



























@@ -7,16 +7,13 @@ read_when:
77title: "Testing"
88---
9910-# Testing
10+OpenClaw has three Vitest suites (unit/integration, e2e, live) and a small set
11+of Docker runners. This doc is a "how we test" guide:
111212-OpenClaw has three Vitest suites (unit/integration, e2e, live) and a small set of Docker runners.
13-14-This doc is a “how we test” guide:
15-16-- What each suite covers (and what it deliberately does _not_ cover)
17-- Which commands to run for common workflows (local, pre-push, debugging)
18-- How live tests discover credentials and select models/providers
19-- How to add regressions for real-world model/provider issues
13+- What each suite covers (and what it deliberately does _not_ cover).
14+- Which commands to run for common workflows (local, pre-push, debugging).
15+- How live tests discover credentials and select models/providers.
16+- How to add regressions for real-world model/provider issues.
20172118## Quick start
2219@@ -317,48 +314,83 @@ Think of the suites as “increasing realism” (and increasing flakiness/cost):
317314- Runs in CI
318315- No real keys required
319316- Should be fast and stable
320-- Projects note:
321-- Untargeted `pnpm test` now runs twelve smaller shard configs (`core-unit-fast`, `core-unit-src`, `core-unit-security`, `core-unit-ui`, `core-unit-support`, `core-support-boundary`, `core-contracts`, `core-bundled`, `core-runtime`, `agentic`, `auto-reply`, `extensions`) instead of one giant native root-project process. This cuts peak RSS on loaded machines and avoids auto-reply/extension work starving unrelated suites.
322-- `pnpm test --watch` still uses the native root `vitest.config.ts` project graph, because a multi-shard watch loop is not practical.
323-- `pnpm test`, `pnpm test:watch`, and `pnpm test:perf:imports` route explicit file/directory targets through scoped lanes first, so `pnpm test extensions/discord/src/monitor/message-handler.preflight.test.ts` avoids paying the full root project startup tax.
324-- `pnpm test:changed` expands changed git paths into the same scoped lanes when the diff only touches routable source/test files; config/setup edits still fall back to the broad root-project rerun.
325-- `pnpm check:changed` is the normal smart local gate for narrow work. It classifies the diff into core, core tests, extensions, extension tests, apps, docs, release metadata, and tooling, then runs the matching typecheck/lint/test lanes. Public Plugin SDK and plugin-contract changes include extension validation because extensions depend on those core contracts. Release metadata-only version bumps run targeted version/config/root-dependency checks instead of the full suite, with a guard that rejects package changes outside the top-level version field.
326-- Import-light unit tests from agents, commands, plugins, auto-reply helpers, `plugin-sdk`, and similar pure utility areas route through the `unit-fast` lane, which skips `test/setup-openclaw-runtime.ts`; stateful/runtime-heavy files stay on the existing lanes.
327-- Selected `plugin-sdk` and `commands` helper source files also map changed-mode runs to explicit sibling tests in those light lanes, so helper edits avoid rerunning the full heavy suite for that directory.
328-- `auto-reply` now has three dedicated buckets: top-level core helpers, top-level `reply.*` integration tests, and the `src/auto-reply/reply/**` subtree. This keeps the heaviest reply harness work off the cheap status/chunk/token tests.
329-- Embedded runner note:
330-- When you change message-tool discovery inputs or compaction runtime context,
331- keep both levels of coverage.
332-- Add focused helper regressions for pure routing/normalization boundaries.
333-- Also keep the embedded runner integration suites healthy:
334-`src/agents/pi-embedded-runner/compact.hooks.test.ts`,
335-`src/agents/pi-embedded-runner/run.overflow-compaction.test.ts`, and
336-`src/agents/pi-embedded-runner/run.overflow-compaction.loop.test.ts`.
337-- Those suites verify that scoped ids and compaction behavior still flow
338- through the real `run.ts` / `compact.ts` paths; helper-only tests are not a
339- sufficient substitute for those integration paths.
340-- Pool note:
341-- Base Vitest config now defaults to `threads`.
342-- The shared Vitest config also fixes `isolate: false` and uses the non-isolated runner across the root projects, e2e, and live configs.
343-- The root UI lane keeps its `jsdom` setup and optimizer, but now runs on the shared non-isolated runner too.
344-- Each `pnpm test` shard inherits the same `threads` + `isolate: false` defaults from the shared Vitest config.
