fix(tokenizer): align edge-case parsing behavior with HTML spec
Merged
FB
7mo ago+611+611+012345678901234567890123456789-18-18-01234567890123456789
Handle the following tokenizer edge cases to match the HTML spec:
- Allow comment to close on
--!>, - Recognize closing tags that end with
/before>(e.g.</div/>), .reset()special-tag state correctly, and- Wait for
?>in XML processing instructions.