retransmitted, add items to its tree describing those retransmissions (frame
where the retransmission is and RTO). Limit this to 100 retransmissions to
avoid running out of memory in pathological cases.
This adds the filters "sctp.retransmitted" (TSNs that were retransmitted)
and "sctp.retransmitted_count" (count of number of times the TSN was
retransmitted).
The RTO is intentionally not added to the retransmitted TSN tree as it is
already added to the retransmission(s). The RTO is displayed, however.
svn path=/trunk/; revision=21081
remove all compiler warnings:
a) prevent wrong malloc/free definitions by lex/yacc generated files
b) add int/time_t casts - MSVC2005 is more "sensitive" about this than MSVC6
svn path=/trunk/; revision=21078
add sccp_info to struct _packet_info (Sorry but the way private_data works and the fact that TCAP uses it and BSSAP/RANAP can be tunnelled on GSMMAP over TCAP makes it impossible to avoid)
SCCP
- Have SCCP to have a TAP,
- Fix associations so that every message belongs to the association.
- Export message type values so that they can be used by a tap listener
RANAP
- Have RANAP information attached to the sccp_info
BSSAP + GSM_A
- Have DTAP, BSSMAP and BSSAP info attached to the sccp_info
svn path=/trunk/; revision=21076
A new function is introduced, prefs_register_protocol_subtree(), that allows the subtree the protocol should appear under to be specified. The subtree is specified as a string, with a '/' delimiting each subtree element. For example,
prefs_register_protocol(proto_dap, prefs_register_dap);
becomes
prefs_register_protocol_subtree("OSI/X.500", proto_dap, prefs_register_dap);
The function will create all the intermediate subtree nodes that are required, if they don't already exist.
This allows the grouping of procotols which should make the list of protocols more manageable as even more are added. The current aim is to group by protocol family e.g.
+ OSI
+ X.400
X.411
X.420
+ X.500
DISP
DAP
DOP
DSP
+ X.509
X509AF
X509CE
...
but others grouping could be envisioned (e.g. by first letter).
As the intermediate nodes may already have preferences (e.g. OSI), then modules are now allowed to have submodules. Previously each node was either a subtree or held preferences. This is consistent with the "User Interface" node.
The subtree structure has no effect on how the preferences are saved to file, and the "Protocol Preferences..." menu option will bring up the preferences expanded to the correct node.
In addition, a new "blank page" has been introduced for intermediate nodes that have no preferences (and is also used when the "Protocols" node itself is chosen). This prevents confusion when the user moves from a node with preferences to a node without preferences, but the page old page is still shown.
There is also a change to prevent '#' characters in a value being treated as a comment when parsing the preferences file. (There is nothing that adds comments when writing the preferences file.)
svn path=/trunk/; revision=21066
- Change "sctp.retransmitted" to "sctp.retransmission" since that field
is set on messages that are retransmissions, not messages that were
retransmitted.
- Change some formatting to make it more consistent.
svn path=/trunk/; revision=21065
- instead of wrongly using the h248 call counter use tapinfo's counter
graph_analysis.c:
- beautify the code (I was trying to the bug that got fixed by bzeroing m3ua tap data however this looks better!).
svn path=/trunk/; revision=21063
http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1336
Since mtp3_stat_draw() returns immediately if passed a NULL tapdata pointer,
call it with a pointer to the stat data (instead of NULL) in
mtp3_stat_gtk_cb().
svn path=/trunk/; revision=21042
screen
Changed IP address label to fit into an event box which is in the ip
address hbox. This allows the label to expand to fit the capture dialog
window, but no further. Expanding the capture dialog window will show
more IP addresses. Left align the label. Add ... in the middle of a
long list of IP addresses to show the first few and the last few with
the gtk_label_set_ellipsize() function in GTK 2.6 and up.
svn path=/trunk/; revision=21038
The capture file the user supplied had a HTTP chunked response
in it with no actual chunks other than the zero length chunk
indicating the end of the chunks. The fix is to only create
a new_tvb and copy it over the tvb going into the
chunked_encoding_dissector() function if the chunk size is > 0.
svn path=/trunk/; revision=21034