Log file Viewer
#pypa-dev logs for Tuesday the 31st of March, 2020
[09:16:51] <devesh_> Hi pradyunsg: Could I get these two PR merged please. https://github.com/pypa/pip/pull/7927 and https://github.com/pypa/pip/pull/7938
[09:18:27] <devesh_> 7938 has a linter check error out with "Workflow Error" https://github.com/pypa/pip/pull/7938/checks?check_run_id=548100262, not sure if we restart all tests or if there is a possibility to restart just that
[09:30:31] <travis-ci> Change view : https://github.com/pypa/pip/compare/d79632a7a9aa...827539544fc9
[10:19:03] <travis-ci> Change view : https://github.com/pypa/pip/compare/0f7fac3a3ddb...534561cbecb6
[11:13:55] <cnx> (I'm McSinyx) ugh gotta drop that half-working matrix-irc bridge, it somehow kicked me out of irc completely
[11:15:34] <cnx> pradyunsg: I've just taken a look at pip option parsing logic and relized you weren't joking about that partial thing
[11:16:48] <cnx> maybe I understate the ``just'' part, it's been quite a while and still I'm struggling to parse the logic as a human
[14:33:00] <cnx> pradyunsg: I've just updated the proposal, although I don't think you need to check any time sooner than when it arrives to you via gsoc
[14:34:57] <cnx> I've added some vague description on the option parsing thing there is no clear direction
[14:36:03] <cnx> also to sumanah, do we have any sort of backup for pypa centrialized bug tracker on github?
[14:38:14] <cnx> cython has one, ironically hosted on github also: https://github.com/cython/cython-issues
[14:42:41] <EWDurbin> fwiw: CPython Github has been using backhub.co PyPA doesn't currently have anything configured though afaik
[14:45:55] <EWDurbin> backhub is pretty reasonably priced and supports archives to S3... so if the PyPA can agree on what key repos could use backing up... probably worth the cost for the PSF
[14:51:45] <techalchemy> yeah it's not too difficult, the new guidelines just mean everyone else is now doing what i normally do
[14:52:21] <sumanah> You said: "i'll likely just make the test do what it says it does.... fix news entry(ies?), figure out PR wording ... add bug for missing test for pep517 backend"
[14:55:34] <techalchemy> but then i found another bug in the pipenv re-vendoring since i needed to relock requirementslib dependencies so it's jsut a bit of a cascade there, should be ok now
[15:04:18] <techalchemy> sumanah, i'm not going to bother trying to sort through issues atm to figure out what this will close
[15:04:46] <sumanah> techalchemy: you mean which issues your current PR on requirementslib will close?
[15:05:06] <techalchemy> I think it mostly really does focus on revendoring which likely has a lot of unintended closure effects
[15:12:50] <sumanah> techalchemy: I think you will also want to merge the docs update https://github.com/pypa/pipenv/pull/4167 because it fixes a broken link in pipenv/cli/command.py
[15:13:27] <pradyunsg> EWDurbin: would it be OK to do that backup on a per-repo basis, since each project in pypa is basically a close group of fairly independently run projects?
[15:16:13] <pradyunsg> Functionally, the difference would be whether I write an email to pip maintainers vs start a thread on d.p.o.
[15:24:12] <techalchemy> whatever permissions you want i grant you them in words plz dont ask me to talk on discourse
[15:29:12] <sumanah> techalchemy: ok I edited the TODO list at the top of the release issue https://github.com/pypa/pipenv/issues/3369 so you have a somewhat organized TODO list for today
[15:48:41] <devesh> hi pradyunsg: Finally the check passed after I added a new commit to the pr: https://github.com/pypa/pip/pull/7938 :)
[17:24:19] <sumanah> hey techalchemy do you need to make that new requirementslib release and incorporate it into the PR you just made?
[17:35:19] <devesh> Hi sumanah: Thanks for suggesting to take a look at twine yesterday, the feature-set provided is pretty focused on uploading packages
[17:35:49] <devesh> Apart from the standard bug fixes/issues, are there other new features which are at scope in twine
[17:35:56] <sumanah> as I recall, you mentioned wanting to work on something in the world of Python packaging but with lower complexity
[17:36:45] <sumanah> devesh: I'm a little confused .... how many different projects do you want to contribute to simultaneously?
[17:37:18] <sumanah> or are you doing sort of an inventory of existing projects to find The Perfect Project for you to contribute to right now?
[17:38:22] <devesh> no inventory actually, just an exploration I would say to get the lay of the land
[17:38:25] <sumanah> devesh: there are several issues in https://github.com/pypa/twine/issues that will take substantial time to work on
[17:39:59] <sumanah> And, in the future, once PyPI implements TUF https://github.com/pypa/warehouse/issues/5247 , it will be important to add TUF metadata handling to twine
[17:41:03] <devesh> Thanks sumanah: I think the point of your question was that it's better to find a few projects and work on them with focus
[17:42:02] <sumanah> if you want to do a big survey of "what is it like to contribute to open source projects in 2020?" and then write up a report about it later, then learning and contributing to many different projects can be a good way to get that data!
[17:42:44] <sumanah> or if you have a lot of curiosity about what different codebases look like, or how different projects do code review, or something
[17:43:13] <sumanah> if you are looking for pure timepass then I suppose it doesn't matter whether you go for breadth or concentrate on depth!
[17:43:52] <sumanah> devesh: a year from now, what do you want to have or know that you currently don't?
[17:45:05] <devesh> So my aim is to improve my python knowledge, hopefully in projects used by a lot of people.
