perf: measure large trackers and reduce native mutation overhead

.agents/pm/history/pm-rust-bench-workflow.jsonl

@@ -0,0 +1,3 @@
+{"hash_algorithm":"sha256","ts":"2026-10-22T20:38:30.250Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_instance":"9cd644682823b57766dff4f3","agent_provenance":{"model":null,"effort":null,"role":{"value":"implementer","source":"argv"},"topic":null},"op":"create","patch":[{"op":"add","path":"/metadata/id","value":"pm-rust-bench-workflow"},{"op":"add","path":"/metadata/title","value":"Generalize native benchmark workflow beyond the performance branch"},{"op":"add","path":"/metadata/description","value":"Orchestrator review identified branch-specific workflow gating, a fixed historical BENCH_BASE_SHA and snippet-sensitive negative controls. The current PR keeps this historical benchmark protocol; generalize future source-PR and dispatch behavior separately without confusing ordinary benchmarks with implementation-specific revert controls."},{"op":"add","path":"/metadata/type","value":"Task"},{"op":"add","path":"/metadata/status","value":"open"},{"op":"add","path":"/metadata/priority","value":3},{"op":"add","path":"/metadata/tags","value":["ci","follow-up","performance"]},{"op":"add","path":"/metadata/created_at","value":"2026-10-22T20:38:30.250Z"},{"op":"add","path":"/metadata/updated_at","value":"2026-10-22T20:38:30.250Z"},{"op":"add","path":"/metadata/author","value":"codex-agent"},{"op":"add","path":"/metadata/acceptance_criteria","value":"Ordinary source PRs can run an explicitly defined base/head benchmark; revert controls apply only to matching implementations and fail clearly when stale; workflow and regression evidence cover future branches and changed implementation snippets."},{"op":"add","path":"/metadata/parent","value":"pm-rust-wdj4"}],"before_hash":"3cc22dff72be7b14824654a7a64ea62b04799939b2fee54c1b5f52ca60bf6df0","after_hash":"2414119e53e268cb1accd0052004bebc62d7adbabaeff8b4402ac79e7e9b92ff","item_hash_version":3,"message":"","context":{"agent_provenance_outcomes":{"model":{"status":"failed","reason":"resolver_failed","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"failed","reason":"resolver_failed","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"4d0b2b837febd30c97582e9f9a1f01c5cb6d1b800caaff944de39fc125459662"}
+{"hash_algorithm":"sha256","ts":"2026-10-22T20:38:36.835Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_instance":"9cd644682823b57766dff4f3","agent_provenance":{"model":null,"effort":null,"role":null,"topic":null},"op":"files_add","patch":[{"op":"replace","path":"/metadata/updated_at","value":"2026-10-22T20:38:36.835Z"},{"op":"add","path":"/metadata/files","value":[{"path":".github/workflows/native-bench.yml","scope":"project"}]}],"before_hash":"2414119e53e268cb1accd0052004bebc62d7adbabaeff8b4402ac79e7e9b92ff","after_hash":"cc4477ac842f965fce11f643193a36ddbbd58303f7e18da2e393427128029b74","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"failed","reason":"resolver_failed","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"failed","reason":"resolver_failed","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"maintenance","record_hash_version":1,"record_hash":"460529780190163fa6bbb4feb7cc44de5e2bbe8231ddaa436ceeace2b7db27e2"}
+{"hash_algorithm":"sha256","ts":"2026-10-22T20:38:43.394Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_instance":"9cd644682823b57766dff4f3","agent_provenance":{"model":null,"effort":null,"role":null,"topic":null},"op":"files_add","patch":[{"op":"add","path":"/metadata/files/1","value":{"path":"scripts/perf-revert-proof.mts","scope":"project"}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-22T20:38:43.394Z"}],"before_hash":"cc4477ac842f965fce11f643193a36ddbbd58303f7e18da2e393427128029b74","after_hash":"e5a46abe225a22a8d380f5701ec4c8404e0b31e16d3f7e94c3a05d4e537686e7","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"failed","reason":"resolver_failed","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"failed","reason":"resolver_failed","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"maintenance","record_hash_version":1,"record_hash":"8eac2e6480cf7959f661a49f28fe63c86e5eb3b072b36d56e2cbc0ce40415b6c"}

