http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=2017
jyoung@xxxxxxx changed:
What |Removed |Added
----------------------------------------------------------------------------
Attachment #1312 is|0 |1
obsolete| |
------- Comment #34 from jyoung@xxxxxxx 2008-01-13 04:45 GMT -------
Created an attachment (id=1376)
--> (http://bugs.wireshark.org/bugzilla/attachment.cgi?id=1376&action=view)
Examples of Graph Analysis report differences for same VoiP call
Hello Steve,
With your Option #3 patch applied I can no longer recreate the crash. ;-)
The only "bad" thing is when your Option #3 patch applied (as opposed to my
obsolete Option #1, v2 patch) it makes it harder to see the corruption that
sometimes occurs to the VoIP Graph Analysis reports if RTP_packet_draw() is
called multiple times (see my remarks at the bootom comment #28 about the
problem with the inconsistent VoIP "Graph Analysis" reports.)
I've just uploaded the file "voip.graphs.exports.txt" that shows three examples
of the VoIP graph produced from the smaller of the two pcap files attached to
this bug (the "Subset of initial trace file - call #2 frames" file).
If RTP_packet_draw() is called only once when the VoIP Calls window is
populated then the resulting Voip Calls "Graph Analysis" reports are apparently
consistent. But if RTP_packet_draw() gets called multiple times when the VoIP
Calls window is getting populated then the resulting VoIP Calls "Graph
Analysis" reports will generally have some apparently bad entries.
I included two examples of the bad Graph Analysis reports in the newly attached
file. The first of the two bad graph reports was produced with the Option #3
patch applied, the second with the obsolete Option #1 v2 patch applied.
Interestingly both of these "bad" reports are consistently broken. The
difference between these two reports is that the first graph entry in the
report generated with the Option #3 patch applied has the codec "(g711U)"
instead of the codec "(<Missing>)" that appears when the Option #3 patch
applied. A
I think the Option #3 patch should be applied to fix and bug 2017.
I also think a new bug should be opened to deal with the "Graph Analysis"
report discrepencies that occur when RTP_packet_draw() is called more than once
during the tap.
Comments?
--
Configure bugmail: http://bugs.wireshark.org/bugzilla/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the assignee for the bug, or are watching the assignee.
