Yes. I'd prefer to see the core library split further into a few smaller libraries, defined by functionality they expose:
kaffe/corejar : jar and zipfile handling
kaffe/verifier : the verifier should be easy to split off, and contains a few files in its own.
and presumably others along the same lines. We can get there by looking at what core files are shared between kaffeh and kaffevm and trying to separate them into functional libraries. I'd welcome patches in that direction. I believe that Guilhem was doing some work to separate the different engines more clearly, for example.
I have still some #define around in one of my local trees for kaffe but I may check it in because it should be stable (basically I had to rename and a few symbols). There will be a slight performance loss for JIT though: currently the JIT is putting the pointers to translated methods directly in the vtable. To be compatible with intrp and other way of handling methods we must put a pointer to Method. This has the consequence to have to dereference it and get the appropriate field (always the same) in the structure. So basically it is around 3 instructions more by function call. Once all defines has been removed I think we should try to build a common engine interface where the core is totally ignorant of how is built the engine and reciprocally. This will be a great step towards modularity. :)
Cheers,
Guilhem.
_______________________________________________ kaffe mailing list [EMAIL PROTECTED] http://kaffe.org/cgi-bin/mailman/listinfo/kaffe
