[17:45:30] <dapatrick>	 bd808: Re. OAuth approval permission, Jamesofur can give you that.
[18:37:08] <ostriches>	 I wonder if mediawiki_mysql2postgres.pl even works.
[18:37:27] <ostriches>	 (was just looking at the 2 perl scripts we have in MW :))
[18:38:34] <paladox>	 ostriches when will we bump mw version to 2.x?
[18:38:45] <paladox>	 so instead of 1.x it can be 2.x :)
[18:38:50] <ostriches>	 Heh, never.
[18:38:53] <ostriches>	 I tried years ago :p
[18:39:12] <paladox>	 Oh
[18:39:16] <ostriches>	 Actually, my preference would be to drop the 1.x and start doing like 28.0, 29.0, 30.0
[18:39:22] <paladox>	 Oh :)
[18:39:32] <ostriches>	 Otherwise we'd just stick with 2.x forever like we've had 1.x forever
[18:39:35] <paladox>	 I guess we could do that ^^ in 1.30
[18:39:40] <ostriches>	 Primary version #s that never change are pointless :)
[18:39:41] <paladox>	 so instead of 1.30 it will be 30
[18:39:51] <ostriches>	 I've considered it, never wrote up an RfC though
[18:40:00] <paladox>	 Oh
[18:45:47] <MaxSem>	 ostriches, I say namespace the whole of MW and go 2.0 because compatibility is successfully broken anyway
[18:51:47] <ostriches>	 So we'll stick with 2.x forever, yay
[18:52:11] <ostriches>	 I'd rather stick with 1.x and avoid the headache
[18:56:28] <MaxSem>	 how did Linux decide to go from 3.x to 4.0?
[18:59:49] <ori>	 https://lkml.org/lkml/2015/2/22/203
[19:02:25] <ostriches>	 'On the other hand, the strongest argument for some people advocating 4.0 seems to have been a wish to see 4.1.15 - because "that was the version of Linux skynet used for the T-800 terminator".'
[19:28:15] <paladox>	 ostriches 2.x is a perfect chance to remove all deprecated things
[19:28:19] <paladox>	 and clean up.
[19:28:36] <Reedy>	 lmfao
[19:28:42] <Reedy>	 And break 99% of the extensions
[19:28:45] <paladox>	 all broken things, and stuff that is not needed any more.
[19:28:51] <paladox>	 yeh, thats why the 2.0
[19:28:59] <ostriches>	 We have a deprecation cycle already
[19:29:07] <ostriches>	 Mark it deprecated, remove after a few releases.
[19:29:13] <ostriches>	 No need for a 2.x for that
[19:29:16] <Reedy>	 we'd have to make sure all the WMF stuff were 2.0 ready.. Which is bad enough
[19:29:18] <paladox>	 Ok
[19:29:27] <ostriches>	 "Stuff that is not needed" is subjective too :)
[19:29:47] <ostriches>	 I have lots of things in mind but others would probably disagree ;-)
[19:30:12] <paladox>	 Oh
[19:30:21] <paladox>	 ostriches your probaly thinking about db's
[19:30:24] <paladox>	 LOL
[19:30:38] <ostriches>	 I'm thinking of lots of things!
[19:30:43] <paladox>	 Oh
[20:39:25] <anomie>	 MaxSem: Re Linux versioning, IIRC they decided that the 2 in version numbers like 2.6.X.x would never change, so instead of 2.8.X.x they did 3.X.x. Then I guess 4.X.x was what would have been 2.10.X.x under the old system.
[21:11:53] <Reedy>	 AaronSchulz: Any chance you could fix https://phabricator.wikimedia.org/T148957 so we don't have to revert numerous patches in REL1_28?
[21:18:36] <hoo>	 Reedy: Whoops
[21:18:44] <hoo>	 The global is working
[21:18:56] <Reedy>	 hoo: Been broken for a month :P
[21:19:04] <Reedy>	 It meant no useful errors on beta last night too
[21:19:16] <hoo>	 but you also need to set $wgShowDBErrorBacktrace
[21:19:40] <hoo>	 I stumbled upon that many times, also once during an outreach event
[21:19:49] <hoo>	 and it's on my to do list… but that's the thing with to do lists…
[21:20:07] <bd808>	 TODO: make hoo work on this TODO list
[21:21:18] <Reedy>	 hoo: well, the error message is wrong too then :P
[21:21:25] <hoo>	 Reedy: Indeed, it is
[21:21:36] <Reedy>	 Is that really the fix?
[21:21:38] <hoo>	 especially nice if you hit that on a newly installed wiki
[21:21:43] <paladox>	 Reedy yes
[21:21:46] <hoo>	 It worked for me once or twice
[21:21:46] <paladox>	 I saw that too
[21:21:56] <Reedy>	 Hmm.. So we need to set wgShowDBErrorBacktrace on beta
[21:22:21] <paladox>	 Yep
[21:22:27] <paladox>	 Reedy i know where the problem is
[21:22:29] <paladox>	 i think
[21:22:33] <paladox>	 let me check again
[21:22:48] <Reedy>	 				} else {
[21:22:48] <Reedy>	 					$message .= "Exception caught inside exception handler.\n\n" .
[21:22:48] <Reedy>	 						"Set \$wgShowExceptionDetails = true; at the bottom of LocalSettings.php " .
[21:22:48] <Reedy>	 						"to show detailed debugging information.";
[21:22:49] <Reedy>	 				}
[21:22:53] <Reedy>	 That one wants changing
[21:23:07] <paladox>	 Found it
[21:23:27] <paladox>	 Submiting a patch
[21:23:27] <bd808>	 the exceptions should really log there too I think.
