"Sebastian Kreft" <[EMAIL PROTECTED]> writes: > Besides the use of int instead of size_t, what is the problem of portability > with my patch?
Nothing. That was the problem I was referrring to. For example, the patch you proposed mishandles a string of length 2147483648 on hosts with 32-bit int, because 'length' wraps around to -2147483648. > And finally, why do you use mempcpy instead of strncpy? A minor speed increase. mempcpy need not check for '\0'. _______________________________________________ Bug-coreutils mailing list [email protected] http://lists.gnu.org/mailman/listinfo/bug-coreutils