.agents/pm/history/pm-rust-qhem.jsonl

@@ -122,3 +122,4 @@
 {"hash_algorithm":"sha256","ts":"2026-10-21T17:34:03.261Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"e7e2d1ef918344626241964e","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":null,"topic":null},"op":"comment_add","patch":[{"op":"add","path":"/metadata/comments/70","value":{"created_at":"2026-10-21T17:34:03.261Z","author":"codex-agent","text":"Full Linux just release-check passes at source head 0029cf2e22e247b0cb806cf12a7dc6f022351482 in CI run 39641179755: 578 stable plus 578 instrumented tests, 44 binary executions, zero failed or ignored. Required coverage remains 100 percent: lines 14025/14025, functions 1466/1466, regions 23438/23438, branches 3296/3296. Full macOS aggregate already passes 567 plus 567 at 100 percent. Windows standalone passes 561 tests; its full aggregate remains the only pending validation job. No source, test, build-control or threshold changes after the certified source commit."}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-21T17:34:03.261Z"}],"before_hash":"203ffa693d3cd1bac353a7affcd139892e5c39fce14508553ffdc9a9c4ba1727","after_hash":"595778a24e9be3f5389a8c1a4fc8597da2c80a0489386977b3c4c038851bff51","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"809057efd632ee4bc6a32d505baaf4a87449dda06da109cabd35c7a1eddc6660"}
 {"hash_algorithm":"sha256","ts":"2026-10-21T17:40:35.932Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"e7e2d1ef918344626241964e","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":null,"topic":null},"op":"comment_add","patch":[{"op":"add","path":"/metadata/comments/71","value":{"created_at":"2026-10-21T17:40:35.924Z","author":"codex-agent","text":"Final certification at source head 0029cf2e22e247b0cb806cf12a7dc6f022351482: all nine jobs in CI run 39641179755 and CodeQL run 39641179753 pass. Full just release-check totals: Linux 578 stable + 578 instrumented, lines 14025/14025, functions 1466/1466, regions 23438/23438, branches 3296/3296; macOS 567 + 567, lines 13940/13940, functions 1453/1453, regions 23288/23288, branches 3258/3258; Windows 561 + 561, lines 13936/13936, functions 1452/1452, regions 23286/23286, branches 3260/3260. Each aggregate executes 44 test binaries, zero failed or ignored, unchanged 100 percent thresholds in all four dimensions. Both targeted regressions pass directly and through pm test. Two isolated function reverts and the requested path-expansion experiment fail at runtime with exit 101, restoring sources afterward. Final review fetch confirms only PRRT_kwDOTwgMRs6u3EZE remains unresolved, complete comments, latest review unchanged and no outside-diff findings. Refuse that path rewrite using oracle evidence; canonical comment advice is fixed. The published alias contract confirms CRUD root names remain unchanged and existing collection helpers already emit canonical names. No new pm CLI/SDK defect. Existing history issue pm-rust-b4yl and three health advisories remain outside scope. Release the claim without closing; final evidence commit changes only the receipt, this item and its appended history. No thread reply, resolution, merge, software release or history rewrite."}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-21T17:40:35.932Z"}],"before_hash":"595778a24e9be3f5389a8c1a4fc8597da2c80a0489386977b3c4c038851bff51","after_hash":"3889749774fc4c20ca9b47b142107becae7778599f640ffaa33b435339e8003e","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"fcb5bc66e026a8acd2ddec0cb60f63c5f81fb5627bf189d4e15fd26d7083ede8"}
