- support for multiple input files
- #.MODULE_IMPORT replaced with #.MODULE (#.MODULE_IMPORT is supported for backward compatibility)
- #.MODULE is used for just compiled modules too
- more debug outputs
svn path=/trunk/; revision=18111
we need this in order to be able to provide proper itlq structures to the scsi dissector so that response data from scsi is dissected properly.
svn path=/trunk/; revision=18104
Fix bug 833 (remove 'dead' code flagged by Coverity);
Update defines based upon current version of FreeTDS tds.h; Reformat defines for readability;
svn path=/trunk/; revision=18103
it does not seem like it it is possible to trigger this bug since the condition where st is dereferenced and null should be impossible to create.
Still handle this impossible situation better if the guarantee that st can not be null here is changed in the future.
svn path=/trunk/; revision=18101
Make teh msb of the cmdset variable indicate whether we detected the commandset from the trace or whether we used the default value from preferences.
indicate in the dissection of the packet whether the command set is "known" or whether we are using the default one.
make scsi srt stats work even for when we are using the "default" dommandset. Previously scsi srt would ignore all pdus for itl sessions where the "default" command set was used.
svn path=/trunk/; revision=18098
In summary these patches do the following:
1. GTK+ 2.x builds will now use Tree View to display the expert items, expanded items are the individual packet numbers.
2. GTK+ 1.x builds will continue to use the older clist functionality
3. Implements the equivilant of expert info (actually calls expert_dlg.c)
svn path=/trunk/; revision=18091
Most of the time AssertionValue will contain an ascii string so make it always display as a string to make the display "correct" most of the time insterad of being "wrong" most of the time.
There are situations when AssertionValue contains binary data though and in those cases the display will be "wrong" (but not more wrong than the old dissector anyway)
What someone really should do (someone interested in ldap that is) to make it more correct would be to implement a dissector for AssertionValue in the template file and having the dissector check if any of the bytes of the octet string has a value <32 and if so display it in hex as 0x.....
It all bytes have values >=32 then it shoudlk display it as a string "...
instead.
Someone interested in ldap may spend time on this refinement.
svn path=/trunk/; revision=18089
Two more Kerberos error codes where it has been witnessed that the payload contains a PA-DATA structure with the magic salt containing an nt status code
svn path=/trunk/; revision=18088
authesserre samuel <sauthess@gmail.com> kindly pointed out an issue with session renegotiation in the current ssl decryption code.
Encrypted handshake message are decrypted, but the dissector try to interpret the encrypted code. Renegotiation messages are therefore ignored. The attached pcap trace and key can be used to trigger the issue.
The attached patch fix the problem storing the decrypted version of encrypted handshake message and dissecting it when available. The patch also fix bad issue with des cipher (alike the issue fixed in my previous post)
svn path=/trunk/; revision=18081
I attached patch to add preferences in SDP for RTP stream detection. By default SDP decodes RTP stream but now I can disable it.
svn path=/trunk/; revision=18080
sets errno to 0 when we can't open a file. This resurrected the annoying
"C:\Program Files\Ethereal\asn1/default.tt" error in the ASN.1 plugin,
since we depended on errno being set to ENOENT. Move some logic around
to account for this.
svn path=/trunk/; revision=18073