Currently there module supports only simple set syntax. But it is possible that in future it will support extended syntax: nested sets and set operations. Unfortunately that syntax is not fully compatible with the current syntax. In particular open bracket'[' in a character set starts a nested set. The code of html5lib contains a regular expression that will be broken if the new syntax will be accepted.

ascii_punctuation_re = re.compile("[\u0009-\u000D\u0020-\u002F\u003A-\u0040\u005B-\u0060\u007B-\u007E]")

It would be good to guard the code from possible future breakage. It is enough to add a backslash before[. Replace\u005B with\u005C\u005B,\\\u005B or\\[.

See Python issue:https://bugs.python.org/issue30349.