pmxbot IRC Log Viewer

[11:58:35] <pmxbot> jaraco pushed 1 commit to setuptools (https://bitbucket.org/pypa/setuptools/) :

[11:58:35] <pmxbot> Bypass dist/build/egg dirs when running tests

[12:13:46] <pmxbot> jaraco pushed 1 commit to setuptools (https://bitbucket.org/pypa/setuptools/) :

[12:13:46] <pmxbot> Started work on a developer guide.

[13:27:53] <dstufft> xafer: that is... strange

[13:28:22] <xafer> dstufft, did you reproduce it ?

[13:28:35] <dstufft> no, just woke up and read your message

[13:29:42] <xafer> and if I change print('KO'* 10) to print('KO') it gets stuck also ...

[13:30:55] <xafer> I then put print(1/0) print('KO') and test passed again...

[13:31:21] <xafer> 100% reproducible on my current setup T.T

[13:49:47] <ionelmc> ever seen pip segfault ?

[13:53:08] <ionelmc> anyone take a look ? https://gist.github.com/ionelmc/91602c271507d02a2fb5#file-output-log-L103-L113

[13:59:22] <xafer> ionelmc, can you reproduce it with fewer wheels ?

[13:59:44] <ionelmc> doesn't repro after recreating the virtualenv

[13:59:58] <ionelmc> i guess apt upgraded some libs - or something weird

[17:02:26] <Alex_Gaynor> dstufft: RFC https://github.com/pypa/pip/pull/1877

[17:04:45] <Alex_Gaynor> dstufft: thanks

[17:04:55] <dstufft> np

[17:35:07] <Alex_Gaynor> dstufft: can you retry https://travis-ci.org/pypa/pip/jobs/27701139 , it's a network failure unrelated to the patch

[18:25:59] <Alex_Gaynor> dstufft: besides the timed out builder, I think https://github.com/pypa/pip/pull/1878 is RFC

[18:26:45] <dstufft> it timed out again?

[18:27:01] <dstufft> restarted it again

[18:27:39] <Alex_Gaynor> dstufft: thanks

[18:47:40] <Alex_Gaynor> dstufft: https://travis-ci.org/pypa/pip/jobs/27701142 :-(

[18:48:23] <dstufft> there was someone in here earlier saying that they had a test where removing a function (which just did a print statement) caused the tests to lock up

[18:48:30] <dstufft> maybe there's some sort of race condition being exposed

[18:48:43] <dstufft> also on python 3.3

[18:50:19] <Alex_Gaynor> I don't see how anything I changed could have cuased a problem :/

[18:53:31] <dstufft> Alex_Gaynor: http://bpaste.net/show/sLbkCCvZigQRk336TLaH/

[18:53:44] <Alex_Gaynor> wat

[18:54:27] <dstufft> I wonder if this is pydist's fault

[18:55:15] <dstufft> Alex_Gaynor: can you add a py33 here https://github.com/pypa/pip/blob/develop/.travis/run.sh#L10-L12

[18:55:20] <dstufft> on your branch

[18:55:25] <dstufft> just like py32

[18:55:28] <dstufft> and see if that fixes it

[19:32:27] <Alex_Gaynor> dstufft: https://travis-ci.org/pypa/pip/jobs/27708873

[19:33:19] <dstufft> so

[19:33:25] <dstufft> it's not pydists fail

[19:35:23] <Alex_Gaynor> py.test you mean?

[19:35:38] <dstufft> pytest-xdist yea

[19:35:44] <dstufft> my brain is frazzled

[19:36:01] <dstufft> i'm poking

[19:39:10] <dstufft> of course it doens't freeze for me

[19:39:44] <dstufft> wonder if this is an ubnutu 12.04 thing

[20:48:48] <jimbaker> dstufft, hey, just wanted to let you know we fixed the blocking cjk encoding support in jython. with that, i was able to update https://github.com/html5lib/html5lib-python/pull/150

[20:48:57] <jimbaker> hopefully we can get that approved soon

[21:18:24] <dstufft> jimbaker: awesome :)

[21:18:36] <dstufft> jimbaker: so that html5lib PR is the only thing holding back pip on Jython?

[21:18:45] <jimbaker> dstufft, correct

[21:19:54] <jimbaker> so next on my list is ensurepip support, as well as virtualenv support via venv

[21:20:49] <jimbaker> dstufft, btw, can one install pypa virtualenv and venv in parallel, or is that disallowed some way?

[21:21:05] <dstufft> well venv comes with Python 3.4, you can't install it seperately

[21:21:12] <dstufft> and virtualenv runs fine on 3.4

[21:22:25] <dstufft> jimbaker: if html5lib merges that and releases it within the next month or so it'll make it into pip 1.6

[21:22:30] <jimbaker> ok, just want to understand any possible implications. anyway, i believe it continues to make sense to backport venv to jython, assuming that's not a crazy amount of work. (the runtime considerations we can readily handle)

[21:23:18] <dstufft> jimbaker: basically backporting venv is primarily, modifying the startup of the intepreter so that it'll make a sys.path which is relative to a particular file if that file exists

[21:23:20] <jimbaker> dstufft, mind adding a comment like that to https://github.com/html5lib/html5lib-python/pull/150 ?

[21:23:47] <dstufft> the rest of it is just some pure python code, but it's really that intense, it primarily just creates a directory structure and adds that file

[21:24:00] <jimbaker> dstufft, yeah, seems straightforward

[21:30:50] <dstufft> jimbaker: commented

[21:31:17] <dstufft> fwiw we can't actually test that pip works with jython without virtualenv working

[21:31:25] <dstufft> so it'll be unsupported until they both work

[21:35:16] <jimbaker> dstufft, yeah, i noticed some of these cross dependencies ;)

[21:39:40] <dstufft> jimbaker: techincally virtualenv (and venv) don't have a dependency on pip

[21:39:51] <dstufft> you can create isolated environments without installing pip into them

[21:40:05] <dstufft> but those envs aren't very useful w/o pip

[21:40:56] <jimbaker> right. also we cannot leverage tox without this support. but soon we will :)

[23:45:21] <dstufft> Alex_Gaynor: strange thing, it always times out on the same test

[23:45:57] <Alex_Gaynor> dstufft: is it in a subprocess? Did I maybe break something not executed directly?

[23:46:00] <Alex_Gaynor> but wtf breaks only on 3.3

[23:46:49] <dstufft> it's subprocess in that it invokes pip

[23:47:05] <dstufft> it does pip install initools==0.2 && pip freeze && pip freeze --local

[23:48:18] <Alex_Gaynor> dstufft: ran it locally, that works fine :/

[23:50:11] <Alex_Gaynor> dstufft: How does that PR differ form mine?

[23:50:20] <Alex_Gaynor> oh, the -v

[23:50:27] <dstufft> yea I'm just getting more info

[23:50:41] <dstufft> figured it'd be easier to just do it this way then ask you to keep adding stuff

[23:51:04] <Alex_Gaynor> sure, though I'm around if I can help

[23:58:29] <dstufft> ok, so it does appear to be test_freeze_with_local_option

[23:59:19] <dstufft> https://github.com/dstufft/pip/blob/cleanup-compat/tests/functional/test_freeze.py#L279-L306

[23:59:26] <dstufft> I love that the test has one of the checks commented out

[23:59:33] <dstufft> apparently from so long ago it was still under ian's names

[23:59:43] <dstufft> and on bitbucket

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Monday the 16th of June, 2014