--- Begin Message ---
On Fri, 11 Sep 2020 16:59:19 +0200
George Hopkins via tcpdump-workers <tcpdump-workers@lists.tcpdump.org>

> Hello
> A few years ago there was a request to assign a DLT for RFtap [1]:
> https://lists.sandelman.ca/pipermail/tcpdump-workers/2016-August/000619.html
> Somehow some of the messages weren't kept by the archive. As I

There are four messages in the archive page above, which is the same as
what I see in the messages the mailing list sent me at the time, so it
looks like whatever had made it to the list is there. From the last
message contents it looks like some of the discussion could be off-list.

> understand it, it would only be possible to assign a DLT if the RFtap
> is tied to an actual link-layer type (e.g. RFtap followed by AX25).
> Does this assumption still hold? If so, is there an other,
> recommended way to store such information?

Guy had made a point that a variable-structure link level header is a
bad idea, that was likely based on his prior experience with other
DLTs. I do not have enough experience in this specific area to
elaborate on that, but my own observations in similar areas tend to be
such that when an encoding allows a human to make a mistake, mistakes
will happen, if not in one implementation then in another. So it would
make sense to look into this with attention if the goal is to produce a
good design.

    Denis Ovsienko

--- End Message ---
tcpdump-workers mailing list

Reply via email to