// Run tape-six's own test suite in this repo. Use when invoking tests, picking between Node/Bun/Deno or parallel/sequential variants, or debugging a failing run.
[HINT] Download the complete skill directory including SKILL.md and all related files
| name | run-tests |
| description | Run tape-six's own test suite in this repo. Use when invoking tests, picking between Node/Bun/Deno or parallel/sequential variants, or debugging a failing run. |
Invoke tape-six's own test suite or a subset of it. The repo already has all the npm scripts wired — this skill is about choosing the right one.
tape-six test files are directly executable — node tests/test-foo.js (or bun tests/test-foo.js, deno run -A tests/test-foo.js) works without any runner, no transpilation, no harness wrapping. This is a defining property of tape-six and the fastest path when iterating on one file.
The tape6 runner / npm test script exists for orchestration: parallel execution via worker threads, glob-based discovery, per-environment filtering (Node / Bun / Deno / browser), and aggregated reporting. Use them for full-suite runs and CI; bypass them when iterating on a single file.
package.json#scripts)npm test — Node, parallel (default).npm run test:bun / npm run test:deno — Bun / Deno, parallel.npm run test:seq / npm run test:seq:bun / npm run test:seq:deno — sequential, in-process variants. Easier to debug; preserves stack traces.npm run ts-test / npm run ts-test:bun / npm run ts-test:deno — runs only the TypeScript test files (tests/test-*.*ts). Useful when iterating on .ts test changes.npm run ts-check — TypeScript-only typecheck pass (tsc --noEmit); complements the runtime test runs.npm run lint — prettier check.A single file can be run directly without the runner: node tests/test-foo.js (or bun run, deno run -A). Fastest path when iterating on one file; the worker-thread runner adds overhead for single-file runs.
All test scripts pass --flags FO by default.
--flags FO — Failures only, fail Once (quiet on green, stop on the first red). Lowercase disables, so passing --flags fo (via direct invocation) inverts both for a single run (shows passes, continues past failures). Override per-invocation with TAPE6_FLAGS=.... The full flag list lives at wiki/Supported-flags.md.
Before publishing, run all six matrix points to confirm parity (this is what release-check step 15 does):
npm test && npm run test:bun && npm run test:deno && \
npm run test:seq && npm run test:seq:bun && npm run test:seq:deno
If Node passes but Bun fails, run the failing file directly under each runtime to bisect:
bun tests/test-foo.js
deno run -A tests/test-foo.js
Browser tests live in tests/browser/ (configured under tape6.browser in package.json). Start the harness:
npx tape6-server --trace
Then open http://localhost:3000/ (all browser tests) or http://localhost:3000/?q=/tests/browser/test-foo.js (single file). Query parameters: ?flags=FO, ?par=N.
npm run test:seq) gives clean stack traces and avoids worker-thread overhead.npx tape6 --info.beforeAll (Docker spawn, large fixture loads): set TAPE6_WORKER_START_TIMEOUT=60000 (default 5 s).node ./bin/tape6.js --flags fo (lowercase shows passes too).wiki/Set-up-tests.md — full tape6 config block reference.wiki/Supported-flags.md — every flag, every override layer.wiki/Utility-‐-tape6.md — env vars (TAPE6_FLAGS, TAPE6_PAR, TAPE6_WORKER_START_TIMEOUT), parallelism control.wiki/Utility-‐-tape6‐server.md — browser harness query parameters, dev-time path mounts.wiki/Environment-‐-{Node,Bun,Deno,Browsers}.md — runtime-specific guidance.TESTING.md — top-level testing guide (also shipped to consumers).