Fix snapshots missing iterations

[JonathanPlasse]

Until now the snapshots only contained the first iteration of diagnostics.
This PR adds the diagnostics from the other iterations.
The duplicates between the iterations are removed.

[github-actions]

PR Check Results

Ecosystem

✅ ecosystem check detected no changes.

Benchmark

Linux

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.02     15.6±0.34ms     2.6 MB/sec    1.00     15.4±0.30ms     2.6 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.02      4.1±0.07ms     4.0 MB/sec    1.00      4.0±0.08ms     4.1 MB/sec
linter/all-rules/numpy/globals.py          1.03   572.1±11.62µs     5.2 MB/sec    1.00   555.9±10.03µs     5.3 MB/sec
linter/all-rules/pydantic/types.py         1.00      6.9±0.14ms     3.7 MB/sec    1.00      6.9±0.10ms     3.7 MB/sec
linter/default-rules/large/dataset.py      1.00      8.2±0.16ms     4.9 MB/sec    1.00      8.2±0.16ms     4.9 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00  1845.8±43.56µs     9.0 MB/sec    1.00  1848.3±40.19µs     9.0 MB/sec
linter/default-rules/numpy/globals.py      1.07    215.9±3.73µs    13.7 MB/sec    1.00    202.3±5.54µs    14.6 MB/sec
linter/default-rules/pydantic/types.py     1.01      3.9±0.10ms     6.5 MB/sec    1.00      3.9±0.07ms     6.6 MB/sec

Windows

group                                      main                                   pr
-----                                      ----                                   --
linter/all-rules/large/dataset.py          1.00     17.9±0.48ms     2.3 MB/sec    1.02     18.3±0.53ms     2.2 MB/sec
linter/all-rules/numpy/ctypeslib.py        1.00      4.9±0.26ms     3.4 MB/sec    1.01      5.0±0.19ms     3.3 MB/sec
linter/all-rules/numpy/globals.py          1.02   629.2±23.47µs     4.7 MB/sec    1.00   616.0±23.12µs     4.8 MB/sec
linter/all-rules/pydantic/types.py         1.01      8.1±0.26ms     3.1 MB/sec    1.00      8.0±0.23ms     3.2 MB/sec
linter/default-rules/large/dataset.py      1.00      9.8±0.42ms     4.2 MB/sec    1.00      9.8±0.37ms     4.2 MB/sec
linter/default-rules/numpy/ctypeslib.py    1.00      2.1±0.07ms     7.8 MB/sec    1.02      2.2±0.08ms     7.6 MB/sec
linter/default-rules/numpy/globals.py      1.00   234.6±11.30µs    12.6 MB/sec    1.02    239.4±8.91µs    12.3 MB/sec
linter/default-rules/pydantic/types.py     1.00      4.6±0.19ms     5.5 MB/sec    1.01      4.6±0.23ms     5.5 MB/sec

[charliermarsh]

Hmm, this may have the potential to confuse, since the diagnostics that are generated after the first iteration will contain locations that don't map to the input file (assuming the input file is changed via autofix).

Is there an example rule or set of rules for which we missed something before, but have improved diagnostic coverage after this change?

[JonathanPlasse]

• Yes, I separated in two commits Fix autofix conflict between D209 and D400 #3564, where there is the before and after snapshot where the former had three fixes compared to the latter that had only two.
• The format of the snapshots, could be changed to contain an array of the iterations.

[JonathanPlasse]

• Here is the diff of Fix autofix conflict between D209 and D400 #3564 between the two commits.

  ---
  source: crates/ruff/src/rules/pydocstyle/mod.rs
  expression: diagnostics
  ---
  - - kind:
        name: NewLineAfterLastParagraph
        body: Multi-line docstring closing quotes should be on a separate line
        suggestion: Move closing quotes to new line
        fixable: true
      location:
        row: 2
        column: 4
      end_location:
        row: 3
        column: 72
      fix:
        content: "\n    "
        location:
          row: 3
          column: 69
        end_location:
          row: 3
          column: 69
      parent: ~
    - kind:
        name: EndsInPeriod
        body: First line should end with a period
        suggestion: Add period
        fixable: true
      location:
        row: 2
        column: 4
      end_location:
        row: 3
        column: 72
      fix:
        content: "."
        location:
          row: 3
          column: 69
        end_location:
          row: 3
          column: 69
      parent: ~
- - - kind:
-       name: NewLineAfterLastParagraph
-       body: Multi-line docstring closing quotes should be on a separate line
-       suggestion: Move closing quotes to new line
-       fixable: true
-     location:
-       row: 2
-       column: 4
-     end_location:
-       row: 4
-       column: 8
-     fix:
-       content: "\n    "
-       location:
-         row: 4
-         column: 5
-       end_location:
-         row: 4
-         column: 5
-     parent: ~

[JonathanPlasse]

I do not know why the CI is failing, the tests are passing locally.

[MichaReiser]

I do not know why the CI is failing, the tests are passing locally.

The failing tests are gated behind the logical_lines feature [source].

You can run them locally with

cargo insta test --all-features

@charliermarsh I wonder if we should remove the feature and instead rely on cfg(debug_assertions) or disable the feature for release builds stack overflow to avoid this confusion in the future.

[JonathanPlasse]

@charliermarsh, do you want to merge this PR or should we close it?

[charliermarsh]

@JonathanPlasse - I do see the value, but thinking on it... I'd prefer to close, I'm just worried about making the snapshots more difficult to follow since they now represent a sequence of iterative changes. I'm sorry for not saying so sooner.

[JonathanPlasse]

I understand, no problem. I could understand better how the tests work.