Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

- remove setuptools dependency #7687

Closed
wants to merge 20 commits into from
Closed

- remove setuptools dependency #7687

wants to merge 20 commits into from

Conversation

franzhaas
Copy link

@franzhaas franzhaas commented Dec 23, 2021

similar to numba/llvmlite#801 (comment), this is an effort to remove usage / dependency of setuptools.

The motivation is that it makes it easier to use tools like pyoxidizer, replaces a library dependency with a builtin depenency, decreases import latency and reduces overall dependancies.

also like the mentioned pull request, not a killer feauture not a bug fix, but it is expected to speedup import times.

At some point i noticed the PR fix/4927-slow-import and used the developoments from there.

This is not a successfull simplification. A dependency of a backport and two implementations how to find the entrypoints have been added. On the other side the dependancy of setuptools and 1 implementation got removed.

If this yields improvements on importtime, it propably is still be worth it, especially for python 3.8 and newer. Also for me personaly droping setuptools as a requirement is a big plus as packages tend to use it if it is there, which ocationally causes issues...

I used input from: python/importlib_metadata#298
And again, the original PR from @svrakitin.

These code changes would need to be touched when suport for 3.7 and 3.8 gets removed.

Unfortunately I was unable to test it with real extensions.

@franzhaas franzhaas marked this pull request as ready for review December 24, 2021 20:55
@gmarkall
Copy link
Member

Many thanks for the PR! I've queued it for review.

I do need to go back and get #5209 over the line - it got a bit fiddly so I didn't yet find a good moment to get everything as it should be.

@stuartarchibald
Copy link
Contributor

@franzhaas Is this PR still applicable now #5209 is merged? It seems like it makes a similar set of changes to achieve similar goals.

@franzhaas
Copy link
Author

Sry for the late reply.

I did not check, but this would be extremely unlikely as this is a derived work of the mentioned pull request.

@franzhaas franzhaas closed this Mar 29, 2022
@stuartarchibald
Copy link
Contributor

@franzhaas thanks for checking back and electing to close. #5209 should gain the same performance improvement, please do check 0.56 when it's released and report any issues in relation to this. Thanks again!

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Projects
None yet
Development

Successfully merging this pull request may close these issues.

None yet

3 participants