chore: address maintainer review feedback (round 4)

- delete unused tokenizeOursSimple and the now-redundant
  tokenizer-vs-csstools.mjs (ours already delegates to
  @csstools/css-tokenizer, so the comparison was circular)
- dedupe corpus loading between benchmark.mjs and show-divergences.mjs
  into scripts/lib/corpus.mjs
- drop the type-coverage dependency; strict tsc already covers this and
  CI never ran the quality script anyway
- rename src/lib/type.js to convertUnits.js to stop it reading as a
  TS-types or CSS-<type> file
- document each script's purpose in scripts/README.md

Author: dkryaklin

package.json

@@ -24,9 +24,7 @@
   "scripts": {
     "lint": "eslint . && tsc",
     "test": "node --test",
-    "test:mutation": "stryker run",
-    "quality:type-coverage": "type-coverage --at-least 97.5",
-    "quality": "pnpm quality:type-coverage && pnpm lint"
+    "test:mutation": "stryker run"
   },
   "author": "Andy Jansson",
   "license": "MIT",
@@ -46,7 +44,6 @@
     "fast-check": "^4.7.0",
     "postcss": "^8.5.16",
     "prettier": "^3.9.5",
-    "type-coverage": "^2.29.7",
     "typescript": "~6.0.3"
   },
   "dependencies": {

pnpm-lock.yaml

@@ -0,0 +1,21 @@
+# scripts/
+
+None of these run in `pnpm test` or CI — run directly with `node scripts/<name>.mjs`.
+
+- **`harvest-github.mjs`** — scrapes real-world `calc()` expressions from
+  public GitHub into `test/corpus/github/expressions.txt`.
+- **`split-corpus.mjs`** — splits that file into `github-pure.txt` (feeds
+  `benchmark.mjs`/`show-divergences.mjs` below), `preprocessor.txt`, and
+  `invalid.txt` (the latter two are used by real CI resilience tests).
+- **`lib/corpus.mjs`** — shared loader for `github-pure.txt`.
+- **`benchmark.mjs`** — times our pipeline against `@csstools/css-calc` over
+  the pure corpus.
+- **`show-divergences.mjs`** — buckets where our output disagrees with
+  `@csstools/css-calc` over the pure corpus, for manual triage.
+- **`tokenizer-compat.mjs`** — shared helpers for diffing token streams
+  (not runnable on its own); used by `tokenizer-suite.mjs`.
+- **`tokenizer-suite.mjs`** — runs the official `@rmenke/css-tokenizer-tests`
+  corpus through our tokenizer and reports pass/fail per category.
+- **`randomizer.mjs`** — long-running fuzzer: generates `calc()` inputs at
+  increasing depth, compares against `@csstools/css-calc`, logs finds to
+  `reports/randomizer-finds.jsonl`.

scripts/README.md

@@ -1,19 +1,12 @@
 // Benchmark: postcss-calc (pratt) vs @csstools/css-calc on the harvested
 // real-world corpus.
-import { readFileSync } from 'node:fs';
-import { dirname, join } from 'node:path';
-import { fileURLToPath } from 'node:url';
 import { tokenize } from '../src/lib/tokenizer.js';
 import { parse } from '../src/lib/parser.js';
 import { simplify } from '../src/lib/simplify.js';
 import { serialize } from '../src/lib/serialize.js';
 import { calc as csstoolsCalc } from '@csstools/css-calc';
-const ROOT = dirname(fileURLToPath(import.meta.url));
-const CORPUS = join(ROOT, '..', 'test/corpus/github-pure.txt');
-const corpus = readFileSync(CORPUS, 'utf8')
-  .split('\n')
-  .map((l) => l.trim())
-  .filter(Boolean);
+import { loadCorpus } from './lib/corpus.mjs';
+const corpus = loadCorpus();
 const ours = (s) => {
   try {
     return serialize(simplify(parse(tokenize(s))), { precision: false });

scripts/benchmark.mjs

@@ -0,0 +1,16 @@
+// Shared corpus loader for the exploratory scripts (benchmark.mjs,
+// show-divergences.mjs): reads the harvested real-world calc() corpus and
+// splits it into trimmed, non-empty lines.
+import { readFileSync } from 'node:fs';
+import { dirname, join } from 'node:path';
+import { fileURLToPath } from 'node:url';
+
+const ROOT = dirname(fileURLToPath(import.meta.url));
+const CORPUS = join(ROOT, '..', '..', 'test/corpus/github-pure.txt');
+
+export function loadCorpus() {
+  return readFileSync(CORPUS, 'utf8')
+    .split('\n')
+    .map((l) => l.trim())
+    .filter(Boolean);
+}

scripts/lib/corpus.mjs

@@ -1,18 +1,11 @@
 // Bucket github-pure corpus divergences against @csstools/css-calc.
-import { readFileSync } from 'node:fs';
-import { dirname, join } from 'node:path';
-import { fileURLToPath } from 'node:url';
 import { tokenize } from '../src/lib/tokenizer.js';
 import { parse } from '../src/lib/parser.js';
 import { simplify } from '../src/lib/simplify.js';
 import { serialize } from '../src/lib/serialize.js';
 import { calc as csstoolsCalc } from '@csstools/css-calc';
-const ROOT = dirname(fileURLToPath(import.meta.url));
-const CORPUS = join(ROOT, '..', 'test/corpus/github-pure.txt');
-const lines = readFileSync(CORPUS, 'utf8')
-  .split('\n')
-  .map((l) => l.trim())
-  .filter(Boolean);
+import { loadCorpus } from './lib/corpus.mjs';
+const lines = loadCorpus();
 const ours = (s) => {
   try {
     return serialize(simplify(parse(tokenize(s))), { precision: 10 });

scripts/show-divergences.mjs

@@ -2,7 +2,6 @@
 // @rmenke/css-tokenizer-tests JSON) onto one shape and diffs the streams.
 // Compares decoded values, normalizing the two designed differences:
 // function-token ≡ ident + `(`, and signed numeric ≡ punct sign + numeric.
-import { tokenize as ourTokenize } from '../src/lib/tokenizer.js';
 const PUNCT_DELIMS = new Set(['+', '-', '*', '/']);
 export class OutOfSubsetError extends Error {
   tokenType;
@@ -93,9 +92,6 @@ export function fromOurs(tokens) {
   }
   return out;
 }
-export function tokenizeOursSimple(css) {
-  return fromOurs(ourTokenize(css));
-}
 const isNumeric = (t) => t.type === 'number' || t.type === 'dimension';
 const tokenEq = (a, b) =>
   a.type === b.type &&

scripts/tokenizer-compat.mjs

@@ -5,13 +5,13 @@
 // rad/grad/turn, …). Buckets with `base === null` (relative or unknown
 // units) keep their own slot.
 
-const { convert } = require('../type.js');
+const { convert } = require('../convertUnits.js');
 
 /**
  * @typedef {object} UnitBucket
  * @property {string} unit
  * @property {number} total
- * @property {import('../type.js').BaseType | null} base
+ * @property {import('../convertUnits.js').BaseType | null} base
  * @property {number} order
  */