Check exclusions prior to resolving pyproject.toml files

[charliermarsh]

Summary

This PR inverts the order of the "Check if this directory contains a pyproject.toml" and "Check if this path is included" operations.

Right now, build contains a subdirectory that intentionally includes a pyproject.toml with a syntax error (but no other files). You'd expect that:

[tool.ruff]
exclude = ["path/to/directory"]

...would ignore the directory, and thus avoid a parse error. However, right now, when we visit that directory, we check for a pyproject.toml immediately, then skip recursing any further based on the current exclusions.

This PR reverse the order of operations. This is a change in behavior, as previously, it was impossible to exclude a directory that itself contained Ruff configuration. But if a user is excluding a directory as above, this seems like the expected and intuitive behavior.

See #3569.

[charliermarsh]

\cc @henryiii

[github-actions]

PR Check Results

Ecosystem

✅ ecosystem check detected no changes.

Benchmark

Linux

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.00     13.7±0.07ms     3.0 MB/sec    1.01     13.8±0.04ms     2.9 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      3.6±0.01ms     4.6 MB/sec    1.00      3.6±0.00ms     4.6 MB/sec
linter/all-rules/numpy/globals.py          1.00    501.3±3.84µs     5.9 MB/sec    1.01    504.0±1.42µs     5.9 MB/sec
linter/all-rules/pydantic/types.py         1.00      6.0±0.01ms     4.2 MB/sec    1.01      6.1±0.01ms     4.2 MB/sec
linter/default-rules/large/dataset.py      1.00      7.5±0.02ms     5.4 MB/sec    1.01      7.6±0.01ms     5.4 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00   1691.7±1.55µs     9.8 MB/sec    1.01   1701.4±5.91µs     9.8 MB/sec
linter/default-rules/numpy/globals.py      1.00    191.7±0.38µs    15.4 MB/sec    1.02    196.0±0.68µs    15.1 MB/sec
linter/default-rules/pydantic/types.py     1.00      3.5±0.01ms     7.3 MB/sec    1.01      3.5±0.01ms     7.2 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.00     15.7±0.19ms     2.6 MB/sec    1.01     15.9±0.35ms     2.6 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      4.2±0.05ms     3.9 MB/sec    1.01      4.3±0.07ms     3.9 MB/sec
linter/all-rules/numpy/globals.py          1.01   543.2±13.34µs     5.4 MB/sec    1.00   538.8±10.80µs     5.5 MB/sec
linter/all-rules/pydantic/types.py         1.00      6.9±0.09ms     3.7 MB/sec    1.01      6.9±0.10ms     3.7 MB/sec
linter/default-rules/large/dataset.py      1.00      8.7±0.10ms     4.7 MB/sec    1.01      8.8±0.09ms     4.6 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00  1911.5±36.88µs     8.7 MB/sec    1.00  1920.2±27.99µs     8.7 MB/sec
linter/default-rules/numpy/globals.py      1.00    210.0±3.71µs    14.0 MB/sec    1.02   214.2±10.42µs    13.8 MB/sec
linter/default-rules/pydantic/types.py     1.00      4.0±0.06ms     6.4 MB/sec    1.02      4.1±0.08ms     6.3 MB/sec

[henryiii]

Looks reasonable to me!