On Thu, 24 Jul 2003 10:56:00 EDT, [EMAIL PROTECTED] said: > prior to that in the same e-mail. I'm not a programmer by any stretch of > the imagination but I found it curious that the C library had no kind of > error checking for valid values in the protocol field.
1) The "C library" has no read way of telling which protocols are "valid" for any given packet - if the packet is going over the network, it won't know for sure that protocol 50 is or isn't valid at the destination until it gets there. What usually happens if it's invalid is you get back an ICMP packet with type=3 (Dest Unreachable) with Code=2 (Protocol Unreachable). 2) C has type-casting and type-coercion rules that you can get bit by. Yes, you may have an "int" going from 0 to 1023, and you're assigning it to a 8-bit quantity. This will silently truncate and assign the low-order 8 bits. This is probably what bit you... > It basically "did what it was told". That's the basic design philosophy of C, actually.
pgp00000.pgp
Description: PGP signature