[00:05:00] Krinkle: In your inline comment, you say to "Use .text(), not .html()!", but in this case the message contains HTML. Is there a better way to handle this? [00:14:09] Krinkle: got your change fixed [00:14:53] I had to cherry-pick the old patch and make sure the changeID was the same [00:14:55] kaldari: well,yes, but the short version is, DontDoThat [00:15:13] there is no justifiable exception imho. [00:15:24] let me check out the value, I haven't seen where the value comes from [00:16:03] I assume this is a bad idea because someone could exploit it with a local MediaWiki: message? [00:16:57] it is bad because it is an exploit, yes, but more importantly because localizers and sysops don't know about html escaping [00:17:08] so if they insert a special character that needs escaping, it will be inserted raw [00:17:22] e.g. If they do Red is color. [00:17:28] then bang goes your page [00:18:13] kaldari: In this case I'd say: either truly seperate the color in presentation, or don't at all. Just is so classic imho. [00:18:42] i.e. a sentence above and then a big value below [00:19:18] sounds reasonable [00:19:20] kaldari: a bit like this: http://aux2.iconpedia.net/uploads/3951818241085781941.png [00:19:32] top-line and shot out [00:19:53] its interesting how tech issues sometimes dribble down into good habits for other reasons. [00:21:15] I usually don't do any styling inline, but was trying to keep it simple. Anyway, good to know this is a bad idea :) [00:22:07] kaldari: Do you know LanguageQqx ? [00:22:21] (where it shows all the message keys as values in the interface) [00:22:26] like ?uselang=qqx [00:22:28] yeah [00:22:32] right [00:22:39] I've got a LanguageQqy locally [00:22:50] what's that do? [00:23:09] where it returns [00:23:12] as message value [00:23:17] to spot injection vectors [00:23:40] There's a few that always log something (like the footer) [00:24:07] that's handy [00:24:27] Yep [00:24:49] I imagine TimStarling might have something like that too. I don't remember where I got it from actually, might have gotten it from someone else, not sure. [00:25:19] I'm gonna grab some lunch and then I'll fix that up [00:25:33] well breakfast really :) [00:25:50] I need to remember to eat [00:26:08] qqx and qqy are not exactly good mnemonic names for language codes [00:26:29] I complained about qqq as well, when it was introduced [00:26:39] BCP 47 reserves all language codes starting with x- for private use [00:26:46] so you can have x-message-keys or something [00:26:50] yeah [00:26:52] * Krinkle renames file [00:26:56] :) [00:28:07] class LanguageX_script extends Language { function getMessage( $key ) { return ''; [00:28:15] http://krinkle.dev/wm-git/mediawiki/core/index.php/Special:JavaScriptTest/qunit?uselang=x-script [00:28:35] Interesting, just "LanguageQqy retrievedfrom" [00:28:43] improvement, there used to be more [00:29:11] ah, "lastmodifiedat" as well no wiki pages [00:29:39] eh, s/LanguageQqy/x-script [00:39:48] kaldari: That last patch set to https://gerrit.wikimedia.org/r/#/c/13902 didn't do anything [00:39:52] only the commit date [00:39:57] not even the parent [00:41:37] it changed it back to your previous version [00:41:58] for example, the files that are deleted are different than in patchset 4 [00:42:26] or rather the files that are added... [00:42:38] no BoilerPlate/BoilerPlate.i18n.magic.php [00:43:04] but yeah, there are no changes _to_ files [00:45:03] basically, the last 2 patchsets are: [00:45:11] me accidently reverting to patchset 2... [00:45:18] and then reverting to patchset 3 [00:45:39] so patchset 5 should be identical to patchset 3 [00:46:25] sorry for the mess [00:47:04] aha [00:47:06] ok [00:47:31] mediawiki history has its advantages - compared to gerrit patch set hisotry [00:49:52] definitely :) [01:08:03] since I made a mess of the dependancy, I'm just going to abandon my 'Example revision' change and resubmit it fresh (to avoid accidently making any further patch reverts on 13902). [01:11:04] all done [02:56:18] If anyone in the Bay Area is looking for something to do on the Fourth, I'm hanging out at home, I have a bit of food, I'd welcome some company :) PM me if you'd like an address! [10:00:11] Krinkle: I have sent Tomasz an email so we get a jquery.tipsy git repo on github [10:00:16] Krinkle: I am definitely going to fork it [10:00:20] apply our local hack [10:00:24] tag a new version [10:00:40] and then apply the patches laying down in bugzilla [10:00:43] How does Tomasz fit in the picture? [10:00:56] AFAIK he is the admin for http://github.com/wikimedia [10:01:12] Oh, someone on staff manages it? [10:01:18] though I might be wrong ahaha [10:01:32] that is definitely our official github page [10:01:38] great, that means we can finally ask someone who will likely listen to add other wmf staff to the org [10:01:47] hehe [10:01:52] I have no idea why we aren't in there [10:01:55] anyway you are on cc: [10:02:00] so we will be able to argue with Tomasz :-] [10:02:17] I asked him to create a virtual team made of you and I to maintain a new jquery-tipsy git repo [10:02:37] (not asking you to actually take over that plugin maintenance but I feel it is better to have you there just in case) [10:02:50] Just add us to the org in general, and then as 'repo collab' to that repo [10:03:00] make sure we do keep the courtesy of the author [10:03:05] fork it from https://github.com/jaz303/tipsy [10:03:08] well I am not familiar with github project rights, I guess Tomasz will sort it out for us [10:03:10] using the github "fork" method [10:03:14] yup definitely [10:03:18] so that we stay in the network and others kind find us too [10:03:33] I did a fair bit of gh rights with jquery [10:03:34] I dislike forking [10:03:44] but jaz303 never answered my email sollicitaitons [10:03:47] nor my github messages [10:03:57] so? [10:04:00] nor does he handle any pull issue requests, some being very minor / easy [10:04:08] so last resort is to fork it :-] [10:04:30] probably nothing personal [10:04:53] I see https://github.com/mediawiki has been emptied finally [10:04:59] there used to be some svn mirroring [10:05:12] yeah that was created by Avar to host a 'git svn' clone of MediaWiki trunk [10:05:17] and extensions as well iirc [10:05:24] Chad is managing it now [10:05:40] it is going to be a clone of the git repos we have in Gerrit [10:06:09] I think it is Chad going to work on that, probably using Gerrit build-in replication system [10:12:13] hashar: it would go to /wikimedia/mediawiki-core though, right? [10:12:27] makes more sense [10:12:50] (github doesn't support slashes in repo names) [10:13:46] and core.git is very bad for github as all forks/prs will be "core" [10:13:49] e.g. Krinkle/core.git [10:14:29] and avoids gh-org-managing duplication [10:16:34] core is going to be in mediawiki github project [10:16:41] we want it to be outside of wikimedia repo [10:16:46] to be able to add volunteers / non-staff [10:17:09] ? [10:17:14] for what reason [10:17:19] its a one way mirror [10:17:45] we might eventually honor github pull requests and reported issue [10:17:53] yes PR [10:17:59] but we wouldn't merge them into github [10:18:03] and will definitely have some volunteers to handle them [10:18:04] and issue tracker should be disabled [10:18:13] I agree with you :-] [10:18:16] github doesn't allow restriction of code review [10:18:26] anyone can leave inline comments and fork or ammend [10:18:28] anyway, there is nothing setup yet and the version 0 will just be a git mirror [10:18:32] AFAIK [10:18:40] the only thing that needs rights is to close it when its merged. [10:18:47] I am not even sure that is Chad being assigned to it nor if that project is in the roadmap yet [10:19:22] which github can even do automatically if you include "closes GH-123." in the commit message in gerrit. [10:19:40] or even just if the git commit hash remains the same [10:19:58] which it will since gerrit doesn't rebase but merge commits, so the commit will go into the repo [10:20:54] I'm just saying if we can, it'd be ideal to not maintain 2 github repos and have other stuff to set up. and core.git is a no go imho. [10:21:28] yeah I fully agree we should not have an issue tracker nor pull requests in our github page [10:21:30] I can set up a mirror from cron on toolserver to run every 30 minutes. Easy enough. And allow only PR, no issues or wiki. [10:21:39] but we might want to do that to try to get more volunteers involved [10:21:45] cause github as a huuuuugeeeee audience :-] [10:21:56] Sure, and github allows that. But you don't need an org for that at github [10:22:23] its like a wiki on github. you can mess around pretty big time in people's repos [10:22:24] and mirroring will be either done with Gerrit (preferred, instant replication) or Jenkins (through change-merged event). [10:22:54] and I am pretty sure we will use Gerrit replication system which make mirroring way easier (it would just be gerrit interacting with github) [10:23:03] hashar: fine by me, but if that takes until christmas, I'm asking a repo and set up 30 min replication. nobody cares about how its done. [10:23:19] will want to talk about it with Chad I guess [10:23:19] only 10 minutes work with the proper tools [10:23:28] we don't want to use the tool server for that [10:23:30] !b github [10:23:30] https://bugzilla.wikimedia.org/show_bug.cgi?id=github [10:23:31] ;-D [10:23:33] !bs github [10:23:33] https://bugzilla.wikimedia.org/buglist.cgi?quicksearch=github [10:23:40] getting fresh air brb [10:24:31] what the .. [10:24:32] !bs github [10:24:33] https://bugzilla.wikimedia.org/buglist.cgi?quicksearch=github [10:24:33] !bs [10:24:34] https://bugzilla.wikimedia.org/buglist.cgi?quicksearch=$1 [15:19:34] It looks like someone or something has disabled vector collapseNav? [15:19:36] (on wmf) [15:20:47] <^demon> wfm. [15:20:57] Hm.. just on en.wikipedia, resetting prefs fixed it [15:21:04] maybe I disabled it at some point [15:22:18] New patchset: Hashar; "(bug 34815) Use QUnit ?module= query parameter." [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/11830 [15:22:35] New review: Hashar; "rebased" [integration/testswarm] (master); V: 1 C: 2; - https://gerrit.wikimedia.org/r/11830 [15:22:38] Change merged: Hashar; [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/11830 [15:24:20] New patchset: Hashar; "update testswarm submitter" [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/14309 [15:24:56] New review: Hashar; "Included in Jenkins with I4766d16c. Deploying on gallium." [integration/testswarm] (master) - https://gerrit.wikimedia.org/r/11830 [15:25:08] Change merged: Hashar; [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/14309 [15:26:34] New review: Hashar; "Deployed on gallium." [integration/jenkins] (master) - https://gerrit.wikimedia.org/r/14309 [21:32:28] Reedy: around? [21:35:10] Ya [21:38:17] Reedy: could you pls set up bug 16655 and deploy it? (why did you choose just b & s actually?) [21:39:27] Per Tims comment [21:40:29] Certainly on enwiki (and probably the top 10 wikipedias) it's not safe to enable [21:41:19] just enable everywhere and disable on big wikis? [21:41:28] or enable on wikinewses at least [21:41:41] Wikinews I can do [21:42:07] wiktionaries? ;-) [21:43:00] I can do all non wikipedias without much issue I think [21:45:51] awesome, please do at least that, thank you [21:52:06] grrr [21:53:21] Why doesn't setting default to true and wikipedia to false have the desired effect? :( [21:53:49] <^demon|away> +wikipedia? [21:54:07] <^demon|away> Dunno, been awhile since I mucked about in InitialiseSettings. [21:54:19] Doesn't work in this section [21:54:24] That's for group overrides and stuff [21:54:27] Meh, just do it another way [21:54:41] <^demon|away> Ah, see it's been too long. [21:56:29] if ( $site !== 'wikipedia' ) { [21:56:29] require_once( "$IP/extensions/RandomRootPage/Randomrootpage.php" ); [21:56:29] } [21:56:33] Danny_B|backup: fixed [21:56:53] <^demon|away> Wow, if that's not a hack :p [21:57:45] 'wmgEnableRandomRootPage' => array( [21:57:45] 'default' => true, [21:57:45] 'wikipedia' => false, [21:57:45] ), [21:57:49] I did try that originally :( [21:58:05] <^demon|away> Maybe just 'wiki'? [21:58:07] <^demon|away> Dunno [21:58:33] 'wikipedia' => array( 'Wikipedia' => NS_PROJECT ), [21:58:33] 'wiktionary' => array( 'Wiktionary' => NS_PROJECT ), [21:59:23] there's worse code in those files ;) [21:59:56] I'll have a look in wgConf [22:00:03] Cause it sorta feels like a bug [22:00:58] // Fall back to default. [22:00:58] if( array_key_exists( 'default', $thisSetting ) ) { [22:01:03] That was easy - default overrides all [22:01:38] <^demon|away> ...that sounds backwards? [22:02:07] indeed [22:02:11] default is at the bottom [22:02:21] there are breaks in some of the above conditionals, not all of them [22:02:48] I'm wondering if a few more are needed [22:03:17] brb, drink [22:05:41] default => false, [22:05:46] dbname => true, [22:05:55] ^ we know those work, so it's the grouping that's the issue [22:13:59] Seriously [22:14:09] [0]=> [22:14:09] string(9) "wikipedia" [22:14:09] [1]=> [22:14:10] string(4) "wiki" [22:15:05] ^demon|away: I think you're right [23:03:48] <^demon> Gerrit about to go down for the upgrade to 2.4.2. Will let everyone know when it's back up. [23:06:04] \o/ [23:06:49] <^demon> brion: You're going to be in DC with us next week, right? [23:06:59] yep [23:07:31] <^demon> Yay. We'll have to grab a beer at some point :) [23:07:41] You're going to be drinking with us next week, right? [23:08:05] :) [23:08:18] <^demon> Indeed. I know quite a few places to drink in DC :p [23:09:06] excellent [23:55:57] <^demon> Gerrit's back up. Took a bit longer than I would've liked, sorry for the delay. [23:56:50] whee [23:57:02] let's all go spam click the rebase button [23:57:03] * Reedy grins [23:57:26] Reedy: You haven't been, you mean, for the past ten minutes? [23:57:53] I did it on one of my revs, knowing it'd make absolutely no difference