Some minor parsing issues for GitHub markdown

[EdmundKorley]

When parsing: a header named

this identifier

(Note the code formating) doctoc yields:

#this-identifier

... when it should yield:

#-this-identifier

This is because, -this-identifier, rather than this-identifier, is the actual id that GitHub renders when it converts the markdown to HTML for display. As a result the generated doctoc link will not work as a relative link.

There are a few other examples of this like:

For

Header & noise

doctoc produces

#header-&-noise

which will lead to a dead link in GitHub markdown rather than the appropriate

#header-noise

[rm-hull]

For the #header-&-noise case, I think the result should be a double dash: #header--noise

[jez]

The #thisidentifier example currently works correctly on GitHub.

The # Header & noise example is indeed a bug though.

The fix is almost identical to thlorenz/anchor-markdown-header#29.

[richbodo]

For the #header-&-noise case, the result should be a single dash: #header-noise.

Tested on github and gitlab - the link generated ignored the ampersand entirely, and uses a single hyphen.

[karfau]

I would guess that it converts all special chars to - and merges multiple sequenced - into a single one. But maybe it is possible to look at the code behind this to be sure?