[Guy Harris <gharris@xxxxxxxxx>]

On Wed, Jun 11, 2003 at 03:38:54PM -0700, Guy Harris wrote:
> OK, I've checked that version in - with some fixes to get rid of other
> compiler warnings (such as a "possibly uninitialized variable" warning
> due to the code checking whether there's additional command data,
> constructing a subtree if there is, and then adding to that subtree
> whether there was additional command data or not - I changed it to move
> all that code under the "is there additional command data?" check).
> 
> You should probably get the updated version from anonymous CVS, and sent
> all subsequent updates as patches to that version, rather than as the
> entire source file.

By the way, the dissector (both the checked-in version, and the previous
version, without the changes in question) shows frame 49 of your sample
capture as a "Malformed Packet" - that packet shows an "Application
Reply Size" of 0x09 in the command-specific data of the second item in
the top-level command-specific data, but there isn't any application
reply data in the packet.

Is that packet, in fact, incorrectly constructed?