Eric Blake <[EMAIL PROTECTED]> wrote: ... > - The implementation was naively quadratic in the worst-case complexity. I > lifted Bruno's KMP ideas in mbsstr to make it linear+delayed allocation. > glibc > still uses the naive implementation - should we file a bug with them to fix > it?
This is the part I was replying to when I said it's worth a try. > Should we update memmem.m4 to reject known-quadratic implementations? I think this is definitely worthwhile, assuming (given your interest) that m4 is a heavy enough user of memmem that it makes a difference in practice.
