[00:11:27] <robla>	 AaronSchulz: any migration related reason for  "Internal error: Server failed to store temporary file"?  http://commons.wikimedia.org/wiki/Commons:Village_pump#Internal_error:_Server_failed_to_store_temporary_file_-_when_trying_to_upload_new_image
[00:13:35] <AaronSchulz>	 works for me
[00:14:49] <robla>	 well, I suppose we should keep an eye out for sporadic reports of problems with that
[00:15:15] <chrismcmahon>	 robla: I was getting that error from UW, afaik Roan fixed it some hours ago
[00:15:30] <RoanKattouw>	 I didn't
[00:15:35] <RoanKattouw>	 Aaron was going to take a stab at it
[00:15:41] <RoanKattouw>	 It looks like Tim has now fixed some of the GRANTs
[00:15:52] <chrismcmahon>	 robla: take it back, Tim may have made the fixes
[00:15:59] <AaronSchulz>	 wikiadmin grant was broken
[00:16:05] <AaronSchulz>	 and here I though it was deliberate
[00:16:17] <RoanKattouw>	 TimStarling	| 10.64.0.0/255.255.252.0 | wikiadmin |
[00:16:18] <RoanKattouw>	 	TimStarling	mysql doesn't do CIDR last time I checked
[00:16:29] <RoanKattouw>	 TimStarling	I missed one of the masters, I got it just now
[00:16:30] <RoanKattouw>	 	TimStarling	so we should have no more errors after 00:05
[00:16:43] <RoanKattouw>	 Also, for your amusement:
[00:16:44] <RoanKattouw>	 TimStarling	| GRANT ALL PRIVILEGES ON `%a%`.* TO 'wikiadmin'@'10.0.%'
[00:16:51] <RoanKattouw>	 	TimStarling	right, so wikiadmin gets access to any database with "a" in its name
[00:17:10] <AaronSchulz>	 huh, php supports try-catch-finally
[00:17:23] <TimStarling>	 it basically doesn't matter what database it gets access to, the only one you might want to deny access to is the mysql database
[00:17:32] <RoanKattouw>	 AaronSchulz: I suppose the appropriate response is either "finally?" or "finally!" :D
[00:17:40] <TimStarling>	 so as long as mysql doesn't match %a% then it doesn't really matter
[00:17:56] <RoanKattouw>	 Well, except that enwiki doesn't match %a% either
[00:18:23] <TimStarling>	 there was also a grant for %wik%, that's the traditional pattern
[00:18:35] <TimStarling>	 someone probably added %a% to that master only for some test database
[00:19:15] <TimStarling>	 it would make things easier if all our databases had some fixed prefix
[00:19:21] <robla>	 anywaaaaay  :)  so, what you're telling me is that the "Server failed to store temporary file" problem is probably  the db problem that Tim fixed, and that it shouldn't be a problem anymore.  right?
[00:20:00] <RoanKattouw>	 I believ eso
[00:20:40] <robla>	 chrismcmahon: you aren't seeing the problem anymore, and you're running the same tests that used to repro?
[00:21:47] <chrismcmahon>	 robla:  I have not checked recently on commons, but I can.   I have a feeling test2 didn't make it across to eqiad.
[00:22:08] <robla>	 ?
[00:22:49] <robla>	 chrismcmahon: " I have a feeling test2 didn't make it across to eqiad."....what do you mean by that?
[00:22:52] <ori-l>	 chrismcmahon: "I may not get there with you", etc.
[00:23:15] <chrismcmahon>	 robla: I mean UW works fine on test2 and was broken on commons.
[00:26:27] <robla>	 I ventured a reply on commons.  I think it would be good to make sure things are still working, so yeah, chrismcmahon, if you've got time for another test run, it'd be greatly appreciated.
