Minor regexp oversight for setext headings

[John Gruber]

Michel Fortin <michel.fortin at michelf.com> wrote on 10/9/06 at 
12:47 AM:


>  As much as I agree with you, I tend to believe it'll break 

>backward  compatibility for a couple of people. I've seen this 

>a couple of times:

>       Paragraph...

>       Header

>     ======

>     Paragraph...

>   although I don't remember exactly where. So I think this 

>should be kept  working.


OK. I don't feel strongly about it, I suppose.

But what about this:

     #this#
     that

Right now that gets turned into

     <h1>this</h1>

     <p>that</p>

and I think that's just plain wrong. The only reason this works is
by accident; so I plan to require a blank line here.



> About this:

>       this

>     hello

>     =====

>     that

>   I wouldn't encourage anyone to write a two-line heading in

> most circumstances, but given ambiguous input, it does seem

> the most logical thing to do is to not separate contiguous

> text into parts. So I'd say it's one header and one paragraph:

>       <h1>this

>     hello<h1>

>       <p>that</p>


After reading other comments in this thread, I'm inclined to say
that Setext-style headers should not have embedded newlines. In
which case the above would translate to:

     <p>this
     hello
     =====
     that</p>

I think.

-J.G.