On 1/10/07, Joerg Mayer [EMAIL PROTECTED] wrote:
On Tue, Jan 09, 2007 at 10:15:55PM +0100, Thomas Anders wrote:
Plenty, but I'm not sure they're related:
packet-snmp-template.c: In function 'snmp_usm_priv_des':
packet-snmp-template.c:1169: warning: ISO C90 forbids mixed declarations
and
PROTECTED] [EMAIL PROTECTED] wrote:
http://anonsvn.wireshark.org/viewvc/viewvc.cgi?view=revrevision=20353
User: lego
Date: 2007/01/09 06:38 PM
Log:
SNMPv3 USM decryption/authentication phase 1
Directory: /trunk/epan/crypt/
ChangesPath Action
+3 -1
MD5 DES
80001F888059DC486145A26322 pippo2 plutoxxx PLUTOxxx
#defaults to MD5 DES
0002 test1 maplesyrup XXX
# example from RFC3414
On 1/9/07, LEGO [EMAIL PROTECTED] wrote:
Here we go!
what's missing is:
* auth SHA1 -- RFC3414 section 7
* crypt AES -- RFC3826
* localized
try move them after net-snmp's headers
On 1/9/07, Thomas Anders [EMAIL PROTECTED] wrote:
LEGO wrote:
However the most important test to be done is that it does no harm
while disabled. Weird runtime linking problems disallow me to build
with net-snmp so I cannot test it that way. I do
any warnings afterwards?
On 1/9/07, Thomas Anders [EMAIL PROTECTED] wrote:
LEGO wrote:
might be a problem caused by a macro defined in one of the following
71 #include epan/crypt/hmac.h
72 #include epan/expert.h
73 #include epan/report_err.h
76
meddelande-
Från: [EMAIL PROTECTED]
[mailto:[EMAIL PROTECTED] För LEGO
Skickat: den 9 januari 2007 23:52
Till: Developer support list for Wireshark
Ämne: [Wireshark-dev] SHA1 in airpdcap_sha1.c fails on big endian boxen
As per the subject, the sha1 implementation in airpdcap_sha1.c does
not work
After spending (wasting) several hours debugging my SNMPv3 USM
authentication code I found out that our md5 implementation is
broken!
It returns just plain wrong digests!
12345 should yield 827ccb0eea8a706c4c34a16891f84e7b
but instead it yields cad0f574e6b876f5248314742bb09b7c
I'll see if I can
You can use lua for that...
do
local tcp_port_table = DissectorTable.get(tcp.port)
local http_dissector = tcp_port_table:get_dissector(80)
for port in {4888,4889,4890,4891} do
tcp_port_table:add(port,http_dissector)
end
end
On 1/5/07, Douglas
take a look to http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1223
for a workarround.
On 11/17/06, Ravi Kondamuru [EMAIL PROTECTED] wrote:
Hi,
I just build the latest revision of wireshark rev: 19915. I am seeing a
weird display issue. I cant see any text. I am attaching a screenshot to
few months later both issues are solved... (revision 19903)
On 5/15/06, Jacques, Olivier (OpenCall Test Infra)
[EMAIL PROTECTED] wrote:
Hello,
I am working on implementing MSCML support in Ethereal
(http://ietfreport.isoc.org/idref/draft-vandyke-mscml/ or
may be we should have EP_COPY_ADDRESS, SE_COPY_ADDRESS and
PE_COPY_ADDRESS instead.
On 11/13/06, Kukosa, Tomas [EMAIL PROTECTED] wrote:
Hi,
does not it make sence to change g_malloc() in COPY_ADDRESS to
se_alloc() ?
It seems that it is not freed in most cases of usage.
Do wee need
to the forbidden-function list?
Luis
On 11/9/06, Joerg Mayer [EMAIL PROTECTED] wrote:
On Tue, Nov 07, 2006 at 12:57:59PM +, LEGO wrote:
what about #defining them so they trigger an error?
#define ntohl error() won't work, a g_ntohl would match as well. But how
about creating our own ntohl
However I believe that what Gerald is talking about is using threads
for asyncronously executing upgrades.
I do not think that the very linking to the thread libraries causes
performance degradation.
If on a single processor you must continiuously switch contexts
between the main thread and the
what about #defining them so they trigger an error?
