<hello> - wikidata-test

<hello>

[02:27:08] 03(FIXED) Make donate.wikimedia.org (FundraiserLandingPage) detect the browser language (in addition to the country) - 10https://bugzilla.wikimedia.org/37397 (10Robin Pepermans (SPQRobin)) [02:27:12] That's because it goes through the parser [02:27:15] Stuff from wikitext is different because of ^ [02:27:16] That's a whooooole nother story [02:27:23] Have you read Parser.php xD [02:27:33] wfMessage()->text() vs. ->plain() related [02:28:15] New patchset: Trevor Parscal; "(bug 38545) Fix RTL position of the drop-down arrow" [mediawiki/extensions/VisualEditor] (master) - https://gerrit.wikimedia.org/r/16218 [02:29:30] Interesting. [02:29:34] Well, I feel better, then. [02:29:52] html only requires escaping for < and & (at least according to TimStarling - iirc he wrote that, and omitted other common special chars as a way of optimization - instead of calling htmlspecialchars() [02:29:59] New review: Trevor Parscal; "This is a hack, but it's fine for now. Using an SVG or a character might work in the future." [mediawiki/extensions/VisualEditor] (master); V: 1 C: 1; - https://gerrit.wikimedia.org/r/16218 [02:30:00] I don't think I wrote that [02:30:14] TimStarling: You did write the javascript variant of that though, right ? [02:30:16] I would have thought htmlspecialchars() would be faster than strtr() [02:30:23] yes [02:30:28] 010c4568 (Aryeh Gregor 2010-01-15 01:16:52 +0000 163) # There's no point in escaping quotes, >, etc. in the contents of [02:30:31] but then again, js doesn't have htmlspecialchars [02:30:34] okay [02:30:52] Probably as part of the HTML5 work [02:31:30] Aryeh wrote the whole of Html.php I think [02:32:16] that "etc" isn't quite correct [02:32:16] Right, he was the one that split it off from Xml.php [02:32:35] since htmlspecialchars() literally escapes only four characters: <>"& [02:32:43] Change merged: Catrope; [mediawiki/extensions/VisualEditor] (master) - https://gerrit.wikimedia.org/r/16218 [02:33:20] I would have thought escaping > would be necessary for standards compliance [02:33:27] but I guess Aryeh would know that better than me [02:33:45] maybe that's why he did it, Html.php was mostly aimed at taking advantage of some of the more arcane features of HTML 5 [02:34:05] like omitting end tags [02:34:11] The validator doesn't seem to complain about > not being escaped. [02:34:24] It complains a lot about the ISO codes in hreflang. [02:34:35] features that became legal in HTML5 spec that were done by browsers for ever [02:34:48] I assume ^ since it wasn't in an HTML5 conditional [02:35:31] well, the idea of XHTML was to make HTML easier to parse [02:35:37] Brooke: Validator is not very useful, because browsers are not standards compliant. [02:35:37] which I think was a noble goal [02:35:48] it removed a lot of the stranger syntax features from HTML [02:36:05] what does the validator say in html5 mode? [02:36:22] (which is basically how browsers have behaved since forever) [02:36:29] just because all browsers have thousands of lines of HTML parsing code doesn't mean MediaWiki should assume that all clients have such code [02:36:37] right? [02:36:41] I agree [02:37:00] 03(FIXED) VisualEditor: Keys for increasing reducing indentation of lists (TAB, SHIFT+TAB) and finishing lists - 10https://bugzilla.wikimedia.org/33142 +comment (10Roan Kattouw) [02:37:11] there's lots of html5parser libs by now though [02:37:27] yes, no doubt that's what Aryeh would have said if I complained at the time [02:37:29] which still are more compact than what "real" browsers do (much less tolerant, which I think is good) [02:37:40] and he had a few other anti-XHTML arguments lined up I think [02:37:49] :) [02:38:24] 03(FIXED) VisualEditor: Wrong rendering of levels of lists - 10https://bugzilla.wikimedia.org/33143 +comment (10Roan Kattouw) [02:39:57] 03(FIXED) VisualEditor: Undo of indentation is not properly displayed - 10https://bugzilla.wikimedia.org/33225 +comment (10Roan Kattouw) [02:40:53] 03(FIXED) VisualEditor: Indent and outdent buttons do not exist - 10https://bugzilla.wikimedia.org/37906 +comment (10Roan Kattouw) [02:41:02] have you ever read about SGML? [02:41:35] SGML has lots more crazy features than HTML [02:42:05] like empty end tags,

