There's another case to account for... if $a or $b isn't integer, and their fractional parts differ (i.e. ($a - int $a) != ($b - int $b)) and $a < $b, then

$a = $b unless $error or $a >= $b;

(or similar examples in the thread) has a different result than the OP's code.

I think it muddies the water to test $error every time through the loop, as some of these solutions do, when you only need to test it once. If $a or $b might be non-integer, here's one way to do it:

if ($a < $b and !$error) {

  $a = int $b + ($a - int $a);
  $a++ if $a < $b;
}
[download]

I left out doing the floating point comparisons correctly...

Updated: Fixed a dumb error in the code.