[01:25:08] who's watching the election in SF? [01:27:06] werdna: see pm [01:41:58] kaldari, does the change to modules/overlay/ext.echo.overlay.css actually do anything? [01:42:08] the other file seems to fix the overlap issue [01:43:01] yes, it makes sure that the 'All notifications' link is always flush to the left and cleared of any other content [01:43:21] it's minor, but it does matter in some situations [01:43:55] Ah yes, I see the difference [01:44:21] Krenair: These are just stopgap measures though... we're going to be doing more extensive CSS/layout changes as soon as Vibha finalizes the designs [01:44:59] merged. now, where was that bug... [01:46:40] kaldari, is this fixed now? or do you want to keep it open until these future changes are done? https://bugzilla.wikimedia.org/show_bug.cgi?id=41814 [01:46:59] Oh it looks like you closed it while I was typing :) [01:47:10] that one is fixed now [01:49:38] One of the other important things was https://gerrit.wikimedia.org/r/#/c/32084/ [01:49:55] I gave that +1, however [01:50:08] If we want it to be configurable, perhaps a good idea would be to do [01:50:15] $wgLiquidThreadsNotificationTypes[] = 'echo'; [01:51:23] That way you can $wgLiquidThreadsNotificationTypes = array( 'standard' ); in config if you want it to use the normal LQT notifications system as well as echo, or array() to just use echo [01:53:36] (posted that on the change) [02:01:09] just looking at ExtensionDistributor [02:02:13] you can run "git ls-remote refs/heads/master" to get the current master version to see if it is different to the cached version [02:02:21] but guess how it is implemented [02:04:19] git clone, git rev-parse head [02:04:27] what do i win? [02:05:32] no, not quite that bad [02:05:41] but it does appear to download every ref [02:05:56] it just filters them on the client side [02:07:12] TimStarling: https://github.com/mmueller/supybot-git/blob/master/README.md [02:07:26] see "how notification works". headdesk. [02:07:35] "Warning #1: If the repository is big and/or the network is slow, the first load may take a very long time!" [02:08:26] yes, pretty much the same as how ED is already [02:08:47] but I want to make something that's a bit less fragile than how ED is already [02:10:06] then don't use --porcelain! [02:10:10] couldn't resist. sorry. [02:11:00] don't know what you're talking about [02:12:00] "a bit less fragile" / git porcelain. it's a terrible pun. [02:12:18] right [03:07:36] it's actually faster to get the current head ref from the github mirror than it is to get it from manganese [03:08:05] github has an API for that, and I guess it isn't implemented in the stupidest possible way [03:11:36] not suprising [03:11:53] it's quicker to do blames on github, take the sha1 and go back to gerrit to find the commit... [09:56:44] I wonder if there is a way to extract the article's abstract by using the api? [10:19:18] Vaio: what abstract? [10:19:34] The first paragraph of the article [10:20:47] no [13:52:05] zeljkof: https://www.mediawiki.org/wiki/Git/aliases [13:59:56] hashar: zeljkof: https://github.com/Krinkle/dotfiles [14:03:40] zeljkof: or https://github.com/hashar/alix/blob/master/bashrc [15:20:10] hashar: zeljkof: http://etherpad.wmflabs.org/pad/p/CiSprintNL12 [16:01:28] chrismcmahon: around ? :-] [16:01:28] chrismcmahon: we are going to call you [16:01:41] hi hashar [16:01:47] zeljkof calling via hangout [18:12:00] chrismcmahon: j^ : we need to get to the bottom of the IE7 stuff. chrismcmahon, is there a bug for this? [18:12:21] robla if there's not I'll make one [18:12:44] thanks [18:13:04] mdale: you're not seeing any problems with IE7 in your testing? [18:13:37] right now IE7 asks if you want to save the file or open with a different application [18:13:57] with webm files .. or with all files? [18:13:57] will be back in a bit. [18:14:05] thats if you have no java installed and you get a notice [18:14:10] webm, checking ogg [18:15:09] but my IE7 plays the Folgers.ogv file [18:15:38] ok....so, presumably transcoding should fix that, right? [18:16:02] j^: should the thumbtime bug be fixed now? https://bugzilla.wikimedia.org/show_bug.cgi?id=39014 [18:16:39] right. new webm file is not yet transcoded [18:17:55] New patchset: Stefan.petrea; "Fixed debianization." [analytics/udp-filters] (master) - https://gerrit.wikimedia.org/r/32137 [18:19:58] so IE7 is behaving as expected live on commons [18:20:40] robla: for that i have to disable the temporary workaround, did not want to take that out in case we have to role back [18:20:41] chrismcmahon: ++ [18:20:52] but as things look ok can remove that now [18:45:59] bsitu, I'm not sure what you mean by " the extension global variable initialization to application config file" [18:46:41] Initialise $wgLiquidThreadsNotificationTypes in the extension itself, default to