Log file Viewer

[09:27:03] <Kuba> hello! :)

[09:28:05] <Kuba> I'm encoutering strange issue with tox and pip starting at 6.0.7: AssertionError: Egg-link /home/schoen/src/lets-encrypt-preview does not match installed location of letsencrypt (at /home/schoen/src/lets-encrypt-preview/.tox/lint/lib/python2.7/site-packages) (https://github.com/letsencrypt/lets-encrypt-preview/issues/235)

[09:28:48] <Kuba> anyone met this bug before? couldn't find anything in the tracker

[11:29:57] <xafer> hello Kuba, looks like the distribution is installed twice: we can find letsencrypt-0.1-py2.6.egg-info and letsencrypt.egg-link in site-packages

[11:32:54] <Kuba> xafer: *.egg-info being the result of "setup.py install" and *.egg-link being the result of "setup.py develop"?

[11:33:15] <xafer> yes

[11:33:56] <Kuba> I don't know much about how tox works internally, but I've set the command to be roughly "setup.py develop"

[11:34:19] <Kuba> https://github.com/letsencrypt/lets-encrypt-preview/blob/master/tox.ini#L10

[11:35:04] <Kuba> hm

[11:38:07] <Kuba> xafer: could "setup.py develop easy_install mypkg[extra]" be the problem?

[11:38:29] <Kuba> xafer: and do you suggest that it was just a coincidence that 6.0.6 worked well?

[11:39:25] <xafer> nope, I'd say 6.0.6 pip used a different uninstall order

[11:40:04] <xafer> Kuba maybe you should use "pip install -e .[extra]" instead of python setup.py develop

[11:46:53] <Kuba> i'm not sure what is the dot "." relative to

[11:47:04] <Kuba> in the tox command context

[11:47:18] <Kuba> xafer: but first tests seem promising

[11:49:55] <xafer> I'd guess it is the current directory and should contain the setup.py (just like in "python setup.py develop")

[12:06:19] <Kuba> xafer: thanks, a lot

[12:06:31] <Kuba> xafer: FYI https://github.com/letsencrypt/lets-encrypt-preview/pull/237

[12:21:27] <xafer> great :) glad I could help

[12:36:32] <mitsuhiko> so what's the replacement for --no-install inpip?

[17:41:25] <dstufft> mitsuhiko: what are you using --no-install for?

[20:37:33] <mitsuhiko> dstufft: i was attempting to find a reasonable way to download somethign and unpack it

[20:37:46] <dstufft> mitsuhiko: pip install -d . thing

[20:37:46] <dstufft> ?

[20:38:03] <mitsuhiko> that does not unpack

[20:38:33] <dstufft> oh, i see

[20:39:03] <dstufft> We don't have anything to unpack like that. there was a ticket about it but we weren't sure if it would be useful or not

[20:39:42] <mitsuhiko> well, i have a hacky way to unpack that stuff now which is good enough for my purposes

[20:39:49] <dstufft> https://github.com/pypa/pip/issues/906

[21:59:15] <ronny> dstufft: ping?

[21:59:26] <dstufft> ronny: hi

[22:00:12] <ronny> dstufft: recall the idea about moving some parts of the stdlib into wheels (mutlirpocessing, sqlite, tk, ...) in particular those with external dependencies

[22:00:34] <ronny> how likely would a pep pass that proposed maing a minimal python, and a sumo python on top of it?

[22:02:11] <dstufft> ronny: I have no idea, I feel like it's probably not very likely, but I'm not sure

[22:04:53] <ronny> it would turn distributions and library updates into a breeze at least

[22:09:04] <dstufft> I think it's a great idea nd I wish that's how Python was

[22:09:15] <dstufft> but i feel like python-dev is unlikely to accept that PEP

[22:19:49] <ionelmc> they have to at least acknowledge that dealing with various breakage in stdlib is damn annoying for users