[02:43:33] !log re-enabled snaprotate, bug 27471 does not involve data loss [02:43:33] --elephant-- I don't know anything about "log". [15:23:31] do we have somebody to deploy r82307? [15:23:41] it's pretty annoying as it breaks a lot of SVG images [15:33:25] Without poking the ops guys... I'm not sure [15:38:28] Bryan, might be able to poke maybe mark to do it [15:38:31] Has it been reviewed? [15:38:49] I signed it off [15:38:57] haven't tested it though [15:39:00] !r 82307 [15:39:00] --elephant-- http://www.mediawiki.org/wiki/Special:Code/MediaWiki/82307 [15:39:21] We do need to tie down the review procedure [15:39:41] Needs merging out also [15:55:06] hi folks [15:55:45] hi everybody [15:57:45] notes prepped here: http://eiximenis.wikimedia.org/TestFrameworkDiscussion2011-02-17 [15:57:46] robla: Hi Rob, I have completed 170 sites from the list http://eiximenis.wikimedia.org/1-17-allwikis [15:58:07] Nice [15:58:41] nadeesha_calcey: great, thanks! any of them broken badly? [16:00:06] (as an aside, Reedy and g uillom both pointed me to better sources of those comprehensive lists, which I'll use next time) [16:00:18] heh [16:00:35] robla, always remember programmers are lazy, and there's usually a better/faster/easier way ;) [16:01:02] robla: I found this http://sq.wikipedia.org/wiki/, when yoou go to recent changes, the content appears once and then some text displayed [16:01:31] Seems fine for me on Chrome [16:01:41] robla: 'Make a donation to Wikipedia and give the gift of knowledge!' text [16:01:55] robla: I tested this in IE7 [16:04:20] hmm....yeah, I'm not seeing a problem in Chrome either. I don't have IE handy [16:04:52] nadeesha_calcey: is everything you found filed in Bugzilla? [16:05:47] meanwhile: mglaser, how are things going for you? [16:06:15] robla: I have mark the list(pass and fail) in http://eiximenis.wikimedia.org/1-17-allwikis. I can file the bugs in bugzilla [16:06:20] robla: i have updated the documentation about test styleguides (action item) [16:06:50] (fyi, Priyanka probably isn't going to be around for a little while. she's helping out with the Virginia data center setup, then she'll be on vacation) [16:07:26] ah, ok [16:07:54] mglaser: great! [16:08:48] robla: next thing i want to do is push the clean database mechanism a little further [16:09:33] i.e. testing and also provide the possibility to start with a freshly installed mw database [16:09:58] (which doesn't need to be in svn as a file) [16:12:01] nadeesha_calcey: that would be good. it will get more eyes looking at the problems. [16:12:34] mglaser: that sounds great. need any guidance on anything while we're here? [16:13:10] robla: I ll log the bugs [16:13:14] @robla @hexmode I think we could also have a look at using the framework with phpundercontrol, but i don't know what would be reasonable next steps. [16:16:24] hexmode's main focus now has shifted toward his Bugmeister work. he can probably provide advise, but probably won't be pitching in much [16:16:36] robla: shall we continue testing the remaining items in the list ? [16:16:44] nadeesha_calcey: yeah, that would be great [16:17:37] related to this meeting: we're going to be hiring a Senior QA Engineer: http://wikimediafoundation.org/wiki/Job_openings/Senior_QA_Engineer [16:18:57] as you've seen lately, the current batch of WMF employees haven't had as much time to focus on this stuff. once we get this new person in place, they'll be in a good spot to push things [16:19:06] @robla: is there someone else i could work with in implementing it? basically, I have server access, so i could try my luck myself, but i don't know this is reasonable... [16:19:55] robla: ok we will mark the test results as we did today and will log the bugs if found [16:20:58] *robla ponders mglaser's question [16:22:26] mglaser: I don't want to put anyone on the spot right this instant, but I've got an idea or two [16:23:24] as I was saying before, hexmode may very well be able to at least provide guidance [16:23:38] @robla ok [16:24:37] mglaser: could you send me email w/ what you're having trouble with? [16:25:14] @hexmode yes, I will [16:28:45] alright, I guess one last thing we should talk about before we go [16:29:19] it looks like we've sorta drifted into two separate meetings with this one [16:29:58] ...and maybe we should split it up into two, since from a timezone perspective this time isn't really great for anyone [16:30:27] except me! [16:30:44] well, I guess there is you :) [16:30:52] "Western Hemisphere" and "Eastern Hemisphere" [16:31:29] we're also probably running low on stuff to discuss for a while until we hire the new QA person [16:31:38] hm.... time is ok for me. [16:31:58] what about taking