pmxbot IRC Log Viewer

[00:08:14] <lifeless> dstufft: ?

[00:25:44] <ErikRose> Are we still thinking about a January-ish release for pip 8?

[00:31:16] <lifeless> StevenK: so - why does localextras-0.0.2 use dependency links ?

[00:32:48] <StevenK> lifeless: I wanted something very close to the existing localextras, I didn't try and trim it.

[00:37:07] <lifeless> StevenK: ah

[01:35:05] <lifeless> StevenK: ok, sorry it took so long to get you a review, bunch of stuff to page in

[01:36:12] <lifeless> StevenK: please ping me when you update it

[01:36:25] <lifeless> StevenK: (I know waiting around without input suuuucks)

[01:36:58] <lifeless> StevenK: and github either wants me to be subscribed to *everything* or get spotty notifications

[01:38:03] <StevenK> lifeless: I prodded you via IRC when I pushed it

[01:39:10] <StevenK> lifeless: I've been poking at setuptools and translations stuff, so I have to do the same paging.

[01:39:38] <StevenK> lifeless: I can drop localextras 0.0.2 if you don't think it provides value at the moment

[01:39:58] <StevenK> Since it doesn't fail, it just picks one and deals.

[01:44:33] <lifeless> StevenK: yeah, I don't care about localextras 0.0.2 either way

[01:44:45] <lifeless> StevenK: I know you did, I'm just asking you to do it again :)

[01:44:58] <lifeless> StevenK: because I feel bad about how long I left you hanging

[01:45:25] <lifeless> StevenK: I have I think however identified the confusion between you and Nakato and proposed a change that will resolve things

[01:45:36] <lifeless> StevenK: as well as describing in detail the xfail test we need

[04:27:16] <StevenK> lifeless: You're totally right, re: adding existing_req.extras == install_req.extras for the error case, I can drop my code path and then Nakato's fires anyway.

[04:28:52] <StevenK> Which changes the diff of req_set.py to +5/-1. \o/

[04:30:31] <StevenK> lifeless: I'm not following your information dump about the tests, you also say it might be overkill, so I can push this up after I rewrite the prose in the commit message if you wish.

[05:17:27] <lifeless> StevenK: push it up for sure

[05:18:10] <lifeless> StevenK: the thing missing is a single failing test

[05:18:18] <lifeless> e.g.conflictingextras depends on simple==1.0, but conflictextras[bad] depends on simple==2.0

[05:18:22] <lifeless> pip install conflictingextras conflictextras[bad]

[05:18:24] <lifeless> today with your patch will install simple==1.0, in future with the resolver it should error.

[05:18:27] <lifeless> just that bit

[05:55:06] <StevenK> lifeless: All done, pushed up.

[05:55:22] <StevenK> Now to get out of my study again before I melt into a puddle of goo.

[06:41:04] <lifeless> StevenK: the heeeeat

[06:41:09] <lifeless> StevenK: you'll like summer here

[06:43:49] <StevenK> lifeless: If it doesn't mean dying in 42C, yes.

[06:45:46] <lifeless> StevenK: one tiny tweak

[06:45:55] <lifeless> StevenK: and then we start nagging dstufft for reviews

[06:46:19] <StevenK> Does dstufft want to be nagged, really? :-P

[06:47:35] <lifeless> yes, yes he does

[06:48:52] <StevenK> lifeless: I don't understand the code well enough to work out why pip is happy enough with simple==1.0

[06:49:06] <StevenK> Which makes it difficult to formulate a comment that makes sense.

[06:51:13] <lifeless> StevenK: So, explain the situation

[06:51:20] <lifeless> StevenK: there's nothing in the test that explains it

[06:51:24] <lifeless> StevenK: then tack this on

[06:51:52] <lifeless> StevenK: "without a resolver, pip does not detect the conflict between simple==0.0.1 and simple==0.0.2. Once a resolver is added, this conflict should be detected.

[06:52:30] <lifeless> StevenK: you should also use

[06:52:33] <lifeless> https://pytest.org/latest/skipping.html

[06:52:37] <lifeless> @pytest.mark.xfail

[06:53:13] <lifeless> and have your test assert that the result code is non-zero, or something

[06:53:23] <lifeless> StevenK: I believe the run helper lets you pass in an expected result code

[06:54:27] <StevenK> pip_install_local() throws an AssertionError, see test_install_distribution_duplicate_extras

[06:55:56] <lifeless> StevenK: pass expect_error=True

[06:58:30] <StevenK> lifeless: http://paste.openstack.org/show/479538/

[06:59:37] <lifeless> StevenK: yes, except your assert is now wrong- you need to check that an error happened :)

[07:00:23] <lifeless> StevenK: e.g. assert 'installed' not in result.stdout \n assert 'Conflict' in result.stderr

[07:00:34] <lifeless> StevenK: or something like that, which will fail now, and thus the xfail will kick in

[07:10:44] <StevenK> lifeless: Pushed up.

[07:22:35] <lifeless> cool, will review monday

[14:37:39] <dstufft_test> Deployed test-app at 0b5fb19

[14:54:18] <pypabot> Deployed 0b5fb19 to Warehouse (Staging)

[15:04:50] <pypabot> Deployed 0b5fb19 to Warehouse (Production)

[15:05:31] <pypabot> Deployed 70de350 to Warehouse (Production)

[15:25:02] <pypabot> Deployed cced45e to Warehouse (Production)

[17:45:48] <xafer> hello lifeless, I'm the train and unable to check easily before sunday: does 'pip install . pip==6' work ? (where . would be a 8.0.dev0 workspace)

[18:02:51] <lifeless> xafer: yes

[18:03:00] <lifeless> xafer: where works == installs pip 6.0 without any error

[18:03:13] <lifeless> pip$ pip install . pip==6

[18:03:13] <lifeless> Processing /home/robertc/work/pip

[18:03:14] <lifeless> Collecting pip==6

[18:03:14] <lifeless> Downloading pip-6.0-py2.py3-none-any.whl (1.3MB)

[18:03:14] <lifeless> 100% |████████████████████████████████| 1.3MB 239kB/s

[18:03:16] <lifeless> Installing collected packages: pip

[18:03:19] <lifeless> Found existing installation: pip 7.1.2

[18:07:32] <xafer> 7.1.2 ? Does pip list show 7.1.2, 6 or 8.0.dev ?

[18:10:31] <xafer> any way, point taken, pip is already inconsistent 😢

[18:32:57] <lifeless> xafer: it installs 6.0

[18:33:02] <lifeless> xafer: so pip install shows 6.0

[18:33:12] <lifeless> xafer: (after the operation)

[18:33:26] <lifeless> xafer: and yeah, it is :)

[19:21:36] <xafer> pip has a bunch of PR waiting for review... any idea on how we could reduce the queue ?

[19:32:11] <xafer> I'm wondering if a process could help... With a few guidelines on how to deal with new feature requests or abandoned PR

Log file Viewer

Help | Karma | Search:

#pypa-dev logs for Friday the 20th of November, 2015