fix: decode html entities correctly (#8047)

fixes #8026 1. replace the big entities list with entities that have a ; at the end where valid (there are some exceptions) 2. construct two regexes from that entities list: one for HTML where it's strictly matched, one for attributes where it tries to match browser behavior by also allowing = / digit / character afterwards 3. decode character references with one of these regexes depending on this is an attribute value or not --------- Co-authored-by: Yuichiro Yamashita <[email protected]>
sveltejs · Feb 28, 2023 · f34abc5 · f34abc5
1 parent aa15a64
commit f34abc5
Show file tree

Hide file tree

Showing 8 changed files with 2,218 additions and 1,936 deletions.
diff --git a/src/compiler/parse/state/tag.ts b/src/compiler/parse/state/tag.ts
@@ -519,7 +519,7 @@ function read_sequence(parser: Parser, done: () => boolean, location: string): T
 
 	function flush(end: number) {
 		if (current_chunk.raw) {
-			current_chunk.data = decode_character_references(current_chunk.raw);
+			current_chunk.data = decode_character_references(current_chunk.raw, true);
 			current_chunk.end = end;
 			chunks.push(current_chunk);
 		}

diff --git a/src/compiler/parse/state/text.ts b/src/compiler/parse/state/text.ts
@@ -19,7 +19,7 @@ export default function text(parser: Parser) {
 		end: parser.index,
 		type: 'Text',
 		raw: data,
-		data: decode_character_references(data)
+		data: decode_character_references(data, false)
 	};
 
 	parser.current().children.push(node);