Log file Viewer

[06:33:25] <dstufft> 45% unit test coverage :[

[06:34:33] <Alex_Gaynor> dstufft: gotta start somewhere!

[06:34:50] <dstufft> 44% actually once I exclude pip/compat.py

[06:34:52] <dstufft> Alex_Gaynor: yea :/

[06:35:02] <dstufft> least we'll be able to actually measure it soon

[06:36:13] <dstufft> later on I might try and figure out how to get coverage to measure inside of a sub process, though not sure how I feel about that, I kind of like that by not covering inside of a subprocess we incentivize unit tests

[06:38:01] <dstufft> So good news, pip.index (interacting with PyPI) actually has 81% coverage

[06:38:46] <dstufft> bad news, req/* files all have 41%, 26%, 25%, and 13% .. and that's where most of the complex logic is :(

[06:58:51] <Ivo> dstufft: so https://github.com/pypa/pip/pull/1902 looks awesome

[07:00:30] <Ivo> Alex_Gaynor: you know of any secret sauce to make pypy run tests fast?

[07:01:01] <Ivo> I wonder if its the constant subprocess invocation that slows things down the most

[07:01:05] <dstufft> Ivo: I've asked him a lot yea :(

[07:01:08] <dstufft> and basically it is

[07:01:19] <dstufft> the thing pypy is the worst at, is starting up

[07:01:30] <dstufft> and we basically do that like 400+ times

[07:02:03] <Ivo> So the answer is just the hard one, we should get better at unit testing

[07:02:20] <Ivo> I wanna be lazy damnit

[07:03:19] <dstufft> Ivo: I was gonna merge that xmlrpc one, but I hadn't pulled it down yet to add a changelog item

[07:03:53] <Ivo> Could just be lazy and merge then add changelog

[07:05:09] <Ivo> dstufft: reading logs... you said yesterday that 3.3 was the one timing out on the compat refactor commits, I thought it was 3.2 tho

[07:05:25] <dstufft> it was 3.3 I think

[07:05:45] <dstufft> 3.2 was timing out on my change

[07:05:48] <dstufft> for sessions