pmxbot IRC Log Viewer

[00:09:12] <lifeless> Nakato: ping

[00:13:23] <Nakato> lifeless: pong

[00:13:59] <lifeless> Nakato: so this pbr bug :)

[00:14:05] <lifeless> Nakato: tchaypo has said you're tackling it now

[00:14:11] <lifeless> Nakato: I prepped some patches to make testing it easier

[00:14:29] <lifeless> https://review.openstack.org/#/c/230243/

[00:14:51] <lifeless> Nakato: I'm really really keen on getting it fixed asap, since its hurting folk daily

[00:15:04] <lifeless> Nakato: wondering if it is what you're hacking on, and if so how I can help

[00:15:52] <Nakato> I'm working on the test, though I'm not sure how that patch makes things easier

[00:16:58] <lifeless> Nakato: tchaypo said he got stuck setting up a fully isolated environment without pbr in it

[00:17:19] <lifeless> Nakato: so this makes making a venv with just setuptools/wheel installed easy

[00:18:00] <lifeless> Nakato: but perhaps you're doing a narrower test? This was a from-the-outside style test we'd previously discussed

[00:31:44] <Nakato> lifeless: Is there anything making a sdist of pbr, or something I can get easy_install to install?

[00:38:45] <lifeless> Nakato: an sdist is probably made in the other integration tests, but I think we want a little mini-index

[00:40:01] <lifeless> Nakato: easiest way is to do '$venv/bin/pip wheel $pbr_root_dir -w tempdir/wheel' within the test, whic

[00:40:40] <lifeless> Nakato: oh or yeah, sdist - I guess because easy_install sadface

[00:42:09] <lifeless> Nakato: then follow https://pythonhosted.org/setuptools/easy_install.html#creating-your-own-package-index and https://pythonhosted.org/setuptools/easy_install.html#configuration-files

[00:42:26] <lifeless> specifically in ~/.pydistutils.cfg

[00:42:43] <lifeless> (which should be in a TemporaryHomeDir fixture so as not to pollute any global stuff)

[00:42:48] <Nakato> setup.cfg also works too, which is nice.

[00:43:04] <lifeless> Nakato: oh cool, ok yeah

[00:43:11] <Nakato> Is there anything already grabbing pbr_root_dir, as that's what's got me choked right now.

[00:43:59] <lifeless> Nakato: in the test file - os.path.abspath(os.path.join(__file__, '..', '..', '..'))

[00:44:03] <lifeless> Nakato: or something like that

[00:45:07] <Nakato> lifeless: Awesome!

[01:28:17] <lifeless> Nakato: where does that get you to?

[01:28:46] <Nakato> lifeless: Pretty much done, cleaning up a bit and then I'll push

[01:29:02] <lifeless> \o/

[01:29:12] <lifeless> Nakato: does it demonstrate the problem ?

[01:34:43] <Nakato> lifeless: I'm goign to go grab some lunch. Let me know how it looks. https://review.openstack.org/#/c/226088/

[01:38:18] <Nakato> tchaypo: ^

[02:06:30] <lifeless> Nakato: I have some cosmetic improvements

[02:14:19] <lifeless> Nakato: but the patch you have seems to be installing pbr from pypi, not the local pbr - I see

[02:14:24] <lifeless> /tmp/tmp5ZtMpv/tmp17U7sf/.eggs/pbr-0.0-py2.7.egg/pbr/pbr_json.py

[02:14:30] <lifeless> if not cmd.distribution.pbr:

[02:15:17] <lifeless> yeah - Downloading https://pypi.python.org/packages/source/p/pbr/pbr-1.8.0.tar.gz#md5=bc7173f5ccec8e6125e3b68f0285eb8a

[02:20:24] <Nakato> /tmp/tmp2ym8rfr8/tmppad1p6ha/.eggs/pbr-0.0-py3.5.egg/pbr/pbr_json.py

[02:20:24] <Nakato> 23 if not hasattr(cmd.distribution, 'pbr') or not cmd.distribution.pbr:

[02:20:30] <Nakato> Let me try with 27

[02:23:38] <lifeless> I'm going to push up my tweaks

[02:24:09] <lifeless> Nakato: they are - don't use six (no need), rebase on the virtualenv fixture (so as to avoid future conflicts)

[02:24:37] <lifeless> Nakato: some textwrap format tweaks (so that the left hand side is visually aligned

[02:25:06] <lifeless> Nakato: and putting you in the co-authored-by

[02:27:54] <lifeless> Nakato: to run just your test I'm doing: TOX_TESTENV_PASSENV=PBR_INTEGRATION PBR_INTEGRATION=1 tox -epy27 -- test_install_without_pbr

[02:31:26] <Nakato> Thanks, taking a look

[02:36:29] <lifeless> I put a print statement in pbr_json.py, and it isn't there in the traceback in the nopbr-stderr attachment

[02:37:01] <lifeless> so, pretty sure we're running off of PyPI or similar

[02:37:37] <lifeless> Nakato: oh, I have a nice trick for you I'll put in another push

[02:39:24] <Nakato> Yea, 2.7 is definitely downloading from pypi

[02:39:31] <lifeless> Nakato: what were you testing with ?

