pmxbot IRC Log Viewer

[01:57:19] <agronholm> any virtualenv developers around? need a PR taken care of

[01:57:27] <agronholm> this: https://github.com/pypa/virtualenv/pull/747

[02:01:14] <jimbaker> dstufft, any chance we can get this changed merged in?

[02:01:56] <jimbaker> i'm pretty sure we had something similar in virtualenv in the past for JYTHONPATH vs PYTHONPATH

[02:03:47] <jimbaker> lifeless, re JYTHONPATH vs PYTHONPATH - it's something i'm mulling we should support in the next release. don't know implications however for users

[02:04:18] <agronholm> but jython 3.5 will respect PYTHONPATH at least

[02:04:25] <jimbaker> we do support most of the other PYTHON*** environment variables

[02:05:21] <lifeless> jimbaker: so I can understand not changing older releases

[02:05:29] <lifeless> jimbaker: the updated patch is fien by me, but I'm not a committer

[02:05:33] <jimbaker> (next release of 2.7, namely 2.7.1, probably with some sort of compatibility setting... for 3.5, yes, we have to change)

[02:05:37] <lifeless> dstufft: is a little busy right now changing jobs

[02:06:01] <jimbaker> lifeless, ahh, that's not so good for us at rackspace to see dstufft go...

[02:06:24] <lifeless> jimbaker: I don't know if he's said where he's going yet

[02:06:36] <jimbaker> lifeless, i assume i will see you next month in vancouver?

[02:06:43] <lifeless> jimbaker: yes indeed

[02:07:11] <jimbaker> lifeless, very much enjoyed that dinner at the kurdish restaurant

[02:07:48] <lifeless> that was fun

[02:07:51] <lifeless> and tasty

[02:08:22] <jimbaker> good company. good kurdish wine and food. in paris!

[02:09:15] <lifeless> best thing

[02:09:29] <lifeless> its been another 6 months of reasonable regular duolingo french practice

[02:09:32] <tomprince> lifeless jimbaker: dstufft is going to HP, working on pypa fuil time.

[02:09:37] <jimbaker> lifeless, not to hijack #pypa-dev, but i think i will finish up my (pseudo) greenlet support for next month. just because.

[02:09:53] <lifeless> tomprince: yes I know, just wasn't sure if he'd advertised the HP bit :)

[02:09:55] <jimbaker> lifeless, and you gain a great colleague at hp :)

[02:11:22] <jimbaker> in any event, if we can get a committer to expedite reviewing agronholm's PR that would be great

[02:11:36] <jimbaker> we are planning to release jython 2.7.0 *tomorrow*

[02:11:43] <lifeless> I suggest looking through the git info about who the committers are

[02:11:47] <jimbaker> it would be awesome if virtualenv worked

[02:11:49] <lifeless> and nagging them directly

[02:11:52] <lifeless> cause I agree

[02:20:10] <lifeless> dstufft: huh, I didn't realise pip *does* support eggs - as_egg in req_install.py.

[03:37:57] <dstufft> jimbaker: there is no Jython 3.x release right now right?

[03:38:14] <dstufft> or more importantly, there is/will be no Jython 3.x release that needs JYTHONPATH?

[03:38:59] <dstufft> lifeless: as_egg is sort of a hack, we won't download a egg and install it, we only install sdists, however there is a parameter, --egg which says "when you install this, install it as an egg". It's a tremendous hack that I want to get rid of

[03:41:58] <dstufft> https://github.com/pypa/pip/issues/1749

[03:47:33] <lifeless> dstufft: oh hi

[03:47:41] <lifeless> dstufft: question for you

[03:47:57] <lifeless> dstufft: why don't we use the orig directory to build from for non-editable file:/// urls ?

[03:48:38] <lifeless> dstufft: is --egg deprecated already?

[03:53:57] <dstufft> lifeless: it's not deprecated yet, mostly for reasons of namespace packages in 2.x, I'm not sure what we're going to do about that exactly yet, but I want to get rid of it

[03:54:10] <dstufft> lifeless: and I'm not sure re: file:// urls

[03:55:10] <lifeless> dstufft: ok, so you wouldn't object to us running sdist in the specified dir ?

[03:55:36] <lifeless> dstufft: I can copy the existing behaviour if you want

[03:55:48] <lifeless> dstufft: I'm shuffling around ownership of files on disk to enable caching

[03:55:58] <lifeless> dstufft: (of processed requirements)

[03:56:33] <dstufft> lifeless: it might exist so that pip install . doesn't shove a bunch of egg files inside of . (or a build dir)

[03:56:42] <lifeless> dstufft: ok, I'll keep it

[04:02:21] <lifeless> dstufft: the other thing was teh build dir change

[04:02:38] <lifeless> dstufft: from req.name to a mkdtemp(req.name-)

[04:25:07] <jimbaker> dstufft, correct, although jython 3.x is currently just an idea in our heads

[04:25:16] <jimbaker> with a sandbox branch to try out

[04:25:24] <jimbaker> some changes in grammar, etc

[04:26:03] <jimbaker> i would conjecture on the order of two years away, perhaps more

[04:29:04] <jimbaker> dstufft, also, thanks for merging in! 2.7.0 final was just announced (soft build, just means we have one last chance to check build correctness), so we are in really good shape

[04:29:11] <jimbaker> for jython

[04:36:35] <agronholm> awesome

[09:18:57] <lifeless> hmm

[09:18:58] <lifeless> 1 passed, 559 deselected, 1 warnings in 82.19 seconds

[09:19:04] <lifeless> wonder where that time is going

[09:29:47] <ionelmc> lifeless: is the new setup_req support in pip opt-in or opt-out?

