[krbdev.mit.edu #2738] krb5int_sendto should look for expected message types

[Ken Raeburn via RT]

On Oct 12, 2004, at 18:02, Sam Hartman wrote:
> Why is this a bug?  It seems like you may get unpredictable behavior
> if you configure things as the tests are doing.  Why do we want to
> introduce complexity to work around this situation?  ALso, won't
> introducing this complexity mask reflections in other less explainable
> circumstances.?

I can think of one or two real-world situations where we might benefit 
from this change, but generally it's about certain active 
denial-of-service type attacks.  Not the intense "swamp the application 
with packets" type, but the "drop in a magic message and the client 
breaks" type.  Neither of which we really do anything about right now, 
so I don't expect it to be a priority.  But config file errors aren't 
the only way this could come up.  Well, reflections, probably, but 
unexpected or unknown message types can come up for other reasons.

Ken