 {"hash_algorithm":"sha256","ts":"2026-10-21T17:40:45.645Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"e7e2d1ef918344626241964e","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":{"value":"implementer","source":"argv"},"topic":null},"op":"release","patch":[{"op":"remove","path":"/metadata/claim_principal"},{"op":"remove","path":"/metadata/assignee"},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-21T17:40:45.645Z"}],"before_hash":"3889749774fc4c20ca9b47b142107becae7778599f640ffaa33b435339e8003e","after_hash":"bdeba7778696b9020c5705c9c053055fd1acc2b768a394a6eabd0ea404a078d6","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"maintenance","record_hash_version":1,"record_hash":"30305505ef356d3e5ec9459c44762b1794a560ecc96ab24990195af33f15d2b0"}
+{"hash_algorithm":"sha256","ts":"2026-10-21T19:27:21.696Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"0bbe066a9ad8a3dc5a99415c","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":null,"topic":null},"op":"comment_add","patch":[{"op":"add","path":"/metadata/comments/72","value":{"created_at":"2026-10-21T19:27:21.696Z","author":"codex-agent","text":"Fresh benchmark follow-up after merged parity baseline: the default hot-path invocations match pm-cli 2026.10.21 output bytes in 2 of 27 generated fixture cases (1k context and 1k search). Other default read/write outputs and all nine target item/history oracle-storage comparisons record differences without normalization; native before/after outputs and storage remain identical. This is compatibility evidence for the existing parity task, not a claim that every difference is a newly introduced defect. Review the raw flags and documented command scope in docs/benchmarks/native-hot-paths-2026-10-21.json and docs/NATIVE_HOT_PATHS.md before claiming equivalent current-oracle performance. Unsupported next/list-open names remain separately tracked in pm-rust-wc8b; large-context refusal remains with pm-rust-s6j4."}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-21T19:27:21.696Z"}],"before_hash":"bdeba7778696b9020c5705c9c053055fd1acc2b768a394a6eabd0ea404a078d6","after_hash":"cdbf83971d698c2bd6e815a8e9174cef0517bb46b083bf4064a67a4799b2b59d","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"6894b91cf76b385eb8c894d5af5d88f942218d4a6199653b1cf7375377a0c750"}

.agents/pm/history/pm-rust-s6j4.jsonl

@@ -117,3 +117,4 @@
 {"hash_algorithm":"sha256","ts":"2026-10-20T03:51:43.390Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"6ea8d289f989319d394eda9f","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":null,"topic":null},"op":"comment_add","patch":[{"op":"add","path":"/metadata/comments/64","value":{"created_at":"2026-10-20T03:51:43.390Z","author":"codex-agent","text":"Complete Bun orientation acceptance PASSED on exact source 1463a49: PM_NODE_INTERPRETER=bun PM_RUST_REQUIRE_PUBLISHED_CLI=1 cargo +1.90.0 test --locked --test orientation_differential -- --test-threads=1 finished 30 passed, 0 failed, 0 ignored, 0 filtered, in 446.30 seconds. Includes the huge-edge-limit timeout regression, 30 exact reference-filter comparisons, and retained root/evidence budget-floor parity. This is separate from the failed broader Bun aggregate tracked by pm-rust-9bci. Corrective CI is 8/9 green; Windows full aggregate remains in progress."}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-20T03:51:43.390Z"}],"before_hash":"c1f706a908f935e6d25c33f4567dfd74859c4ad752277aec24c3f7422cf20af4","after_hash":"55a8991ccec2a56478965ef3c214fc235f68a96f8458344e39b4de2252c60c23","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"247d37fcade229580f61c4a283cb3546409ce65589ba3e00fa0d29cd1db61755"}
