Re: [MiNT] XaAES / GEM memory issues

[Sven Karlsson <svenka@it.kth.se>]

Hi again,

"Konrad M. Kokoszkiewicz" wrote:
> Except the XaAES requires the same hack as N.AES, i.e. F_OS_SPECIAL flag,
> to work with memory protection. Once I realized that, you may be sure
> that I also was extremely disappointed.

The F_OS_SPECIAL flag is there IIRC only because of Craigs performance
hacks that bypassed the pipe and he wanted to be able to debug XaAES as
a single process. The first XaAES could have been splitted into a very
clean design although the implementation with pipes prevented that
performance wise. I'm sure Johan or Thomas Binder knows more about this.
IIRC they were on the XaAES list even before I joined it. And I joined
very early.

There is a lot to be said about some parts of MiNT, and I'm partly
responsible for some of them, where the design is sound but the
implemententation is lousy. MiNT isn't perfect and  you cannot really
blame someone for making the best of the situation. But that's a
personal opinion.

Regards
 Sven