conductor

ascii.test.js (3286B)

---

 // test/ascii.test.js - enforces the repository wide ASCII rule
 //
 // Every source and documentation file must be plain ASCII. This is a hard
 // project requirement rather than a style preference, so it is tested rather
 // than left to review. Smart quotes, dashes and arrows pasted in from a
 // browser are the usual cause of a failure here.
 
 import test from 'node:test';
 import assert from 'node:assert/strict';
 import fs from 'node:fs/promises';
 import path from 'node:path';
 import { fileURLToPath } from 'node:url';
 
 const ROOT = path.resolve(path.dirname(fileURLToPath(import.meta.url)), '..');
 
 // The rule is about code we write, not about a build output or a
 // dependency. This walks the filesystem, so .gitignore does not apply.
 const SKIP_DIRS = new Set(['.git', 'node_modules', 'dist']);
 const CHECK_EXT = new Set([
   '.js', '.mjs', '.cjs', '.json', '.md', '.sql', '.yml', '.yaml',
   '.html', '.css', '.sh', '.txt',
 ]);
 const CHECK_NAMES = new Set(['.editorconfig', '.gitignore', '.dockerignore', 'post-receive']);
 
 // Dockerfile, Dockerfile.worker, and anything else in that family.
 const CHECK_PREFIXES = ['Dockerfile'];
 
 function shouldCheck(file) {
   const base = path.basename(file);
   if (CHECK_EXT.has(path.extname(file))) return true;
   if (CHECK_NAMES.has(base)) return true;
   return CHECK_PREFIXES.some((prefix) => base.startsWith(prefix));
 }
 
 async function* walk(dir) {
   for (const entry of await fs.readdir(dir, { withFileTypes: true })) {
     if (entry.isDirectory()) {
       if (SKIP_DIRS.has(entry.name)) continue;
       yield* walk(path.join(dir, entry.name));
     } else if (entry.isFile()) {
       yield path.join(dir, entry.name);
     }
   }
 }
 
 function describeChar(ch) {
   const code = ch.codePointAt(0);
   return `U+${code.toString(16).toUpperCase().padStart(4, '0')} ${JSON.stringify(ch)}`;
 }
 
 test('every tracked text file is pure ASCII', async () => {
   const offences = [];
 
   for await (const file of walk(ROOT)) {
     if (!shouldCheck(file)) continue;
 
     const text = await fs.readFile(file, 'utf8');
     const lines = text.split('\n');
     for (let i = 0; i < lines.length; i += 1) {
       for (const ch of lines[i]) {
         const code = ch.codePointAt(0);
         // Tab, and the printable range. Newlines are already stripped.
         if (code === 9 || (code >= 32 && code <= 126)) continue;
         offences.push(`${path.relative(ROOT, file)}:${i + 1}: ${describeChar(ch)}`);
         break;
       }
     }
   }
 
   assert.deepEqual(offences, [], `non-ASCII characters found:\n${offences.join('\n')}`);
 });
 
 test('no file uses CRLF line endings or trailing whitespace', async () => {
   const offences = [];
 
   for await (const file of walk(ROOT)) {
     if (!shouldCheck(file)) continue;
 
     const text = await fs.readFile(file, 'utf8');
     const rel = path.relative(ROOT, file);
     if (text.includes('\r')) offences.push(`${rel}: contains a carriage return`);
     const lines = text.split('\n');
     for (let i = 0; i < lines.length; i += 1) {
       if (/[ \t]+$/.test(lines[i])) offences.push(`${rel}:${i + 1}: trailing whitespace`);
     }
     if (text.length > 0 && !text.endsWith('\n')) offences.push(`${rel}: missing final newline`);
   }
 
   assert.deepEqual(offences, [], `formatting problems found:\n${offences.join('\n')}`);
 });