[00:06:41] Dev available? [17:17:03] neilk_, is the upload campaign stuff all live? Or was it supposed to be in the aborted deployment earlier this week? [17:19:10] D'oh! [17:19:15] I was supposed to schedule a deployment fro that [17:19:18] That has to be like today [17:19:44] Seems it isn't [17:19:45] I can replace my HTTPS fixes deployment with UW I guess [17:20:28] neilk_: So what's the status of the UW deployment? Do you want me to do it today? I have an opening at noon-1pm [17:21:18] RoanKattouw, the upload campaign stuff looks broken [17:21:26] I've seen bug reports yeah [18:15:05] Nikerabbit: I'm doing Translate review, could you EXPLAIN a query on the TWN DB for me? [18:15:08] Nikerabbit: explain select page_id, page_namespace, page_title from page, revtag, revtag_type where page_id=rt_page and rtt_id=rt_type and rtt_name='tp:mark' group by page_id; [18:17:34] neilk_: Ping re UW deploy [18:34:52] RoanKattouw: sure [19:08:34] Oh, interesting [19:08:48] $wgInternalServer has been broken on secure ever since the introduction of prot rel URLs [19:10:13] RoanKattouw: Any news on the upload wizard deploy? [19:10:35] Neil is working on a patch [19:10:39] ok [19:11:02] multichill: I want to fix it so you can add multiple licenses that aren't in default config, and a couple of other things need fixing [19:12:20] Ah, ok [19:12:41] And we're getting some useful feedback at the wlm/uw page [19:19:22] haha, in secure.php: [19:19:24] // In future we may want a separate item for preferred external interface URLs [19:19:25] *RoanKattouw feels vindicated [21:16:03] neilk_: So, last time, I promise: how close are you? Cause I'm about to go to bed [21:16:19] RoanKattouw: I was waiting for your review [21:16:22] or someone's review [21:16:26] Oh that [21:16:29] That code looks fine [21:16:46] *RoanKattouw goes back and looks at it again [21:17:46] OKed [21:17:51] RoanKattouw: danke [21:18:14] So what's the status of UW now? Is it deploy-ready? [21:25:24] neilk_: ? [21:25:36] I'm trying to make the patch now [21:25:36] I'm asking cause I'm about to go to bed [21:25:38] sigh [21:26:00] I did something bad to my 1.17wmf1 sorry [21:26:04] starting over [21:27:32] Hmm [21:27:42] awjr: Would you mind helping Neil deploy his stuff? [21:27:53] I wanna go to bed [21:27:55] it's just merging 5 changes [21:27:57] sigh [21:28:01] I apologize [21:28:11] awjr -- I'll have it for you in 10-15 min [21:28:26] It's OK -- I would usually stay up later but I'm tired [21:28:28] np [21:28:40] I have the same problem unfortunately I am coming in later and this makes it worse [21:28:41] neilk_, RoanKattouw: i've got to make sure i get the fundraising deployments done first [21:28:56] but assuming i get those done soonish i'm happy to help [21:29:05] neilk_: I don't start work at 7am when I'm in the office either :) [21:29:24] well, I only got here at 11am which sort of sucks for coordinating a window with you [21:29:47] awjr: Alright, you and Neil can figure it out then. Your part will just be svn up and scap, Neil has commit access to the branch and is preparing a commit [21:30:00] oh [21:30:06] right [21:30:11] that's easy [21:30:12] neilk_: Yeah. 11am, noon and 1pm work as deployment times, but that pretty much means you have to already have finished the patch/merge when you come in [21:30:18] there is no schema change this time, but there are core changes. [21:30:19] neilk_: just lmk when you're ready [21:30:25] awjr: sure thing [21:30:30] awjr: Except if his code breaks something and you have to sync out followups :) [21:31:33] neilk_: your code better not break anything :p [21:31:54] Well today is a