[Jakub Zawadzki <darkjames-ws@xxxxxxxxxxxx>]

On Thu, Nov 21, 2013 at 08:46:06AM -0500, Evan Huus wrote:
> I don't see any specific reason to keep the proto_ variables private,
> so I have no objection to making them available. All of the protocols
> involved in this case already have separate header files anyways.

I'd rather use proto_get_id_by_filter_name() in some init function, and later compare.

Making variables extern is somehow tempting, like we could extern dissector handles and stop using find_dissector() ;]
But I think it's wrong way.

Kuba.