Hi,
You've done this the right ways, for a single instance representation.
If this is a representation that would be needed more often (as in, is commonly
used in various dissectors) it can be added to the ftypes system (see
epan/ftypes), in this case ftype-time.c. I could image it being handled via a
new FIELDDISPLAY for FIELDTYPE FT_RELATIVE_TIME.
Thanks,
Jaap
On 06-05-16 01:59, Alan Partis wrote:
> In the DIS dissector (packet-dis.c), I have constructed an nstime_t var
> and populated its .secs and .nsecs fields with calculated values from the
> protocol buffer. I'd like to have this display in the packet dissection
> pane in so-called formatted time i.e. "mm:ss.nnnnnn" where 'nnnnnn'
> displays to microsecond accuracy.
>
> Right now, the dissector simply calls proto_tree_add_time() and the time
> is displayed as a number of seconds with nanoseconds i.e. "ssss.nnnnnnnnn"
> with the final 3 digits always showing '000' (because the nstime_t
> structure only hold a microsecond accuracy).
>
> To achieve what I want, I've changed the dissector to instead call
> proto_tree_add_time_format_value(..., "%02u:%02u.%06u", mm, ss, usec);
>
> While that produces the desired result, it doesn't feel quite right to me.
> Is there another way to get a proto_tree_add_time_xxx() function to print
> automagically the format I want? Or did I actually do it right in the
> first place?
>
> Thanks.
>
> _______________________________________________________
> Alan Partis
> thundernet development group