[00:48:43] <Nemo_bis>	 https://github.com/jayphelps/git-blame-someone-else
[04:41:50] <ori>	 // Profile 1 in 100 requests.
[04:42:02] <ori>	 'profiler.enable' => function() { return rand(1, 100) === 42; },
[04:42:20] <ori>	 err, mistype.
[04:42:43] <ori>	 xhgui has rand(0,100), which is 1:101, I'm just submitting a PR to fix that.
[04:53:04] <ori>	 bd808: https://performance.wikimedia.org/xhgui/run/view?id=56ca93c47ed6ccf971f24e25 -- monolog stuff is pretty expensive
[04:53:42] <ori>	 but then again, this is an mw1017 request, which has some special logging config, right?
[04:54:44] <bd808>	 yeah mw1017 handles all log messages
[04:54:55] <bd808>	 so it would be expected to cost a lot of cycles
[04:56:33] <ori>	 right
[19:58:39] <MaxSem>	 bd808, do you still need https://gerrit.wikimedia.org/r/#/c/270459/ ?
[20:13:12] <bd808>	 MaxSem: I don't think they have tried to do the extension.json conversion again yet.
[21:07:46] <MaxSem>	 bd808, nope - they reintroduced it again, without my fix: https://gerrit.wikimedia.org/r/#/c/268897/
[21:08:17] <bd808>	 MaxSem: that's the one I reverted isn't it?
[21:08:25] <MaxSem>	 dunno? :P
[21:08:39] <MaxSem>	 definitely it doesn't have my fix
[21:08:42] <bd808>	 This patchset was reverted in change: I84e2ba310c425e2d1db1e7032dc014f88ebef087
[21:08:51] <bd808>	 yeah that was the need for you fix
[21:09:47] <MaxSem>	 arr
[21:10:24] <MaxSem>	 due to silly gerrit ordering by latest change instead of commit time, I thought there was a new change bd808 
[21:10:38] <MaxSem>	 all seems good:P
[23:03:54] <tgr>	 anomie: Setup.php calls Session::renew() unconditionally, even if it is not persistent
[23:04:08] <tgr>	 can that be the reason for some of the extra perf burden?
[23:11:24] <ori>	 anomie: filed https://phabricator.wikimedia.org/T127772