Wireshark-dev: Re: [Wireshark-dev] Wireshark PIDL generated dissectors
From: Joerg Mayer <[email protected]>
Date: Tue, 8 Oct 2013 01:34:00 +0200
On Mon, Oct 07, 2013 at 03:38:34PM -0700, Guy Harris wrote:
> > I think that wireshark has been used a de-facto fork for PIDL the tool
> > since many many years. But a fork that is occasionally synced back
> > with upstream.
> > That is the only way we can make sure that we will always be able to
> > even compile the IDL in wireshark to a working dissector.
> 
> If you mean "the tool as a whole", not just "the back end", then
> 
> 	1) there needs to be a spec for PIDL to which both PIDL processors would conform, and it should be sufficiently complete that any of Samba's PIDL files can be processed by a (non-buggy) conforming processor ("non-buggy" meaning that making it actually work is the responsibility of the maintainers of the processor);
> 
> 	2) the maintainers of the spec, i.e. the Samba people, should inform us when the spec changes;
> 
> 	3) the maintainers of the Wireshark PIDL processor, i.e. the Wireshark team, are responsible for changing that processor to follow the spec.

IMO, the Samba team has done a rather good job maintaining the pidl source
(including the wireshark backend) - we just did a not make very good use
of it :-( I also think that the maintainership of the pidl sources including
the Wireshark backend is in good hands where is currently is. We "only" need
to get back to synching the sources and to some lesser extent the idl files.
What we should do is:
Determine whether our pidl (tool) sources contain any changes not upstreamed.
Create patches against current samba pidl, get it working again and get them
upstreamed.
Replace our pidl tree with samba's tree (+idl_types.h).
Regularly copy over the changes from the samba tree.

Ciao
   Jörg
-- 
Joerg Mayer                                           <[email protected]>
We are stuck with technology when what we really want is just stuff that
works. Some say that should read Microsoft instead of technology.