[00:02:16] RoanKattouw: right [00:04:02] RoanKattouw: 2 is also easily fixed I guess [00:04:07] Yeah [00:04:11] RoanKattouw: I can just show the success box on every change [00:04:14] They're both easy [00:04:20] Aye that would work too [00:04:24] RoanKattouw: is 1 not fixed already? [00:04:34] I don't know [00:04:42] RoanKattouw: earlier today I commited stuff that only checks the domain [00:04:44] Erik just showed the broken behavior to me an hour ago [00:04:46] Oh, OK [00:04:48] Good [00:04:51] Maybe Erik isn't up-to-date then [00:05:14] So if you put in anyprotocol://gitorious.com/lalalala it accepts that? [00:05:34] (I think Erik ran into the facts that it required https and required the tilde) [00:07:19] It never required a tilde [00:07:28] In fact, it would reject them on most places [00:07:29] Oh well, whatever [00:07:32] I'll tell Erik to svn up [00:07:36] That's prob what he ment [00:07:46] And if it's still not fixed then, I can address it [00:07:55] not any protocol [00:07:59] only http and https [00:08:09] But yeah, anything on the domain is valid [00:08:28] Oh OK, that's fine [00:08:31] (Why not git:// BTW?) [00:12:22] Oh, could add that [00:12:31] Not sure if we want that though [00:12:41] gregdek: ^^ [00:29:52] RoanKattouw: The email thing is confusing indeed, and so is your email about it :p How do you propose I handle sending the email to a whole bunch of people? [00:30:05] Use Enotif, or create some other job for it? [00:30:16] And how to split the emails over the jobs? [00:33:43] Create a job for it [00:33:49] No splitting needed [00:34:12] I was using Enotif as an example, pointing out it doesn't do any splitting [00:34:19] One Enotif job might very well send 3 million e-mails [00:35:31] ( JeroenDeDauw ---^^ ) [00:37:45] RoanKattouw: isn't sending 3M emails to much work for a job? [00:37:56] No, we don't care how long jobs run [00:38:04] On WMF at least [00:38:25] On 3rd party installs I guess you're right, it would be a problem [00:38:32] But we don't really have time to worry about that right now [00:38:36] Oh, they are not executed on page loads at all? [00:38:54] One job seems silly [00:39:05] In that case, all the work can just be done directly no? [00:39:07] On WMF we have dedicated job runner processes [00:39:23] No, because if you do it directly, some web client (user) is waiting for its completion [00:40:04] Ok [00:40:10] That's why stuff like template invalidation is put in the job queue, because you don't want to force the poor soul that edits [[w:en:Template:!]] wait for 2M pages to be reparsed [00:40:16] *to wait for [00:40:54] :D [00:41:20] I don't like creating something that only works well for the WMF setup if this can be avoided - any idea what would be a reasonable amount of mails to send per job? [00:42:06] I guess 50 would be fine? [00:42:09] *RoanKattouw is just guessing [00:42:22] I appreciate that sentiment though :) [00:45:36] Right - I'll do the jobs stuff tomorrow then, it's not needed right away after all [00:45:45] OK [00:45:52] And the # of mails can easily made configurable ofc [00:46:17] When is the deploy planned? [00:46:22] ASAP [00:46:29] Hopefully some time during the NOLA hackathon [00:46:35] Erik wanted to do it today, but I said no [00:46:54] Most of our ops staff is on a plane somewhere over Grand Canyon right now [00:46:59] Yeah, today would have been a bit rushed :) [00:47:10] So in NOLA we'll have everyone available [00:47:15] NOLA H is this zeekend right? [00:47:17] After NOLA, I'm on vacation for a week [00:47:19] Yeah [00:47:22] The ops folks are flying now [00:47:25] I fly out tomorrow morning [00:47:38] Are you in SF? [00:47:46] Currently yes [00:47:50] Cool [00:48:08] Then after NOLA I come home, and will be on vacation for the rest of the week [00:48:16] So I won't be working until like Oct 24 [00:48:21] That's why Erik wants to rush this deploy :) [00:48:28] Right [00:48:32] But if it slips, someone that's not me can deploy it instead while I'm on vacation [00:50:21] JeroenDeDauw: Why is there an API module for sending reminder e-mails? Would that be called from a maintenance script? Cron job? JavaScript? [00:51:49] RoanKattouw: the idea was to be able to handle all those. Greg initially wants some cron job, but UI integration later on, such as a "send reminder email" button, is also possible like this [00:51:57] OK [00:52:57] RoanKattouw: I'm quiting with dev work for today - it'd be awesome if you could list the issues I should or can still tackle somewhere, so I know what do to tomorrow. Right now I'm not sure what to do next. [00:53:17] Did my issue #2 get fixed? [00:53:26] RoanKattouw 2) When you submit that form, there is no thank you message or ANY indication that your data has been saved, you just go back to the form (which retains your data, so nothing changes) [00:54:46] No, will do that tomorrow, unless you want to take care of it already [00:55:14] OK, that's fine [17:58:48] I need a jQuery plugin to be added to the repo: resources/jquery/jquery.cycle.all.js for generating slideshows with Selenium Unit testing [17:59:02] should this be added to: resources/jquery/ [19:40:14] lo