Changes are only for protocol version 2.
The changes are:
- dissect "TIPC Bundler Protocol" messages correctly
- search for other dissectors which want to dissect encapsulated data according to the TIPC user or TIPC type of a message. The data dissection is difficult since a TIPC data message does not necessarily a "type" set. So for the moment - while TIPC is not widely used - just triggering for the user of a message will be sufficient for people looking into the TIPC protocol.
- "Dissect TIPC data" in the preferences is now switched on by default
- to show undissected data, the "data" dissector is now used.
- corrected some typos
svn path=/trunk/; revision=22183
a SASL encapsulated ldap blob can contain more than one LDAP message so
the rest_is_pad parameter is bogus and thus removed.
make dissect_ldap_pdu handle when we have more than one LDAP message
inside one sasl blob
svn path=/trunk/; revision=22181
receiving a SES MAJOR SYNC POINT, as this indicates the end of the
COTP DT Data stream. Previous the RTSE dissector was called when
receiving a COTP DT Data fragment with the "last data unit" bit set,
but this does not work with messages fragmented in RTSE. Reassembly
can be turned off in the preferences.
svn path=/trunk/; revision=22176
fragmented data without adding an empty data fragment.
This is used by the RTSE dissector which can't identify the
last fragment until after it has been added.
svn path=/trunk/; revision=22174
COTP DT Data (class 1) fragment in the same frame. Also added the
generated dst_ref and a reference to the "COTP segment data" to the
tree.
svn path=/trunk/; revision=22173
This new code adds a save button to the Statistics IO Graphs window and
is also reusable by any other code that uses GDK Pixmaps to draw graphs.
The Gdk-pixbuf library included in GTK is used for this save function.
svn path=/trunk/; revision=22166
process WEP keys. Allow the "wep:" prefix for WEP keys even when
HAVE_AIRPDCAP isn't defined. Add a NULL pointer check to
hex_str_to_bytes(). Fixes bug 1584.
Fixup indentation.
svn path=/trunk/; revision=22151
Sometimes edonkey packets comes joined, and wireshark display only last
one on column Info. Now everyone is displayed.
svn path=/trunk/; revision=22143
- the Facility is not bellow the facility itself, but at the end of the tree.
It's not a problem, but it looks strange.. - Fixed
svn path=/trunk/; revision=22141
For the Cancel Location message, you can replace the
old_CancelLocationArgV2 by the Identity message.
So the old syntax for Cancel Location is not used anymore.
For the message Send Authentication Info Res, there is a typo error. The "item" should not be used.
- one with the decoding of "processUnstructuredSS-Data", the IA5 string
is not decoded/displayed.
This should go into 0.99.6.
svn path=/trunk/; revision=22135
packet-smb.c(5479) : error C2220: warning treated as error - no object file generated
packet-smb.c(5479) : warning C4244: '=' : conversion from 'unsigned short ' to 'unsigned char ', possible loss of data
packet-smb.c(5480) : warning C4244: '=' : conversion from 'unsigned short ' to 'unsigned char ', possible loss of data
svn path=/trunk/; revision=22134
- Remove ethertype preference from recently added FCoE dissector
Me:
- Add Joe to the AUTHORS list
- Change previous line in AUTHORS list from @ to [AT] in e-mail address
svn path=/trunk/; revision=22133
Please find attached a patch to fix some bugs in the armagetronad
dissector. More specifically:
* Update the descriptors to armagetronad-0.2.8.2.1, the file
src/network/Makefile.in does not exist anymore in the armagetronad
source code.
* Ensure the printed data is null terminated.
* Avoid guint16 overflow because of the '* 2'.
* Remove a statement with no effect (data = NULL).
svn path=/trunk/; revision=22131
Attached is a small patch that adds a preference option to allow the SQL
Query string to be appended to the INFO column display. This makes life
a little bit easier when scanning a trace to find the packet associated
with a specific query. I thought it was better to implement it as a
preference option with default set FALSE as it can clutter up the
display for applications that generate large amounts of queries.
svn path=/trunk/; revision=22130
Currently the MySQL dissector assumes that an OK-Packet Response
contains the Server_Status field. Having checked the MySQL protocol page
I can't say conclusively whether it should or it shouldn't, however I've
come across a couple of MySQL Java clients that receive the OK-Packet
without the Server_Status field set in response to a "SET AUTOCOMMIT" call.
The attached patched simply adds a check to ensure the Server_Status
field is present before calling function mysql_dissect_server_status.
svn path=/trunk/; revision=22129