Right now, the code will be public domain due to U.S. Government law 
since only government employees will be working on the software.  The 
problem is that we are also trying to look to the future when we open up 
the development to other industry experts.

We would like to find an elegant solution to use the Wireshark 
dissectors without having to link in the libraries, if possible.  We are 
totally redesigning some software that previously used a Tshark 
interface.  This was clunky and a very brute force method.  It started a 
Windows process, scripted a line in Tshark, collected the standard 
output, then had to perform text processing on the output to determine 
the results.  The text processing was very cumbersome and led to all 
sorts of data errors and sometimes overflow crashes. 

What I am looking for is a way to have a way to filter a capture file 
for specific packets and then pull particular pieces of data out of 
those packets.  The data that I need to pull out is not always what is 
displayed in the "single-line" packet display that Wireshark and Tshark 
display.  Most of the data we need is only displayed in the full packet 
view.  I've tried to use Wireshark/TShark to convert these files to 
PDML, but then they explode to multiple hundreds of Megabytes.  I have 
not found a good way to process these large files.

My project involves doing performance analysis on industrial Ethernet 
devices.  Right now, I am working on cyclic jitter analysis of the 
EtherNet/IP protocol (CIP and ENIP).  I am using a commercial network 
analyzer to collect the data, then I post-process the data in Tshark and 
some custom software.  I would like to eliminate the Tshark step because 
of the reasons I described above.  I would like to find a way under 
Windows to connect to Wireshark via a socket interface (or Tshark if 
absolutely necessary) that could maintain the binary nature of the data 
and allow me access to the specific data I need.

-- Jim

Joerg Mayer wrote:
> On Wed, Mar 05, 2008 at 01:11:19PM -0500, James Gilsinn wrote:
>   
>> I am working on a software package and would like to use the Wireshark 
>> packet dissectors instead of writing my own.  I am a U.S. Government 
>> employee, so any code that I write is public domain and not subject to 
>> the GPL.  I am trying to find out if there is a way to interface to 
>> Wireshark without having to compile the Wireshark libraries into my 
>> software.  I would like to avoid the licensing issues of GPL vs. public 
>> domain vs. company proprietary software that may arise as part of the 
>> project I'm involved.
>>     
>
> Actually: If your program is in the public domain, then it should be
> possible to link against the Wireshark sources. The resulting binaries
> would be covered by the GPL but the sources could remain in the PD
> (if it really is PD - I'd need to see the license of course).
>
>  ciao
>     Joerg
>   

-- 

~ James D. Gilsinn                ~               Phone: 301-975-3865 ~
~ Electronics Engineer            ~                Cell: 301-706-9985 ~
~ NIST, Manufacturing Eng. Lab    ~                 Fax: 301-990-9688 ~
~ 100 Bureau Drive, Stop 8230     ~     Email: [EMAIL PROTECTED] ~
~ Gaithersburg, MD 20899-8230 USA ~ Web: http://www.isd.mel.nist.gov/ ~


_______________________________________________
Wireshark-dev mailing list
[email protected]
http://www.wireshark.org/mailman/listinfo/wireshark-dev

Reply via email to