Re: New file system IDs and others.

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

kellis@primenet.com%INTERNET wrote:
> 
> Well I have been reading a lot lately here. And I don't quite agree
> with many people comments, such as: "Like in magic."
> Like if it EVEN correctly worked.
> Shared lib? like magic? why would MiNT need the magic one?

I don't think anybody said that MiNT *needs* the MagiC shared lib
scheme.

> compatibility reason? We haveto stay clear, and know what we want, do
> we want magic compatibility with their own way of doing? or more
> likely to be closest to unix shared lib? Are the unix shared lib even
> thinkable on atari? if yes, why did magic adopt their own? What is
> stoping us to use more standard way of doing?

Nothing. Do it. Go ahead.

> Handle Background DMA like in magic? "magic does it already blah blah
> with Hddriver blah" , Does it even work in the so perfect Magic OS?
> OS of the excellence. I have read and did try myslef. That ain't
> working well at all. Many / most of the people have this option
> disabled as it is way too slow.

That might be the case. However it's definitively not the fault
of the interface MagiC is using for this.

>...
> 
> Why is pexec reacting differently from magic to mint? what is the

Does it? Some modes are missing, but besides that?

> purpose? won't that create compatibility problem? YES it will, it has.
> 
> I mean come on, shouldn't we try to direct MiNT kernal toward
> something a little bit more robust? following posix threads for
> example, EVEN if it has to be NOT compatible with magic. I don't care
> about magic.

MagiC's threads rely on the MagiC AES, so it's obviously not a good
choice for MiNT. However I don't get why compatibilty between MiNT
and MagiC isn't useful. I sure want to run my code on both operating
systems.

> Ok, Made MiNTsetter to support Dcntl() instead of Sssytem, 1) coz it
> is better, more logical etc, 2) it will be magic compatible. I wonder
> why It even needs to be magic compatible. Magic application are way to
> tight up with magic kernal that just 1 call won't make it compatible
> anyway.
> Example? run vfatconf.prg of magic under newer mint kernal, It
> looks/writes whatever into the magx.inf file. Or will do some special
> Magic AES call.

So what? I am happy that I can go ahead and write a vfatconf ttp 
that will run on both systems.

> Magic can be good? ok prove it to me. coz I have 5.1x and it is really
> not convincing at all. Please help me. This is driving me nuts. Why
> Magic if it doesn't bring anything good? am I missing something
> obvious?

Seems so.