Minor regexp oversight for setext headings

[Allan Odgaard]

On 7. Oct 2006, at 06:27, Allan Odgaard wrote:

Replying to myself here, as I didn’t get any other replies than the  
latest from A. Pagaitzis (so checked the archive).


> Noticed the patterns for setext style headings are:

>     ^(.+)[ \t]*\n=+[ \t]*\n+


Here (.+) is greedy and thus will match as much as possible,  
including trailing tabs or spaces, thus leaving the [ \t]* with  
nothing (which will then do zero repeats).

Example:

     % Markdown.pl <<< $'Test    \n=\n'
     <h1>Test    </h1>

While it doesn’t matter, I would expect the intended output to be:

     <h1>Test</h1>

And if not, well, then the above regexp could just as well be:

     ^(.+)\n=+[ \t]*\n+