in reply to Re: Re: Re: highlight and line breaks
in thread highlight and line breaks

Assuming you never have an empty <span></span> set, I believe the following should work, should it not?

s{((?:(?:<span>)?.(?:</span>)?){80})}{$1\n}g

bbfu
Black flowers blossum
Fearless on my breath

  • Comment on (counting chars but ignoring SPANs) Re4: highlight and line breaks
  • Download Code

Replies are listed 'Best First'.
Re: (counting chars but ignoring SPANs) Re4: highlight and line breaks
by sauoq (Abbot) on Jan 16, 2003 at 08:42 UTC
    Assuming you never have an empty <span></span> set, I believe the following should work, should it not?
    s{((?:(?:<span>)?.(?:</span>)?){80})}{$1\n}g

    No.

    For illustrative purposes, assume that you want to match 10 characters rather than 80. That won't change the behavior. Let's try it on something simple like "<span>1</span>":

    #!perl $_ = "<span>1</span>"; s{((?:(?:<span>)?.(?:</span>)?){10})}{$1\n}g; print "$_\n"; __END__ <span>1</s pan>
    What happened? Well, perl tries hard to make the expression match. When it doesn't succeed at first by matching the initial span tag with the first optional group, it backtracks and matches that first "<" with the dot. This isn't restricted to one or two edge cases either. It may work for "<span>foo</span>" but only because "foo</span>" is exactly 10 characters. It won't work for "<span>fo</span>" or "<span>foob</span>". Of course, the examples would be different if you are trying to count 80 characters but the bug is the same.

    -sauoq
    "My two cents aren't worth a dime.";
    

      Of course you're right. It felt like I was missing something last night when I was testing it but all of my tests ended with enough non-span characters that it seemed to work. :/

      bbfu
      Black flowers blossum
      Fearless on my breath