nostics-migration

Migrate errors and warnings to nostics

Turn ad hoc warnings and errors into a catalog of stable diagnostic codes without changing runtime behavior. API details: references/nostics-api.md.

What to migrate

Inventory console.warn, console.error, warn(...) helpers, throw new Error(...), Promise.reject(new Error(...)). Skip tests, fixtures, snapshots, generated output, and debug console.logs (not user-facing).

Migrate:

• dev warnings that report and keep going
• warnings followed by recovery/fallback (replace the report, keep the recovery)
• plain user-facing thrown or rejected Errors (the diagnostic becomes the thrown/rejected value)
• deprecation notices
• build/config errors caused by a user's file: pass both the original error as cause and the file as sources, because the JS stack points inside the library and is useless to the user

Do not migrate:

• structured errors other code inspects (type fields, private symbols, isXxxError() guards, instanceof checks): control flow, not reporting. Leave unchanged. If such an error is also user-facing, add a separate dev-only report with the error as cause; never replace the error object itself.
• catch blocks that only log a native/platform error and fall back when the library can't name a likely cause or concrete fix: the native error is the best info, keep the plain log. This covers platform APIs failing (history.pushState, storage quota), not errors from the user's own files: a caught parse/config error on a user file becomes a diagnostic carrying the original error as cause and the file as sources.
• anything where the diagnostic would only restate "an operation failed". A diagnostic earns its place by naming a likely user-code problem or a concrete fix.

Preserve behavior exactly

A project's dev guard may be process.env.NODE_ENV !== 'production' or its own build-time constant (recognize whatever the codebase uses). Written as DEV below; treat all forms the same.

• Keep existing dev guards exactly. nostics stripping is additive, not a replacement: a throw/reject that only happened in dev must still only happen in dev.
• Never add a guard the original lacked. A throw/report that fired in production keeps firing in non-stripped production builds. Migrating an unguarded report-only call makes it strippable, so it disappears once nosticsStrip runs. That is usually the goal, but if the library deliberately reports in production, surface that instead of changing it silently.
• Keep throw vs reject, timing, recovery code, and returned fallbacks.
• Keep structured error shapes (fields, symbols). Migrating a throw replaces the message with the diagnostic's why: if tests assert the old text, update them deliberately; never weaken the message to dodge a test.

Catalog shape

One catalog file per area (a single src/diagnostics.ts is fine for small libraries), exported directly. No factory wrappers, no deep barrel re-exports: the strip plugin tracks the export across one relative import. nostics is a runtime import: add it to dependencies, not devDependencies (library bundlers refuse or inline it otherwise).

import { createConsoleReporter, defineDiagnostics } from 'nostics'

export const diagnostics = /*#__PURE__*/ defineDiagnostics({
  docsBase: (code) => `https://example.com/e/${code.toLowerCase()}`,
  reporters: [/*#__PURE__*/ createConsoleReporter()],
  codes: {
    LIB_R0001: {
      why: (p: { hook: string }) => `${p.hook}() must be called at the top of a setup function.`,
      fix: 'Move the call into setup() or a composable called by setup().',
    },
  },
})

• Codes are PREFIX_XNNNN. Pick the category letter by area, not severity: B build, R runtime, C config, D deprecation (reserve D for deprecations; a runtime warning is R). Published codes are permanent: never rename or reuse one.
• why says what happened, with runtime values interpolated through typed param functions (both why and fix accept them; params merge and are required at the call site). fix is the concrete next action, never a restatement of the problem.
• Don't lose extra console.warn/console.error arguments: an error value becomes cause; data values interpolate into why (e.g. JSON.stringify(p.value)).
• cause and sources ('file:line:column' strings pointing at user code) go inside the params object (first argument), merged with message params. The second argument is reporter options only, e.g. { method: 'error' }.
• docsBase is optional. If the project has no documented error-page URL scheme, propose one and surface it rather than inventing pages that don't exist.

Call-site patterns

Calling a handle always runs the reporters, so throw diagnostics.CODE(params) reports and throws. For a warn-then-throw site that's the same double output it already had; for a bare throw new Error(...) it adds a console report before the throw, normally fine, but mention it if the library is strict about console output.

Report-only calls must stay bare expression statements (including DEV && diagnostics.LIB_R0001(p)) so nosticsStrip can remove them. throw/return/assigned diagnostics are behavior and stay.

Stripping diagnostics from production takes two pieces: /*#__PURE__*/ annotations on the catalog (defineDiagnostics(...) and each reporter factory call) so an unused catalog tree-shakes away, and a DEV guard on every report-only call site. Write both manually, or let the nosticsStrip build plugin add them (import { nosticsStrip } from 'nostics/unplugin/strip-transform', then the matching adapter: nosticsStrip.rolldown(), .vite(), .rollup(), ...). When every report-only site is already dev-guarded, manual annotations suffice and avoid a build transform; reach for the plugin when call sites are unguarded and stripping is wanted. Either way: if the library deliberately reports unguarded in production, don't silence it; ask the maintainer.

Documentation pages (optional)

Only when the repo already has a documentation site (docs/ with VitePress, Docusaurus, or similar): create one page per code and make docsBase resolve to them. No docs site → skip pages; just propose a docsBase URL scheme and surface it.

• Derive directory and URL shape from the existing site, not preference: pages live where docsBase resolves (docsBase → https://example.com/errors/<code> means docs/errors/<code>.md), and the URL must match the site's link style (clean URLs vs .html) so every reported see: link lands.
• If the site serves raw markdown alongside HTML (an llms integration: vitepress-plugin-llms, llms.txt tooling, or similar), end docsBase URLs in .md: diagnostics links are read by AI agents as much as humans, and the raw page spares them the HTML. Without such an integration, never assume .md resolves; match the HTML link style.
• One file per code, lowercase (docs/errors/lib_r0001.md), plus an index.md listing all codes grouped by area with one-line summaries. Register the section in the site's sidebar/nav.
• Page template: title CODE: short summary; level line (warn/error, dev-only or also production); What happened; How to fix it with a short wrong/right code example; Common causes.
• Write for the user who just hit the message: name the user-code mistake and the next action. A page that only restates why isn't worth a click: pull in context the one-line message couldn't carry (recovery behavior, related options, links to relevant guides).
• Document only user-facing codes. Internal assertions and companion tooling (devtools panels, debug plugins) don't get public error pages.

Verify

• Tests for warnings, throws, guards, and error shapes still pass; tests asserting exact message text are updated consciously, not accidentally.
• Dev-only gates are present everywhere the source had them, and no new gates were added.
• Report-only diagnostics remain strippable expression statements. Thrown/returned diagnostics keep their message in production by design: they're behavior, not reports.
• If doc pages were created, the docs site builds and the sidebar/index links resolve.