[T-256]

Summary

Same as what's done on httpx: encode/httpx#2922

Checklist

• I understand that this PR may be closed in case there was no previous discussion. (This doesn't apply to typos!)
• I've added a test for each change that was introduced, and I tried as much as possible to make a single atomic change.
• I've updated the documentation accordingly.

[T-256]

If remove this section, result will be:

>ruff check
httpcore\_sync\connection.py:1:1: I001 [*] Import block is un-sorted or un-formatted
httpcore\_sync\connection_pool.py:1:1: I001 [*] Import block is un-sorted or un-formatted
httpcore\_sync\socks_proxy.py:1:1: I001 [*] Import block is un-sorted or un-formatted
tests\_sync\test_connection.py:1:1: I001 [*] Import block is un-sorted or un-formatted
tests\_sync\test_connection_pool.py:1:1: I001 [*] Import block is un-sorted or un-formatted
Found 5 errors.
[*] 5 fixable with the `--fix` option.

Maybe we also could use in-file rule ignore for them?

[karpetrosyan]

Do we really need to avoid excluding sync directories? When we need to format/lint them? they are auto generated.

[T-256]

They are pre-generated. it means currently, they are already available in sdist, wheels and repository.
It doesn't look good that readers navigate through unformatted sources after installation.
For example, after installation, I'll navigate to site-packages/httpcore then in _sync folder's contents I open a single file, then I'll see linter/formatter warnings in my IDE.

[T-256]

When it's possible to have them, why excluding them?

[karpetrosyan]

I'll navigate to site-packages/httpcore then in _sync folder's contents I open a single file, then I'll see linter/formatter warnings in my IDE.

I believe it is typical to see such situations because IDE rules and rules used for linting source files are not always in sync.

When it's possible to have them, why excluding them?

We can, for the time being, ignore I001 for those files, but we don't know how many additional rules may be broken when introducing a new feature. We need add new rules that break sync files into the configuration every time, which makes the maintenance process more difficult.

[T-256]

we don't know how many additional rules may be broken when introducing a new feature. We need add new rules that break sync files into the configuration every time, which makes the maintenance process more difficult.

The codebase is almost stable, I think the only violation rule would be import sorting violation. by checking unasync.py script I don't except new violation happening when adding new feature other than I001 (and F811 which can be in-line ignoring this without changing pyproject.toml)

IMO, we could give it a try, since always doable in future to revert this change.

[stale]

This issue has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. Thank you for your contributions.