[17:45:28] <devesh> Try to give back to the community by learning the ropes of OSS, and perhaps it might help me in future job opportunities
[17:47:05] <devesh> I have observed that folks who actually have good OSS contribution have slightly better chances at gaining roles at good companies,
[17:47:19] <devesh> atleast that what I heard from a lot of places I interviewed, not sure how true is that
[17:47:44] <sumanah> devesh: ok. So, you want to have more knowledge of Python in general, and of the internals/codebases of widely used projects. And you want to have a legacy of useful contribution, and more hireability.
[17:51:01] <sumanah> devesh: you're being careful about how much of a scarce resource (my time) you are using
[17:51:51] <devesh> ohh okay, yes I have been asking so many questions and you have answered all of them, so I think I shouldn't overdo it
[17:52:11] <sumanah> devesh: so, given what you've said about your goals, yes, I think that concentrating on learning and contributing to 1-3 projects would be a better use of your time than spreading your efforts among several more
[17:52:23] <sumanah> techalchemy: ah ok! you don't need to make a fresh requirementslib release for PyPI?
[17:52:49] <techalchemy> sumanah, I do, but I just copied the changes over beforehand so I could get both moving in parallel
[17:53:23] <techalchemy> meeting in 10, also fixing python 2 failures on pipenv probably during the meeting
[17:53:51] <techalchemy> potentially figuring out why tests are so slow on ci, i think pip uses a ramdisk now -- maybe i'll steal that
[17:54:13] <devesh> yes, I thought the same sumanah, I found a lot of projects which had a lot of stars but no commit for almost 6 months, atleast in groups like PyPA you have good turnaround time for contributions
[17:55:07] <sumanah> devesh: yeah. You learn more. And you have a better chance of becoming a co-maintainer, which is more impressive to employers.
[17:56:35] <devesh> I had some packages I developed in my last company, but I think I cannot use them in public
[17:56:35] <sumanah> Brian Rutledge became a co-maintainer of Twine within ... maybe 6 months from his first Twine contribution
[17:56:43] <sumanah> devesh: I suggest you go through this tutorial https://packaging.python.org/tutorials/distributing-packages/
[17:57:19] <sumanah> it will help you understand what it's like to be a package maintainer, so that when you go on to work on any PyPA stuff, you'll better understand how it is meant to fit into the whole process
[17:58:36] <devesh> Will do thanks. I think I will let you be on your way now, thanks for the suggestions. I will let you know how my interaction with twine goes
[17:58:50] <sumanah> setting up a little toy package on Test PyPI will help you so you can make new releases to test stuff like Twine or Setuptools .... https://test.pypi.org/project/Forms990-analysis/ is mine
[18:00:06] <devesh> thanks sumanah, I should say that the advice I get here is better than some corporate environments I have seen, pretty honest
[18:01:19] <sumanah> devesh: btw if you want to hear me talk for hours on end, https://www.harihareswara.net/talks.html has links to recordings of talks I've given and and interviews with me
[18:01:42] <sumanah> I should stress: this is absolutely not obligatory. No one has to listen to these
[18:04:22] <devesh> I think I have seen one of your talks before, about http pycon 2016, when I was preparing for interviews a few months back
[20:10:43] <techalchemy> sumanah, i think some of them won't work due to being set for something azure doesn't support now
[20:12:03] <sumanah> cool. I just re-re-re-re-re-edited the checklist at https://github.com/pypa/pipenv/issues/3369#issue-389160510
[20:13:37] <sumanah> so at this point techalchemy I think it would be good to manage expectations about when the pre-release will actually happen. Which sounds like: tomorrow
[20:25:29] <sumanah> techalchemy: so, there's a maybe .... on the outside, if the runs are slow, then it could be the end of the week that you release it. Right?
[20:28:10] <sumanah> techalchemy: right. So it could be as early as today! But, if you end up having to play whack-a-mole, it could be much slower.
[20:28:30] <sumanah> I figure part of responsible communication to the users is being open to the range of possibilities here
[20:28:57] <sumanah> techalchemy: in case you are waiting for the CI -- there are a few other things you can do in the meantime:
[20:34:03] <techalchemy> but yes if you have more things for me to do i'm certainly not focusing on what they are so a reminder would be great
[20:35:18] <sumanah> two in particular: file a bug about the pep517 backend virtualenv Windows thing, and update your packaging/release checklist
[20:35:52] <sumanah> (I'm sorry techalchemy .... I know that sometimes when I am waiting for a process like that I get distracted, but I think you have more practice than I do in multitasking regarding that sort of thing)
[20:39:01] <techalchemy> sumanah, it's less the practice thing and more the amount of stuff i have to do at any given time, i dont really have a choice
[20:41:52] <techalchemy> sumanah, when you say update the checklist, do you mean, with the things that are required to release?
[20:42:37] <sumanah> techalchemy: yeah, I mean update https://demo.codimd.org/wY3hW3f2SuGXBfd0ZsHsag with any steps you realize you're gonna need to do
[21:12:37] <sumanah> techalchemy: I updated the release issue with a new comment basically saying: might be today, might be a few days from now, please react here if you can test on Windows, and please come to IRC if you can help with Azure pipelines templates.
[21:21:52] <sumanah> Thanks techalchemy. Feel free to edit the comment I made on GitHub in case it is inaccurate. I'm gonna head off for the night now ... would you like for me to ping you tomorrow or do you feel like you have everything in hand (in terms of staying on task) now?
[22:27:59] <travis-ci> Change view : https://github.com/pypa/twine/compare/5482d6877df0cd90545c4708e5abe13b359f8a08...f7402e0d2e1c1e6ecde61dc8bcba515b807a48a0
[22:30:33] <travis-ci> Change view : https://github.com/pypa/pip/compare/0f7fac3a3ddb123ff03f24c44a749c2abc694d4b...534561cbecb6d1cb8773b3bf545badfa122ece25