Log file Viewer

[13:18:51] <AndreasLutro> is there a reason twine uses "pypitest" instead of "testpypi" which is what it's been in setuptools forever? https://github.com/pypa/twine/blob/master/twine/utils.py#L59

[13:19:27] <AndreasLutro> https://github.com/pypa/python-packaging-user-guide/blob/master/source/guides/migrating-to-pypi-org.rst#using-testpypi

[21:40:01] <techalchemy> dstufft: if you get a sec and have the context for it I'm having a weird issue converting pipenv to pip 10 due to packaging=>SpecifierSets

[21:40:07] <techalchemy> wouldn't mind an extra set of eyes

[21:50:08] <StevenK> techalchemy: I did some of the work on packaging, I can peer at it

[21:50:59] <StevenK> techalchemy: However, keep in mind I just did 22+ hours on planes and my body thinks it's Sunday when it's apparently Saturday, so no promises

[22:21:19] <techalchemy> StevenK: https://github.com/techalchemy/pipenv/blob/drop-pip9/pipenv/patched/piptools/resolver.py#L153 is where things are failing

[22:21:51] <techalchemy> https://github.com/techalchemy/pipenv/blob/drop-pip9/pipenv/patched/notpip/_vendor/packaging/specifiers.py#L589 is my modified specifier set (i had to modify it because piptools does &= operations to specifiersets)

[22:22:13] <techalchemy> note that this worked fine in pip 9 and i don't see any relevant changes in pip 10 so i have no clue what is wrong

[22:22:53] <techalchemy> and a failing build log with the current error would be here: https://ci.appveyor.com/project/techalchemy/pipenv/build/1.0.228/job/bx3u6dv7log5de7h

[22:23:24] <techalchemy> the initial failures were TypeError: unsupported operand type(s) for &=: 'SpecifierSet' and 'SpecifierSet'

[22:24:29] <techalchemy> now when I try to use copy.deepcopy(InstallRequirement) I get: AttributeError: 'list' object has no attribute 'split'

[22:29:38] <StevenK> techalchemy: Are you able to hook up pdb there, I'd be very curious what specifiers actually is

[22:29:54] <techalchemy> ''

[22:29:56] <techalchemy> :p

[22:30:38] <StevenK> Huh, maybe that needs to be if specifiers.strip()

[22:30:55] <StevenK> But split should return an empty array anyway

[22:31:28] <techalchemy> does deepcopy somehow include the local context of the object? like will 'specifiers' be turned into a list still?

[22:32:44] <StevenK> Maybe it does?

[22:39:16] <StevenK> Although I can't reproduce that with a few examples

[22:42:38] <techalchemy> oh

[22:42:39] <techalchemy> hm

[22:43:50] <techalchemy> StevenK: reproduction should just be `pip install -e git+https://github.com/techalchemy/pipenv.git@drop-pip9#egg=pipenv && mkdir /tmp/testpipenv && cd /tmp/testpipenv && curl -o Pipfile http://dpaste.com/366CM3D.txt && pipenv install`

[23:00:16] <StevenK> techalchemy: It isn't '', it's []

[23:00:34] <StevenK> There are a bunch of calls previously which are correct

[23:03:12] <techalchemy> hmmmm

[23:03:28] <techalchemy> so there's a specific requirement getting passed down incorrectly?

[23:03:37] <techalchemy> StevenK: do you know which one?

[23:03:45] <techalchemy> presumably the one with the index specified

[23:04:07] <techalchemy> also thanks, the help is much appreciated

[23:05:34] <StevenK> I'm not sure of the call chain there, since the exception turns up from deepcopy

[23:07:30] <techalchemy> ok... I have logged all of the specifiers and haven't seen a list yet

[23:07:59] <StevenK> It turns up after Locking [packages] dependencies…

[23:08:28] <StevenK> There are 19 transforms before the problematic one

[23:09:26] <techalchemy> hm yeah i got it this time around

[23:09:41] <techalchemy> logged the wrong things i guess

[23:09:43] <techalchemy> thanks!

[23:09:55] <StevenK> q (as in the module) is amazing

[23:10:06] <techalchemy> i've never heard of it...