pyperf Python performance guide (v0)pyperf rule ids (slugs)pyperf corpus targets pyperf evidence policy

pp-deepcopy-hot — measured, still hint pp-json-hot — measured, still hint pp-list-plus-rebind — xs = xs + [i] vs append pp-open-hot — measured, still hint pp-print-hot — print in loop vs join once (stdout discarded)pp-str-plus-hot — str += vs str.join

pp-datetime-format-in-loop pp-datetime-format-in-loop — removed 2026-10-23 (Asia/Shanghai)pp-deepcopy-in-loop — renamed 2026-10-27 (Asia/Shanghai)pp-json-in-loop — renamed 2026-10-27 (Asia/Shanghai)pp-open-in-loop — renamed 2026-10-27 (Asia/Shanghai)pp-print-in-loop — renamed 2026-10-27 (Asia/Shanghai)pp-regex-compile-in-loop — removed pp-regex-compile-in-loop — removed 2026-10-23 (Asia/Shanghai)pp-str-plus-in-loop — renamed 2026-10-27 (Asia/Shanghai)PP004 — removed PP004 — removed 2026-10-22 (Asia/Shanghai)

Project

Contributing Changelog Security Policy Contributor Covenant Code of Conduct AGENT.md — XPerf agent operating contract

LanguagesPythonEvidence_removed

pp-regex-compile-in-loop — removed 2026-10-23 (Asia/Shanghai)

Status: removed. CPython caches re.compile. The checked-in microbench was 1.22× (CPython 3.14, 5000 ops). A later rerun on a short haystack was 2.05× and on a longer haystack 2.62× (CPython 3.9). The search dominates the compile.

Slug never reused.
Historical JSONL and markdown stay beside this note. Fixtures moved under _removed/fixtures/.
Gate: a fair microbench that stays near 1× cannot remain an active lint.

pp-regex-compile-in-loop — removed

Previous Page

pp-str-plus-in-loop — renamed 2026-10-27 (Asia/Shanghai)

Next Page