Log file Viewer

[04:18:16] <lifeless> dstufft: so, you want pip install PATH to always reinstall it right ?

[06:12:01] <ronny> btw, hoow does one add extras to pip install PATH?

[06:37:25] <lifeless> ronny: sec, looking up an issue for you

[06:39:10] <ronny> (currently i use find-links, it just wouldnt work with filename[extras]

[06:39:38] <tomprince> ronny: It doesn't work currently (unless you use -e)

[06:40:06] <tomprince> Not for any good reason, that I am aware of, though.

[06:41:30] <ronny> tomprince: i need it for gumby elf, for develop installs using it, i generate a wheel with a local version, thant installs the metadata and a bootstrapper

[06:42:03] <ronny> pip wheel installing already handle so many details of the installer side, its insane not to make use of that ^^

[06:42:34] <tomprince> Ok ...

[06:43:54] <ronny> for example exe generation on windows

[06:45:19] <tomprince> I'm not sure why you are telling me?

[06:48:00] <lifeless> ronny: found it. https://github.com/pypa/pip/issues/1236

[06:48:29] <lifeless> ronny: This should be a fairly straight forward fix. I might even have fixed it in my issue-988 branch, come to think of it. Give it a go.

[06:55:52] <ronny> lifeless: at first glance i dont see a test for it, and i cant test atm, i dont have the relevant computer with me

[06:58:08] <lifeless> ronny: I haven't added a test, but I've changed a tonne of plumbing that could have been responsible for discarding the extras

[07:00:07] <tomprince> There also appears to be a PR for that issue, with some tests.

[07:03:16] <lifeless> https://github.com/pypa/pip/issues/1161 <- some core might want to consider closing that

[07:04:45] <lifeless> tomprince: well, that will conflict with my branch, but not too badly, and it looks like I won't have fixed it accidentally, given the nature of the fix

[08:32:49] <lifeless> hmm, there's memory leak of some sort in the test suite

[08:33:14] <lifeless> or maybe my new code :<.

[11:03:11] <lifeless> dstufft: ok its still very rough, but its now ~ feature complete

[11:03:55] <lifeless> dstufft: worth a nitpicky review - you'll likely find lots of stuff I already know about- dead code not removed, refactoring half done etc - but it will at least let you see the broad thing intact

[11:24:39] <dstufft> lifeless: ok

[11:24:50] <dstufft> lifeless: will review after breakfast :D

[18:45:38] <lifeless> dstufft: https://github.com/pypa/pip/issues/1161 - thoughts on my last comment

[19:18:16] <sigmavirus24> I figure someone here might have better ideas about https://gitlab.com/pycqa/flake8/issues/56 than I do, so your input would be both welcomed and appreciated

[19:38:00] <lifeless> sigmavirus24: hmm, so you dynamically generate a script?

[19:38:16] <lifeless> sigmavirus24: rather than just linking/copying in something generated by pip?

[19:38:20] <sigmavirus24> lifeless: it should just be a separate asset that we copy into place

[19:38:32] <sigmavirus24> But right now it's a static string because before it used to be a lot different

[19:38:32] <lifeless> yeah

[19:38:39] <sigmavirus24> Yeah this is also per-git repo

[19:38:51] <sigmavirus24> I'm not sure how we'd link something generated by pip

[19:39:05] <lifeless> on unix ln -s :)

[19:39:19] <sigmavirus24> Except we work on many a platform =P

[19:39:21] <lifeless> use which to find the version on the path

[19:39:23] <lifeless> on windows

[19:39:35] <lifeless> use the path to find it, then copy

[19:39:54] <sigmavirus24> Right, the problem is more along the lines of "We don't yet have an entry-point for the git or mercurial stuff"

[19:41:18] <sigmavirus24> So copying whatever pip would generate for that would be fun

[19:41:51] <sigmavirus24> I'm also wondering how doing that would affect downstream redistributors

[20:30:23] <dstufft> sigmavirus24: we use distlib's script support to miunge paths right now

[21:21:21] <lifeless> dstufft: did you have a look at my eyesore?

[21:55:40] <dstufft> lifeless: not yet, got distracted by poking at s3ing PyPI again

[21:55:44] <dstufft> will look in a moment

[22:02:35] <dstufft> lifeless: do you have the same hobby I do, going to read old opened issues in bug trackers :V

[22:14:50] <lifeless> dstufft: well, I look for things related to the thing I'm hacking on :)

[22:20:58] <lifeless> dstufft: 102 failing tests, I'm starting to burn those down now, before getting into code cleanup

[22:32:44] <dstufft> lifeless: small style nit, can you add whitespace around if statements? like https://bpaste.net/show/0669da10c9af I wouldn't hold up a review for that issue, but it always catches me that I have to go back and double check if it's a bunch of if statements or an if else chain

[22:40:10] <dstufft> lifeless: oh man, I jsut got into the meat of this and I think my head hurts

[22:40:16] <lifeless> dstufft: sure

[22:40:19] <lifeless> I can do that

[22:41:20] <lifeless> dstufft: its just your everyday run of the mill totally rework something that took years to get where it was using a random sampled walk in a multi-graph NPC satisfiability problem patch :).

[22:43:08] <dstufft> lifeless: heh, yea :D

[22:45:27] <dstufft> I'm trying to figure out the best way to review this heh

[22:45:33] <lifeless> hammer and tongs

[22:45:37] <lifeless> I was trying to make it all modular

[22:45:46] <lifeless> I only got so far before the rage took over

[22:45:52] <lifeless> I am happy to back-modulate it

[22:46:11] <lifeless> I suggest taking req_cache as a block

[22:46:13] <lifeless> look through that

[22:46:15] <lifeless> see if it makes sense

[22:46:25] <lifeless> compare it to your mental model of what req_install did

[23:09:20] <dstufft> lifeless: I think I'm going to have to punt on this until later, I've got a bunch of warehouse/pypi stuff in my RAM atm and I feel like I'm gonna need to carve out a decent chunk of time to really understand this

[23:09:28] <lifeless> dstufft: ok

[23:09:38] <lifeless> dstufft: I'll keep plugging on making tests pass

[23:09:44] <lifeless> and perhaps do some profiling

[23:09:53] <lifeless> dstufft: thanks for running at the wall :)

[23:10:27] <dstufft> that's mostly because you're changing a pretty core part of pip some of the worst areas of pip's code base with something that is complex (and needs to be I think to really solve the underlying problem)

[23:10:28] <dstufft> lifeless: no problem!

[23:16:07] <lifeless> dstufft: when should I next ping?

[23:16:53] <dstufft> lifeless: I wrote myself a note to do it first thing tomorrow

[23:18:18] <lifeless> dstufft: oh cool, thanks