Comments::
[I should have better explained my reasons for certain parts of the patch]
Re: "Get rid of the old file header definition." (in the SVN 15606 note)
I left the old header in under '#ifdef 0' so the references
to 'xxb' and 'xxc' in several comments would continue to make sense.
I can submit a patch changing the (small number of) comments with
these references.
> I have a change to base the usec/nsec decision on the time units value
> (if it's > 10^6, use nsec).
I made the change to display nanosecs only for the gigpod because,
of the captures I have, only the gigpod captures display in the snffer
with nanosecs precision for time.
It seems to me that displaying nanosecs for the Tps=1193000, 1193182,
and 1250000 cases suggests too high an accuracy.
Also:
>> a. Define a hdr field which appears to be 'time-zone'
> Perhaps that's the way to fix bug 299:
I suspect that 299 is "not a problem" :) in that the capture
sniffer appliance time was the 'local' time which Ethereal then
converts to a UTC time.
(The capture file was for v2.1 and had zero in the timezone field).
re new comment in code: "[Does 0xff 0xff mean "time zone unknown?]"
I should have defined the timezone field as gint16 with negative values
to mean timezone hours *east* of UTC.
(I have one capture with FFF8 in this field; Based upon circumstantial
evidence the capture could very well have been made in the far East
which could match UTC +8 as the timezone).
I'll submit a patch to make this a signed integer;
Bill Meier
