> Please put srand/rand to be the last one. IIRC, it is always worse
> than the alternatives.
Encryption code should not be using any of these. If it is trying to
mix data obtained from elsewhere, it should be using a cryptographically
strong hash, like MD5. If it is trying to obtain a starting point for
the random number, it MUST use truly random sources. On linux, there is
a driver which attempts to be a truly random source, although one cannot
rely on distributions setting it up properly (you need to carry state
across reboots by including code in the shutdown and restart scripts).
; To UNSUBSCRIBE: Send "unsubscribe lynx-dev" to [EMAIL PROTECTED]
