Trying to understand parsing of overlapping code and link

1

I’m trying to understand the difference in parsing two different cases where the document could be parsed as either inline code or link - but it feels the choice is not consistent. The two cases are

[`]()`
[](`)`

These parse as:

[](<code>)</code>
<a href="%60"></a>`

i.e. in the first case, the inline-code interpretation overrules the link interpretation, while in the second case the link interpretation overrules the inline-code interpretation.

Is there a specific rule in the spec that explains why this happens?

2

In the second case [](`)` the first backtick occurs in the URL part of the link which gets parsed literally and hence, can’t start a markdown element. It simply gets %60.

In the first case, the dingus parser parses this code to [<code>]()</code>. That’s because the backtick occurs at a point where markdown elements can start. Hence, it starts an inline-code which takes everything up to the next backtick and ‘destroys’ the link.

3

@jo-so is correct - I had a typo in the original post: the two strings were parsing as

[<code>]()</code>
<a href="%60"></a>`

not

[](<code>)</code>
<a href="%60"></a>`

as I originally had - that would be extra weird. I’ve corrected the original post now.