 {"hash_algorithm":"sha256","ts":"2026-10-20T03:56:16.882Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"6ea8d289f989319d394eda9f","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":null,"topic":null},"op":"comment_add","patch":[{"op":"add","path":"/metadata/comments/65","value":{"created_at":"2026-10-20T03:56:16.882Z","author":"codex-agent","text":"Source 1463a49 CERTIFIED by exact-head CI run https://github.com/unbraind/pm-rust/actions/runs/39363908803: all nine jobs and CodeQL passed. Full unchanged just release-check totals: Linux 515 ordinary + 515 instrumented, macOS 511 + 511, Windows 506 + 506; each platform completed 42 suite executions with zero failed/ignored tests and 100% lines/functions/regions/branches. Covered totals in that order: Linux 13137/1355/21945/3152; macOS 13184/1367/22032/3154; Windows 13183/1368/22033/3156. Formatting, strict Clippy, private-item rustdoc, identity, audit, changelog and clone-local strict check-only health passed. Audit retains the inherited allowed RUSTSEC-2024-0449 advisory. All three latency jobs passed unchanged 50/150/250 ms bounds. Bun orientation passed 30/30; full Bun remains blocked at the inherited date-overflow runtime gap on pm-rust-9bci and is not claimed green. Final evidence-only head will receive its own CI and fresh-clone health verification."}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-20T03:56:16.882Z"}],"before_hash":"55a8991ccec2a56478965ef3c214fc235f68a96f8458344e39b4de2252c60c23","after_hash":"49f3db12a80dc443ce1866d0004668d23598233de7573805c1574a98c54f4a23","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"979bbd7db4e8eb5bdc86c8cccc302b4167bdb929ebda68aa5a622c27ed29aa18"}
 {"hash_algorithm":"sha256","ts":"2026-10-20T03:56:17.679Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"6ea8d289f989319d394eda9f","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":{"value":"implementer","source":"argv"},"topic":null},"op":"release","patch":[{"op":"remove","path":"/metadata/claim_principal"},{"op":"remove","path":"/metadata/assignee"},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-20T03:56:17.679Z"}],"before_hash":"49f3db12a80dc443ce1866d0004668d23598233de7573805c1574a98c54f4a23","after_hash":"217cbb5c41b7f5874d441b6db3e300432f0fe432b3ace3ee26be2ec1c63b3ed6","item_hash_version":3,"message":"Review fix certified on 1463a49; leave in progress for orchestrator verification. Ownership released after recording revert proof, exact-head gates and the inherited Bun follow-up. Final evidence-only head receives independent CI and fresh-clone health verification.","context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"maintenance","record_hash_version":1,"record_hash":"a32a7d68fd27b48098fa9325952e32720589b255e4bcbc70792819476b855036"}
+{"hash_algorithm":"sha256","ts":"2026-10-21T17:08:49.638Z","author":"codex-agent","author_source":"asserted","agent_harness":"codex","agent_model":"gpt-6.1-sol","agent_model_source":"probe","agent_instance":"0bbe066a9ad8a3dc5a99415c","agent_provenance":{"model":{"value":"gpt-6.1-sol","source":"probe"},"effort":{"value":"high","source":"probe"},"role":null,"topic":null},"op":"comment_add","patch":[{"op":"add","path":"/metadata/comments/66","value":{"created_at":"2026-10-21T17:08:49.637Z","author":"codex-agent","text":"The new generated hot-path benchmark explicitly includes 10k and 50k trackers. Main currently refuses context at those sizes for unsupported workspace-memory rollups, so those native rows are compatibility gaps and excluded from successful speed claims. Retain this outstanding scale boundary on the existing orientation item; benchmark procedure and measurements are linked from docs/NATIVE_HOT_PATHS.md."}},{"op":"replace","path":"/metadata/updated_at","value":"2026-10-21T17:08:49.638Z"}],"before_hash":"217cbb5c41b7f5874d441b6db3e300432f0fe432b3ace3ee26be2ec1c63b3ed6","after_hash":"dedd93ab127ce63c794806d8409eb7d7025e5e59a179dac6f3aa861fd4489a04","item_hash_version":3,"context":{"agent_provenance_outcomes":{"model":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"effort":{"status":"resolved","resolver":"codex_session_file","rule_version":"v1"},"version":{"status":"unavailable","reason":"harness_unavailable","resolver":"ai_agent_version","rule_version":"v1"}}},"event_class":"substantive","record_hash_version":1,"record_hash":"7c269dc0015b18da3e25857eefdcfd4cced985421698f1738dffc23ef11f45bc"}