> I know this sort of mirrors recent conversation... but it there a reason
> these packet counts are kept using independent sets of functions
> instead of piggybacking on some of the dissector stuff ( other than the
> obvious which is that we don't want to do a full dissection if we
> don't have to and we acheive not doing a full dissection by not passing
> a proto tree and thus have not proto tree to query for this info ).
That's pretty much it. We'd like to have a way of getting some simple
capture statistics that doesn't do very much work, as the machine could
be a bit busy shuffling packets around while capturing, so we don't want
to add to its burden too much.
(Microsoft Network Monitor normally has all sorts of meters and displays
and 8x10 glossy color photographs with circles and arrows and paragraphs
on the back running during a capture, but if you put it in "Dedicated
Capture Mode" it gets rid of its main window (not just minimizes, gets
rid of; I think that, in X11 terms, that'd be "unmapped" rather than
"iconified") and has only a window with a count of frames captured, and
"Stop", "Stop and View", "Pause", and "Normal Mode" buttons. I don't
know whether we'd want to consider having a mode like that, which does
even less work during a capture than we currently do, or not.)
