Gates
A gate is a check that gives a verdict on a change. These are the kinds the stack runs, sorted by how soon each reports, as the oracles are.
The holds whose instances this table lists: Lints for the mechanical rules (Design drifts to the model's average); Drift checks and twin suites (Two copies drift apart); Oracles, mutants and control cases (Green but wrong); Suppressions that expire (A check is silenced instead of satisfied); Suites that fail on a skip (A gate stops gating); CI reruns the gate (A slow check gets skipped); Link and drift checks (A stale statement is trusted); Committed size limits (A predicted speedup that is not one); The gate (A known mistake comes back); Checks at the edges (Private content reaches a public repo)
Where a row speaks of a class of repos, the repos under it are the class's members, computed from this site's snapshot of each repo's files. A repo that reaches the same check another way, such as through a script its check script calls, is not listed unless the row names it. Read from the repos on .
A row may carry probes: a string that a named file, check script or CI step of the repo must have, or must not have, for the row to still be true in the way a file can show. The tests of this site check the ones about a check script or a CI step against its snapshot of the repos, and the evidence task checks the ones about a file. A probe that passes says the file, script or step still has the text the row depends on. It does not say the check is sound. A probe reads only the text it names, so one that says a step is absent does not see the same thing reached through another script or switched on by an environment variable.
- Every change In the gate a session runs before landing, and again in CI.
- CI In CI on a push to the main branch or a pull request, and not in the local gate.
- Release Before a release, and not on each change.
- On demand When someone chooses to run it.
| Gate | Catches | Runs | Where | Source |
|---|---|---|---|---|
| Types, tests, lints and format | Type errors, regressions, lint errors and style drift | Every change | The repos whose check script runs gro's check, with lints where the repo has a lint config. pglet's check script reaches it through a script of its own, for its TypeScript The repos whose check script runs gro check: fuz_template, zzz, fuz_forge, fuz_app, fuz_css, fuz_ui, fuz_code, mdz, fuz_blog, gro, tsv.fuz.dev, svelte-docinfo, fuz_repos, docs.fuz.dev, corpora, fuz_util, fuz_schema, zap | check task 2 probes
|
| Drift in generated files | A generated file that no longer matches its source | Every change | The repos whose check script runs gro's check, and pglet's TypeScript the same way The repos whose check script runs gro check: fuz_template, zzz, fuz_forge, fuz_app, fuz_css, fuz_ui, fuz_code, mdz, fuz_blog, gro, tsv.fuz.dev, svelte-docinfo, fuz_repos, docs.fuz.dev, corpora, fuz_util, fuz_schema, zap | checking generated files 2 probes
|
| Format, lints as errors, tests and doc links | Lint and test failures and broken doc links | Every change | The Rust workspaces, less the CI gaps listed below, and less zervo, whose lints only warn and which checks no doc links | Rust gate 1 probe
|
| Release audit | A forbidden crate or a test-only feature in a release build, a broken layering rule, and a copied lint table that has drifted | Every change | fuz, fuz_forge and zap. pglet runs an audit of its own, for its test-only crate alone | release audit 9 probes
|
| Pinned pass counts | Conformance coverage that quietly shrinks or shifts | Every change | pglet, and tsv's fixture validation | pinned counts 5 probes
|
| Required programs and suites | A suite that passes by skipping: a program a test needs is missing, or a database suite is left off the CI list | Every change | zap for programs, fuz for the CI list | testing rules 4 probes
|
| Links and anchors in Markdown | A dead relative link or anchor in any Markdown file of the repo | Every change | tsv | conformance audit 2 probes
|
| Size limits in the gate set | A WebAssembly build that has grown | Every change | pglet | gate runner 2 probes
|
| A budget for the startup snapshot | A startup snapshot that has grown past its committed budget | Every change | gro | baseline suite 3 probes
|
| Drift in the package exports map | An exports map that no longer matches the source, which the local gate rewrites instead of failing on | CI | The repos with an exports map whose CI runs gro's check with its workspace flag | check task |
| Build | A broken build | CI | The repos whose CI runs gro's check with its build flag The repos with a CI step, on a push or a pull request, that runs gro check with --build: fuz_template, zzz, fuz_forge, fuz_app, fuz_css, fuz_ui, fuz_code, mdz, fuz_blog, gro, tsv.fuz.dev, svelte-docinfo, fuz_repos, docs.fuz.dev, fuz_util, fuz_schema, zap | check task |
| A clean tree after the run, and no TODO docs at the root | Uncommitted output and a TODO Markdown file at the repo root | CI | The repos whose CI runs gro's check with its workspace flag The repos with a CI step, on a push or a pull request, that runs gro check with --workspace: fuz_template, zzz, fuz_forge, fuz_app, fuz_css, fuz_ui, fuz_code, mdz, fuz_blog, gro, tsv.fuz.dev, svelte-docinfo, fuz_repos, docs.fuz.dev, corpora, fuz_util, fuz_schema, zap | check task |
| Size limits in CI | A WebAssembly build or a native addon that has grown | CI | tsv and fuz_schema | artifact validation 3 probes
|
| Mutation gate | A test that passes whether or not the code is right | CI | fuz's install script | mutation gate 1 probe
|
| A snapshot that reproduces from its pins | A corpus that no longer matches its pinned upstreams, byte for byte, or whose license no longer reads as declared | CI | corpora | CI workflow 1 probe
|
| Conformance against committed ledgers | A parser or formatter difference that no ledger records | Release | tsv | audits 6 probes
|
| Mutation run | A test that passes whether or not the code is right | On demand | fuz_schema | mutants 3 probes
|
| Evidence links on this site | A link in this site's data whose file or heading is gone | On demand | This site | evidence check 2 probes
|
| Wall time and instruction counts against a base | A performance regression: tsv fails a row past a committed percentage, and gro asks for a written reason | On demand | gro and tsv | measure 5 probes
|
CI gaps
Where CI runs less than the local gate, or does not run:
- The Rust gate of zzz and its cross-backend tests run only locally, and its CI is Node-only
- zervo's CI builds and tests nothing, because its browser engine is too heavy to build on each push
- blake3 has no CI, so its size limits and test vectors run only by hand and before a release
- tsv-format has no CI, so its typecheck, build and tests run only locally
- tsv's CI skips the two legs of its gate that read sibling checkouts, the wider Prettier round trip and corpus discovery
- pglet's CI leaves out the WebAssembly and concurrency suites of its gate set
- zap's CI has no build of gro's runtime, so the tests that evaluate config through it are skipped there, by an explicit allowance