aio-libs / multidict / 35579952505

87%

push

github

Run mutating benchmarks as copy plus operation in a loop (#1494)

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

## What do these changes do?

Reorganizes the mutating CodSpeed benchmarks to copy a base multidict
and apply the operation 100 times per round, the way the `add()` and
`extend()` benchmarks already work.

The insertion, `update()` and `clear()` benchmarks mutated one multidict
shared across rounds, so only the first round measured the intended
operation (later rounds replaced existing keys or cleared an empty
dict). The `pop()`, `popitem()`, `__delitem__()`, `__setitem__()` and
same-key `add()` benchmarks did a single copy per round, which let
per-round overhead dominate. Working-set sizes per copy were reduced so
the pure-Python legs stay at a similar per-round cost.

## Are there changes in behavior for the user?

No, benchmarks only. Reported numbers for the touched benchmarks will
shift since they now measure a different amount of work.

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

No.

## Related issue number

N/A

## Checklist

- [x] I think the code is well written
- [ ] Unit tests for the changes exist: N/A, benchmark-only change
- [ ] Documentation reflects the changes: N/A
- [ ] If you provide code modification, please add yourself to
`CONTRIBUTORS.txt`: N/A
- [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`

Drafted with Claude Code (Claude Opus 5); reviewed by @asvetlov.

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

Tests: `pytest tests/test_multidict_benchmarks.py
tests/test_views_benchmarks.py -q` on 3.14t with the in-place C
extension build: 269 passed (both `c` and `py` backend params).
CodSpeed walltime local run of touched benchmarks: all pass; pure-Python
`add_same_key` ~0.4 s/round after shrinking to 100 values.
Lint: pre-commit hooks pass; mypy skipped because it c... (continued)

762 of 1524 branches covered (50.0%)

Branch coverage included in aggregate %.

77 of 77 new or added lines in 1 file covered. (100.0%)

6393 of 6661 relevant lines covered (95.98%)

1.92 hits per line