Comment # 9
on bug 8741
from chas williams (CONTRACTOR)
> I don't think the payload of the UDT should be a subtree of UDT, it should
> be part of the "main" tree. UDT is the payload of UDP (which is a payload
> of IP), and they all start at the "main" tree. I didn't see any "Data
> (dissector)" packets in your sample trace to use as an example of the
> difference.
That is why I called the data dissector directly. You didn't see any because I
need to call the data dissector explicitly.
> > 5. It looks like the ""Missing Sequence Numbers" would be better as
> > expert
> > info
>
> > The NAK can contains a single sequence number or a list of
> > sequence numbers. That seems to make it clumsy to represent.
>
> I'll provide a patch for what I'm thinking to see if you like it.
UDT's NAK explicitly indicates in the packet what is missing (either a single
packet, or a list of ranges). This is in the protocol itself not a part of the
dissector trying to detect the missing data. That is why I don't think it is
an expert option.
Try this latest version. It fixes the [Data] issue and should heuristically
detect the UDT stream based on the handshake packet.
You are receiving this mail because:
- You are watching all bug changes.
