Add basic Oniguruma comparison tester

README.md

@@ -591,10 +591,10 @@ Notice that nearly every feature below has at least subtle differences from Java
     <td align="middle">✅</td>
     <td align="middle">✅</td>
     <td>
-      ✔ Error for variable-length quantifiers within lookbehind (JS allows)<br>
+      ✔ Error for variable-length quantifiers within lookbehind (allowed in JS)<br>
       ✔ Allows variable-length top-level alternatives<br>
       ✔ Allows following quantifier (unlike JS in any mode)<br>
-      ✔ Values captured within min-0 quantified lookbehind remain referenceable (unlike JS)<br>
+      ✔ Values captured within min-0 quantified lookbehind remain referenceable<br>
     </td>
   </tr>
 

package.json

@@ -21,6 +21,7 @@
     "build": "npm run bundle:global && npm run bundle:esm && npm run types",
     "pretest": "npm run build",
     "test": "jasmine",
+    "compare": "node spec/compare-oniguruma.js",
     "prepare": "npm test"
   },
   "files": [
@@ -47,6 +48,7 @@
   "devDependencies": {
     "esbuild": "^0.24.0",
     "jasmine": "^5.4.0",
-    "typescript": "^5.6.3"
+    "typescript": "^5.6.3",
+    "vscode-oniguruma": "^2.0.1"
   }
 }

spec/compare-oniguruma.js

@@ -0,0 +1,74 @@
+import {toRegExp} from '../dist/index.mjs';
+import {r} from '../src/utils.js';
+import {readFileSync} from 'node:fs';
+import path from 'node:path';
+import oniguruma from 'vscode-oniguruma';
+
+// Help with improving this tester or moving it into Jasmine specs would be very welcome!
+// Note: vscode-oniguruma 2.0.1 uses Oniguruma 6.9.8
+
+const tests = [
+  [r`\Aa`, 'a'],
+];
+
+async function compare() {
+  let numOk = 0;
+  let numErr = 0;
+  for (let i = 0; i < tests.length; i++) {
+    const [pattern, str] = tests[i];
+    const libMatch = toRegExp(pattern).exec(str);
+    const onigMatch = await onigurumaExec(pattern, str);
+    const libValue = libMatch && libMatch[0];
+    const libIndex = libMatch && libMatch.index;
+    const onigValue = onigMatch && onigMatch[0];
+    const onigIndex = onigMatch && onigMatch.index;
+    if (libValue !== onigValue) {
+      numErr++;
+      err(`${i}. Results differ ["${pattern}" with str "${str}"] lib: ${libMatch && `"${libValue}"`}, onig: ${onigMatch && `"${onigValue}"`}`);
+    } else if (libIndex !== onigIndex) {
+      numErr++;
+      err(`${i}. Positions differ ["${pattern}" with str "${str}"] lib: ${libIndex}, onig: ${onigIndex}`);
+    } else {
+      numOk++;
+      ok(`${i}. Results match ["${pattern}" with str "${str}"]`);
+    }
+  }
+  console.log(`\nFinished: ${numOk} OK, ${numErr} error${numErr === 1 ? '' : 's'}`);
+}
+
+compare();
+
+async function loadOniguruma() {
+  const wasmPath = path.join(import.meta.dirname, '..', 'node_modules', 'vscode-oniguruma', 'release', 'onig.wasm');
+  const wasmBin = readFileSync(wasmPath).buffer;
+  await oniguruma.loadWASM(wasmBin);
+}
+
+async function onigurumaExec(pattern, str) {
+  await loadOniguruma();
+  // See https://github.com/microsoft/vscode-oniguruma/blob/main/main.d.ts
+  const re = new oniguruma.OnigScanner([pattern]);
+  const match = re.findNextMatchSync(str, 0);
+  if (!match) {
+    return null;
+  }
+  const m = match.captureIndices[0];
+  return {
+    '0': str.slice(m.start, m.end),
+    index: m.start,
+  };
+}
+
+function err(msg) {
+  console.log(ansiEscape.red, `❌ ${msg}`, ansiEscape.reset);
+}
+
+function ok(msg) {
+  console.log(ansiEscape.green, `🆗 ${msg}`, ansiEscape.reset);
+}
+
+const ansiEscape = {
+  green: '\x1b[32m',
+  red: '\x1b[31m',
+  reset: '\x1b[0m',
+};

spec/helpers/matchers.js

@@ -19,7 +19,7 @@ function getArgs(actual, expected) {
   };
 }
 
-function matchedFullStr(match, str) {
+function wasFullStrMatch(match, str) {
   return !!match && match.index === 0 && match[0].length === str.length;
 }
 
@@ -30,12 +30,12 @@ function matchWithAllTargets({pattern, flags, strings, targets}, {exact, negate}
   for (const target of targets) {
     const re = toRegExp(pattern, flags, {target});
     for (const str of strings) {
-      // In case `flags` included `y`
+      // In case `flags` includes `g` or `y`
       re.lastIndex = 0;
       const match = re.exec(str);
       const failed = negate ?
-        ((exact && matchedFullStr(match, str)) || (!exact && match)) :
-        ((exact && !matchedFullStr(match, str)) || (!exact && !match));
+        ((exact && wasFullStrMatch(match, str)) || (!exact && match)) :
+        ((exact && !wasFullStrMatch(match, str)) || (!exact && !match));
       if (failed) {
         return {
           pass: false,

src/unicode.js

@@ -214,7 +214,7 @@ const PosixClassesMap = new Map([
   ['graph', r`[\P{space}&&\P{cntrl}&&\P{Cn}&&\P{Cs}]`],
   ['lower', r`\p{Lower}`],
   ['print', r`[[\P{space}&&\P{cntrl}&&\P{Cn}&&\P{Cs}]\p{Zs}]`],
-  ['punct', r`[\p{P}\p{S}]`],
+  ['punct', r`[\p{P}\p{S}]`], // New value from Oniguruma 6.9.9
   ['space', r`\p{space}`],
   ['upper', r`\p{Upper}`],
   ['word', r`[\p{Alpha}\p{M}\p{Nd}\p{Pc}]`],