Infinite loop in import sorting for long variable names with force-single-line (ruff >= 0.0.256)

[pstjohn]

The following code snippet causes ruff to hang for ruff >= 0.0.256.

pyproject.toml

[tool.ruff]
select = ["I"]
fix = true
line-length = 88

[tool.ruff.isort]
force-single-line = true

test.py

from mypackage.subpackage2 import (  # long comment that seems to be a problem
    a_long_variable_name_that_causes_problems,
    item2,
)

Doing some bisection, it works in 0.0.255 and hangs in 0.0.256 onwards.
Probably related to #3530? But it happens regardless of the combine-as-imports setting

In the newer versions, it seems to get stuck in an infinite loop

[2023-04-21][20:08:42][ruff_cli::commands::run][DEBUG] Identified files to lint in: 521.625µs
[2023-04-21][20:08:42][ruff_cli::diagnostics][DEBUG] Checking: /Users/pstjohn/Misc/ruff_issue/test.py
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:08:42][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)

The older versions run through twice

[2023-04-21][20:15:05][ruff_cli::commands::run][DEBUG] Identified files to lint in: 381.25µs
[2023-04-21][20:15:05][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:15:05][ruff::rules::isort::categorize][DEBUG] Categorized 'mypackage' as ThirdParty (NoMatch)
[2023-04-21][20:15:05][ruff_cli::commands::run][DEBUG] Checked 1 files in: 1.910292ms
Found 1 error (1 fixed, 0 remaining).

resulting in this formatted import

from mypackage.subpackage2 import (  # long comment that seems to be a problem
    a_long_variable_name_that_causes_problems,
)
from mypackage.subpackage2 import item2

[charliermarsh]

Thanks for the clear issue, taking a look :)

[charliermarsh]

I've identified the issue (introduced in #3521), which is that we're duplicating the comment with an exponential blowup. Figuring out a proper fix, if I can't fix it properly I'll revert that change prior to the next release.

[pstjohn]

You're welcome, thanks for the quick reply (and great package)!