345-- The shared `scripts/run-vitest.mjs` launcher now also adds `--no-maglev` for Vitest child Node processes by default to reduce V8 compile churn during big local runs. Set `OPENCLAW_VITEST_ENABLE_MAGLEV=1` if you need to compare against stock V8 behavior.
346-- Fast-local iteration note:
347-- `pnpm changed:lanes` shows which architectural lanes a diff triggers.
348-- The pre-commit hook runs `pnpm check:changed --staged` after staged formatting/linting, so core-only commits do not pay extension test cost unless they touch public extension-facing contracts. Release metadata-only commits stay on the targeted version/config/root-dependency lane.
349-- If the exact staged change set was already validated with equal-or-stronger gates, use `scripts/committer --fast "<message>" <files...>` to skip only the changed-scope hook rerun. Staged format/lint still run. Mention the completed gates in your handoff. This is also acceptable after an isolated flaky hook failure is rerun and passes with scoped proof.
350-- `pnpm test:changed` routes through scoped lanes when the changed paths map cleanly to a smaller suite.
351-- `pnpm test:max` and `pnpm test:changed:max` keep the same routing behavior, just with a higher worker cap.
352-- Local worker auto-scaling is intentionally conservative now and also backs off when the host load average is already high, so multiple concurrent Vitest runs do less damage by default.
353-- The base Vitest config marks the projects/config files as `forceRerunTriggers` so changed-mode reruns stay correct when test wiring changes.
354-- The config keeps `OPENCLAW_VITEST_FS_MODULE_CACHE` enabled on supported hosts; set `OPENCLAW_VITEST_FS_MODULE_CACHE_PATH=/abs/path` if you want one explicit cache location for direct profiling.
355-- Perf-debug note:
356-- `pnpm test:perf:imports` enables Vitest import-duration reporting plus import-breakdown output.
357-- `pnpm test:perf:imports:changed` scopes the same profiling view to files changed since `origin/main`.
358-- `pnpm test:perf:changed:bench -- --ref <git-ref>` compares routed `test:changed` against the native root-project path for that committed diff and prints wall time plus macOS max RSS.
359-- `pnpm test:perf:changed:bench -- --worktree` benchmarks the current dirty tree by routing the changed file list through `scripts/test-projects.mjs` and the root Vitest config.
360-- `pnpm test:perf:profile:main` writes a main-thread CPU profile for Vitest/Vite startup and transform overhead.
361-- `pnpm test:perf:profile:runner` writes runner CPU+heap profiles for the unit suite with file parallelism disabled.
317+<AccordionGroup>
318+<Accordion title="Projects, shards, and scoped lanes"> - Untargeted `pnpm test` runs twelve smaller shard configs (`core-unit-fast`, `core-unit-src`, `core-unit-security`, `core-unit-ui`, `core-unit-support`, `core-support-boundary`, `core-contracts`, `core-bundled`, `core-runtime`, `agentic`, `auto-reply`, `extensions`) instead of one giant native root-project process. This cuts peak RSS on loaded machines and avoids auto-reply/extension work starving unrelated suites. - `pnpm test --watch` still uses the native root `vitest.config.ts` project graph, because a multi-shard watch loop is not practical. - `pnpm test`, `pnpm test:watch`, and `pnpm test:perf:imports` route explicit file/directory targets through scoped lanes first, so `pnpm test extensions/discord/src/monitor/message-handler.preflight.test.ts` avoids paying the full root project startup tax. - `pnpm test:changed` expands changed git paths into the same scoped lanes when the diff only touches routable source/test files; config/setup edits still fall back to the broad root-project rerun. - `pnpm check:changed` is the normal smart local gate for narrow work. It classifies the diff into core, core tests, extensions, extension tests, apps, docs, release metadata, and tooling, then runs the matching typecheck/lint/test lanes. Public Plugin SDK and plugin-contract changes include extension validation because extensions depend on those core contracts. Release metadata-only version bumps run targeted version/config/root-dependency checks instead of the full suite, with a guard that rejects package changes outside the top-level version field. - Import-light unit tests from agents, commands, plugins, auto-reply helpers, `plugin-sdk`, and similar pure utility areas route through the `unit-fast` lane, which skips `test/setup-openclaw-runtime.ts`; stateful/runtime-heavy files stay on the existing lanes. - Selected `plugin-sdk` and `commands` helper source files also map changed-mode runs to explicit sibling tests in those light lanes, so helper edits avoid rerunning the full heavy suite for that directory. - `auto-reply` has three dedicated buckets: top-level core helpers, top-level `reply.*` integration tests, and the `src/auto-reply/reply/**` subtree. This keeps the heaviest reply harness work off the cheap status/chunk/token tests.