[09:33:56] <lifeless> ionelmc: setup_req?

[09:34:24] <ionelmc> lifeless: +uires

[09:34:31] <lifeless> oh

[09:34:41] <lifeless> so the declarative stuff

[09:34:45] <lifeless> packages need to opt in

[09:35:33] <ionelmc> lifeless: i mean, if they already have that metadata (eg: d2to1)

[09:36:44] <lifeless> ionelmc: then it should just work

[09:38:44] <ionelmc> lifeless: so it's an opt-out

[09:39:59] <ionelmc> :-)

[09:44:24] <lifeless> ionelmc: I wasn't planning on making it optional beyond having the metadata or not

[09:44:58] <lifeless> so I'm not sure that I'd even describe it as opt-out

[09:52:12] <ionelmc> lifeless: the point was that all the packages that currently use d2to1 (or pbr) will automatically get opted into the new behavior

[09:54:01] <lifeless> ionelmc: yup

[09:54:17] <lifeless> ionelmc: it will affect them all, IFF they have setup_requires

[09:54:22] <lifeless> (which all those probably do)

[11:34:01] <agronholm> what new setup_req

[11:35:20] <ronny> agronholm: there was some pip support added for pbr style setup_requires metadata in setup.cfg so implicit easy_install can be avoided

[11:36:21] <agronholm> can I read more about this on distutils-sig?

[11:37:53] <ronny> not sure atm, i just caught on it via here and some pull requests

[12:30:01] <agronholm> it would really help if someone blogged about developments on packaging

[12:31:19] <ronny> agronholm: i suppose, feed would be nice with interesting changes

[12:32:26] <agronholm> I suggested that at pycon but I was mostly laughed at

[12:34:14] <Wooble> agronholm: were they laughing at the blogging, or the notion that there are developments? :)

[12:34:41] <agronholm> at the blogging

[12:34:46] <ronny> agronholm: who did the laughing?

[12:35:10] <agronholm> several people whose names you would recognize :)

[12:35:42] <ronny> so spil? i want to scold them ^^

[12:35:54] <agronholm> I'm not going to name names

[12:36:10] <Wooble> I'd blog about packaging, but that would invite everyone to look at my packages and tell me all the horrible things I've done wrong. :)

[12:36:31] <ronny> Wooble:use a artist name?

[12:36:33] <agronholm> I can do that for you

[12:38:08] <Wooble> oh neat, metadata 1.2 had its 10th birthday yesterday and I still can't use it.

[12:38:55] <ronny> Wooble: wheels use it

[12:39:33] <Wooble> ok, "I still can't get setup.py register to use it".

[12:40:02] <ionelmc> Wooble: you need to use twine to upload

[12:40:15] <ionelmc> it's goddamn annoying imho

[12:40:20] <agronholm> indeed

[12:40:26] <Wooble> ionelmc: will twine set the author and maintainer correctly in pypi?

[12:40:55] <ionelmc> not sure about that, shouldn't be any difference?

[12:41:22] <ionelmc> my pain is with dependency metadata (setup.py register/upload doesn't add it)

[12:41:28] <agronholm> does pypi have any download stats anywhere which would let me determine how much python 2.6 is still being used?

[12:41:52] <agronholm> well, at least twine uploads dependency metadata

[12:42:21] <ionelmc> agronholm: i think you have to nag dstufft to give you the data :)

[12:42:58] <ionelmc> but he already published some general stats

[12:43:01] <agronholm> where

[12:43:14] <ionelmc> agronholm: https://caremad.io/2015/04/a-year-of-pypi-downloads/

[12:43:24] <agronholm> how did you find that?

[12:44:13] <ionelmc> agronholm: he posted on some mailing list, lol

[12:44:33] <ionelmc> not very reachable

[12:44:37] <Wooble> I just tracked down the URL 2 seconds too late by going to his twitter feed.

[12:44:46] <ionelmc> but some people posted that on twitters

[12:45:22] <ionelmc> agronholm: you know rss aint dead yet

[12:45:26] <ionelmc> :-)

[12:45:44] <agronholm> I do use RSS a lot

[12:46:04] <agronholm> I don't really use twitter though

[12:46:05] <ionelmc> he's not even listed here http://planetpython.org/opml.xml

[12:46:06] <ionelmc> meh

[12:46:23] <Wooble> that caremad feed has had 3 posts since the start of 2014 :)

[12:47:10] <ionelmc> but the planet.python.org is at the lowest visitor count ever i think

[12:47:24] <agronholm> hm, good to know python 3.2 is effectively dead

[12:47:40] <agronholm> strangely so is 3.3

[12:47:52] <ionelmc> early adopters

[12:47:59] <ionelmc> they upgrade faster

[12:48:09] <ionelmc> only early adopters use py3 xD

[12:48:11] <agronholm> "early"

[12:53:16] <ionelmc> LOL

[14:10:17] <tomprince> dstufft: It appears that https://copr.fedoraproject.org/coprs/pypa/pypa/ isn't up-to-date.

[14:11:26] <dstufft> tomprince: oh right, I forgot to update it because it still relies on my doing it manually :(

[21:12:53] <lifeless> ok

[21:13:02] <lifeless> today, more refactoring towards this resolver

[21:18:52] <sigmavirus24> !m lifeless

[21:18:52] <pmxbot> you're doing good work, lifeless!

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Wednesday the 29th of April, 2015