On 11/7/06, Ulf Lamping [EMAIL PROTECTED] wrote:
Jaap Keuter wrote:
Hi,
Checked in.
Thanx,
Jaap
On Mon, 6 Nov 2006, Albert Chin wrote:
Patch attached to convert usage of ntohl() - g_ntohl(). On HP-UX,
ntohl() isn't
You can create an emty dir on c:\ and rename it in config.nmake:32
WIRESHARK_LIBS=C:\wireshark-win32-libs
On 11/7/06, Robert Trybis [EMAIL PROTECTED] wrote:
On Windows XP when trying to run the Automated library dowload
the command
nmake -f Makefile.nmake setup
fails if the directory
That's odd... it should have been caught by the buildbot, (I do not
see that error with perl v5.8.6 on MacOS X).
Can you try this patch and see if that fixes the problem for you.
Luis
Index: make-taps.pl
===
--- make-taps.pl
: malformed patch at line 10:
I have attached the revision of the make-taps.pl I have.
# $Id: make-taps.pl 19579 2006-10-17 18:20:44Z lego $
--
Michael Lum Principal Software Engineer
4600 Jacombs Road+1.604.276.0055
Richmond, B.C.
Canada V6V 3B1
UTStarcom Canada
You beat me by few seconds (not kidding)...
There's an issue here:
The h248v3 asn hasn't yet been tested...
I believe it would be better to revert the changes to the generated
dissector and make them by hand.
Then let the v3 asn for the next release.
Anders, what do you think?
On 10/25/06,
That means that mate was installed with 0.99.3 and it was not
installed with 0.99.4.
What we could do is to add a dummy plugin that only registers the
obsoleted preference and install that when MATE is disabled.
Is it worthwhile?
On 10/24/06, Ulf Lamping [EMAIL PROTECTED] wrote:
Hi List!
I think 19670 should go as well, it has Lua disabling itself by default.
--
This information is top security. When you have read it, destroy yourself.
-- Marshall McLuhan
___
Wireshark-dev mailing list
Wireshark-dev@wireshark.org
19662 should be there too, it alows editcap -A and -B args to take DST
into account.
On 10/24/06, LEGO [EMAIL PROTECTED] wrote:
I think 19670 should go as well, it has Lua disabling itself by default.
--
This information is top security. When you have read it, destroy yourself.
-- Marshall
There's probably other duplicate libs (I've seen the same behaviour we
discussed befor when with a duped openssl).
Luis
On 10/23/06, Andreas Fink [EMAIL PROTECTED] wrote:
0.99.4pre1 still doesnt work under MacOS X unless -without-zlib is being
choosen as configure option even though no other
On 10/21/06, Ulf Lamping [EMAIL PROTECTED] wrote:
LEGO wrote:
Can you check if wslua_datafile_path is added to the generated
WSLUA_DECLARE_FUNCTIONS() ( in declare_wslua.h ) and to
wslua_register_functions() (in register_wslua.c) ?
Yes for .h, no for .c (.c does include datafile_path
Gerald,
This one shoud go in 0.99.4
L
On 10/23/06, [EMAIL PROTECTED] [EMAIL PROTECTED] wrote:
http://anonsvn.wireshark.org/viewvc/viewvc.cgi?view=revrevision=19662
User: lego
Date: 2006/10/23 02:57 PM
Log:
As proposed by Tim Furlong:
set starttm.tm_isdst to -1 to have mktime
On 10/23/06, Ulf Lamping [EMAIL PROTECTED] wrote:
LEGO wrote:
the missing prefix is added by the macro... that's ugly and I'll fix
it but it is not our problem.
I do not have access to a windows box on which to test this so I need help.
Can you:
I cannot reproduce the problem any
the following message box:
Lua: Error during loading:
[string C:\Program Files\Wireshark\init.lua]:230: attempt to call
global 'datafile_path' (a nil value)
Lego, could you have a look to fix it before the release?
Regards, ULFL
___
Wireshark-dev
The docbook files have a description of the API and are generated from
comments in the code using wslua/make-doc.pl, I did not knew about
doxygen, I'll take a look into it.
The wiki pages is out of sync, I'll leave just examples and poiters to
documents about Lua. They are useless as such...
