Re: GEM and memory protection

[Julian Reschke-FJR010 <Julian_Reschke-FJR010@email.mot.com>]

q-funk@megacom.net%INTERNET wrote:
> 
>  >> I know, that it is not Gemini's fault, but why not making it better? BTW
>  >> I have not hered of any Desktop that does that though I have written that in
>  >> a Newsgroup a few months ago. Maybe I've missed something...
>  >
>  >Well, that's certainly a good idea.
>  >
>  >BTW; it would also be a good idea when -- on a protection violation -
>  >the kernel would also display the process name of the owner of the
>  >memory block that the process whas trying to access (if available).
> 
> Iv've always felt AV servers should be rewritten to make use of /pipe
> or /shm so that data can be passed without violating protection. This
> would not change the protocol itself, only how the server handles it.

If you haven't done it properly in the first place, somebody has to
change the program flags - a no-brainer.

Why invent yet another protocol?

If you want to use a protocol which does *not* use pointer passing,
use the MultiTOS D&D protocol. That's why it was invented.

Regards, jr