Thanks for the great work on Pygments! I use it regularly and am amazed by the breadth of languages it supports.

The current HTML lexer marks HTML comment as the token.Comment class, instead of the more specific token.Comment.Multiline. It also pushes into a sub-state which marks various flavors of embedded comments as token.Comment as well.

Would it be possible to replace this line with (r'<!--(.|\n)*?-->', Comment.Multiline), to mark all HTML comments as multi-line? The comment sub-state seems (to me) mostly useless, since it seems to mark everything as a token.Comment. I assume it can be removed?