[08:16:40] hello [17:18:45] RoanKattouw: .loader.register supports isFunction for dependancies, still used ? [17:19:19] I dont think it's used but I'm holding off removing it because I have no idea why it was added in the first place. [17:19:35] Well guess what..... [17:19:40] Michael Dale asked for this [17:19:47] (I think( [17:20:02] It's for loader scripts that want to set dynamic dependencies [17:20:08] Allows said dependencies to be lazy-evaluated [17:21:07] I figured that, but I'm not sure how stable that is since in order to check what is already loaded and when callbacks need to be called it needs to know the dependancies. [17:21:35] Is it in use in mdale's extension ? verified to work ? [17:21:44] You only really need to know a module's deps when loading that module, right? [17:22:13] Right. [17:22:35] But how would that be handled serverside ? [17:22:35] I have no idea whether it's used or confirmed to work [17:22:50] only php spits out mw.loader.register calls. [17:23:04] Loader scripts [17:23:13] Yeah, from third party domains [17:23:22] If a module has a loader script, that loader script is responsible for calling register() [17:23:25] No, not necessarily [17:23:32] TimedMediaHandler uses loader scripts for some modules [17:25:01] 'loader scripts', hm.. [17:29:01] RoanKattouw: I've added a second implementation concept and a json format proposal on etherpad/RL2 [17:29:25] OK, will look later [17:29:36] Turns out there's also a GSoC projects for gadgets [17:29:48] And MaxSem (who co-mentors that project) intends to rewrite Gadgets too [17:30:06] So I talked to Sumana about this and we agreed we should all get together some time to coordinate [17:30:34] The GSoC student is in Spain (or Italy, don't remember; Southern Europe at least), so timezone wise it shouldn't be hard to make that happen [17:30:49] You mean physically ? [17:31:03] No, virutally [17:31:15] k [18:23:09] Anyone knows kaldari's IRC handle ? [18:23:43] kaldari [18:30:18] he's probably in a plane over the atlantic right now [18:30:47] or sleeping [18:31:38] Or both [18:31:54] Hm.. so he hasn't registered it, then ? [18:35:05] I was about to say 'or both' :) [20:07:16] RoanKattouw: There's a weird issue going on in IE7. Apparently both the success and the error callback are fired :O [20:07:21] (for mw.loader.using) [20:07:33] http://toolserver.org/~krinkle/testswarm/user/KrinkleBot [20:08:13] http://toolserver.org/~krinkle/testswarm/?state=runresults&run_id=125&client_id=4#test-output3 mw.loader.implement + mw.loader.using in IE7 [20:08:14] lol [20:13:03] perhaps we can provide a bit more info (perhaps as argument when the error callback is called) [20:13:55] although that would require changing the modules object kept internally. [20:15:19] Is it reproducible? [20:16:11] Yes. [20:16:20] TestSwarm doesnt' continue after re-trying 3 times. [20:16:35] It's been red for 2 commits, so it has been reproduced 6 times. [20:20:01] RoanKattouw: http://pastebin.com/9Wfi0kNu [20:20:42] OK... are you sure this doesn't mess up line numbers though? [20:20:56] Also, if it's reproducible you can just grab IE7 yourself and go debug it [20:21:40] Well, QUnit is supposed to catch/record any and all debug info so that we dont have to boot up VMs to get the info. [20:21:47] Although we'd still have to do so to verify the fix. [20:22:25] The line numbers should stay clean, and even if it doesn't. Right now the e.message is not recorded either because it's only sent to log. [20:22:42] OK [20:23:04] a throw is catched by Qunit and sent to the server. [20:28:30] RoanKattouw: I connected an IE6 to the swarm. Looks like autoEllipis test isn't going well in there. [20:28:38] Could be a problem with the test or perhaps a real bug. [20:28:40] Meh [20:28:49] I have to set up an IE6 VM soon anyway [20:29:16] http://toolserver.org/~krinkle/testswarm/?state=runresults&run_id=127&client_id=384 jquery.autoEllipsis.js / r88392 / IE6 [20:31:04] IE6 is actually fairly easy to run from wine [20:31:48] Meh, I prefer just running IE in a VM [20:33:28] bawolff: I'm not sure how accurate my info is, but from what I understand running these buggy apps in Wine may make render the test results less representative. [20:33:59] hmm, I suppose it would [20:34:31] I only have it setup because thats the version I found instructions for, and its certainly a lot easier then finding a windows computer [21:41:52] Reedy, TimStarling, ^demon -- is there an engineering call today? [21:41:55] hi Ryan_Lane [21:42:00] I've no idea [21:42:10] I'm guessing no one has much to say [21:42:14] yeah [21:42:14] And most are still travelling [21:42:55] <^demon> Tim's been idle for ~1d11h, so I don't think he'll be showing soon. [21:43:07] ^demon and I were there for Mondays bug triage [21:43:08] No one turned up [21:43:12] I think y'all are right and I am free to go to dinner [21:43:32] blergh, I wish Mark had explicitly said "no bug triage this week" so you didn't wonder [21:43:41] <^demon> If anyone asks, we had a quorum and decided not to meet since the majority wasn't here :) [21:43:43] Or cancelled the Calendar appointment ;) [21:43:46] hahaha [21:44:00] *Reedy inserts comment about technology [21:44:12] or about logy techno [21:45:18] <^demon> I had a moment like that last night. I had parked for dinner and was trying to confirm which block the restaurant was on. While waiting for Google Maps to load I looked out the window and saw it right next to me. [21:45:25] ha [21:45:40] <^demon> I've become too reliant :) [21:45:46] don't even ask Rob Halsell about maps and the old Berlin Wall