Avoid unnecessary parser lookahead for operators#25290
Merged
Merged
Conversation
|
Merging this PR will improve performance by 10.75%
|
| Mode | Benchmark | BASE |
HEAD |
Efficiency | |
|---|---|---|---|---|---|
| ⚡ | Simulation | ty_micro[complex_constrained_attributes_1] |
74.6 ms | 70.6 ms | +5.69% |
| ⚡ | Simulation | ty_micro[many_tuple_assignments] |
65.8 ms | 62 ms | +5.99% |
| ⚡ | Simulation | ty_micro[complex_constrained_attributes_2] |
74.4 ms | 70 ms | +6.17% |
| ⚡ | Simulation | ty_micro[pandas_tdd] |
255.3 ms | 238.9 ms | +6.89% |
| ⚡ | Simulation | ty_micro[complex_constrained_attributes_3] |
80.8 ms | 76.9 ms | +5.03% |
| ⚡ | Simulation | ty_micro[many_tuple_assignments] |
74.4 ms | 70.1 ms | +6.18% |
| ⚡ | Simulation | ty_micro[gradual_vararg_call] |
74.9 ms | 70.7 ms | +6.03% |
| ⚡ | Simulation | ty_micro[many_enum_members] |
100.1 ms | 95.3 ms | +4.99% |
| ⚡ | Simulation | ty_micro[vararg_parameter_type_accumulation] |
64.1 ms | 60.4 ms | +6.17% |
| ⚡ | Simulation | ty_micro[many_enum_members_2] |
95.9 ms | 91.5 ms | +4.88% |
| ⚡ | Simulation | ty_micro[very_large_tuple] |
76.7 ms | 72.3 ms | +6.05% |
| ⚡ | Simulation | ty_micro[many_string_assignments] |
85 ms | 81 ms | +5.02% |
| ⚡ | Simulation | parser[large/dataset.py] |
4.7 ms | 3.6 ms | +30.44% |
| ⚡ | Simulation | parser[numpy/ctypeslib.py] |
883.8 µs | 713.6 µs | +23.85% |
| ⚡ | Simulation | parser[numpy/globals.py] |
103.7 µs | 90.4 µs | +14.77% |
| ⚡ | Simulation | parser[pydantic/types.py] |
1.8 ms | 1.4 ms | +27.42% |
| ⚡ | Simulation | parser[unicode/pypinyin.py] |
308.6 µs | 251.3 µs | +22.78% |
Tip
Curious why this is faster? Comment @codspeedbot explain why this is faster on this PR, or directly use the CodSpeed MCP with your agent.
Comparing charlie/parser-bench-delimiter-depth (9764ee1) with main (6cbd59b)
|
I would do another /goal round looking at the buffers that peek() is using. My mind's eye has loading whole cache lines when possible so it's only doing memory fetch or OS calls when absolutely needed. |
thejchap
pushed a commit
to thejchap/ruff
that referenced
this pull request
May 23, 2026
## Summary We only need to peek ahead two tokens if the first token is `not` or `in`. We can avoid the `peek` in the majority of cases, which apparently speeds up the parser by 20-30%.
anishgirianish
pushed a commit
to anishgirianish/ruff
that referenced
this pull request
May 28, 2026
## Summary We only need to peek ahead two tokens if the first token is `not` or `in`. We can avoid the `peek` in the majority of cases, which apparently speeds up the parser by 20-30%.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
4 participants
Add this suggestion to a batch that can be applied as a single commit.This suggestion is invalid because no changes were made to the code.Suggestions cannot be applied while the pull request is closed.Suggestions cannot be applied while viewing a subset of changes.Only one suggestion per line can be applied in a batch.Add this suggestion to a batch that can be applied as a single commit.Applying suggestions on deleted lines is not supported.You must change the existing code in this line in order to create a valid suggestion.Outdated suggestions cannot be applied.This suggestion has been applied or marked resolved.Suggestions cannot be applied from pending reviews.Suggestions cannot be applied on multi-line comments.Suggestions cannot be applied while the pull request is queued to merge.Suggestion cannot be applied right now. Please check back later.
Summary
We only need to peek ahead two tokens if the first token is
notorin. We can avoid thepeekin the majority of cases, which apparently speeds up the parser by 20-30%.