As
Today I finally got the generated docbook files for the wslua ref man
to compile (There's still work to do on the content but formally they
seem OK).
My question is:
How should I handle the compilation of this docbook?
the docbook xml files are in epan/wslua/doc
should I have docbook/Makefile
This may be what is happening with:
http://bugs.wireshark.org/bugzilla/show_bug.cgi?id=1158
I'm waiting for the reporter's answer to verify that anyway I'll try
to set starttm.tm_isdst to -1 and see whether it works .
Luis
On 10/18/06, Tim Furlong [EMAIL PROTECTED] wrote:
Hi folks,
I've run
remove rdm.so or rdm.dll from the plugins directory, as the dissector
was recently moved to epan.
On 10/19/06, Stephen Fisher [EMAIL PROTECTED] wrote:
Ever since RDM became a normal dissector, I am getting this message when
running from the latest SVN:
10:44:35 Err Duplicate
I would love to have some (at least minimalistic) documentation for
the lua API for the next release, I'll throw myself into it in the
next few days.
Can we hold for that?
Luis
On 10/10/06, Gerald Combs [EMAIL PROTECTED] wrote:
I'd like to release 0.99.4 next Wednesday (the 18th). If you're
I got the tools menu hidden if there's no items in it...
Now the problem is that if there's are menu items in it it appears
last (i.e. after Help) which is not how it should be...
I been trying to figure out how to move it, but so far without
success, any hints?
On 10/16/06, Ulf Lamping [EMAIL
The menu is there just when compiled with Lua, and the current
init.lua adds one item for it.
Howhever it would be nice to keep it hidden until it gets populated
(I'll take a look at it soon).
A second item for that menu (DTD generator), is coming, a tool that
parses xml and adds the fields
/viewvc.cgi?view=revrevision=19507
User: lego
Date: 2006/10/12 05:00 PM
Log:
- add persconffile(opt_str)
- add datafile_path(opt_str)
- add Dir class -- a Directory Iterator
Directory: /trunk/epan/wslua/
ChangesPathAction
+25 -4 wslua.h Modified
+149 -3
Nope!
I was running it from a TextWindow... where the button callback data
was ep_allocated...
ep_allocated memory got reused after cf_read() and so it crashed...
On 10/2/06, LEGO [EMAIL PROTECTED] wrote:
Hi folks,
while writing an open_capture_file() function for lua I came up with
an odd
Is it ok to have the preference and register a port (once).
What can cause problems is to register a port instead of creating a
conversation, think in what would happen if it starts to use ports
used by other protocols.
On 10/3/06, Stephen Fisher [EMAIL PROTECTED] wrote:
On Wed, Oct 04, 2006 at
As I added that menu for Lua's use I thought that having an empty menu
would not be nice so I excluded it when Lua isn't there.
On 10/2/06, Martin Mathieson [EMAIL PROTECTED] wrote:
Hi,
I notice that the whole Tools menu is not available unless HAVE_LUA_5_1
is defined (it isn't for me, I
, it is a little confusing to see this empty menu. Would the
attached patch be OK (until some non-Lua item is added to the Tools menu)?
Regards,
Martin
LEGO wrote:
As I added that menu for Lua's use I thought that having an empty menu
would not be nice so I excluded it when Lua isn't there.
On 10/2
Hi folks,
while writing an open_capture_file() function for lua I came up with
an odd situation.
When executing the following code more than once, during the cf_read
call the stack gets overwritten.
if (cf_open(cfile, fname, FALSE, err) != CF_OK) {
return FALSE;
On 9/25/06, ronnie sahlberg [EMAIL PROTECTED] wrote:
Are these zero length constructs actually allowed by the standard?
If they are not it might be better to just abort dissection completely
with a [malformed packet] message.
I honestly do not know if the standard allows for that, however,
I
all that H.248 says about sequences is:
NOTE 2 – The ASN.1 specification below contains a clause defining
TerminationIDList as a
sequence of TerminationIDs. The length of this sequence SHALL be one,
except possibly when used in contextAuditResult.
Is that our case?
Luis.
