orca/config/performance-audit.md

Performance regression checks

pnpm --silent audit:perf > performance-audit.json scans production src/ with the existing app-store and buffer-concatenation rules. Warnings are advisory in this full inventory; tool/parser failures fail. New findings from rules enabled in the changed-code configuration fail pnpm check:code-quality:changed.

The app-store-performance/no-whole-hot-map-selector warning is advisory only: it appears in this inventory and the performance-contracts artifact, and its findings cannot fail the changed-code gate or other required checks. It catches selectors returning the whole runtimeStatusByEnvironmentId map, which is replaced on meaningful host observations. Select a stable derived answer inside the selector (see paired-runtime-parking-capabilities.ts). Fleet consumers can optionally silence the inventory warning with // oxlint-disable-next-line app-store-performance/no-whole-hot-map-selector -- <why every host is needed>. It resolves direct returns, module-scope named selectors and selector-local const aliases/destructuring; it does not infer consumer lifetime or follow imported helpers. Legitimate fleet consumers also match, so behavior-level render-isolation assertions remain the safeguard against actual fanout regressions.

Tests, generated files, mobile/ and cloud/ are outside this source audit.

The ast-grep rule config/ast-grep/rules/no-repeated-collator.yml (pnpm lint:ast-grep, an error everywhere) detects optioned localeCompare and Intl.Collator construction inside inline sort/toSorted callbacks. Construct one collator outside the callback, preserving locale, options and tie-breakers. If the locale changes at runtime, reconstruct at the next sort or key the cache by locale. Bare comparisons and standalone equality checks are allowed. There is no autofix or interprocedural analysis: named comparators, aliases, custom methods and deferred callbacks need manual review. A finding identifies repeated setup, not proof of visible lag.

pnpm test:perf:contracts runs the explicit selection in vitest.performance.config.ts: SQLite statement reuse and schema parity, relay filesystem concurrency, incremental Codex scans and reader lifetimes, pinned JSONL readers and snapshots, linear tool-result pairing and attribution allocation, tokenizer rejection, highlighting cache, queued cancellation, terminal backing-memory retention and detector fixtures. Missing listed files fail configuration loading. Tests run serially, without retries, and inherit the full suite's setup and forced-GC support. This makes existing regression coverage easy to run and attribute; it does not create new workload coverage by itself.

Incremental Claude scans and persisted Claude cache checkpoints are also included.

.github/workflows/performance-contracts.yml runs daily and manually on Linux, macOS and Windows, and on PRs changing this tooling or any listed contract file. It uploads per-OS JSON test results, plus the source inventory once from Linux because that scan is OS-independent. Its schedule starts after merge. Run the existing test:e2e:terminal-perf:scale:report (guide) for rendered typing/frame budgets. Unit tests do not measure SSH RTT, WSL scheduling or a packaged Electron renderer.

To extend coverage, select a production-path regression with an operation-count, identity, queue-admission or retained-memory oracle. Confirm it fails with the old behavior. Use controlled, counterbalanced benchmark samples for timings; avoid new machine-dependent millisecond gates in the normal unit suite. A green source scan and these contracts cannot establish that the whole app is fast.