Tom,

I am still tracing the issue. I seem to have a problem with the transfer of
the context between child 1 and child 2 (e.g. child 2 gets the original
parent context).
In that case it wouldn't be a gss library issue, but a inter process
communication issue.

Thank you
Markus

"Tom Yu" wrote in message
news:ldvprm9v8gi.fsf@cathode-dark-space.mit.edu...
> "Markus Moeller" writes:
>
>> I understand that you need to keep the sequencing and in my case it is
>> the
>> case. Child 2 will always start after child 1 has finished. So there is a
>> clear flow order which is just handled by different processes, why I
>> thought
>> I can re-export/re-import the context. Unfortunately it is an existing
>> application and not that easy to completely rewrite.

>
> It sounds like you are experiencing a problem with sequence number
> state when you are doing (in serial order):
>
> * Exporting context1 from a parent
> * Importing context1 into child1
> * Sending wrap token from child1
> * Exporting context2 from child1
> * Importing context2 into child2
> * Sending wrap token from child2
>
> and child2 is producing an incorrect sequence number.
>
> If this is the case, I would consider it to be a bug. Would you be
> willing to provide additional details about this problem? What GSS
> library and what release?
> ________________________________________________
> Kerberos mailing list Kerberos@mit.edu
> https://mailman.mit.edu/mailman/listinfo/kerberos
>