a break for a few weeks? [16:31:58] QA? w00! [16:32:40] mglaser: yeah, I think that may be best. I'll figure out a different time to check in with nadeesha_calcey and crew [16:33:06] (since it's late for them, and early enough that I'm not 100% awake) [16:34:17] How about we plan on talking on March 10? [16:34:26] (at this time for now) [16:34:57] that's ok for me [16:35:07] robla: current meeting time is 9:30 PM - 10:30 PM for us, which is manageable for us [16:35:48] ok....we'll leave it at the same time, then [16:36:31] ok...anything else we should talk about here? [16:37:14] not from me. [16:37:18] robla: if you are proposing a better time, we can switch to that [16:39:10] robla: do we have any task after finishing the current task? [16:40:15] nadeesha_calcey: not right away. let me have a chat with Matt and Mangala about that [16:40:47] robla: okay [16:41:39] anything else to discuss here? [16:42:45] robla: nothing from my end [16:44:49] alright...thanks everyone! [16:45:21] thanks [16:45:28] thanks [18:12:40] what is carrie's IRC nick? [18:12:52] ArtemisRich [18:13:02] RoanKattouw_away: so I thought... [18:13:26] damn those people and their unintuitive / difficult to spell IRC nicks [18:13:33] She's not online [18:13:43] yeah [18:14:09] *RoanKattouw_away goes back to watching TV [18:28:31] alolita1, robla ? [18:28:59] guillom: oh...hey, sorry, I should have cancelled today [18:29:11] sgrmbl [18:29:15] oh, well [18:30:02] dinner, then [19:03:04] jorm: hey - turns out Kul has all the firefly episodes on Netflix streaming [19:03:18] thank you again for offering to lend him those DVDs [19:03:28] harh. i brought the discs in and they're on your desk now. [19:03:36] but no worries. i forgot i was at home yesterday. [19:07:01] dude, thanks! [22:50:46] hi RoanKattouw and all! [22:50:57] *aude is on http://wikimaniateam.wikimedia.org/ private wiki and it's very broken [22:52:30] Hmph [22:52:36] Seems to be on 1.16 [22:52:38] http://dpaste.org/vOjJ/ [22:52:39] Broken how? [22:52:52] no, special version says 1.17wmf1 (r82223) [22:52:55] That's not 1.16 code [22:53:10] Dammit I've seen this before, and I knew what caused it once [22:53:12] it's in the sidebar, and there are two search boxes [22:53:18] two logos [22:53:48] RoanKattouw: I forgot to get that in BZ: http://mediawiki.pastebin.com/eiiaVQdn [22:53:52] http://i.imgur.com/NrYys.png [22:53:57] aude: I need a user account to be able to debug [22:54:12] ok, i'll see who can give access [22:54:48] There's probably something wrong with MediaWiki:Sidebar [22:54:54] Alternatively, you can help me by pastebinning the content of that page [22:55:35] the error is on every page [22:56:08] I know that [22:56:19] I was asking for the contents of [[MediaWiki:Sidebar]], sorry if that was unclear [22:56:31] brion: ping [22:56:42] brion could get you access [22:56:59] I can give myself access if I really need to [22:57:12] But before I do that, could you get me the content of the sidebar page first? [22:57:36] here's the sidebar: http://dpaste.org/iI7g/ [22:58:21] ** 2011|2011 [22:58:28] Try changing that to ** 2011 [22:59:47] If that doesn't work I have a workaround [22:59:50] i'm not an admin so can't [23:00:01] Hmph [23:00:04] I'll just deploy the workaround then [23:00:13] ok [23:05:57] RoanKattouw: problem fixed! [23:06:10] Yay [23:06:16] thanks! :) [23:36:56] RoanKattouw: I see Bryan did a cross-wiki search for document.write. Any idea how he did that ? [23:37:02] http://toolserver.org/~bryan/stuff/document.write.html [23:37:09] No idea [23:37:31] Data is not available on TS and he doesn't have direct DB access [23:37:35] And AFAIK he's not a TS root [23:37:46] (Even then, I think that doesn't help for getting content) [23:37:53] indeed [23:37:55] So maybe he just grabbed the content of each page through action=raw [23:38:01] I do recall the script taking forever to run [23:38:27] And since search is able to do that kind of search, looping through api-calls isn't possible either [23:38:34] is _not_ able* [23:38:41] Yeah [23:39:05] Hm.. indeed. Let php download all wikis MediaWiki:*.js pages. get titles from database [23:39:07] possible [23:42:47] RoanKattouw: I've seen things before but I am still stunned by the age of some javascript. .wikitable class and functions like importScript or appendCSS are defined in a lot of wikis' local resources. [23:43:00] Yeah some stuff is really old [23:43:25] I think my/our Tour de Wiki will save de servers some bandwidth as well. [23:43:37] average 10 kb removed of which 7 for all anonyous users. [23:43:45] by my actions. [23:43:59] * all wikis * number of hits. [23:44:04] gotta be noticable in stats perhaps. [23:44:31] although mw.config.get() in