On Wed, 2005-08-10 at 02:14 -0700, Austin Schutz wrote:
> So.. was the original question regarding sub-identifiers,
> or bytes?
Subidentifiers.
> I have seen objects in the wild with greater than 128 bytes,
Yes - that's perfectly legal
> but not 128 sub-identifiers.
and that's not.
> It seems like maybe MAX_OID_LEN used to refer
> to the maximum byte count.
Quite possible. In which case, it's a bug.
> So.. the max oid size should be MAX_OID_LEN * sizeof(u_long),
> which is a much bigger string than MAX_OID_LEN * sizeof(u_char): 512 vs. 128.
Well strictly speaking, it should probably be
MAX_OID_LEN * sizeof(oid)
so it'll have the correct size, regardless of the definition of 'oid'.
> Btw, according to the gcc docs int is always 32 bits, long can
> be 64 on 64 bit platforms. If the spec says the max val is 2 ** 32 - 1,
> then shouldn't that be u_int instead of u_long?
Quite possibly.
We've had a problem with a variety of approaches to integer ranges
for a long time now. Ideally we'd pick a particular set of fixed
size types (e.g. int8, int32, int64 + uint versions) and convert
everything to use those.
But the implications for backward compatability are so painful,
that we keep ducking the issue, and hoping it'll go away. :-)
Dave
-------------------------------------------------------
SF.Net email is Sponsored by the Better Software Conference & EXPO
September 19-22, 2005 * San Francisco, CA * Development Lifecycle Practices
Agile & Plan-Driven Development * Managing Projects & Teams * Testing & QA
Security * Process Improvement & Measurement * http://www.sqe.com/bsce5sf
_______________________________________________
Net-snmp-coders mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/net-snmp-coders
