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 <xydybaseball@gmail.com>

Author: xxkl1

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);
 }

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);