pmxbot IRC Log Viewer

[03:03:25] <sontek> Hey, with pypi is there a way to get it to follow the same lookup semantics that pip does? I released a new release candidate of a lib and the rc shows as the default version

[03:03:34] <sontek> Even though pip still pulls the stable version

[03:12:26] <sontek> Also, did something change with how restructured text is used in the docs? My README renders just fine on github and with python setup.py --long-description | rst2html.py > output.html but its completely broken on ppi

[03:12:31] <sontek> pypi*

[03:13:12] <sontek> https://pypi.python.org/pypi/pyramid_celery

[03:46:15] <ionelmc> sontek: use collective.checkdocs

[03:46:44] <ionelmc> probably your missing the empty line under the headings

[03:46:49] <ionelmc> you're

[03:48:07] <ionelmc> you can hide the rc under https://pypi.python.org/pypi?:action=pkg_edit&name=pyramid_celery i think

[03:49:56] <sontek> ionelmc: Yeah, I was just wondering why it can't do it by default

[03:53:18] <sontek> ionelmc: collective.showdocs/checkdocs both seem to think the README.rst is fine as well

[03:55:17] <ionelmc> sontek: no idea, tho you have a typo: Yyou

[03:56:39] <sontek> hah, good catch

[03:56:48] <sontek> Maybe it refuses to render if spellcheck fails

[03:58:45] <ionelmc> for sure =)

[03:59:43] <sontek> Tried using https://pypi.python.org/pypi/restview to test it as well and everything thinks its fine

[03:59:57] <ionelmc> sontek: maybe it's a problem with the codeblocks (eg, invalid syntax)

[04:00:04] <ionelmc> install pygments and run checkdocs again

[04:00:47] <sontek> Just wondering what pypi is using that collective.showdocs, restview, and rst2html aren't using

[04:00:56] <sontek> since they all have no issue with it and are built to detect pypi problems

[04:01:36] <sontek> Or is there at least a way to view logs to see the error pypi is seeing?

[04:04:23] <ionelmc> sontek: no idea :|

[04:04:34] <ionelmc> maybe where some output when you do setup.py register

[04:04:46] <ionelmc> otherwise your best shot is https://bitbucket.org/pypa/pypi/issues or http://sourceforge.net/p/pypi/support-requests/

[04:05:01] <ionelmc> s/where/there's/

[04:54:13] <sontek> I e-mailed the list and commented on a ticket similar to my problem

[04:54:33] <sontek> I'm sure its something I'm doing wrong but its hard to debug with no tool failing but pypi

[06:29:53] <tomprince> sontek: I think it uses something custom. dstufft plans to change it to readme (and warehouse already uses it.

[06:30:11] <dstufft> yea it's a custom mess of code

[06:30:26] <dstufft> that bit of code is sort of standalone-ish in PyPI itself though

[06:30:55] <dstufft> normally when I have a problem I just copy/paste it out of PyPI and add some print statements

[06:30:56] <dstufft> :/

[06:53:14] <sontek> dstufft: Want to link me to the code? Or is there a way to get someone to find out why its failing?

[06:54:03] <dstufft> https://bitbucket.org/pypa/pypi/src/48f1add0748e59746f0e98ca89cc6f0dbef2fc7e/description_utils.py?at=default#cl-102

[06:54:04] <dstufft> processDescription function

[07:05:19] <sontek> dstufft: thanks

[07:42:39] <sontek> dstufft: Could there be a chance something is broken on pypi right now?

[07:42:49] <sontek> dstufft: I ripped that source code out and ran it against my README and it worked perfectly

[07:49:14] <dstufft> sontek: always possible, nothing that i'm aware of though

[07:49:51] <dstufft> sontek: does warehouse.python.org render it?

[07:52:28] <sontek> dstufft: I figured out what was breaking it at least (just kept deleting lines until it worked)

[07:52:40] <dstufft> what was it

[07:53:13] <sontek> dstufft: warehouse renders it fine though

[07:53:25] <sontek> its a .. codeblock:: without a language added to it

[07:54:07] <dstufft> I just need to switch pypi over to using readme

[07:55:38] <sontek> dstufft: https://github.com/sontek/pyramid_celery/commit/fb69dd44143a5b030b48346e6322845194a690fe thats the change

[09:50:38] <apollo13> dstufft: around?

[09:50:55] <dstufft> apollo13: yea

[09:51:36] <apollo13> dstufft: so for pip6 is there anything you'd recommend as pip.conf?

