Log file Viewer

[18:36:02] <lifeless> \o/ merges

[18:46:22] <lifeless> dstufft: 2709 updated

[18:47:21] <dstufft> lifeless: looks good, will merge once tests

[18:50:13] <lifeless> dstufft: ok;how does 2710 look ?

[18:51:45] <dstufft> lifeless: looking

[19:15:04] <lifeless> dstufft: see https://github.com/rbtcollins/pip/blob/5456a293d15ce0f030322a58ec4397935e3bccab/pip/wheel.py#L50 for the next change to the Cache

[19:15:32] <lifeless> (thats inside 2699)

[19:15:37] <dstufft> lifeless: ok cool

[19:30:12] <sigmavirus24> !m lifeless

[19:30:12] <pmxbot> you're doing good work, lifeless!

[19:48:09] <lifeless> dstufft: 2710 pushed up with changes, waiting on tests

[20:02:50] <lifeless> ah travis, you so silly

[20:03:16] <lifeless> dstufft: can you tickle travis for 2710? test result is for the push before...

[20:34:17] <dstufft> hm

[20:34:19] <dstufft> that's strange

[20:34:23] <dstufft> not sure what happened

[20:34:32] <dstufft> you might need to git commit --amend and git push -f

[20:43:07] <lifeless> I've done that

[20:45:31] <sigmavirus24> you mean people who work on openstack haven't all aliased amend='commit --amend' ?

[20:45:35] <sigmavirus24> =P

[20:46:25] <lifeless> travis definitely have a race in there

[20:46:26] <lifeless> sheese

[20:51:22] <sigmavirus24> lifeless: tweet at Travis. My friend will respond and I'll teach you to troll her

[20:51:29] <sigmavirus24> Won't fix our problems, but it'll be funny

[20:53:03] <lifeless> trolled

[21:32:39] <lifeless> sigmavirus24: https://twitter.com/rbtcollins/status/590991216033959936

[21:32:45] <lifeless> dstufft: can you try tickling travis again ?

[21:33:26] <lifeless> in fact, let me do that.

[21:33:40] <sigmavirus24> lifeless: I'm waiting for them to respond

[21:34:10] <lifeless> https://github.com/pypa/pip/pull/2712 is 2710 in a new PR

[21:34:19] <lifeless> and AHAHAHAHAH FUCK

[21:34:21] <lifeless> excuse me

[21:34:25] <lifeless> but travis

[21:35:06] <lifeless> (it picked up the same wrong result...)

[21:40:25] <pixelfog> I'm thinking that the "Ensuring Repeatability" section of the pip user guide needs to include a caveat about how setup_requires allows a package to ignore pip's configuration and pull down fresh bits from the Internet. At minimum, the "Ensuring Repeatability" section should have a link to the "Controlling setup_requires" section in the Reference guide. Opinions?

[21:41:45] <lifeless> pixelfog: seems reasonable to me

[21:41:59] <lifeless> pixelfog: I'd propose a link

[21:42:09] <lifeless> pixelfog: if only so that I don't need to edit more places in my setup_requires work

[21:43:00] <pixelfog> lifeless: edit more places? I don't know what you mean.

[21:44:50] <lifeless> pixelfog: once I fix setup_requires, if there's duplicate stuff about setup_requires in the manual, I'll need to make duplicate edits in my patch :)

[21:47:35] <pixelfog> You are working on a patch to the setup_requires documentation? In setuptools or in pip? Can I see it?

[21:48:35] <lifeless> no, to setup_requires support

[21:48:50] <lifeless> let me see if I hve a branch up atm

[21:49:27] <lifeless> this is alittle stale

[21:49:35] <lifeless> I've been yak shaving for a fortnight

[21:49:57] <lifeless> pixelfog: here https://github.com/rbtcollins/pip/tree/declarative

[22:02:19] <lifeless> man

[22:02:26] <lifeless> I just don't see how this - https://travis-ci.org/pypa/pip/jobs/59628689 is possible

[22:02:31] <lifeless> I fixed stub_parse_requirements

[22:02:41] <sigmavirus24> lifeless: it's all an illusion

[22:02:50] <sigmavirus24> lifeless: could also be GitHub's API being total shit again

[22:02:52] <lifeless> and the test passes locally

[22:02:58] <sigmavirus24> You know, like they do every so often

[22:03:10] <lifeless> the commit being tested is https://github.com/pypa/pip/commit/04b8f8c25ecb85dc850f3363e8bc766a1382d672

[22:03:25] <lifeless> which is a merge of 258, which is my commit

[22:04:23] <lifeless> orI'm being a muppet.

