RE: [SQU] NTLM on Solaris

> Chemolli Francesco (USI) wrote:
>
> > The problem is that the Solaris libc doesn't define some
> integer types with
> > known bitlength, which we need to have.
>
> Maybe a non-problem by have you guarded yourself against byte ordering
> in the network exchanged data structures?

It's not a choice I did, but Microsoft.
IIRC those integers are sent in little-endian order, because that's what
i386 does htonl() and friends are no help in that case (they are
optimized away as noops on big-endian systems such as HPPA).
It is quite likely that the current code is portability-impaired (basically,
it might be restricted to i386-land only).

-- 
	/kinkie 
--
To unsubscribe, see http://www.squid-cache.org/mailing-lists.html