unpin setuptools

[minrk]

while ipykernel technically pins setuptools 60 in setup.py, it doesn't actually need it and the pin causes installation issues

[conda-forge-linter]

Hi! This is the friendly automated conda-forge-linting service.

I just wanted to let you know that I linted all conda-recipes in your PR (recipe) and found it was in an excellent condition.

[blink1073]

Thank you!

[github-actions]

Hi! This is the friendly conda-forge automerge bot!

I considered the following status checks when analyzing this PR:

• linter: passed
• azure: passed

Thus the PR was passing and merged! Have a great day!

[bollwyvl]

Without a patch of setup.py to remove the pin, this will break all downstream recipes with pip check.

[blink1073]

I was planning to release a new ipykernel on Monday, should we do it now?

[blink1073]

Or make a new build here with the pin restored?

[bollwyvl]

Yeah I dunno: it's been a rather confusing release at this point, and different builds will be broken for various people in different ways. At worst, folk have to pin, but if nobody but me is complaining, don't let it take up more of your time.

[blink1073]

Sounds good.

[minrk]

Sorry, I should have done the patch on setup.py as well, to remove the pin from pip.

this will break all downstream recipes with pip check.

I don't think that's true in general. pip check will only complain when some other package has pinned down setuptools, preventing the current version from being installed. The current version satisfies the pin, older versions work, so only pip check will complain, nothing will actually be broken. Only the subset of downstream packages that

1. run pip check, and
2. depend on both his package and another package that pins setuptools<60

should see any issue. The rest should not be affected.

Before this: envs couldn't be installed
After this: pip check may complain (only in those cases where envs couldn't be installed at all)

[minrk]

Forgot to mention: I think Monday's fine for 6.11.1. If this causes any downstream issues, I'll do the patch for setup.py here (that's better than restoring the pin, I think).

[bollwyvl]

all downstream recipes with pip check.

totally right, not all. but certainly possible as..

1. run pip check, and
2. depend on both his package and another package that pins setuptools<60

yeah, i found one, hence making #121 in the first place! bad stuff afoot there, anyway, so just skipped pip check for now.

certainly i can offer no argument that runtime use of setuptools (or any of the build backends) is bad news, at any version, and look forward to the next bump... but its not breaking anything in particular at the moment.