Log file Viewer

[00:03:11] <techalchemy> dstufft, if you are saying the json api is stable enough i'm going to use that and cache json metadata locally to speed up resolution

[00:06:02] <techalchemy> that will be substantial because I don't think it would be necessary to invalidate the version-specific endpoints right?

[00:43:53] <pradyunsg> dstufft: I hear what you're saying, and parsing what you're saying, but not necessarily understanding it.

[00:44:13] <dstufft> pradyunsg: about what

[00:44:39] <pradyunsg> about pip as a forcing function.

[00:44:51] <pradyunsg> dstufft: ^

[00:54:14] <pradyunsg> dstufft: could you clarify? :)

[01:33:33] <dstufft> pradyunsg: sorry, kinda afk. for a large number of people t hey dont care about anything but if it works in pip. This puts us in a useful position where we can use our behavior to help push the ecosystem towards a better, more correct path-- in this case ensuring that building as an sdist works for all projects

[01:39:32] <pradyunsg> dstufft: yea, I understand.

[01:45:55] <dstufft> pradyunsg: in this case, building in place is a speed improvement, so people arent going to react well if we later switch to building sdists cuz it will be slower and they wont directly see a benefit from it, since it is a ecosystem health thing not an individual benefit. With enhancements to the pep517 api we could make building an sdist then building a wheel faster than it is but never as fast as not doing that. Were unlikely to add those apis

[01:45:55] <dstufft> without a project demonstrating need which pip could have done, but cant now without regressing on perf first (and then never getting back to the in place perf).

[01:47:30] <dstufft> that isnt a wrong tradeoff to make, but it doesnt look like we thought about that, anf backed ourselves into a corner for a more immediate gain

[02:26:18] <Raito_Bezarius> encountered a strange issue with pip, but that most likely due to tensorflow

[02:26:33] <Raito_Bezarius> ERROR: Could not find a version that satisfies the requirement tensorflow-estimator<2.2.0,>=2.1.0rc0 (from tensorflow==2.1.0) (from versions: none)

[02:26:43] <Raito_Bezarius> when tensorflow-estimator==2.2.0rc0 is available in the path

[02:26:45] <Raito_Bezarius> and installed

[02:27:08] <Raito_Bezarius> I don't know if x.y.zrcK are valid versions

[02:27:12] <Raito_Bezarius> and if 2.2.0rc0 < 2.2.0

[03:04:15] <toad_polo> Raito_Bezarius: You may need --allow-pre

[03:04:36] <Raito_Bezarius> toad_polo: I see, thanks

[03:05:10] <toad_polo> Or whatever it's called. The thing that enables downloading release candidates / betas, etc.

[03:08:04] <pradyunsg> --pre

[03:19:27] <toad_polo> <toad_polo "Raito_Bezarius: You may need --a"> N

[03:25:30] <toad_polo> Oops. Phone unlocked in my pocket.

[07:41:07] <omry> is there an alternative build backend a user can pick that behaves better than distutils?

[12:17:18] <toad_polo> I don't know what you mean by "behaves better"

[12:18:28] <toad_polo> I think invoking your tests tox instead of directly through pytest should solve the build folder problem, and it's a lot less drastic of a step than switching build backends.

[12:25:46] <toad_polo> invoking your tests through tox, rather.

[12:26:51] <toad_polo> Also once you've invoked them you can activate the virtual environment tox created if for some reason you don't want to invoke through tox every time.

[14:22:50] <Raito_Bezarius> Hello there, thanks for all your valuable inputs yesterday, I'm wondering how can I diagnose or chase why pip tried to use Internet sometimes

[14:22:53] <Raito_Bezarius> WARNING: Retrying (Retry(total=4, connect=None, read=None, redirect=None, status=None)) after connection broken by 'NewConnectionError('<pip._vendor.urllib3.connection.VerifiedHTTPSConnection object at 0x7ffff5f05e90>: Failed to establish a new connection: [Errno -2] Name or service not known')': /simple/setuptools-scm/

[14:23:59] <Raito_Bezarius> Does it signal that pip was not able to find the dependency and need it? Because if yes, it seems like it succeeds to install the dependency afterwards even though setuptools-scm was not downloaded apparently (though it is available in some path)

[14:24:14] <Raito_Bezarius> (Sorry for the vague question, any kind of answer is good to take, I just need to understand more how pip works)

[14:36:50] <toad_polo> Raito_Bezarius: I suspect its this bug: https://github.com/pypa/pip/issues/7863

[14:38:41] <Raito_Bezarius> toad_polo: thanks a lot!