319+</Accordion>
320+321+<Accordion title="Embedded runner coverage">
322+- When you change message-tool discovery inputs or compaction runtime
323+ context, keep both levels of coverage.
324+- Add focused helper regressions for pure routing and normalization
325+ boundaries.
326+- Keep the embedded runner integration suites healthy:
327+`src/agents/pi-embedded-runner/compact.hooks.test.ts`,
328+`src/agents/pi-embedded-runner/run.overflow-compaction.test.ts`, and
329+`src/agents/pi-embedded-runner/run.overflow-compaction.loop.test.ts`.
330+- Those suites verify that scoped ids and compaction behavior still flow
331+ through the real `run.ts` / `compact.ts` paths; helper-only tests are
332+ not a sufficient substitute for those integration paths.
333+</Accordion>
334+335+<Accordion title="Vitest pool and isolation defaults">
336+- Base Vitest config defaults to `threads`.
337+- The shared Vitest config fixes `isolate: false` and uses the
338+ non-isolated runner across the root projects, e2e, and live configs.
339+- The root UI lane keeps its `jsdom` setup and optimizer, but runs on the
340+ shared non-isolated runner too.
341+- Each `pnpm test` shard inherits the same `threads` + `isolate: false`
342+ defaults from the shared Vitest config.
343+- `scripts/run-vitest.mjs` adds `--no-maglev` for Vitest child Node
344+ processes by default to reduce V8 compile churn during big local runs.
345+ Set `OPENCLAW_VITEST_ENABLE_MAGLEV=1` to compare against stock V8
346+ behavior.
347+</Accordion>
348+349+<Accordion title="Fast local iteration">
350+- `pnpm changed:lanes` shows which architectural lanes a diff triggers.
351+- The pre-commit hook runs `pnpm check:changed --staged` after staged
352+ formatting/linting, so core-only commits do not pay extension test cost
353+ unless they touch public extension-facing contracts. Release
354+ metadata-only commits stay on the targeted
355+ version/config/root-dependency lane.
356+- If the exact staged change set was already validated with
357+ equal-or-stronger gates, use
358+`scripts/committer --fast "<message>" <files...>` to skip only the
359+ changed-scope hook rerun. Staged format/lint still run. Mention the
360+ completed gates in your handoff. This is also acceptable after an
361+ isolated flaky hook failure is rerun and passes with scoped proof.
362+- `pnpm test:changed` routes through scoped lanes when the changed paths
363+ map cleanly to a smaller suite.
364+- `pnpm test:max` and `pnpm test:changed:max` keep the same routing
365+ behavior, just with a higher worker cap.
366+- Local worker auto-scaling is intentionally conservative and backs off
367+ when the host load average is already high, so multiple concurrent
368+ Vitest runs do less damage by default.
369+- The base Vitest config marks the projects/config files as
370+`forceRerunTriggers` so changed-mode reruns stay correct when test
371+ wiring changes.
372+- The config keeps `OPENCLAW_VITEST_FS_MODULE_CACHE` enabled on supported
373+ hosts; set `OPENCLAW_VITEST_FS_MODULE_CACHE_PATH=/abs/path` if you want
374+ one explicit cache location for direct profiling.
375+</Accordion>
376+377+<Accordion title="Perf debugging">
378+- `pnpm test:perf:imports` enables Vitest import-duration reporting plus
379+ import-breakdown output.
380+- `pnpm test:perf:imports:changed` scopes the same profiling view to
381+ files changed since `origin/main`.
382+- `pnpm test:perf:changed:bench -- --ref <git-ref>` compares routed
383+`test:changed` against the native root-project path for that committed
384+ diff and prints wall time plus macOS max RSS.
385+- `pnpm test:perf:changed:bench -- --worktree` benchmarks the current
386+ dirty tree by routing the changed file list through
387+`scripts/test-projects.mjs` and the root Vitest config.
388+- `pnpm test:perf:profile:main` writes a main-thread CPU profile for
389+ Vitest/Vite startup and transform overhead.
390+- `pnpm test:perf:profile:runner` writes runner CPU+heap profiles for the
391+ unit suite with file parallelism disabled.
392+</Accordion>
393+</AccordionGroup>
362394363395### Stability (gateway)
364396此内容由惯性聚合(RSS阅读器)自动聚合整理,仅供阅读参考。 原文来自 — 版权归原作者所有。