Only one function is implemented so fat M-Get but it would be trivial to add all the missing ones once there are example captures.
svn path=/trunk/; revision=12521
It worked reasonably well mainly, I suspect, due to implicit tags are reasonably uncommon in the dissectors we have already implemented and that the bugs were masking eachothers.
my regression tests (limited test samples though) decodes this new one exactly the same as the old one.
As a bonus by not changing anythiong in the decode is that now it is possible to get dissection of implice items to work properly, hence CMIP
(and also x509 Extensions work now)
make heaps of dissector helpers implicit_tag aware.
change asn2eth to generate code to call the implicit_tag aware integer dissector helper.
svn path=/trunk/; revision=12520
make ethereal be able to measure the time it took to transfer a PDU atop TCP.
This is great for analyzing performance issues caused by network/frame loss/congestion.
See http://wiki.ethereal.com/TcpPduTime
svn path=/trunk/; revision=12516
buffer will necessarily be sufficient to make it big enough to handle
what's being added to the buffer - continue doubling the size until it
is.
Put the code to maintain the line buffer into a routine.
As that buffer is maintained as being large enough to handle what's
being put into it, there's no need to use "snprintf()" to print into it.
svn path=/trunk/; revision=12508
Contact header in the REGISTER reply contains more then one binding.
Details: up to now each Contact header was counted as one binding. But that
is not correct, because several Contact headers (as several other SIP headers
as well) can be written in one line separated by commatas.
svn path=/trunk/; revision=12507
fix the heuristic code -- sometimes a conversation already
exists;
fix the dissect code to display all the tags in the PDU.
svn path=/trunk/; revision=12504
support 6 additional login message verbs;
correct a problem with displaying proper return code values in
the reply packets.
Clean up white space.
Use "tvb_reported_length_remaining()" instead of
"tvb_length_remaining()", and don't use either one to avoid throwing an
exception on a short or malformed packet.
Use "val_to_str()" rather than "match_strval()", so we don't crash if a
value happens not to be valid.
svn path=/trunk/; revision=12503
error in "adns_init()". (I suspect this is what's causing the crashes
somebody's seeing when they run Ethereal on a Windows system with no
networking protocols installed - no networking protocols -> no Internet
protocols probably -> no ability to set ADNS up.)
svn path=/trunk/; revision=12500
so they show up near the top of the list of fields in the dialog box for
adding a field to a filter - those are probably quite likely to be used
in filter expressions where you don't happen to remember the name of the
field, and those should show up at the top so you don't have to scroll
through the entire list of fields to find them. (I suspect most other
fields either will rarely be filtered on at all, or would be filtered
only mainly with the "Match" or "Prepare" filter items, where you don't
need to know the name or even the text of the field.)
svn path=/trunk/; revision=12489
"history_current" happens to be null (that might not be possible, but
this is still perhaps a bit clearer).
Clear "history_current" after clearing the list, as the current entry no
longer exists, so that "packet_history_add()" doesn't use the now-bogus
value.
svn path=/trunk/; revision=12488
might have 0 in what would be the Ethernet type field.
Also, handle the first 5 octets of the destination address of an ISL
frame being 0C-00-0C-00-00.
svn path=/trunk/; revision=12484
take stuff that has to be done regardless of whether a protocol
tree is being built outside "if (tree)";
handle PDUs with both data and control units.
svn path=/trunk/; revision=12481