here are the highlights of what's fixed:
1) Decoding of vendor-defined types
2) Decoding of some event notification parameters (not all)
3) Decoding of ABSTRACT-SYNTAX&Type
4) Decoding of BACnetTimeStamp
5) Some problems with context-tagged values (like booleans)
6) Continuation segments - don't try to decode them since you can't
start mid-PDU
7) Removed some excessive levels of subtrees
svn path=/trunk/; revision=15226
Some changes that I made to flesh out GPRS message parsing. More information is displayed about the various frame formats. I have also added some code to parse XID parameters in the U frame. I have also fixed a couple of display bugs in the GSM and GPRS LLC parser.
svn path=/trunk/; revision=15224
- not only look for the cygwin tools, but also check the MSVC tools required like cl.exe, link.exe and nmake.exe
- I don't know why we should use /usr/bin/find but simply find, check for it instead. If there's a reason to use /usr/bin/find, we should use $(FIND) instead but I currently don't see a reason for this
svn path=/trunk/; revision=15220
"make distclean", even though they come with the distribution, and
make the one in plugins/xml/Makefile.nmake match the one in some other
files.
svn path=/trunk/; revision=15214
Among the improvements are:
- fixes to call-tracking (it's now less likely to confuse two separate
calls, for instance)
- improvements to Information Element dissection (clearer dissection,
dissects more IE types, easier to extend)
- you can now filter on the content of DTMF packets
- Analysis of timestamps (calculation of absolute timestamp, and packet
lateness).
- fixed a couple of assertion failures in subtle corner-cases.
negative relative times:
- get_timedelta()
- addtime()
- ftype-time.c:relative_val_from_unparsed()
I've also moved get_timedelta() and addtime() out of calldata.c into a
new file, epan/nstime.c, as I needed to use them in a dissector I'm
working on (and they therefore needed to go into libethereal).
svn path=/trunk/; revision=15201
the strings mmse got from tvb_get_string were never released anyway
so this would resolve significant memory leakage in mmse.
there are still memleaks in mmse but one step at a time
svn path=/trunk/; revision=15197
Unfortunately, this is *very* safe now as it will mark a lot of DCE/RPC packets as malformed, thus the reassembling code won't work with most packets :-(
I've replaced this with the correct check and used THROW(ReportedBoundsError) instead which is hopefully the right error in this case.
BTW: could someone please add a description of the ReportedBoundsError vs. BoundsError to the code as I still don't get it.
In addition: Use a toplevel tree item to display the DCE/RPC defragmentation just as it's done in the TCP desegmentation code.
svn path=/trunk/; revision=15188