aio-libs / multidict / 35507866989

87%

push

github

Scale up syscall-dominated CodSpeed benchmarks (#1485)

<!-- Thank you for your contribution! -->

## What do these changes do?

Scales up the working set of the `pop()`, `popitem()`, `__delitem__()`,
`add()` and item-insertion CodSpeed benchmarks in
`tests/test_multidict_benchmarks.py`. Comparing repeated CodSpeed runs
on the same master commit (via the CodSpeed MCP) showed these
particular pure-Python-backend benchmarks consistently flagged as
"dominated by syscalls": instrumentation mode cannot account for
syscall time, so it is excluded from the measured value, which both
understates the real cost and introduces run-to-run noise depending
on when a syscall happens to land during the measured window. Giving
each benchmark more instrumented work to do amortizes that overhead,
which is CodSpeed's own guidance for this class of measurement issue.

## Are there changes in behavior for the user?

No. Only the size of the fixtures inside these benchmark functions
changed; no library code is touched.

## Is it a substantial burden for the maintainers to support this?

No, it is a mechanical change to existing benchmark fixture sizes.

## Related issue number

N/A

## Checklist

- [x] I think the code is well written
- N/A Unit tests for the changes exist -- this PR only touches
benchmarks
- N/A Documentation reflects the changes
- N/A If you provide code modification, please add yourself to
`CONTRIBUTORS.txt`
- [x] Add a new news fragment into the `CHANGES/` folder
- [x] `make doc-spelling` passes and any new technical words are added
to `docs/spelling_wordlist.txt`

<details>
<summary>Agent run details (optional, for reviewers)</summary>

Analysis: used the CodSpeed MCP to compare two runs of the same master
commit (`7310298`, runs `6aaf7cd3` vs `6aaf045b`, and `9c08be4`, runs
`6aae2b4f` vs `6aad8528`). Both comparisons flagged the same cluster of
pure-Python `pop`/`popitem`/`delitem`/`add`/insert-style benchmarks as
"dominated by syscalls"; no other ben... (continued)

762 of 1524 branches covered (50.0%)

Branch coverage included in aggregate %.

14 of 14 new or added lines in 1 file covered. (100.0%)

6320 of 6588 relevant lines covered (95.93%)

1.92 hits per line