[09:51:46] <apollo13> from the looks of it, I can just run without one fine

[09:52:14] <dstufft> only thing is if you want to use ``pip wheel`` to make your own wheel cache that'll be used

[09:52:29] <apollo13> nice :)

[09:52:41] <apollo13> and another thing: https://hg.python.org/cpython/file/1c51f1650c42/Lib/asyncio/selector_events.py#l699 -- do you have any idea why python3.5 code has fallbacks for 3.3?!

[09:53:58] <dstufft> if you want to make your own wheel cache you can add the wheel-dir and find-links to your pip.conf, it'll require you to do like ``pip wheel -r requirements.txt`` but after that subsequent ``pip installs`` will use the wheels youj built

[09:54:17] <dstufft> apollo13: probably because asyncio is distributed on PyPI

[09:54:25] <dstufft> and it makes it easier to do that if you use the same code

[09:54:38] <the_drow_> Anyone knows in which directory 6.x caches sdist or wheels?

[09:54:48] <the_drow_> It's not specified in the documentation

[09:54:55] <apollo13> .cache/pip

[09:54:57] <dstufft> the_drow_: what OS

[09:55:00] <the_drow_> linux

[09:55:12] <dstufft> it follows XDG for user cache dir

[09:55:14] <the_drow_> It's for travis-ci

[09:55:32] <the_drow_> XDG?

[09:55:33] <dstufft> http://standards.freedesktop.org/basedir-spec/basedir-spec-latest.html

[09:56:14] <dstufft> which roughly translates to ~/.cache/pip unless someone sets it to something different either via $XDG_CACHE_HOME or the pip config to point it somewhere else

[09:56:19] <the_drow_> Oh so it's the home directory

[09:56:28] <the_drow_> dstufft: Can someone document it somehow?

[09:56:50] <apollo13> it's funny how people have different expectations

[09:57:15] <the_drow_> Also since travis uses pip 1.5 how can I configure the environment to ensure that pip is forward compatible?

[09:57:21] <apollo13> when I saw the cache dir deprecation warning in pip 6.0 I immediatelly checked ~/.cache/pip, cause that is where stuff is supposed to bew :)

[09:57:30] <apollo13> the_drow_: pip install --upgrade pip ? :)

[09:58:09] <the_drow_> But that also takes build time.

[09:58:23] <apollo13> it's a wheel, so…

[09:59:07] <dstufft> the_drow_: open a ticket wrt documentation

[10:05:12] <the_drow_> dstufft: will do

[10:08:34] <the_drow_> dstufft: https://github.com/pypa/pip/issues/2277

[11:25:51] <agronholm> how should I declare a dependency for only python < 3.2?

[11:25:57] <agronholm> in setup.py

[12:13:54] <doismellburning> agronholm: the best way I know - https://github.com/py3minepi/py3minepi/blob/master/setup.py#L10-L12

[12:14:15] <agronholm> doismellburning: that won't work for wheels

[12:23:49] <agronholm> doismellburning: extras_require={':python_version=="2.6"': 'argparse'} works

[12:23:59] <agronholm> but it doesn't seem to work with <

[12:42:09] <ionelmc> list every version :)

[12:42:13] <ionelmc> there aren't that many

[12:42:28] <agronholm> I really shouldn't have to

[12:42:37] <agronholm> since this works perfectly in setup.cfg

[12:44:26] <ionelmc> must be a bug then

[12:44:56] <ionelmc> agronholm: what would you put in setup.cfg?

[12:44:56] <doismellburning> agronholm: interesting, thanks

[12:45:40] <agronholm> argparse; python_version < '2.7'

[12:45:52] <agronholm> in requires_dist

[12:45:55] <agronholm> under [metadata]

[13:04:50] <ionelmc> agronholm: hahah, well, it allows the "in" and "not in" operators

[13:04:59] <ionelmc> https://bitbucket.org/pypa/setuptools/src/e3d6b547ac05cf456ae09c0924f2a16b0d2e6938/pkg_resources/__init__.py?at=default#cl-1411

[13:05:23] <ionelmc> kinda makes you wonder why it doesn't have lt/gt

[13:05:29] <agronholm> indeed

[13:05:43] <ionelmc> so it basically compiles that down to ast

[13:06:03] <ionelmc> and walks it, if there something not in that dict, raises error

Log file Viewer

Help | Karma | Search:

#pypa logs for Sunday the 28th of December, 2014