[02:40:18] <Nakato> I dropped a sleep into the test to give me time to snapshot the environment it built so I could go in and fiddle with it.

[02:42:13] <Nakato> Looks like in py27 it creates a sdist for pbr_testpackage and not pbr

[02:47:04] <lifeless> Nakato: pushed up another iteration - I've grouped the setup stuff so that each package is worked on in one unit - and added comments

[02:47:28] <lifeless> Nakato: any idea /why/ testpackage is being picked up

[02:49:04] <lifeless> Nakato: but I can reproduce it - pbr_root was /tmp/tmpioVKqe/tmpHOIBqa/testpackage for me

[02:49:12] <Nakato> lifeless: __file__ is resolving to pbr/tests/test_integration.pyc in 2.7 and in 3.5 it's /home/nakato/git/openstack-dev/pbr/pbr/tests/test_integration.py

[02:49:12] <Nakato> So I'm guessing (no actual clue) that when abspath, it's switching to using os.getcwd() or something.

[02:50:03] <lifeless> ah

[02:50:04] <lifeless> I know

[02:50:20] <lifeless> >>> import os.path

[02:50:20] <lifeless> >>> os.path.abspath('')

[02:50:20] <lifeless> '/home/robertc/work/pbr'

[02:50:22] <lifeless> you're spot on

[02:51:31] <lifeless> there

[02:52:19] <lifeless> Nakato: pushed up a fix

[02:52:33] <lifeless> Nakato: resolving the path before the unique tempdir isolation kicks in

[02:52:37] <lifeless> Nakato: fixeth it

[02:57:49] <Nakato> lifeless: Just looked at patch 5, much cleaner.

[02:57:50] <Nakato> \o/

[02:58:35] <lifeless> Nakato: 6 is current, but its just a trivial tweak

[02:58:39] <lifeless> Nakato: to fix the path

[02:59:02] <lifeless> Nakato: ok cool, so whats next on our hitlist? tox?

[03:03:48] <lifeless> Nakato: (also nice working getting us to a reproducible patch, super important)

[03:03:54] <lifeless> s/patch/test/

[03:05:05] <Nakato> tox is held up on hpk and suor deciding what to do with regards to wording on how to run factors and sections. And then if it's even going to be worth implemnting once that's done.

[03:06:11] <Nakato> So I've taken to the duplication method. https://review.openstack.org/#/c/228315/1/tox.ini

[03:06:41] <Nakato> The patch chain to get more -constraints jobs in https://review.openstack.org/#/c/228343/

[03:08:09] <Nakato> Currently looking at periodic job for fixing up the URLs for stable branches.

[03:08:40] <Nakato> lifeless: Did you want someone to look at https://bugs.launchpad.net/pbr/+bug/1502692 ?

[03:09:02] <Nakato> I wouldn't mind starting to work on that as well

[03:09:20] <lifeless> I think it would good to fix

[03:09:30] <lifeless> so yes please - I added it to trello

[03:10:10] <Nakato> lifeless: Oh, and I guess we should figure out how to proceede with https://github.com/pypa/pip/issues/3046 so I can finish that off

[03:11:12] <lifeless> Nakato: question for you on the constraints tox.ini

[03:11:26] <lifeless> Nakato: yah

[03:24:43] <Nakato> You know what's absolutely awesome and confusing. I could delete the install_command line from the -constraints sections and it'd use the constraints_install command... But *everything* else would still have to be duplicated.

[03:33:04] <lifeless> yeah

[03:33:10] <lifeless> definately confusing:)

[20:41:13] <ErikRose> dstufft: Hashing is ready for review! https://github.com/pypa/pip/pull/3137

[20:41:23] <ErikRose> Docs are updated.

[20:41:55] <ErikRose> I'm going to go add the 3-liners to add --require-hashes support to pip download and pip wheel. They already check hashes if they're there, since I did the work in RequirementSet.

[21:12:59] <Ivo> sent you a personal mail dstufft cus I'm really good at hitting reply instead of reply-list, soz

[21:46:40] <ErikRose> dstufft: pip wheel and pip download now support --require-hashes.

[21:47:00] <ErikRose> jezdez: If you wanted to review, you're welcome as well.

[22:32:28] <ErikRose> Got "pip hash" working and documented. Just needs a test.

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Wednesday the 7th of October, 2015