats, but I don't think so [02:03:35] cause I was slightly annoyed by seeing it stamped on revisions that no where even possibly going live and not related to wmf deployment (e.g. trunk/tools or trunk/extensions/examples) [02:03:54] enotifs triggered etc. [02:04:54] perhaps crstats can filter on paths more precicly and filter out most stuff that way (e.g. -trunk/tools for instance, or rather whitelist instead of blacklist just trunk/phase3 + whatever extensions need review) [02:05:19] just an idea or curiosity, no problem otherwise [02:06:47] hexmode: so 'nodeploy' is used for revisions that could go live but shouldn't because of something. revisions that aren't scheduled to go live on wmf (e.g. trunk/tools/ToolserverI18N) wouldn't need the tag then? [02:07:36] Krinkle: nodeploy is used for code we don't deploy, that doesn't affect the site [02:07:42] so installer code [02:07:55] or extensions we don't use on the site [02:08:03] oh [02:08:12] Hm.. [02:08:13] so it's not revision related but path related? [02:09:00] then even more reason not to use tags for it imho, except maybe as additional fine-grain control for things inside phase3 (like the installer) [02:09:01] Krinkle: right, kind of [02:09:34] it's status deferred for that though? [02:09:50] (assuming it's to lower number of 'new' revisions) [02:10:05] this way extension authors can use CR without messing up the revision report [02:10:18] yeah [02:10:55] hexmode: Can you document https://www.mediawiki.org/wiki/Special:Code/MediaWiki/tag/important ? [02:11:00] https://www.mediawiki.org/wiki/Talk:Code_review_tags [02:11:20] (or remove it's usage) [02:11:27] its* [02:11:36] Hrm... I haven't used important, have I? I'd have to look at it [02:11:40] * hexmode checks [02:11:48] oh? I thought you did [02:11:58] Hm.. Aaron introduced it, it seems [02:12:54] I'll ping Aaron on his talk page [02:14:56] hexmode: btw, I've added a couple of tags in the documentation for 1.19 and 1.20, basically just copies of their 1.18/1.19 equivelants [02:15:22] pf, just found out I have a week holiday next week :) 5 days off from school. Some kind of "spring holiday" I totally forgot about [02:15:25] all this free time B) [02:16:04] 1.20, it probably wasn't needed. It'll be git-tastic! [02:16:14] well, trunk is 1.20 now [02:16:37] we won't release in 1.20, but we can use the 1.20revert tag [02:17:15] Is there a tag for "revert now, redo when mw-core is in git" ? [02:17:51] I've seen a couple of revs reverted under that motto, we should tag them [02:19:31] Krinkle: yeah, we should have that one [02:19:35] gitrevert [02:20:02] New code comment: Krinkle; This wasn't merged yet (I see no follow-ups). If that is the case, do we still need it? 1.19 was bra; [02:20:17] New code comment: Krinkle; This wasn't merged yet (I see no follow-ups). If that is the case, do we still need it? 1.19 was bra; [02:20:26] !branchpoints [02:20:26] http://www.mediawiki.org/wiki/Branch_points [02:20:38] !1.19 [02:20:38] Working on it, hoping to deploy in February 2012. Branched at r110996 on 9th February, 2011. http://thread.gmane.org/gmane.science.linguistics.wikipedia.technical/56919 [02:21:02] New code comment: Krinkle; This wasn't merged yet (I see no follow-ups). If that is the case, do we still need it? 1.19 was bra; [02:21:32] New code comment: Krinkle; This wasn't merged yet (I see no follow-ups). If that is the case, do we still need it? 1.19 was bra; [02:23:13] !1.19wmf1 [02:23:14] I don't know anything about "1.19wmf1". [02:23:16] !1.18wmf1 [02:23:16] Branched on 19 September 2011 at r97502 from REL1_18 [02:23:56] !1.19wmf1 is Branched on 10 February 2012 at r111189 from REL1_19 [02:23:56] Successfully added keyword: 1.19wmf1 [02:24:06] night, krinkle! [02:24:10] night [02:24:50] Heh... sweet... I managed to hack in a 404 thumbnail handler for nginx without using if () [02:26:14] Reminds me of the xkcd cartoon on "How to manage to finish X without even touching your keyboard?", which was illustratively answered as "By having someone else do it" :P [02:28:32] New code comment: Krinkle; This wasn't merged yet (I see no follow-ups). If that is the case, do we still need it? 1.19 was bra; [02:28:47] New code comment: Krinkle; This wasn't merged yet (I see no follow-ups). If that is the case, do we still need it? 1.19 was bra; [02:35:44] how can I create a link to a path on the web site, without using the domain name? for instance [/downloads/foo.pdf Download the foo-book!] [02:37:47] New code comment: Whym; There is a typo in SpecialContributions.php:258. \ IS: parenteses -> SHOULD-BE: parentheses; [02:40:11] 03demon * 10/trunk/phase3/includes/specials/SpecialContributions.php: Typofix, ping r112427 [02:52:33] 03(mod) API list=logevents does not correctly unserialize log_params - 