On 9/25/06, Anders
Hi folks,
I have been trying to build ws on win32 for a while and I find
myself with some problems:
I build Lua 5.1.1 into a .lib, and then try link that lib into
libwireshark I get some unresolved symbol errors, specifically for:
int __mb_cur_max;
unsigned short* _pctype;
int
,http_data.request_uri)
end
wireshark.open_file(mycap.pcap)
-- end of script
On 9/11/06, LEGO [EMAIL PROTECTED] wrote:
On 9/11/06, Ulf Lamping [EMAIL PROTECTED] wrote:
Both plugins are marked experimental (at least in the Win32 installer) for
some time now.
Yes, they are both kinda
On 9/11/06, Ulf Lamping [EMAIL PROTECTED] wrote:
Both plugins are marked experimental (at least in the Win32 installer) for
some time now.
Yes, they are both kinda experimental...
The LUA scripting support seems to be the way to go, but what's the current
state (implementation,
Found!
there's libz.1.2.3 in both /usr/lib and /opt/local/lib
as soon as I removed that in /opt/local/lib I solved the issue...
diferent library same mis-bahaviour as I used to have with open-ssl
Solved by
[pociccio:~/src/trunk] lego% sudo mv /opt/local/lib/libz.1.2.3.dylib
/opt/local/lib/=libz
XML's dissector is called as an heuristic dissector for http, can't
you do like it does?
Luis
On 9/7/06, Bryant Eastham [EMAIL PROTECTED] wrote:
All-
I have done several plugin dissectors in the past for internal
protocols, all TCP- and UDP-based. I am now trying to add a plugin for a
Committed revision 19172.
I modified so that calling it like
make-sminmpec.pl -gen
is like calling it
make-sminmpec.pl -gen \
http://www.iana.org/assignments/enterprise-numbers \
sminmpec.c
Luis
On 9/7/06, Andrew Feren [EMAIL PROTECTED] wrote:
This adds some features to
On 9/7/06, Guy Harris [EMAIL PROTECTED] wrote:
Gak. Any idea what package installed its own private libz? It probably
shouldn't be doing that, unless it needs 1.2.3 or later and can't work
with earlier versions.
Darwinports does use its own dependencies for a *lot* of things even
if
On 9/7/06, Jeff Morriss [EMAIL PROTECTED] wrote:
Andreas Fink wrote:
I recently compiled wireshark under MacOS X 10.4.7 on a intel machine.
This time I succeeded even with GTK+2 after fiddling with a lot of options.
I'm preparing an installer for it for users without fink or darwin
On 7/27/06, Mike Duigou [EMAIL PROTECTED] wrote:
(repost)
The enclosed patch extends the way in which DTDs are loaded to allow DTD
definitions to be located in both the default installation directory and
in the user's .wireshark preferences folder.
The patch doesn't provide any handling for
On 7/26/06, Martin Mathieson [EMAIL PROTECTED] wrote:
Martin Mathieson wrote:
name [A-Za-z][-a-z0-9_]*[-a-zA-Z0-9_]*
Wouldn't
[A-Za-z][-a-zA-Z0-9_]*
suffice? ([...]* matches zero or more occurrences, and [-a-zA-Z0-9_] is
a superset of [a-z0-9_].)
That would have
I noticed that when I build --with-ssl wireshark cannot read any file.
If I compile without it works OK.
We had a similar report on Solaris where WS crashed when reading the
DTDs as soon as the reporter build it --without-ssl it worked.
I did not further dig into the issue.
My question is
On 7/19/06, Gerald Combs [EMAIL PROTECTED] wrote:
The Debian package approval process turned up several source files in
the Wireshark distribution that don't have explicit licenses. With the
exception of in_cksum.h, is there any reason these shouldn't be GPLed?
tap-funnel.c: no license
can you type
$ lex -V
$ flex -V
and see what comes out. I think you might be using sun's lex (for
which I never tested the code) instead of flex.
On 7/20/06, Jeff Morriss [EMAIL PROTECTED] wrote:
Hi list,
I've been trying to get a running Wireshark 0.99.2 on Solaris 9 for a
couple days now;
I did it a while ago.
On 6/29/06, Bálint Réczey (IJ/ETH) [EMAIL PROTECTED] wrote:
Hi,
Could someone apply the patch to the svn repository?
Regards,
Balint
-Original Message-
From: Bálint Réczey (IJ/ETH)
Sent: Tue 6/27/2006 18:24
To: wireshark-dev@wireshark.org
Subject: [patch]
55 matches
Mail list logo
