Jeetendra Singh wrote:
It seems, I am misunderstood in the first place. I never meant to expose 
the interfaces in the form of APIs as they exist today. They are 
certainly not *well-defined* to be frozen. But I cannot see the reason 
to put the fix (if i may call them) on the back burner. We are anyways 
tied to use the APIs as they exist today and also the fact if they are 
changed, they have to be propagated to all the legs/applications which 
are currently using them (not recommended when we know for certain that 
there are problems with those APIs and they will certainly get 
modified).   So why not work on them to fix them in *near* future.
For suitable definition of "near" ("for the next release" probably 
wouldn't be suitable from my point of view; "in the next six months" 
*might* be possible), that'd be acceptable.
I currently have following requirements which am not sure how well to 
integrate in Ethereal:
1. I would like to use Ethereal to be able to read data from a 
proprietary interface which by nature is not like Ethernet for ex: lets 
say a TDM interface.<>
If by "read" you mean "capture", then
	1) Ethereal already supports interfaces that are "not like Ethernet", 
e.g. ISDN, even if it doesn't currently support live capture on all of 
them (although it *does* support live capture on ATM, for example).
	2) I'd consider that to be something to do, to some degree, in libpcap 
(somebody's already added support for Intel's SS7 cards, and there's 
also support for DAG cards).  Raw TDM bit streams wouldn't fit well with 
Ethereal at all, but something that supplies frames would work.  That'd 
also require Ethereal changes to read packets of that type, but the 
low-level capture is probably best done in libpcap, so *other* 
applications could be written to capture on that interface as well.
2 .I need to include some proprietary user information with a well 
defined protocol but not part of PDU, which could be parsed and captured.
What do you mean by "proprietary user information with a well-defined 
protocol but not part of PDU"?  In particular, what do you mean by "not 
part of PDU"?  Is it part of the network traffic in a capture, but not 
part of the particular PDU you're dissecting (e.g., some traffic prior 
to the PDU you're dissecting governs how to dissect that PDU), or is it 
some configuration information not part of the network traffic?
3. I would like to be able to get the captured data online for some real 
time correlation?
What do you mean by "some real time correlation"?  That *might* be 
something you'd do in a tap.