Hi Maritn,
MR> You don't write what OS you are using.
MR> Have you built Ethereal 0.9.9 yourself or downloaded binaries from the Ethereal site?
MR> Have you done any own modifications or loaded any other plugins that the normal ones?
I am using XP Pro on a Thinkpad X30 with the 3COM XJACK OfficeConnect
wireless card.
Interestingly, my colleague have no such problem although his X30 was
being bought with mine (i.e. theoretically the same platform). But as
you mentioned below, I also believe that it's being triggered by some
packets. In fact, I do not see the problem at home. Anyway, I will do
some captures in the office and try to dig that packet(s) out. (Give
me 2 days. :p I am away from office now.) In fact, the problem seems
to be only with the XJACK card's interface.
On the other hand, even my colleague Felix doesn't encounter this
problem, he encounter another "strange" problem which I don't. I will
suggest him to report that on another email. Put it simply, he would
always have an empty line on the RADIUS decode.
MR> Guy Harris went through the code and looked where the could be a fault leading to
MR> "file proto.c: line 1651: should not be reached" in v0.9.9 just two days ago.
MR> http://www.ethereal.com/lists/ethereal-users/200301/msg00232.html
MR> I also went through part of the code to search for those faults.
MR> (however of course it might be that we missed some places in the code).
MR> It turned out that the problem Ketil Kintel had was due to some packet that was decoded by the
MR> mdshdr dissector (even if the packet he had really wasn't mdshdr, but had ethernet type 0 that triggered the mdshdr dissector)
MR> The reason why Guy could find that fault when he did the search was due to that the problem had already been solved in the
MR> code he was doing the search (latest source code at that time).
: )
Albert