[21:24:59] <Reedy>	 mmmm
[21:24:59] <paladox>	 Reedy https://gerrit.wikimedia.org/r/322739 
[21:25:04] <paladox>	 I really hope that fixes it
[21:25:58] <bd808>	 paladox: that won't fix anything
[21:26:05] <paladox>	 Oh
[21:26:12] <bd808>	 see line 278?
[21:26:38] <bd808>	 you just added a completely random check to the test
[21:26:48] <Reedy>	 Never mind no global call either
[21:26:51] <bd808>	 that is a duplicate of a check that already exists
[21:27:25] <paladox>	 bd808 but $wgShowExceptionDetails && ( !( $e instanceof DBError ) || $wgShowDBErrorBacktrace ) <!-- that will only work if you set $wgShowExceptionDetails and $wgShowDBErrorBacktrace
[21:27:45] <paladox>	 Whereas $wgShowExceptionDetails && ( !( $e instanceof DBError ) || $wgShowDBErrorBacktrace ) || $wgShowExceptionDetails will allow it to work either way
[21:27:51] <bd808>	 paladox: I love you enthusiasm, but reviewing your "fix by untested magic" patches are a bit of a drain on all of us.
[21:28:05] <paladox>	 ok
[21:28:22] <paladox>	 Should i abandon?
[21:29:07] <Reedy>	 hoo: So setting the right variable does fix it?
[21:29:18] <bd808>	 the boolean logic in that test is correct I believe, but the message wording is not quite right.
[21:29:20] <bd808>	 ( !( $e instanceof DBError ) || $wgShowDBErrorBacktrace )
[21:29:33] <Reedy>	 The message seems completely wrong
[21:29:35] <Reedy>	 Said global is true
[21:29:40] <Reedy>	 So it's just a wtf
[21:29:53] <bd808>	 if the exception is not an DBError OR if it is and the global is set, then show the trace
[21:32:05] <Reedy>	 I wonder if the variable being mentioned is wrong in other places
[21:32:31] <bd808>	 wrong docs?! how could that happen Reedy? ;)
[21:32:39] <Reedy>	 !bug 1
[21:33:43] <Reedy>	 https://gerrit.wikimedia.org/r/322742
[21:34:09] <bd808>	 it needs both right?
[21:34:10] <AaronSchulz>	 no reason not to at least include the "public" message with the code for double-exception cases
[21:34:40] <AaronSchulz>	 if the backtrace is too sensitive for DB errors, I don't think the message/SQL would should be public either
[21:34:42] <bd808>	 and maybe only wgShowDBErrorBacktrace in some cases
[21:35:25] <AaronSchulz>	 the two globals never quite got along perfectly afaik
[21:36:05] <AaronSchulz>	 if DB backtraces are on but not exceptiondetails, it's not super obvious what to do (MW requires both in that case to be enabled, which is still odd)
[21:38:18] <Reedy>	 tgr: Do you know where the task about the useless oauth emails went? I can't find it, and am still getting the emails with no useful info
[21:39:24] <bd808>	 Reedy: echo problem?
[21:39:32] <Reedy>	 echo/oauth, yeah
[21:39:51] <bd808>	 the one time I tried to fix echo messages I only succeeded in completely breaking them (on wikitech)
[21:40:02] <Reedy>	 I'm sure I filed a task about it
[21:40:32] <tgr>	 it's still around
[21:40:45] <Reedy>	 I cannae find it :)
[21:40:57] <Reedy>	 It's not tagged as oauth...or echo
[21:41:30] <tgr>	 I hope https://gerrit.wikimedia.org/r/#/c/316302/ will fix it once I get around to finish
[21:42:51] <tgr>	 https://phabricator.wikimedia.org/T61772#2639878
[21:43:00] <tgr>	 not sure if there was a standalone task
[21:43:21] <Reedy>	 Ah, I thought I'd filed a seperate task
[21:43:28] <Reedy>	 The screenshot is there though :)
[21:48:08] <ostriches>	 Reedy: Heh, https://gerrit.wikimedia.org/r/#/c/322771/
[21:48:52] <Reedy>	 lols
[21:49:48] <paladox>	 Reedy  https://gerrit.wikimedia.org/r/322772
[21:50:22] <paladox>	 I just reverted it ^^
[21:50:56] <Reedy>	 Well, if it's just that message was wrong...
[21:51:13] <paladox>	 Oh
[21:51:24] <paladox>	 Reedy i doint know what a good revert commit msg is?
[21:51:27] <Reedy>	 I wouldn't abandon that commit yet
[21:51:54] <paladox>	 Ok
[21:58:13] <Reedy>	 anomie: :)
[21:58:44] <Reedy>	 I think you're right though
[22:00:06] <anomie>	 Reedy: Or, I suppose, showBackTrace() could somehow do it.
[22:01:06] <Reedy>	 That seems to convolute it even futher, I think
[22:01:31] <Reedy>	 There's a few places that mention $wgShowExceptionDetails = true;
[22:01:46] <Reedy>	 but work on showBackTrace() in different ways
[22:08:21] <Reedy>	 Plummed that into the various places it's used
[22:18:03] <Reedy>	 anomie: thanks
[22:46:10] <Krinkle>	 bd808: Regarding the logstash issues with field types. Are there any action items?
[22:46:25] <Krinkle>	 I noticed also things like 'type' and 'source' conflicting, which is quite hard to maintain.
[22:46:36] <Krinkle>	 Should ideally be namespaced as well.
[22:47:07] <Krinkle>	 slow-parse is just one of the dashboards I'd like to re-create, there are others I'd like to add in addition. To find top metrics in various categories for performance team to look at.