return adds a few bytes. [23:45:02] I was also thinking, you should be able to make a wiki quite a bit faster (front-end performance wise) by making all of its gadgets RL-compatible, optimizing site scripts, et [23:45:03] c [23:45:13] vs. what it's like now [23:45:19] RoanKattouw, stop generating yourself projects ;) [23:45:37] haha [23:45:41] I'm generating him a project [23:45:50] *Reedy wonders if RoanKattouw now has 2 todo lists [23:46:03] I already had, still do. only 742 wikis left to inspect throroughly [23:46:10] whatever the spelling of that word maybe [23:46:17] thoroughly [23:46:28] thaurowli :) [23:46:36] well [23:46:40] It's much easier to spell [23:46:45] and means essentially the same thing ;) [23:46:51] *RoanKattouw wonders whether he'll be allowed to use Krinkle as a todo list or whether he'll be assigned to other things instead [23:48:10] There's most like loads of shitty code still around [23:49:09] Yep, many non-wikipedia projects just copy and paste loads of crap [23:49:14] stuff that never ever worked. [23:49:29] or was on the edge of deprecation when copied and then not updated on the copies [23:50:06] ie. wikquote copying a function for featured articles javascript in the interwikis [23:50:11] Certainly wherever it said about having centralised gadgets makes a load of sense [23:50:15] which doesn't even exist on wikiquote afaik. [23:50:36] Right now I'm replacing most gadgets I see with a mw.loader.load call. [23:50:50] Yeah I was thinking about how RL 2.0 should ideally support centralized Gadgets [23:50:59] Or somehow tweak Gadgets to allow it [23:52:01] global centralauth preferences, Meta's MediaWiki:Gadgets-definition-global, ResourceLoader support for loading wiki pages as modules (like it does for user scripts and site scripts, RL support to do this cross-wiki (to a predefined wiki, eg. metawiki) [23:52:57] so that at some point I can set tick a checkbox in global prefernces and RL will load a gadget script from meta on all pages across the wmf wikis, minified in 1 request. [23:53:32] Yeah [23:53:47] I need to talk to Alolita about getting Trevor, you and me to work on ResourceLoader 2.0 :) [23:54:00] I think, although RL minification saves a lot, all importScript stuff is still uninified and extra http requests. [23:54:07] RoanKattouw, can we get v1 out and stable first? :D [23:54:11] Maybe not with *that* feature but at least with the stuff that we talked about when designing 1.0 and went "meh, let's defer this to 2.0" [23:54:13] It's out [23:54:16] And mostly stable :D [23:54:29] heh [23:54:40] Well, it's not like we can say it took the site down to any extent [23:54:43] and client side registering, zero bugs, perfect interface and world peace [23:54:45] anythign else ? [23:54:54] Solve world hunger [23:54:55] Aids? [23:54:59] Errm [23:55:03] Malaria.. [23:55:07] That's old fashioned [23:55:19] Didn't we solve that in another reality ? [23:55:19] Affordable housing in SF. Clearly that's something that'll never happen [23:55:33] Supersonic airplanes [23:55:46] RoanKattouw, I can sell you a cardboard box... [23:55:58] (Bryan says he's working on some aerospace thing that ought to be able to get you from Frankfurt to Sydney in 2h... in 2050) [23:56:03] *Krinkle can't wait to see the new JSLint-clone that will not suck that jQuery developers talked about. [23:56:24] We won't have retired by then [23:56:28] so all is good ;) [23:56:48] MediaWiki version 2.0 [23:56:52] Clearly *that* will never happen [23:57:55] If we call the current 1.X, when X >= some reasonable size, that == 2.0 [23:58:08] You know you'll never get consensus ; [23:58:25] maybe when we move all the classes to a namespace [23:58:37] Probably 2.0 would be a rewrite without backwards compatiblity, a new parser and lots of other things [23:58:44] We've gotta bring PHP support minimums up... [23:58:44] ie. how it would be if we wrote MW today [23:58:45] move configuration variables out of globals [23:58:47] Alright, time to end this week of 13-hour working days and 6-hour nights, and get some serious sleep [23:58:56] but since WMF could never use it, it would never be programmed thus, there will never be a 2.0 [23:58:56] to one super global? [23:58:58] get rid of all global objects in favour of a central context object [23:59:01] Store config in DB! [23:59:07] (or other backends) [23:59:08] like in SecurePoll [23:59:18] maybe we can do all that at once and call it 2.0 [23:59:22] but the only problem is... [23:59:33] when you go to 2.0, users expect to see something different [23:59:34] It'd take ages to do and the deployment would be reminiscent of the past few months [23:59:40] and all of those things are invisible to the user