good day apparently [21:32:05] I deployed almost a thousand lines of core changes [21:32:08] ! [21:32:30] and apart from one PHP warnings that was a stupid merging oversight ($foo['bar'] vs $foo->bar , it had changed in the meantime) it was totally fine [21:32:36] *warning [21:32:44] awesome [21:33:18] ok, well, neilk_, if things go south with the deploy i dont really have time to do much damage control, so i'll likely just revert things [21:33:44] understood [21:33:47] ^awjr [21:33:50] word [22:15:45] awjr: I think 1.17wmf1 is in a good state, finally. I had to resort to brute force that's a bit hard to follow, just a delete & rebranch [22:16:19] neilk_: ok - it's likely going to be a while before i can get to it [22:16:22] maybe an hour or two [22:17:37] heh [22:17:41] I'm about for a bit if necessary [22:17:49] awjr: ok but be aware that 1.17wmf1 has my committed changes [22:18:01] Reedy -- can you deploy? [22:18:08] I can [22:18:32] neilk_, Reedy - i'm in the middle of deploying changes for CentralNotice just fyi [22:18:42] uhhh [22:18:47] was this a bad time to commit then? [22:18:52] Yeah, why can't we scap a specific directory [22:19:03] As long as UW isn't svn upped, should be fine [22:19:04] no that's fine as long as no one svn up'd [22:19:09] cuz i sure didn't :) [22:19:21] it'd take years if you had [22:19:27] heh [22:21:23] Reedy, awjr: I believe there may be an issue. Hold off deploy for a minute. [22:21:34] I think my caches have libraries that aren't really in 1.17wmf1 [22:21:45] deploy of UW that is [22:28:53] well this is fan-fucking-tastic, there's a define() 2x in 1.17wmf1... [22:29:13] :( [22:30:28] so you can't run 1.17wmf1 locally with warnings on [22:30:39] What? :/ [22:31:06] Reedy: I just tried it -- we define the PROTO_* things twice now, it seems -- once in GlobalFunctions and once in Defines [22:31:16] which is the right place? [22:31:42] trunk has it only in defines [22:31:50] odd [22:32:44] neilk_, it looks like they *were* in GF before [22:32:49] http://www.mediawiki.org/w/index.php?path=%2Fbranches%2Fwmf%2F1.17wmf1%2Fincludes%2FGlobalFunctions.php&title=Special%3ACode%2FMediaWiki%2Fstatus%2Fnew&status=new [22:32:59] In neither of those 2 revs is roan adding them to GF [22:33:07] oh, that's only new [22:33:08] bleh [22:33:14] still only 2 [22:33:18] or not [22:33:22] BLEUGH [22:33:25] *Reedy kicks CR [22:35:28] http://www.mediawiki.org/wiki/Special:Code/MediaWiki/93886 [22:35:34] There he added them [22:36:04] Can you kill them from GF and see how it goes? [22:40:03] Reedy: ok [22:40:03] neilk_, they're not in my local GF [22:40:13] *defines [22:40:34] but then [22:40:36] i haven't svn upped it [22:40:39] *Reedy whistles [22:43:09] The one in GF is missing canonical [22:43:16] Want me to remove them? [22:43:52] does scap essentially just run svn up on all of the apaches? i ask because there are a lot of unversioned files and locally modified files in /home/wikipedia/common/php on fenari and that makes me nervous [22:44:53] Doesn't do svn up, no [22:45:19] rsyncs [22:45:24] scheisse [22:45:28] if rsync -a --delete --exclude=**/.svn/lock --no-perms \ [22:45:28] 10.0.5.8::common/ /usr/local/apache/common-local [22:45:32] so i should be worried about all the locall modified files etc/ [22:45:34] I've had this issue before [22:45:48] No, you shouldn't have to be [22:45:55] But that doesn't means you shouldn't have to [22:46:05] hmm [22:46:09] what is the appropriate course of action? [22:46:16] i wish we could just sync individual dirs :( [22:46:23] what I wrote makes no sense [22:46:29] that's true [22:46:32] No, you shouldn't have to be [22:46:35] but i *think* i understoo [22:46:36] d [22:46:36] But that doesn't means you don't have to [22:46:45] heh [22:46:57] In theory, people should leave the directories in a "clean"/working way, where all changes are live [22:47:00] Reedy, awjr: okay I'm back on the 3rd floor [22:47:16] that is a nice theory [22:47:28] neilk_: i'm starting feel like i'm not gonna be able to help you with deployment today [22:47:32] now it seems that catrope added all these PROTO things as part of the https:// crap [22:47:38] Yeah [22:47:38] this CN deployment is taking forever and I have more FR deployments to do [22:47:57] awjr, yeah, but I've been caught by that before, and in that case, there was no outstanding changes according to SVN, but there was some unpushed changes in common/initialise settings :( [22:48:06] doh [22:48:06] neilk_, I think we can kill the GF ones [22:48:07] awjr: well, then I'm leaving the changes in 1.17wmf1 for Roan to do 1st thing tmw? [22:48:14] that's fine with me [22:48:18] it's missing one define also [22:48:29] Reedy Yeah, but the question for me is -- is there anything that loads GF but not defines [22:48:37] Reedy: otherwise why was this done [22:48:52] Can we assume everything must load Defines? [22:49:09] so Reedy, you're saying i should just go ahead and scap? [22:49:22] awjr, pretty much [22:49:42] k [22:49:49] neilk_, yeah, WebStart loads Defines [22:49:53] hold on to your hats [22:50:02] Reedy: sure, but maintenance...? [22:50:31] if ( !defined( 'MW_COMPILED' ) ) { [22:50:31] require_once( "$IP/includes/Defines.php" ); [22:50:31] } [22:50:32] Yup [22:50:34] Reedy: I'm not sure that I know all the entry points. There's index.php, api.php, and then maintenance scripts. [22:50:49] and then img_auth.php and so on [22:50:56] index and api load webstart [22:51:07] as does img_auth [22:51:12] as does load.php [22:51:38] It seems they might have been doubly introduced during merges and moves [22:51:44] ok [22:52:03] I think it's safe to put them only in Defines, just grepped for any entry point that doesn't use it and I see none. [22:53:58] ok kids, i'm /almost/ done on fenari i just hav a few more changes to push out [22:55:06] *Ryan_Lane scaps [22:55:24] *Reedy points at Ryan_Lane [22:55:29] IT'S HIS FAULT [22:55:35] muahahahaha [22:56:48] Ryan_Lane: i'm in the middle of doing deployments [22:57:04] I didn't actually scap ;) [22:57:12] phew [22:57:19] i was about to come down and punch you in the fac [22:57:20] e [22:57:25] :D [22:57:33] hehe [23:10:45] hmm Reedy, i forget - to push out config changes do i just sync-file also or is there some other mgic? [23:11:16] In theory, scap should've done it [23:11:20] oh [23:11:21] ha [23:11:23] im not scapping [23:11:28] just sync-filing a coupla change [23:11:28] heh [23:11:28] s [23:11:35] just use sync-file wmf-config/InitialiseSettings.php [23:11:39] groovy [23:34:48] Reedy: I see you just removed the PROTO_* from GlobalFunctions. I some time figuring out the change that should have removed them, made a nice commit comment. [23:35:35] Reedy: I dunno, I feel like I am just wasting my time when I do careful commits, while you & Roan patch in an eyeblink. Am I just inexperienced or something [23:36:20] also I got distracted for about 10 minutes while someone was talking to me, I guess [23:36:46] do you just use a tool which alerts you that someone else is changing the file you're working on? [23:48:45] awjr: hey, Jeremy P did not give us a preferred username [23:48:51] so we are going with jposte...... [23:49:00] you know, first initial + last name. Hope that's ok awjr [23:49:07] sumanah: that is fine, im sure