fix(tokenizer): Properly handle CR when peeking (#715)

inikulin · Oct 30, 2022 · e6f304c · e6f304c
1 parent bb0129f
commit e6f304c
Show file tree

Hide file tree

Showing 2 changed files with 10 additions and 1 deletion.
diff --git a/packages/parse5/lib/parser/index.test.ts b/packages/parse5/lib/parser/index.test.ts
@@ -78,6 +78,13 @@ describe('parser', () => {
         expect(doctype).toHaveProperty('systemId', '');
     });
 
+    it('Regression - CRLF inside </noscript> (GH-710)', () => {
+        const onParseError = jest.fn();
+        parse('<!doctype html><noscript>foo</noscript\r\n>', { onParseError });
+
+        expect(onParseError).not.toHaveBeenCalled();
+    });
+
     describe('Tree adapters', () => {
         it('should support onItemPush and onItemPop', () => {
             const onItemPush = jest.fn();

diff --git a/packages/parse5/lib/tokenizer/preprocessor.ts b/packages/parse5/lib/tokenizer/preprocessor.ts
@@ -159,7 +159,9 @@ export class Preprocessor {
             return $.EOF;
         }
 
-        return this.html.charCodeAt(pos);
+        const code = this.html.charCodeAt(pos);
+
+        return code === $.CARRIAGE_RETURN ? $.LINE_FEED : code;
     }
 
     public advance(): number {