Simplify linkification filter - check for ASCII/non-ASCII only

js/modules/link_previews.js

@@ -212,134 +212,7 @@ function assembleChunks(chunkDescriptors) {
   return concatenateBytes(...chunks);
 }
 
-const LATIN_PATTERN = new RegExp(
-  '[' +
-    '\\u0041-\\u005A' +
-    '\\u0061-\\u007A' +
-    '\\u00AA' +
-    '\\u00BA' +
-    '\\u00C0-\\u00DC' +
-    '\\u00D8-\\u00F6' +
-    '\\u00F8-\\u01BA' +
-    ']'
-);
-
-const CYRILLIC_PATTERN = new RegExp(
-  '[' +
-    '\\u0400-\\u0481' +
-    '\\u0482' +
-    '\\u0483-\\u0484' +
-    '\\u0487' +
-    '\\u0488-\\u0489' +
-    '\\u048A-\\u052F' +
-    '\\u1C80-\\u1C88' +
-    '\\u1D2B' +
-    '\\u1D78' +
-    '\\u2DE0-\\u2DFF' +
-    '\\uA640-\\uA66D' +
-    '\\uA66E' +
-    '\\uA66F' +
-    '\\uA670-\\uA672' +
-    '\\uA673' +
-    '\\uA674-\\uA67D' +
-    '\\uA67E' +
-    '\\uA67F' +
-    '\\uA680-\\uA69B' +
-    '\\uA69C-\\uA69D' +
-    '\\uA69E-\\uA69F' +
-    '\\uFE2E-\\uFE2F' +
-    ']'
-);
-
-const GREEK_PATTERN = new RegExp(
-  '[' +
-    '\\u0370-\\u0373' +
-    '\\u0375' +
-    '\\u0376-\\u0377' +
-    '\\u037A' +
-    '\\u037B-\\u037D' +
-    '\\u037F' +
-    '\\u0384' +
-    '\\u0386' +
-    '\\u0388-\\u038A' +
-    '\\u038C' +
-    '\\u038E-\\u03A1' +
-    '\\u03A3-\\u03E1' +
-    '\\u03F0-\\u03F5' +
-    '\\u03F6' +
-    '\\u03F7-\\u03FF' +
-    '\\u1D26-\\u1D2A' +
-    '\\u1D5D-\\u1D61' +
-    '\\u1D66-\\u1D6A' +
-    '\\u1DBF' +
-    '\\u1F00-\\u1F15' +
-    '\\u1F18-\\u1F1D' +
-    '\\u1F20-\\u1F45' +
-    '\\u1F48-\\u1F4D' +
-    '\\u1F50-\\u1F57' +
-    '\\u1F59' +
-    '\\u1F5B' +
-    '\\u1F5D' +
-    '\\u1F5F-\\u1F7D' +
-    '\\u1F80-\\u1FB4' +
-    '\\u1FB6-\\u1FBC' +
-    '\\u1FBD' +
-    '\\u1FBE' +
-    '\\u1FBF-\\u1FC1' +
-    '\\u1FC2-\\u1FC4' +
-    '\\u1FC6-\\u1FCC' +
-    '\\u1FCD-\\u1FCF' +
-    '\\u1FD0-\\u1FD3' +
-    '\\u1FD6-\\u1FDB' +
-    '\\u1FDD-\\u1FDF' +
-    '\\u1FE0-\\u1FEC' +
-    '\\u1FED-\\u1FEF' +
-    '\\u1FF2-\\u1FF4' +
-    '\\u1FF6-\\u1FFC' +
-    '\\u1FFD-\\u1FFE' +
-    '\\u2126' +
-    '\\uAB65' +
-    ']'
-);
-
-const HIGH_GREEK_PATTERN = new RegExp(
-  '[' +
-    `${String.fromCodePoint(0x10140)}-${String.fromCodePoint(0x10174)}` +
-    `${String.fromCodePoint(0x10175)}-${String.fromCodePoint(0x10178)}` +
-    `${String.fromCodePoint(0x10179)}-${String.fromCodePoint(0x10189)}` +
-    `${String.fromCodePoint(0x1018a)}-${String.fromCodePoint(0x1018b)}` +
-    `${String.fromCodePoint(0x1018c)}-${String.fromCodePoint(0x1018e)}` +
-    `${String.fromCodePoint(0x101a0)}` +
-    `${String.fromCodePoint(0x1d200)}-${String.fromCodePoint(0x1d241)}` +
-    `${String.fromCodePoint(0x1d242)}-${String.fromCodePoint(0x1d244)}` +
-    `${String.fromCodePoint(0x1d245)}` +
-    ']',
-  'u'
-);
-
-function isChunkSneaky(chunk) {
-  const hasLatin = LATIN_PATTERN.test(chunk);
-  if (!hasLatin) {
-    return false;
-  }
-
-  const hasCyrillic = CYRILLIC_PATTERN.test(chunk);
-  if (hasCyrillic) {
-    return true;
-  }
-
-  const hasGreek = GREEK_PATTERN.test(chunk);
-  if (hasGreek) {
-    return true;
-  }
-
-  const hasHighGreek = HIGH_GREEK_PATTERN.test(chunk);
-  if (hasHighGreek) {
-    return true;
-  }
-
-  return false;
-}
+const ASCII_PATTERN = new RegExp('[\\u0000-\\u007F]', 'g');
 
 function isLinkSneaky(link) {
   const domain = getDomain(link);
@@ -350,12 +223,14 @@ function isLinkSneaky(link) {
     ? nodeUrl.domainToUnicode(domain)
     : domain;
 
-  const chunks = unicodeDomain.split('.');
-  for (let i = 0, max = chunks.length; i < max; i += 1) {
-    const chunk = chunks[i];
-    if (isChunkSneaky(chunk)) {
-      return true;
-    }
+  const withoutPeriods = unicodeDomain.replace(/\./g, '');
+
+  const hasASCII = ASCII_PATTERN.test(withoutPeriods);
+  const withoutASCII = withoutPeriods.replace(ASCII_PATTERN, '');
+
+  const isMixed = hasASCII && withoutASCII.length > 0;
+  if (isMixed) {
+    return true;
   }
 
   return false;

test/modules/link_previews_test.js

@@ -360,6 +360,12 @@ describe('Link previews', () => {
       assert.strictEqual(actual, true);
     });
 
+    it('returns true for ASCII and non-ASCII mix', () => {
+      const link = 'https://www.аррӏе.com';
+      const actual = isLinkSneaky(link);
+      assert.strictEqual(actual, true);
+    });
+
     it('returns true for Latin + High Greek domain', () => {
       const link = `https://www.apple${String.fromCodePoint(0x101a0)}.com`;
       const actual = isLinkSneaky(link);