Another bug in parsing of headers

[eGavr]

I have such md-file

# ##

I've received the tokens of the given markdown by method marked.lexer:

[ { type: 'heading', depth: 1, text: '#' }, links: {} ]

But the given markdown does not have any headers.

[joshbruce]

See #981

[joshbruce]

See Commonmark observation from @Feder1co5oave in #981

[calculuschild]

Commonmark Example 49 seems to indicate that emitting an empty header is the correct behavior, and Example 42 tells us that the number of opening and closing # do not need to match.

Our tests pass both of these Commonmark cases, and the OP's example correctly emits an empty header.

@UziTech I think can we close this?