Re: Formal Grammar — some thoughts

[Michel Fortin]

Le 31 juil. 2006 à 18:25, A. Pagaltzis a écrit :


> No, it’s pretty clear. A backtick starts a sequence in which each

> character is interpreted literally. I don’t see how there can be

> any question: within code spans, there is no markup.


That's true, except that an unpaired backtick does not start  
anything. So if the link syntax takes precedence, whenever it is  
because of some priority or because it was started first, then you  
have an unpaired backtick within the link text and there is no code  
span to take literally.



> It does the right thing according to the syntax definition.


Yes, because the syntax definition leave enough room for both  
interpretations to be valid. I agree with what the implementation  
does, but I still think it is loosely defined in the specification.

Personally, I don't care about what should happen in these cases, I  
just follow what John's Markdown tool does. But with no clear rule to  
bind to, these corner cases will be incompatible between different  
implementations, which could be a problem to those who move from one  
to another. And in the case of

     *this **this* this**

or

     *this [this*](there)

even John's implementation cannot be the reference as it generates  
invalid markup! Personally, in these cases I'd give priority to  
strong emphasis and links over normal emphasis, but what will others do?


Michel Fortin
michel.fortin at michelf.com
http://www.michelf.com/