On Sun, Nov 19, 2000 at 10:37:01PM +0200, Heikki Vatiainen wrote:
> dissect_sap() did not create a new tvbuff before calling the SDP
> dissector. Also, the test for the MIME content specifier was
> inversed.
Checked in. Thanks.
> The broken SAP dissector also broke the SDP dissector which ended
> up looping forever. The little fix for packet-sdp.c makes sure that
> the offset is advanced and thus the loop terminates eventually.
The correct fix is, I think to do
offset = next_offset;
to step to the next line; that's the version I checked in.
> A single packet demonstrating the looping is available at
> http://www.cs.tut.fi/~hessu/sdp-loop.cap
Good - I now have one more beast in my Menagerie of Capture Files on
which I run my regression tests, so if I break those dissectors again
I'll be more likely to catch the breakage....
(BTW, you might want to let the University of Toulose know that they
appear to be putting Windows-specific characters into their SDP data. :-)
If I set LC_ALL to en_US.ISO_8859-1, the Session Information shows up as
Session Information (i): Semaine en Région Midi-Pyrénées de l\222Académie des Sciences de l\222Institut de France
where "\222" is hex 92, which is, I think, one of the places where
Microsoft embraced and extended ISO 8859/1 - I suspect it's one of those
Windows-specific accent marks/apostrophes/etc. that plague those who
browse the Web on OSes or with browsers not from Redmond, i.e. the stuff
that shows up as "?".
Setting LC_ALL to fr_FR.ISO_8859-1 doesn't change that. :-))
