Log file Viewer

[00:49:50] <Ivo> elarson: elarson http://pastie.org/9170449

[01:06:01] <dstufft> Is there anyone in here who happens to be familar with zc.buildout? aclark I think? ;)

[03:22:52] <jaraco> I've got this issue with pyvenv - http://paste.jaraco.com/DFTHx

[03:24:08] <jaraco> I'm getting inconsistent results depending on if the env is 'activated' or not when a package is install. Yet, the activation script does little but update the PATH and VIRTUAL_ENV vars.

[03:24:30] <jaraco> And I can't tell where the issue lies or where I can report it.

[03:24:50] <jaraco> I can work around the issue by not using the activation script, but I'd like to file a report and link to it if possible.

[03:24:57] <jaraco> (unless one is already filed).

[03:25:31] <jaraco> I'm just not sure if this is an issue with Debian or pip or pyvenv (or other).

[03:28:09] <dstufft> um

[03:28:32] <dstufft> when did they unbreak pyvenv

[03:30:31] <jaraco> This isn't with Trusty, but with Precise. I don't know if that makes a difference.

[03:30:57] <dstufft> jaraco: can you type whcih python when you activate the env

[03:32:11] <jaraco> /home/vagrant/env/bin/python

[03:33:32] <dstufft> jaraco: what's the sys.path if you do env/bin/python -c "import sys; print(sys.path)"

[03:34:45] <jaraco> dstufft, I was preparing this - http://paste.jaraco.com/nes3W

[03:35:24] <jaraco> site-packages only appears in the activated mode

[03:35:40] <dstufft> jaraco: if you look in the env, which folder is six installed in

[03:35:47] <dstufft> I know this bug I think

[03:35:51] <dstufft> it's becuase of debian's hacks

[03:36:26] <jaraco> It gets installed to site-packages if the env is activated.

[03:36:29] <dstufft> debian applies hacks to make dist-packages the default installation instead of site-packages

[03:36:32] <jaraco> But to dist-packages if the env is not.

[03:36:45] <dstufft> and those hacks detect if it's inside of a virtualenv

[03:36:51] <dstufft> but they need updated for a Python 3.3+ venv

[03:39:10] <jaraco> So given that this is using fkrull/deadsnakes for Python 3.4, would you report this issue with that PPA or with Debian?

[03:40:05] <dstufft> I think I told barry already about that bug

[03:40:11] <jaraco> Perhaps I'll sleep on it. Being aware of the issue is half the battle.

[03:40:13] <dstufft> so debian might already have it fixed

[03:40:17] <dstufft> I'm not sure

[03:40:32] <jaraco> Cool. Maybe I'll bug Barry about it next time we're hacking together.

[03:40:38] <jaraco> !thanks dstufft

[03:40:38] <pmxbot> you're doing good work, dstufft!

[03:40:50] <dstufft> np

[04:19:28] <Ivo> dstufft: deadsnakes incorporates debian's hacks?

[04:19:37] <dstufft> yea I believe so

[04:21:02] <Ivo> glorious arch master race suffers from no such crappy debian hacks

[04:23:43] <dstufft> Ivo: that just means that pip and pacman will stomp all over each other

[04:24:13] <dstufft> Ivo: basically what i'm saying is, everything is terrible and we should probably just give up and grow crops

[04:25:54] <Ivo> dstufft: such a pity the repos are so out of date with pypi then... ;) http://dpaste.com/3CA3ATV/

[04:26:20] <dstufft> oh hey they updated our url already

[04:26:52] <dstufft> I've never used arch tbh

[04:27:10] <Ivo> come to the configurable side

[04:27:21] <dstufft> linux on the desktop is a terrible thing

[04:27:30] <dstufft> at least every time i've tried it It resulted it sadness

[04:27:50] <Ivo> where people stop fucking with what upstream packaged

[04:28:37] <Ivo> dstufft: you'd strike me as a brilliant candidate for a tiling wm

[04:29:50] <dstufft> never used one

[04:35:59] <Ivo> dstufft: https://www.youtube.com/watch?v=QnYN2CTb1hM

[04:36:42] <Alex_Gaynor> jezdez: not sure if there's anythign to be done, but https://caniusepython3.com/check/54fc9192-b6af-47e2-8c32-c5a700e07f14 has been spinning for likw 5 minutes

[09:52:53] <jezdez> !logs

[09:52:53] <pmxbot> http://chat-logs.dcpython.org/channel/pypa

[09:54:08] <jezdez> Alex_Gaynor: yeah, it's a bummer, the queuing system doesn't behave the same since I moved to heroku. gonna switch to rq that should speed things up

[13:25:09] <dhellmann> dstufft: thanks!

[13:25:37] <dstufft> dhellmann: :)

[13:25:44] <dstufft> I thnk Ivo got a reproceable case now

[13:25:55] <dstufft> not sure if i was determined if it's fixable or fundamental

[13:26:01] <dstufft> it was*

[13:26:17] <dhellmann> Ivo: yes, if we install oslo.config with "pip install" and oslo.messaging with "pip install -e" then "from oslo.config import cfg" fails with a message indicating that the config sub-package can't be found at all

[13:26:26] <dstufft> https://github.com/pypa/pip/issues/3#issuecomment-42906232

[13:28:45] <dhellmann> dstufft: yes, I think that case is at least similar to what we were seeing

[13:29:15] <dhellmann> dstufft: my description above may be remembering the order of installation backwards, or it may not matter which order things are installed

[13:30:11] <dstufft> cool