Jeff Clites <[EMAIL PROTECTED]> wrote: > On Apr 21, 2004, at 10:14 AM, Leopold Toetsch wrote:
>> The "equal" case was missing one thing: if both strings are COWed >> copies, >> the compare can be avoided too - it's equally fast then, as "not >> equal". > That makes sense, as long as we never optimize substring via a COW copy > with a different strlen. The strlen's are compared earlier and are already equal, when the compare with C<strstart> is done. > JEff leo