[22:04:28] <lifeless> HTF did the test pass localy.

[22:04:59] <lifeless> (don't answer that)

[22:06:58] <sigmavirus24> lifeless: I know how

[22:07:01] <sigmavirus24> But I can't tell you

[22:07:27] <lifeless> sigmavirus24: I need alcool

[22:07:33] <lifeless> true story

[22:07:35] <sigmavirus24> alcool?

[22:07:41] <lifeless> l'alcool :)

[22:07:44] <lifeless> acohol

[22:07:50] <lifeless> bah, *now* I typo.

[22:07:50] <sigmavirus24> acohol?

[22:07:54] <lifeless> alcohol

[22:07:58] <sigmavirus24> OOOOO

[22:08:01] <sigmavirus24> I thought you wanted something to do with COBOl

[22:08:01] <lifeless> was in french the first times

[22:08:05] <sigmavirus24> *COBOL

[22:08:11] <lifeless> then I'd need even more

[22:08:21] <lifeless> plus hazard pay

[22:14:24] <pixelfog> lifeless: I want to add something like this to the Ensuring Repeatability section: "pip cannot guarantee a repeatable installation if any of the packages in the requirements files use the setup_requires keyword. See "Controlling setup_requires". If that sounds right, I can make a PR.

[22:15:10] <lifeless> pixelfog: A tweak - if any of the packages to be installed utilise...

[22:15:24] <lifeless> pixelfog: because a partial requirements file (if someone did that) might include a setup_requires thing transitively

[22:18:04] <lifeless> pixelfog: you'll want to rebase on develop too, since ^ will have conflicted

[22:20:48] <pixelfog> lifeless: My pull request should be against develop, not master?

[22:20:50] <lifeless> dstufft: I think I've got patches up (or disagreements on the bug :)) for all the release blockers

[22:20:54] <lifeless> pixelfog: yes

[22:21:00] <lifeless> pixelfog: develop is 'next version' vs stable

[22:21:13] <pixelfog> lifeless: Got it.

[22:21:34] <lifeless> dstufft: so I'm going to move onto the next thing in my queue, other than keeping on top of travis

[22:21:41] <lifeless> dstufft: if you want something please do shout

[22:21:55] <lifeless> dstufft: https://github.com/pypa/pip/pull/2710 is green

[22:22:06] <dstufft> lifeless: will take a look in a bit, thanks!

[22:23:44] <lifeless> dstufft: when yuou get a sec, I'd like to know why --no-deps is specified there, given pip's current behaviour

[22:23:51] <lifeless> (the ensuring repeatability)

[22:25:24] <dstufft> lifeless: um, probably ana rtifact of a time when versions could change

[22:26:49] <lifeless> does github let you put some prose on the bug filing screen

[22:26:57] <lifeless> e.g. to say 'please list python and pip and setuptools versions'

[22:38:04] <sigmavirus24> lifeless: it doesn't

[22:38:11] <sigmavirus24> lifeless: yet another reason I hate GitHub issues

[23:05:04] <lifeless> I need ideas

[23:05:14] <lifeless> I'm working on issue 2687

[23:05:21] <lifeless> I figure we need an escape hatch

[23:05:36] <lifeless> so that if someone doesn't care about co-installability, they can opt out

[23:05:40] <lifeless> perhaps opt-out of the check

[23:05:50] <lifeless> or ???

[23:05:54] <lifeless> and what to call ze option

[23:16:44] <lifeless> also when to consider changing existing

[23:30:46] <citruspi> dstufft: Re: #2670 (https://github.com/pypa/pip/issues/2670) should my pull request make "-qqq" disable all output or just leave it at the error level for now?

[23:32:00] <lifeless> citruspi: FWIW - I think we would want the actual tell-this-to-the-user stuff to be output

[23:32:09] <lifeless> citruspi: e.g. if pip itself throws an exception, that shouldn't be hidden

[23:32:16] <lifeless> IMNSHO

[23:33:18] <citruspi> lifeless: So should "-qqq" map to the critical level?

[23:33:26] <lifeless> would make sense to me

[23:34:17] <citruspi> Okay, I'll add that.

[23:34:22] <citruspi> Do I need to do this - https://github.com/pypa/pip/commit/4d042621e8ba857efca70455af47ed5d2bf0767e?

[23:34:42] <citruspi> Or does the test for a single "quiet" accomplish everything needed?

[23:35:39] <lifeless> I'm not sure; will defer to qwcode / dstufft etc