[David Aggeler <david_aggeler@xxxxxxxxxx>]

After reading quite a bit of documentation and going through different 
dissectors I'm even more confused. How am I now supposed to do the 
reassembly?

tcp_dissect_pdus() sound most promising, but the DICOM protocol lacks an 
overall size. I have a PDU length and a last fragment bit.
To use fragment_add_check() I think I'm missing a sequence number. For 
the later, it is also not clear to me, how I would separate different 
channels and same channel numbers, but different sessions within the 
same capture, in respect to the dissector 'global' variables.

Do I need both? I.e. first the tcp_dissect_pdus() for each PDU combined 
with fragment_add_check() inside or outside? When would I need 
p_add_proto_data()?

I know there's no short answer, but I still hope to get some feedback or 
directions.

Regards
David