array( 'standard' ); [18:47:04] If Echo is enabled, append 'echo' instead of setting it to array( 'echo', 'standard' ); [18:47:23] That way you can do $wgLiquidThreadsNotificationTypes = array(); in the config file to turn off standard notification [18:47:47] Your change forces it to at least use standard while echo is on. This should be configurable. [19:14:30] robla / chrismcmahon / j^: I'm going to have dinner now, and I'll waiting for you guys to give me the green light to publish the announcement about TMH. [19:14:51] j^: chrismcmahon: seeing anything particularly worrying? [19:15:17] guillom: yes, if you don't mind coming back after dinner, it'd be good to have a little more time for the dust to settle [19:15:24] sure [19:15:39] guillom: I'd say after you come back, go for it unless we explicitly ask you to hold back. [19:15:46] sounds good [19:16:07] j^ robla the one thing I'm unclear on is how transcoding jobs work. the Market Street vid shows "Added to job queue " and no sign of processing [19:23:17] chrismcmahon: transcoding all videos on commons will take some time [19:24:57] chrismcmahon: video of the day has one transcode already http://commons.wikimedia.org/w/index.php?title=File%3ARed_Square_military_parade_November_7%2C_1941.ogv [19:29:32] Krenair: Sorry about the late reply. I think either way works, they end up having to modify the same file for configing. I am hesitant about initializing it with an empty array though for turning off standard LQT link, there would be no link at all if echo is disabled [19:30:33] Well then the wiki should only do that if echo is enabled, but let's leave it up to them [19:33:23] chrismcmahon: I'm not seeing anything worrying on commons VP or enwiki VPT related to TMH, but it would be good to investigate this further: http://en.wikipedia.org/wiki/Wikipedia:VPT#Wikimedia_Error_editing_fairly_large_article [19:33:37] chrismcmahon: could you see if you can repro, and then file a bug if you can? [19:37:59] looking [19:44:23] guillom: Don't! :p [19:44:49] There's 1 major fatal being seen in production, another one I spotted and a couple of PHP warnings [19:44:59] https://bugzilla.wikimedia.org/show_bug.cgi?id=41860 [19:53:57] idle question: anyone know what User:Animum/easyblock.js is exactly? [19:54:38] Reedy: one of those looks like we're not handling the case of someone embedding a deleted file gracefully [19:55:25] ...not that that is a great thing, but it's probably not going to be embarrassingly prominent [19:56:03] <^demon> chrismcmahon: A script in someone's user space on some wiki. [19:56:06] <^demon> Could do anything. [19:56:28] <^demon> (You could make [[User:Foo/poop.js]] if you were User:Foo) [20:00:51] ^demon: I often get non-fatal errors in IEs in enwiki from that bit of code, and I don't understand why some user script would show up universally when I have zero custom js as my own user. [20:02:01] there'll be some weird transclusion tree... [20:02:02] Reedy: ok :) I'm holding for now. I'm around for ~1 more hour, so if we can't be sure we want to publish by then, I'll ask Matthew to do it. [20:02:52] thanks guillom. I've gotta head out here for a meeting, but I'm comfortable with whatever plan you all decide on [20:03:19] ok [20:05:28] robla I can reproduce the issue at http://en.wikipedia.org/wiki/Wikipedia:VPT#Wikimedia_Error_editing_fairly_large_article, but I don't think it has anything to do with TMH [20:05:40] or IE8 for that matter [20:06:20] seeing if I can repro in FF [20:07:52] yep, repro in FF too [20:08:19] straight up READ_TIMEOUT [20:10:26] It's a big article [20:10:51] Though, they did say section editing [20:11:05] yep, Reedy there are actually two failure modes, on a super-large save, there is no error at all, the browser just hangs forever even though the page has been saved. [20:11:09] mmm [20:11:19] It's not that big. Obama is twice the size [20:11:19] ;) [20:11:22] Reedy: on the section edit, I get the ERR_READ_TIMEOUT msg. [20:13:22] Reedy: one thing about that page that seems unique to me is the banner on it that says "This article or section is currently undergoing a major edit by the Guild of Copy Editors...." [20:14:04] I'll BZ it, I think that would be nice at least. [21:55:30] chrismcmahon: I have to rush out the door in a few minutes, but I wanted to bring your attention to https://www.mediawiki.org/wiki/User_talk:Cmcmahon%28WMF%29#Global_AbuseFilter_on_beta.wmflabs.org [21:55:39] chrismcmahon: a community member wants to test out Global Abuse Filter [21:55:52] and needs help understanding that probably beta labs is not the best place to do that [21:55:57] chrismcmahon: and probably they should set up a labs instance [21:56:06] chrismcmahon: can I ask you to help with that? [21:56:11] sumanah: looking. but yeah, we don't have AbuseFilter up in beta afaik [21:56:15] sure