[20:47:09] <omry> toad_polo, I am invoking my tests through nox, but my nox tests includes installation (both in editable and regular modes). I already worked around it but I think you can expect more complaints about this as 20.1 is getting wider adoption.

[20:48:10] <omry> pip should have it's own requirements from the build backend, such as the ability to specify a built outpout directory. this seems fairly straight forward.

[20:48:27] <omry> (except of course in the sense that it involves multiple projects).

[21:00:29] <sumanah> techalchemy: CONGRATULATIONS (fireworks) (cake with candles) (trumpet) (confetti) (trophy) on the prerelease of pipenv!

[21:01:13] <techalchemy> thanks sumanah it was quite exhausting lol

[21:02:00] <techalchemy> i need to do some communications about it still

[21:03:52] <sumanah> yes - techalchemy - would you like for me to help you draft that?

[21:04:21] <sumanah> I just wrote a blog post venting a bit about how down-in-the-dumps I feel, and helping you write a bit of pipenv communications might help me feel better

[21:04:36] <techalchemy> sorry to hear that you're feeling down :(

[21:04:40] <sumanah> thanks :\

[21:05:45] <techalchemy> I would appreciate any help you want to give, especially if you think it might help you, but don't feel obligated

[21:06:27] <sumanah> techalchemy: I can help - thanks for being careful. I can disengage -- tap out, as the boxers say -- at any moment if I feel overwhelmed

[21:06:34] <sumanah> techalchemy: shall we work in a HackMD?

[21:07:28] <techalchemy> sure, I can never remember where to edit shared markdown and always wind up googling 'shared markdown editor' so I don't have any actual preferences

[21:09:00] <sumanah> techalchemy: :-) ok https://hackmd.io/w2CHFdszQ7-AyAlf5LwcvA

[21:22:49] <sumanah> ok I have dumped some text in there and we're working together there

[21:33:31] <techalchemy> sumanah, thanks sorry fixing changelog

[21:40:01] <techalchemy> sumanah_, pasted the changelog in there

[21:40:01] <techalchemy> lol

[21:45:21] <sumanah_> so it's like 250 lines, ok

[21:50:35] <sumanah> techalchemy: am I right in presuming that this announcement will go to, basically, the same places we announced the pip beta and are announcing the pip release?

[21:50:48] <techalchemy> that's my thought

[21:51:05] <sumanah> cool

[21:51:16] <sumanah> https://github.com/pypa/pip/issues/7951 has that list in case you want to follow it

[21:54:42] <techalchemy> awesome thank you

[21:55:03] <techalchemy> should i just link to the testing document do you think?

[21:56:50] <techalchemy> also who can add @frostming to the pypa under pipenv, he has been doing most of the work i think

[22:01:07] <sumanah> techalchemy: I think dstufft can do that

[22:01:20] <sumanah> if you cannot, maybe dstufft should give you that power?

[22:01:26] <sumanah> techalchemy: yeah I figure you can link to the testing document

[22:01:28] <techalchemy> i think org owners can

[22:01:33] <techalchemy> i think @ pradyunsg can do it

[22:01:48] <sumanah> it's 3:30 in India so I hope he is sleeping

[22:01:53] <sumanah> but when he comes back, yeah

[22:02:03] <techalchemy> 50/50 but just highlighting so he sees it

[22:02:44] <sumanah> right

[22:03:56] <techalchemy> sumanah, am I still the release manager if I am actively trying to make it so other people can release

[22:04:01] <sumanah> sure

[22:04:05] <techalchemy> now everyone who can push to the repo can release, technically

[22:04:06] <techalchemy> lol

[22:04:15] <sumanah> I figure RM is partly technological and partly social

[22:04:30] <techalchemy> apparently it's easy enough someone released one of my other libraries without me noticing

[22:04:33] <sumanah> hee!