[00:26:41] <chrismcmahon>	 doing that
[00:27:24] <robla>	 it'd be good to make sure we're seeing a clean set of tests now, if we haven't gotten that yet
[00:28:53] <chrismcmahon>	 robla TimStarling on commons right now UploadWizard still shows me "Internal error: Server failed to store temporary file"
[00:29:24] <chrismcmahon>	 robla: it would.  but test2 does seem to be our biggest issue just today
[00:29:55] <robla>	 chrismcmahon: I'm confused.  what's wrong with test2?
[00:30:20] <chrismcmahon>	 robla: UW tests are passing on test2 and they should fail
[00:30:51] <robla>	 well, maybe things are working on test2
[00:31:24] <robla>	 that would make sense for a db problem, since commons and test2 are on different db shards
[00:32:41] <chrismcmahon>	 could be.  test2 is my best oracle for prod right now though.
[00:33:38] <kaldari>	 ori-l: you around?
[00:33:47] <ori-l>	 kaldari: hey
[00:33:56] <ori-l>	 what's up?
[00:34:00] <kaldari>	 I need a hebrew speaker
[00:34:10] <ori-l>	 should I come up? I'm on 3.
[00:34:20] <kaldari>	 if you have a few minutes
[00:34:25] <ori-l>	 np
[00:34:32] <ori-l>	 am I going to be praying for AFT or something?
[00:34:47] <kaldari>	 I won't help
[00:34:50] <kaldari>	 er
[00:34:55] <kaldari>	 I meant, it won't help
[00:34:57] <kaldari>	 lol
[00:35:08] <ori-l>	 heh
[00:35:15] <ori-l>	 be right there
[00:38:12] <robla>	 chrismcmahon: RoanKattouw: AaronSchulz: I've manually reproed the UploadWizard problem....pretty trivial repro
[00:38:24] <chrismcmahon>	 yes
[00:52:53] <Krinkle>	 RoanKattouw: ext.fooBar vs. ext.FooBar
[00:53:04] <Krinkle>	 (module names)
[00:54:40] <RoanKattouw>	 ?
[01:16:22] <robla>	 AaronSchulz: any luck reproing the problem yourself?
[01:16:55] <AaronSchulz>	 well the JS seems to be missing a change I thought it had
[01:17:17] <AaronSchulz>	 in any case I'm moving the hack to core instead
[01:33:01] <AaronSchulz>	 yeah it's the image table like I thought
[01:35:37] <chrismcmahon>	 AaronSchulz: I think I just managed to corrupt an image file on commons, File:Obscure_Anasazi_ruin.jpg.  UW wizard tells me that it already exists, but there is no page for it.
[01:36:44] <chrismcmahon>	 (which is new)
[01:41:15] <Krinkle>	 RoanKattouw: \
[01:41:17] <Krinkle>	 		// Continue
[01:41:17] <Krinkle>	 		return true;
[01:41:20] <Krinkle>	 ARG
[01:41:22] <RoanKattouw>	 ?
[01:41:28] <RoanKattouw>	 Oh
[01:41:31] <Krinkle>	 A comment in templateinfo code
[01:41:32] <RoanKattouw>	 As opposed to return false?
[01:41:36] <Krinkle>	 no, the commenet.
[01:41:40] <RoanKattouw>	 Or from a hook?
[01:46:23] <Krinkle>	 RoanKattouw: Ok, but this definitely is: \
[01:46:25] <Krinkle>	 		// return text
[01:46:25] <Krinkle>	 		return $text;
[23:13:46] <Krinkle>	 Anyone know of a good way to save state somewhere from a parser hook (e.g. <foobar> , as registered from the ParserFirstCallInit hook) so that this information can be read when in the onArticleSave hook?
[23:14:10] <Krinkle>	 The point is to abort the save from the onSaveHook if the parser tag renderer (registered via ParserFirstCallInit hook) determines that the content of <foobar> is invalid.
[23:14:45] <Krinkle>	 The validator in my extension provides a Status object, which I can merge with the one that onArticleSave has.
[23:15:10] <Krinkle>	 But how do I get this Status object from callback registered in Parser::setHook to the callback registered with onArticleSave
[23:58:43] <marktraceur>	 ori-l: Hey, did you ever get a chance to look at the wmf-vagrant bug?