"call_ber_oid_callback()". (Arguably, the caller of
"call_ber_oid_callback()" should check for that, and report that a
presumably-required field is missing.)
svn path=/trunk/; revision=16544
tvb_get_ptr() returns a "const guint8 *"; don't bother casting its
return value to "const guint8 *".
"cms_verify_msg_digest()" can be passed a const pointer for "alg"; make
its argument a "const char *".
svn path=/trunk/; revision=16534
exist in GLib 2.x, but not in 1.2[.x]. Check the version number of
GLIB, not of GTK+, to decide whether to #define them.
svn path=/trunk/; revision=16522
The following diff (based on current SVN) fixes the following...
1. Adds additional return value 0xa901
"Invalid Data Type Flag (outdated return value - replaced in NSS as 0x89aa error)"
for 87/xx NCP's.
2. Fixes decode for NCP 87/20 reply packets
3. Fixes decode for NCP 89/53 request packets
svn path=/trunk/; revision=16509
Here are a number of small patches for asn1 based dissectors:
acse:
release request/response column information (many X.400/X.500 unbinds are
empty)
"standardised" PNAME to "ISO 8650-1 OSI Association Control Service"
fix for crash when using EXTERNAL dissector
rtse:
column information when attempting a resume
x509if:
generation of LDAP-style DNs from RDNSequences
new function x509if_get_last_dn() to get the last DN generated.
x509af:
DSS parameters
certificate extension naming
subject naming of certificate
x509sat:
Guide syntax (as SET now supported)
PDU exports.
cms:
verification of message digest attribute (SHA-1 and MD5)
ess:
enumerated/restrictive/permissive/informative security categories
x411:
generation of string encoding of X.400 addresses, trace information and message identifiers.
s4406:
separate types for primary and copy precedence to allow better filtering (e.g. primary precedence = flash)
priority-level-qualifier
svn path=/trunk/; revision=16508
handled (but not the same, as the prefix length in OSPF is rounded up to
a multiple of 4 bytes, not just a byte). This simplifies the code, and
keeps us from overflowing the buffer we allocate if the prefix length is
> 128 (we now just report that as an error).
svn path=/trunk/; revision=16507
always put the filename in col_info if it is known
for getinfo/setinfo put the class/level in colinfo as well.
this makes it very easy to see in the packet summary which file a icommand applies to and what kind of info is requested
svn path=/trunk/; revision=16496
updates to rlogin
Hi, this patch:
- uses value_string for control message field
- doesn't use a single global time value to mark start of all conversations!
- corrects client-user-name, server-user-name order in initial user-info
message
- adds data (text input/output) into the protocol tree
- makes lots more things filterable
- improves info column text (always calls rlogin_display() now)
- cleans up some whitespace
svn path=/trunk/; revision=16495
updates to smb
A patch for packet-smb.c is attached:
- it improves timeout decoding
- it defines common NT transaction IOCTL functions
- it corrects decoding of resume key in search queries
- it defines a new function dissect_4_2_16_8_unsure() to replace
dissect_4_2_16_8(). I'm unsure if it is correct. As said in
comments, it works for me, but I find strange that nobody noticed
dissect_4_2_16_8() was wrong. So, it is between "#if 0".
Someone else should confirm dissect_4_2_16_8_unsure() works
before activating it.
svn path=/trunk/; revision=16494
add TID tracking. for all TreeConnect requests/resposnes seen, store the name->tid mapping and other metadata.
as a freebee the disswection of the tid in the ehader is aware of this table so when a tid value is dissected in the header and we known the name for this tid then put it in an expansion below the tid.
svn path=/trunk/; revision=16483
we can regenerate from the header
we need to remember between request/response
we need on a per conersation bases
to reduce the amount of data we store in the per req/resp pair since there will be many of them and we want that struct as small as possible.
svn path=/trunk/; revision=16482
make them a guid in smb2 dissector and make tim pottesr excellent conmtext handle tracking also able to do the same thing for GUIDs
( a context handle is just a 32 biut attribute field followed by a uuid anyway)
svn path=/trunk/; revision=16477
Malformed packet IAM, Calling party number
Application transport - some code clean up.
bug 551 Nor Med PC ethereal-bugs@ethereal.com NEW SS7, ISUP, IAM-signal, malformed packet (not)
Bug 554 Min Low PC ethereal-bugs@ethereal.com NEW ETSI codec dissection error in BICC
svn path=/trunk/; revision=16476
we will do service-response-time statistics before other inferior products have even noticed a new protocol is in town.
svn path=/trunk/; revision=16463
the BYTE_ORDER == BIG_ENDIAN values for FMIP_HI_FLAG_ASSIGNED and
FMIP_HI_FLAG_BUFFER are bogus for the way they're used anyway (they're
used to check bits in a one-byte value), so throw away those values, and
get rid of the leading 00 in the BYTE_ORDER == LITTLE_ENDIAN values.
svn path=/trunk/; revision=16455
I just learned how easy it is to reassemble multiple UDP packets, so I
had to implement this for the P_Mul protocol :)
This patch fixes:
* Reassemble multiple data packets
* Add packet length check
* Fix for PDU type in tree entry
svn path=/trunk/; revision=16448
Ohoy sailor, we have spotted the concept of current working directory.
does this mean we also get rid of the silly 200-256 byte path length limitation?
svn path=/trunk/; revision=16445
You don't pull the mask off the old Lone Ranger, and you don't
overflow a guint8. Fixes the infinite loop in bug 576.
Fix up whitespace. Add an entry to the release notes. Apologies
to Jim Croce.
svn path=/trunk/; revision=16436
just supplying the result of "tvb_get_ptr()" with the same tvbuff,
offset, and length as the byte array.
Get rid of an unused variable.
Set svn:keywords and svn:eol-style appropriately.
svn path=/trunk/; revision=16430
In two functions that call proto_tree_add_bytes I forgot the
'tvb_get_ptr' function, the argument was 'FALSE' instead. This gave
a predictable 'NULL' pointer bug...
Fix an ftsanity.py error. Fixup whitespace.
svn path=/trunk/; revision=16428
ptr.
The answer to the question
"??????????????????????????????????????????????????????????????????" is
"No - the return value of tvb_get_ptr() is a reference, not an allocated
copy, and it cannot be freed and doesn't need to be freed."
svn path=/trunk/; revision=16426
Guy Harris