Making Programs Faster

`parse_text`

Old code:

     elsif ( /^([A-Z])(<(?:<+\s)?)$/ ) {
       ## Push a new sequence onto the stack of those "in-progress"
       ($cmd, $ldelim) = ($1, $2);
       $seq = Pod::InteriorSequence->new(
                  -name   => $cmd,
                  -ldelim => $ldelim,  -rdelim => '',
                  -file   => $file,    -line   => $line
              );
       $ldelim =~ s/\s+$//, ($rdelim = $ldelim) =~ tr/</>/;
       (@seq_stack > 1)  and  $seq->nested($seq_stack[-1]);
       push @seq_stack, $seq;
     }

This handles the X< part of a sequence
It builds a new Pod::InteriorSequence and puts it on the stack
Later code takes the remainder, complete sequence> blah blah
- Expands complete sequence if necessary
- Appends it to the Pod::InteriorSequence object
- Puts blah blah back into the input stream
There's a lot of state variable management and stack jiggery-pokery