Re: [MiNT] This fight

[Petr Stehlik <joy@sophics.cz>]

On Thu, 2003-12-11 at 14:53, Frank Naumann wrote:

> Btw. can you explain for me what do you understand under the "FreeMiNT's
> proposed NF syscall" and what these syscall should do?

the syscall would be used by callers instead of pointers obtained from a
cookie. If you can recall the example code for NatFeat shutdown I posted
here on Tue 23:26 then the syscall would replace the NFID and NFCall
macros.

Something along the following lines:

#define SYS_NF  0xABCD       /* to be defined by FreeMiNT */
#define NFID    0
#define NFCALL  1

long shutdown_id = ioctl(SYS_NF, NFID, "NF_SHUTDOWN");
if (shutdown_id) {
    ioctl(SYS_NF, NFCALL, shutdown_id);
}

Does it make sense now?

> > Another option is to write TSR for TOS that would provide emulation of
> > the MiNT syscall in TOS.
> 
> For what? If someone need MiNT he can simply boot it (especially on aranym
> I thought).

MiNTlib is used for easy porting of applications. Such applications can
run also on TOS and MagiC. That's why MiNTlib contains TOS
emulation/compatibility mode (at least it did contain it in pl46).

> > There were proposals by other guys to use NF for multimedia processing,
> > for example (JPEG as the simplest example).
> 
> This is an interesting idea. But why do you want to hardcode and bound
> such nice things into emulator specific code?

> Why not enhance the VDI/AES? The VDI/AES as part of the operating system
> can then optimally handle it (with help of special hardware, or with help
> of the emulator if running on aranym).
> 
> In this way it help all users. Users with normal ATARI hardware, users
> with fast ATARI clones, users with special hardware, users with emulators.

You must have missed about 70% of my mails. Please consult MiNT list
archives. Thanks. Just a small hint: this is exactly what we want to
achieve.

Petr