refactor: Call replaceCodePoint in tokenizer (#1167)

fb55 · Apr 23, 2022 · 8a785b8 · 8a785b8
1 parent 7b8a2f3
commit 8a785b8
Show file tree

Hide file tree

Showing 2 changed files with 7 additions and 17 deletions.
diff --git a/src/Parser.ts b/src/Parser.ts
@@ -1,5 +1,5 @@
 import Tokenizer, { Callbacks, QuoteType } from "./Tokenizer.js";
-import { decodeCodePoint } from "entities/lib/decode.js";
+import { fromCodePoint } from "entities/lib/decode.js";
 
 const formTags = new Set([
     "input",
@@ -258,7 +258,7 @@ export class Parser implements Callbacks {
          */
         const idx = this.tokenizer.getSectionStart();
         this.endIndex = idx - 1;
-        this.cbs.ontext?.(decodeCodePoint(cp));
+        this.cbs.ontext?.(fromCodePoint(cp));
         this.startIndex = idx;
     }
 
@@ -420,7 +420,7 @@ export class Parser implements Callbacks {
 
     /** @internal */
     onattribentity(cp: number): void {
-        this.attribvalue += decodeCodePoint(cp);
+        this.attribvalue += fromCodePoint(cp);
     }
 
     /** @internal */

diff --git a/src/Tokenizer.ts b/src/Tokenizer.ts
@@ -3,6 +3,7 @@ import {
     xmlDecodeTree,
     BinTrieFlags,
     determineBranch,
+    replaceCodePoint,
 } from "entities/lib/decode.js";
 
 const enum CharCodes {
@@ -707,19 +708,8 @@ export default class Tokenizer {
                 this.emitCodePoint(this.entityTrie[this.entityResult + 1]);
                 break;
             case 3: {
-                const first = this.entityTrie[this.entityResult + 1];
-                const second = this.entityTrie[this.entityResult + 2];
-
-                // If this is a surrogate pair, combine the code points.
-                if (first >= 0xd8_00 && first <= 0xdf_ff) {
-                    this.emitCodePoint(
-                        // http://mathiasbynens.be/notes/javascript-encoding#surrogate-formulae
-                        (first - 0xd8_00) * 0x4_00 + second + 0x24_00
-                    );
-                } else {
-                    this.emitCodePoint(first);
-                    this.emitCodePoint(second);
-                }
+                this.emitCodePoint(this.entityTrie[this.entityResult + 1]);
+                this.emitCodePoint(this.entityTrie[this.entityResult + 2]);
             }
         }
     }
@@ -746,7 +736,7 @@ export default class Tokenizer {
             }
 
             this.sectionStart = this.index + Number(strict);
-            this.emitCodePoint(this.entityResult);
+            this.emitCodePoint(replaceCodePoint(this.entityResult));
         }
         this.state = this.baseState;
     }