Re: [hatari-devel] EmuTOS machine support information

["Roger Burrows" <anodyne@xxxxxxxxxxxx>]

On 17 May 2020 at 11:23, Vincent Rivière wrote:

> On 16/05/2020 at 23:06, Roger Burrows wrote:
> > I noticed that Hatari cross-checks the version of Atari TOS against the
> machine
> > type.  It checks EmuTOS as well, but more crudely since EmuTOS version
> numbers
> > (and the version of Atari TOS announced by EmuTOS) are not 100% useful..
>
> The TOS version specified in EmuTOS ROM header is meaningless. I hope that
> Hatari doesn't use it for anything. If it actually does, that's wrong. Any
> Hatari check regarding to the TOS version must be disabled in case of
> EmuTOS.
>
Well, it's slightly more complicated than that (which is one reason why I made
the suggestion).  First, it checks for EmuTOS & 512K ROMs and if so, doesn't do
any more checking.  But otherwise, it does use the fake Atari TOS version from
the header for its checks.

> > EmuTOS now has an extended header which we can put useful information
> into.
> > Would it be helpful to Hatari if I added flag bits to indicate which
> hardware
> > this ROM can run on; for example, runs on ST hardware / runs on STe hardware
> /
> > etc ?  If so, now would be a good time for me to add this.
>
> Good idea.
> So if EmuTOS is detected but the extended header is not present, the above
> rules still applies.
> But if the EmuTOS extended header is present, its information will override
> the above hardcoded rules.
>
> The new EmuTOS flags could be:
> 1) Supported Atari machines: ST, STe, TT, Falcon, ARAnyM
> 2) Supported CPU: 68000, 68010, 68020, 68030, 68040, 68060
>
Yes, I hadn't thought about a CPU flag, but that's a good idea.  That's 11
flags, but we have 31 spare at the moment, so no problem.

>
> So Hatari could reliably issue warnings like:
> "Warning: This EmuTOS version is not designed to run on TT. Do you want to
> try anyway?"
>
> That would be great, and relatively easy to implement.
>
I would think so.

Roger