[22:04:38] <sumanah> yay automation

[22:04:53] <techalchemy> that's probably the most exciting part

[22:06:44] <sumanah> one of the things I find so inspiring about open source is the idea that someone, a stranger I've never met before, who doesn't look like me or think like me, might reuse my work, might have a better idea, might be able to pitch in and make our shared thing better

[22:07:30] <sumanah> and I love that you've made it so easy for that stranger to help all the project's users benefit from everyone's work with faster releases

[22:20:24] <sumanah> (in case anyone is curious, my current work soundtrack is 1 album on repeat: "Soundtrack to the End" by Communist Daughter)

[22:20:36] <techalchemy> now that's a band name

[22:22:59] <sumanah> right?!

[22:32:31] <sumanah> techalchemy: aww and thank you for your kind words about me. I am proud of you for how you have stepped up, not given up, in spite of the bad past interactions, the technical debt, habits you are trying to outgrow

[22:32:53] <techalchemy> sumanah, well deserved

[22:37:38] <techalchemy> ok

[22:37:46] <techalchemy> added links for all the issues

[22:38:01] <techalchemy> oh boy oops

[22:39:16] <dstufft> what am i doing

[22:39:17] <dstufft> oh github perms

[22:39:26] <techalchemy> just add frostming

[22:39:30] <techalchemy> to pipenv committers

[22:40:43] <dstufft> https://s.caremad.io/fKsFDLHhwpr61oiHda0YrFIC/

[22:41:16] <dstufft> oh to the team itself

[22:41:17] <dstufft> i am dumb

[22:41:27] <techalchemy> ^ yeah so he is a member

[22:41:39] <techalchemy> also is that nextcloud or something

[22:42:13] <dstufft> my image thing?

[22:42:19] <techalchemy> ya

[22:42:22] <dstufft> it;s dropshare

[22:42:35] <techalchemy> oh man i dont know that one is that because im old or young

[22:43:20] <dstufft> I think its because it's not well known :p

[22:43:38] <dstufft> it lets me scp to my server tho

[22:43:40] <travis-ci> pypa/twine#1438 (master - da8d62f : Devesh Kumar Singh): The build passed.

[22:43:40] <travis-ci> Change view : https://github.com/pypa/twine/compare/e687b48f3d1eff2b422383fce82f3f71f79d8f3b...da8d62f6e8878900267a8499d4e3cfe49ac11002

[22:43:40] <travis-ci> Build details : https://travis-ci.org/pypa/twine/builds/681718180

[22:44:56] <sumanah> https://blog.python.org/2020/04/pip-20-1-released.html is now up

[22:45:17] <sumanah> I'm gonna wait a moment before emailing python-announce and pypi-announce about it, in case someone points out something super wrong in what I just said

[22:45:22] <dstufft> Im gonna fix these perms a tad

[22:45:38] <dstufft> we have people listed twice with different perms

[22:48:24] <techalchemy> i saw that about at least one person on pipenv

[22:48:31] <techalchemy> but then i was like idk, don't tell me what to do github

[22:48:51] <techalchemy> sumanah, thanks for your help! dinner is ready so im gonna eat before i send

[22:48:51] <techalchemy> back i na bit

[22:48:58] <sumanah> good idea techalchemy

[22:57:16] <travis-ci> pypa/pip#15985 (master - 612970b : Pradyun Gedam): The build passed.

[22:57:16] <travis-ci> Change view : https://github.com/pypa/pip/compare/369565dfefcf8232d1505a863f85860ef34ce53a...612970b1cf08446e7aa06c15931898fd8e064b1c

[22:57:16] <travis-ci> Build details : https://travis-ci.org/pypa/pip/builds/681712216