10https://bugzilla.wikimedia.org/34653 summary (10Tim Starling) [02:55:22] Anyone at WMF have any connections to Google who can forward a pretty technical MediaWiki related question to them? [03:05:05] 03(mod) rebuildImages.php --missing halts with error upon encountering any filename which it wants to rename - 10https://bugzilla.wikimedia.org/34654 +comment (10tstarling) [03:07:34] 03(NEW) Making uselang= stick when there is no chance to login - 10https://bugzilla.wikimedia.org/34748 normal; MediaWiki: Internationalization; (jidanni) [03:16:39] 03(mod) bad performance: reverting file versions takes ages for the browser to "come back" (show the success message) - 10https://bugzilla.wikimedia.org/34717 +comment (10Tim Starling) [03:23:32] New code comment: Santhosh.thottingal; Yes, in r111934; [03:42:18] Does anyone here have an advanced knowledge of the #switch function? [03:46:51] The_Pathogen: Not really but I might be able to help anyway [03:48:31] Thanks for responding, heh. I've been wondering if the #switch syntax supports searching a large string for a specific small string, then replacing that small string with a more comprehensive one. [03:48:36] 03(NEW) [SMW] When mainlabel specified an extra mainlabel parameter is added in the further results link - 10https://bugzilla.wikimedia.org/34749 normal; MediaWiki extensions: Semantic MediaWiki; (fastgoldfish) [03:48:49] Although I not entirely sure if it supports that sort of usage.. [03:49:51] WTF... OutputPage::uncacheableBecauseRequestVars [03:50:15] uh? [03:51:39] The_Pathogen, pretty sure you can't do that [03:52:06] Darn, that's a shame. [03:54:10] there a reason we wouldn't want to host widgets on mw.org? mwwidgets.org is having a lot of problems lately apparently.. [04:03:07] Ok, thanks for the help there Krenair [04:04:49] varnent: We don't even have a real gadgets repo last time I checked [04:06:25] I don't even like the premise of that extension... [04:11:04] Dantman: Gadgets is coming together - but if they want to move the content from MWWidgets.org to MW.org.. [04:11:06] brb [04:12:32] At the very least I don't like the idea of MW.org having an entire namespace full of stuff dedicated to one single extension that WMF and most wikis don't even use. [04:12:39] <^demon> I've never even heard of the site before. [04:12:53] If anything were to happen, id say something more like having the wiki hosted on labs or something. [04:12:55] <^demon> I remember the extension, but don't know anyone using it. [04:13:24] <^demon> http://www.mediawikiwidgets.org/ seems to have become a spam bin. [04:14:52] And I can guarantee you that the extension itself will never be enabled on MW.org. [04:15:07] So all those example pages would not be functioning. [04:15:28] <^demon> Oh, didn't know you were the guy making those decisions nowadays ;-) [04:16:40] ^demon: Generally editable Widgets: namespace so people can upload their widgets + functioning Widgets extension = XSS Injection vector on a WMF wiki [04:17:23] <^demon> I'm being sarcastic. [04:17:40] heh [04:20:37] Oh ^_^ And spam isn't the worst that wiki has to worry about. [04:20:52] I already managed to find a wide open injection vector in one of the widgets. [04:21:30] And here I was only looking for an iframe url injection vector. [04:22:39] ((eg: Say a Facebook widget that used some facebook.com/{something} substitution I could use something like =../../l.php to trick FB's external link redirection code to redirect the iframe to another arbatrary website [04:26:02] How come my edits on mediawikiwiki aren't showing up in Special:RecentChanges? o.O [04:27:35] <^demon> Dantman: Yuck. [04:28:01] Hence. I hate the premise of the extension. [04:30:11] It reduces the creation of things that would usually require an extension or something to the level of "cut and paste this html". At this level completely unfit people are free to create piles of widgets, there is absolutely no proper review of them, there is absolutely no concept of versioning, no concept of any sort of update to people that an XSS vector has been fixed and they should upgrade, ... [04:31:27] Even if I fixed that widget (which I probably don't have the rights to do) doing so would still likely leave a bunch of wiki with an open XSS vector. [04:33:12] You know... if I was black hat, I could probably write a spider that scans wikis with a Widgets: namespace and automatically finds widgets with XSS vectors and exploits them. [04:34:05] And I'm not talking about manually listing widgets with holes. I mean the syntax is such that I could probably write code that would look at Widget: content, find something that opens a hole, and then automatically exploits it [04:38:57] 03santhosh * 10/trunk/extensions/WebFonts/ (5 files in 2 dirs): Add Akkadian font for Cuneiform script. Map