[19:53:38] <Krinkle>	 legoktm: Regarding coverage reports being collected continuously with your tool. I'm thinking we can do the same with benchmarks.
[19:54:01] <Krinkle>	 We'll need to way to make them comparable over time, regardless of infrastructure changes, though.
[19:54:04] <Krinkle>	 Which is tricky.
[19:54:58] <Krinkle>	 For comparing two code versions, I suppose we can simply re-run the old code first during each run, so that it's more comparable, but not sure what to plot on a long-term graph. Any ideas?
[20:44:39] <legoktm>	 Krinkle: I'm not sure either. Moving coverage to PHP 7 will drop core by half a percent, because xdebug calculates something differently now. I was thinking we could add a marker in the graph or something indicating that the drop is expected.
[20:45:13] <legoktm>	 Or the other idea I had was to just re-run old commits on the new environment, but that seemed like a lot of work for little gain
[20:47:52] <Krinkle>	 legoktm: Aye, yeah, I'm referring to benchmarks though. :)
[20:48:03] <Krinkle>	 But suggesting we might be able to aggregate them in a similar way
[20:50:56] <legoktm>	 are cloud services VMs a reliable place to run perf benchmarks?
[20:59:10] <Krinkle>	 legoktm: Well, that's the thing. Over time I don't expect it to be consistent, especially when we change images and what not. But I suppose we can still run before/after which a job on a single image and get reasonably comparable results.
[20:59:14] <Krinkle>	 relative change
[21:00:09] <legoktm>	 makes sense
[21:00:18] <Krinkle>	 legoktm: But I'm not sure how to plot that.
[21:00:33] <Krinkle>	 It'd work for a boolean yes/no we regressed comparing X to Y commit, but not over long-term.
[21:00:57] <Krinkle>	 But maybe that's fine?
[21:01:21] <Krinkle>	 I guess we could plot like +1%, -1% 
[21:01:23] <legoktm>	 we could try charting it and see how off it is in reality?
[21:02:05] <legoktm>	 (also, re doxygen alternatives, I saw https://bugs.debian.org/cgi-bin/bugreport.cgi?bug=888580 a few days ago)
[21:14:49] <subbu>	 legoktm, i know i am being a pest .. but how do i push my linter patch up in your task queue? that work improves ability of editors on wikis like enwiki to more precisely know the scale of work to get tidy replaced.
[21:15:50] <legoktm>	 subbu: I will get to it today, I promise :) just don't have enough time in between classes to do a full review
[21:16:01] <subbu>	 ok. :)
[23:07:54] <Krinkle>	 bd808: Hmm cookiecutter-library still isn't mirrored to github
[23:08:08] <Krinkle>	 Not sure how or why, I assumed that's the default for gerrit repos