lib: remove Latin-1 fast path from TextDecoder

The Latin-1 fast path was incorrectly enabled only for windows-1252
encoding, which differs from ISO-8859-1 (Latin-1) in the 0x80-0x9F
range. Since windows-1252 cannot use the Latin-1 fast path (it requires
different character mappings via ICU), and no other encoding uses it,
the entire Latin-1 fast path mechanism has been removed.

This simplifies the code while fixing the windows-1252 decoding issue.
Windows-1252 now correctly uses the ICU decoder for all characters.

Fixes: #56542

Author: yashwantbezawada

lib/internal/encoding.js

@@ -28,7 +28,6 @@ const kEncoding = Symbol('encoding');
 const kDecoder = Symbol('decoder');
 const kFatal = Symbol('kFatal');
 const kUTF8FastPath = Symbol('kUTF8FastPath');
-const kLatin1FastPath = Symbol('kLatin1FastPath');
 const kIgnoreBOM = Symbol('kIgnoreBOM');
 
 const {
@@ -420,15 +419,9 @@ function makeTextDecoderICU() {
       this[kFatal] = Boolean(options?.fatal);
       // Only support fast path for UTF-8.
       this[kUTF8FastPath] = enc === 'utf-8';
-      // Disable Latin-1 fast path for windows-1252 as it differs from ISO-8859-1
-      // in the 0x80-0x9F range. The fast path uses simdutf which directly maps
-      // bytes to Unicode codepoints (e.g., 0x92 → U+0092), but windows-1252
-      // requires different mappings (e.g., 0x92 → U+2019 '). The ICU decoder
-      // handles these mappings correctly.
-      this[kLatin1FastPath] = false;
       this[kHandle] = undefined;
 
-      if (!this[kUTF8FastPath] && !this[kLatin1FastPath]) {
+      if (!this[kUTF8FastPath]) {
         this.#prepareConverter();
       }
     }
@@ -445,16 +438,11 @@ function makeTextDecoderICU() {
       validateDecoder(this);
 
       this[kUTF8FastPath] &&= !(options?.stream);
-      this[kLatin1FastPath] &&= !(options?.stream);
 
       if (this[kUTF8FastPath]) {
         return decodeUTF8(input, this[kIgnoreBOM], this[kFatal]);
       }
 
-      if (this[kLatin1FastPath]) {
-        return decodeLatin1(input, this[kIgnoreBOM], this[kFatal]);
-      }
-
       this.#prepareConverter();
 
       validateObject(options, 'options', kValidateObjectAllowObjectsAndNull);

test/parallel/test-whatwg-encoding-custom-windows-1252.js

@@ -83,50 +83,3 @@ const assert = require('assert');
     );
   }
 }
-
-// Test that common Windows-1252 encoding aliases work correctly
-// Per WHATWG Encoding Standard, many encodings map to windows-1252
-{
-  const aliases = [
-    'windows-1252',
-    'cp1252',
-    'x-cp1252',
-    'iso-8859-1',  // Per WHATWG spec, iso-8859-1 maps to windows-1252
-    'latin1',      // Per WHATWG spec, latin1 maps to windows-1252
-    'ascii',       // Per WHATWG spec, ascii maps to windows-1252
-  ];
-  const testByte = 0x92; // Right single quotation mark
-  const expected = '\u2019';
-
-  for (const alias of aliases) {
-    const decoder = new TextDecoder(alias);
-    const decoded = decoder.decode(new Uint8Array([testByte]));
-    assert.strictEqual(
-      decoded,
-      expected,
-      `Encoding alias '${alias}' should decode 0x92 to U+2019 (per WHATWG spec)`
-    );
-  }
-}
-
-// Test a realistic Windows-1252 text sample
-{
-  const decoder = new TextDecoder('windows-1252');
-
-  // "It's a "quote" — with €100"
-  const bytes = [
-    0x49, 0x74, 0x92, 0x73, 0x20, 0x61, 0x20, // It's a
-    0x93, 0x71, 0x75, 0x6F, 0x74, 0x65, 0x94, 0x20, // "quote"
-    0x97, 0x20, 0x77, 0x69, 0x74, 0x68, 0x20, // — with
-    0x80, 0x31, 0x30, 0x30, // €100
-  ];
-
-  const expected = 'It\u2019s a \u201Cquote\u201D \u2014 with \u20AC100';
-  const decoded = decoder.decode(new Uint8Array(bytes));
-
-  assert.strictEqual(
-    decoded,
-    expected,
-    'Realistic Windows-1252 text should decode correctly'
-  );
-}