it to akk and sux lang codes. [04:49:47] could someone please tell me of the process to purge a ResourceLoader gadget [04:50:12] it has a ... &smaxage=21600&maxage=86400 [04:50:43] and I want to get that purged before I start exploring other avenues about its not working [05:08:16] 03santhosh * 10/trunk/extensions/Translate/ (3 files in 3 dirs): [05:08:16] Backend logic and storage of language restriction Data (i18n #614) [05:08:16] Generic Table and read/store code for metadata of a group. [05:26:14] any participants of gsoc in here?? [05:26:22] this year or in the past [05:27:24] 03laner * 10/trunk/extensions/LdapAuthentication/LdapAuthentication.php: (log message trimmed) [05:27:24] Seems getCanonicalName is used to also check page names, which causes nastiness [05:27:24] when anons click redlinks. I want to stab our authentication code for reusing [05:27:24] this function in this way. This fix works around an error by only connecting to [05:27:25] LDAP if the user is also using a valid name. That said, this is a major issue [05:27:26] all around. When a user is logged in, if they click a red link it's going to [05:27:27] cause an LDAP lookup unless memcache is being used and there is a key hit, and [05:31:45] 03santhosh * 10/trunk/extensions/Translate/tag/PageTranslationHooks.php: [05:31:45] If the prioritylangs set, show those languages alone. [05:31:45] (i18n #612) [05:36:16] 03(mod) add a summary message for Special:DeletedContributions - 10https://bugzilla.wikimedia.org/34319 +comment (10Srikanth Logic) [05:38:05] 03(mod) API list=logevents does not correctly unserialize log_params - 10https://bugzilla.wikimedia.org/34653 +comment (10bawolff+wn) [05:47:03] !seen hexmode [05:47:05] nischayn22, hexmode is right here! [05:47:13] hexmode: hello [06:05:02] New code comment: Nikerabbit; Don't we usually make shared common protected method in these cases?; [06:09:32] Such fancy Last-Modified handling code... [06:09:36] ...must. destroy! [06:16:11] Dantman: EXTERMINATE? [06:16:40] Exterminieren! [06:20:13] You know what the real shame of killing Last-Modified will be... [06:21:08] More needless HTTP requests? [06:21:17] I can't be tempted to write a WebRequest::getIfModifiedSince() to suppland $_SERVER["HTTP_IF_MODIFIED_SINCE"] that will properly return If-Modified-Since without that IE bug [06:21:58] Joan: ;) No actually my Last-Modified killing plans will be what reduces needless http 200's... neither of them stop requests anywyas [06:22:41] I thought the point of those headers was to see if you need to make a (second) request for the full content. [06:23:28] IMS requests return the full content if it was modified after a certain date [06:23:37] so a second request is not needed [06:23:50] maybe you're thinking of HEAD [06:25:00] anyway I'm not sure why you'd want to get rid of the IMS code [06:25:07] it's a lot faster than a parser cache hit [06:25:13] TimStarling: ;) ETags [06:25:40] If Last-Modified hasn't changed, can't the browser just load from cache rather than make another request? [06:25:44] Proper page caches, without cookie hacks [06:26:04] how would you invalidate it when templates change? [06:26:40] also have you surveyed client support? IIRC the etag support in squid is a bit quirky [06:27:19] TimStarling: We already have that kind of stuff in place for last-modified. What I'm looking to deal with is the fact that for a logged in user and a logged out user the page is different. Hence, ETag [06:28:02] New code comment: Nikerabbit; Forgot autoload?; [06:28:16] I'll need to look closer at that squid bit... [06:28:25] so are you saying you don't want to support If-None-Match etc.? [06:28:34] Huh? [06:28:56] you know, the etag equivalent of If-Modified-Since [06:29:39] When I say ETag, I naturally mean code supporting the ETag + If-None-Match. Instead of Last-Modified + If-Modified-Since. [06:29:49] http://tools.ietf.org/rfcmarkup?doc=2616#section-14.26 [06:29:55] There's no point for ETag without If-None-Match [06:29:59] 03santhosh * 10/trunk/extensions/Translate/_autoload.php: Entry in _autoload.php. Follow up r111629 [06:30:21] so how do you work out whether a given ETag has been invalidated by a template change? [06:30:51] How do you work out whether a given Last-Modified since has been invalidate by a template change? [06:31:00] we update page_touched [06:31:06] ;) There you go. [06:31:47] New code comment: Amire80; It even works in Chrome and IE.; [06:31:55] The point of ETag is not to replace the backend stuff we use for that kind of stuff. But to make it possible to differentiate between the cache for anon, user1, user2, etc... [06:32:19] At the simplest level it might even mean using what we already use for Last-Modified and instead putting it inside ETag with other things. [06:32:32]