... [02:42:16] eeeewwwww! [02:42:18] That's just a bad idea [02:42:28] Careful there [02:42:44] ;) verbose end tags is often an argument used against xml [02:45:18] I'm still trying to wrap my head around > not needing to be escaped. [02:46:24] 03(FIXED) VisualEditor: Placing the cursor (by clicking or through the cursor keys) into sub-lists breaks - it "flashes" in, but then disappears (is no longer drawn anywhere) and cannot be used (e.g. shift-left or etc. doesn't work). - 10https://bugzilla.wikimedia.org/37907 +comment (10Roan Kattouw) [02:47:41] Brooke: That really depends on context [02:48:10] If you're in plain html, so there's no tag for > to close. [02:48:28] New patchset: Mdale; "Sync with kaltura html5 version 1.7" [mediawiki/extensions/TimedMediaHandler] (master) - https://gerrit.wikimedia.org/r/16467 [02:48:50] If you're in a quoted attribute the parsing continues till it hits the end quote, so > doesn't end the tag there either. [02:49:09] That said, you have to make sure the output doesn't go into an unquoted tag. [02:49:24] I thought

was always illegal and wrong. [02:51:40] There's no reason strictly that it "can't" be parsed sanely. And browsers seem to handle it fine. [02:52:07] Strictly speaking it doesn't even require special handling to parse. [02:53:10] Sure, it might if you use write a parser using a strange overly literal system that tries to attach meaning to every special character ignoring it's context [02:53:36] But in general parsing... you hit the end of the

then you search for the next '<' [02:53:58] You have no reason to search for a '>' since it has no meaning in that context so it just ends up as part of the text. [02:55:17] And when you decode the entities and whatnot you don't do it character by character... you just find all the &*;'s and decode them. So there's no reasn for the > to trip up the parser either. [02:58:22] If you disable JavaScript, Wikidata is like "uhhhhhh, nope." [03:02:27] [03:03:26] * Namtnad hates the fact that we didn't have a standard way to do opening menus without JS a long time ago [03:03:31] New patchset: Mdale; "Sync with kaltura html5 version 1.7" [mediawiki/extensions/MwEmbedSupport] (master) - https://gerrit.wikimedia.org/r/16468 [03:07:31] 03(NEW) AFT appeared on "Sorry, the servers are overloaded" page - 10https://bugzilla.wikimedia.org/38627 normal; MediaWiki extensions: ArticleFeedbackv5; (spage) [03:23:16] Brooke: http://www.mediawiki.org/wiki/Special:AbuseLog/14071 apparently has the old "call on member function on non-object" error [03:24:39] Oh dear. [03:24:46] and it appears to be specific to that filter entry in particular. [03:24:48] Is there a bug filed about that? [03:24:55] no, just found it 2night [03:25:34] strangely http://www.mediawiki.org/wiki/Special:AbuseLog/14073 has the same error while the next entry does not [03:30:23] [03:40:28] 03(NEW) Article Feedback - Show only feedback my watchlist - 10https://bugzilla.wikimedia.org/38628 normal; MediaWiki extensions: ArticleFeedbackv5; (fflorin) [03:41:12] 03(mod) Article Feedback - Show only feedback my watchlist - 10https://bugzilla.wikimedia.org/38628 normal->major; +aftv5-1.5 (10Fabrice Florin) [03:55:44] New patchset: Aaron Schulz; "Removed unused $wgAntiLockFlags flags." [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/16469 [04:02:12] 03(FIXED) Add file repo support to prop=duplicatefiles - 10https://bugzilla.wikimedia.org/27567 +comment (10Umherirrender) [04:20:13] 03(mod) Add Extension:ConfirmEdit-Asirra to Bugzilla - 10https://bugzilla.wikimedia.org/38623 summary; +comment (10Purodha Blissenbach) [04:40:36] ext/mysql | Lifecycle | Long term deprecation announced [05:06:36] Namtnad: What? [05:06:46] Jasper_Deng: Did you file a bug about that uncaught exception? [05:06:58] Brooke: php page [05:06:59] Brooke: no...and it appears to have changed. [05:07:05] mysql, mysqli, PDO mysql [05:07:21] https://www.mediawiki.org/wiki/Special:AbuseLog/14071 [05:07:28] > PHP fatal error in /usr/local/apache/common-local/php-1.20wmf8/extensions/AbuseFilter/special/SpecialAbuseLog.php line 284: [05:07:31] Call to a member function getVar() on a non-object [05:07:34] mysql aka: ext/mysql has Lifecycle listed as "Long term deprecation announced" [05:07:39] mysql is the one we're using [05:08:14] I'm not sure what that means. PHP is going to deprecate the use of MySQL? Or the use of a particular MySQL library? [05:08:25] ext/myql [05:08:36] The functional mysql library introduced in PHP 2 [05:08:50] It's pratically been deprecated for awhile [05:08:52] Is the deprecation announcement anywhere? [05:09:13] The other apis are the PDO version of mysql and the mysqli api. [05:09:35] 03(NEW) PHP exception for particular abuse log entries on MediaWiki.org - 10https://bugzilla.wikimedia.org/38629 normal; MediaWiki extensions: AbuseFilter; (jasper) [05:09:44] The mysqli api being oop and having much more support for mysql features than ext/mysql. [05:10:36] idk if mw has migrated, but it might be a good idea to. [05:10:42] lots of find/replacing... [05:10:49] I'm pretty sure we haven't already [05:11:00] Someone tried to once and got rejected [05:11:12] the good thing is that it doesn't need fundamental code work. [05:11:39] Btw, don't exactly need to just drop the current mysql code. A saner way to do this would be to write a new database class for mysqli. [05:11:53] Naturally inheriting from the mysql clas [05:12:03] true... [05:13:50] Heh... now I want a prepared statement API for mw [05:14:57] $prepared = $db->select( 'user', [...], array( 'user_name' => new PreparedVariable('username') ), [...] ); [05:15:14] $prepared->execute( array( 'username' => 'Dantman' ) ); [05:16:09] Of course, you wouldn't use it in general queries... more like when you're doing a lot of separate inserts, etc... [05:17:12] * Jasper_Deng barely can understand OOP [05:23:52] 03siebrand * 10/trunk/tools/ToolserverI18N/language/messages/ (14 files): Localisation updates from http://translatewiki.net. [05:27:10] New patchset: Aaron Pramana; "Watchlist grouping" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/16419 [05:30:32] Change merged: Siebrand; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/15790 [05:30:45] Change merged: Siebrand; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/15873 [05:31:41] Change merged: Siebrand; [mediawiki/extensions/Translate] (master) - https://gerrit.wikimedia.org/r/16027 [05:35:55] New patchset: Aaron Pramana; "Watchlist grouping" [mediawiki/core] (master) - https://gerrit.wikimedia.org/r/16419 [05:39:17] New patchset: Nischayn22; "Implementation of Property count" [mediawiki/extensions/SemanticMediaWiki] (master) - https://gerrit.wikimedia.org/r/16111 [05:57:54] 03(NEW) Bad value ResourceLoaderDynamicStyles for attribute name on element meta: Keyword resourceloaderdynamicstyles is not registered. - 10https://bugzilla.wikimedia.org/38630 normal; MediaWiki: ResourceLoader; (karagota) [05:59:17] 03(mod) [Regression] ExtensionDistributor: SVN-only extensions no longer downloadable - 10https://bugzilla.wikimedia.org/38188 (10Emufarmers) [06:03:04] 03(mod) Bad value ResourceLoaderDynamicStyles for attribute name on element meta: Keyword resourceloaderdynamicstyles is not registered. - 10https://bugzilla.wikimedia.org/38630 (10Olga Butygina) [06:03:24] 03(mod) Bad value ResourceLoaderDynamicStyles for attribute name on element meta: Keyword resourceloaderdynamicstyles is not registered. - 10https://bugzilla.wikimedia.org/38630 +comment (10Olga Butygina) [06:04:24] 03(mod) Bad value ResourceLoaderDynamicStyles for attribute name on element meta: Keyword resourceloaderdynamicstyles is not registered. - 10https://bugzilla.wikimedia.org/38630 (10Olga Butygina) [06:23:06] 03(mod) PHP exception for particular abuse log entries on MediaWiki.org - 10https://bugzilla.wikimedia.org/38629 +comment (10matanya) [06:28:27] 03(mod) Add count parameter to MediaWiki:Contest-welcome-active-contests - 10https://bugzilla.wikimedia.org/38625 (10matanya) [06:32:36] 03(mod) Please Add new component: ConfirmEdit-Asirra to the Extension product - 10https://bugzilla.wikimedia.org/38623 summary; +comment (10matanya) [06:33:57] wtf https://bugzilla.wikimedia.org/show_bug.cgi?id=38630 [06:37:06] 03(mod) MediaWiki:Asirra-edit - message needs better wording - 10https://bugzilla.wikimedia.org/38597 +comment (10Niklas Laxström) [06:38:42] 03(mod) Pinch to zoom is not working on some devices - 10https://bugzilla.wikimedia.org/38348 +comment (10matanya) [06:41:38] 04(REOPENED) MediaWiki:Asirra-edit - message needs better wording - 10https://bugzilla.wikimedia.org/38597 +comment (10Niklas Laxström) [06:43:24] 03(FIXED) Please Add new component: ConfirmEdit-Asirra to the Extension product - 10https://bugzilla.wikimedia.org/38623 +comment (10Niklas Laxström) [06:44:22] 03(mod) Please Add new component: ConfirmEdit-Asirra to the Extension product - 10https://bugzilla.wikimedia.org/38623 (10Niklas Laxström) [06:58:32] New review: Nikerabbit; "Yes I said i18n is ok. One enhancement for qqq would be to add example to weekday-at like "Wednesday..." [mediawiki/core] (master); V: 0 C: 1; - https://gerrit.wikimedia.org/r/15746 [07:06:08] New review: Nikerabbit; "Too many broken links... customised versions, alpha versions, point releases." [mediawiki/core] (master); V: 0 C: -2; - https://gerrit.wikimedia.org/r/16433 [07:06:20] 03(mod) Article Feedback - Show only feedback from my watchlist - 10https://bugzilla.wikimedia.org/38628 summary (10Fabrice Florin) [07:07:26] 03(mod) Images not displaying with IE - 10https://bugzilla.wikimedia.org/5248 +comment (10matanya) [07:08:15] 14(WFM) Images not displaying with IE - 10https://bugzilla.wikimedia.org/5248 +comment (10matanya) [07:15:59] 03(NEW) AFT contributions entries are very ugly and confusing - 10https://bugzilla.wikimedia.org/38631 normal; MediaWiki extensions: ArticleFeedbackv5; (federicoleva) [07:16:00] 03(mod) Article Feedback - My Contributions - 10https://bugzilla.wikimedia.org/36773 (10Nemo_bis) [07:24:48] 03(mod) "oversighter", "oversight", "unoversight", "hide" and "unhide" messages are very unclear - 10https://bugzilla.wikimedia.org/35026 +comment (10Nemo_bis) [07:25:53] 14(WFM) Edit attempt which forces login, does not retain current page - 10https://bugzilla.wikimedia.org/13847 +comment (10matanya) [07:30:58] 03(mod) Search does not find some pages - 10https://bugzilla.wikimedia.org/37578 +comment (10Elitre) [07:39:00] TimStarling, XHTML is easier to parse than HTML, but it also causes pages to break very easily. We've always had hard-to-fix XHTML well-formedness bugs; the only reason we didn't have to waste resources fixing them is because browsers actually use more tolerant parsers. The possibility of a syntax that's as simple to parse as XML but has error recovery is an attractive thought, but it doesn't exist. [07:40:01] Also, there are HTML5 parser libraries for most common languages now that work exactly like browsers. If a client uses an XML library to parse pages instead of an HTML5 library, it will produce different DOMs from the ones browsers see. [07:40:45] E.g., the contents of