Log file Viewer

[18:44:37] <ionelmc> anyone want to review something about packaging i want to blog about ?

[18:46:33] <tomprince> ionelmc: Why don't you post it here, so people can look and comment.

[18:46:47] <ionelmc> https://bitbucket.org/ionelmc/blog.ionelmc.ro/src/master/content/2014-05-25-python-packaging.rst

[18:46:57] <ionelmc> i don't want to publish something very silly

[18:47:49] <ionelmc> would love to hear if i forgot to consider something important

[18:49:08] <tomprince> I'm not sure why you don't put tests in the package.

[18:49:24] <tomprince> If you do that, you can test an installed version, even without the source package.

[18:49:57] <tomprince> Also, you don't motivate using 'src'.

[18:50:01] <ionelmc> tomprince: but then you would be doing test discovery in something that in site-packages

[18:50:08] <ionelmc> i guess i need to explain that better

[18:52:58] <dstufft> tomprince: eh, I don't like putting tests inside the library

[18:55:43] <dstufft> tomprince: I get the motivate for doing so... I just don't think it's really that big of a use case. I've never wanted to run the tests for some random library that I happened to have installed. I have wanted to (although very rarely) run the tests for a newer version against an older version

[19:05:18] <ionelmc> tomprince: i've added some more elaborate explanations after the layout part about the tests location if you want to take a look again

[19:07:00] <tomprince> Almost every test run at least uses the exit code to indicate failure.

[19:07:44] <tomprince> It still isn't convincing to me.

[19:08:09] <ionelmc> tomprince: you're referring to the part about the lack of a standard for test results ?

[19:08:18] <tomprince> My first comment, yes.

[19:09:51] <ionelmc> tomprince: that's still used ... tox returns a non-zero if anything fails in the commands list

[19:10:02] <dstufft> I don't like setup.py test because people need to stop thinking about setup.py as anything but a file for the packaging tools :V

[19:10:25] <ionelmc> not entirely sure about what you disagree with

[19:10:41] <ionelmc> would appreciate if you'd explain more

[19:11:01] <ionelmc> at least dstufft agrees :P