[hatari-devel] Hatari WinUAE vs oldUAE information, and MiNT with MMU?

[Eero Tamminen <oak@xxxxxxxxxxxxxx>]

Hi,

I've removed obsolete WinUAE vs. oldUAE info from
the docs (there was quite a lot of it).  While
doing that, I noticed few things that might need
checking...


MiNT had been properly tested with Hatari v1.x
(oldUAE) when Hatari still didn't have properly
working MMU, so compatibility doc tells to disable
memory protection.

If somebody has a full MiNT setup on a disk image,
some testing with Hatari v2.3.1 MMU emulation +
MiNT memory protection, and reporting the results
would be appreciated!


Nicolas:

* There are still 2 demos that work only with
  oldAUE core.

  Fit's "Hex Pistols" and "Stercus Accidit" Falcon
  demos bomb at startup regardless of what Hatari
  WinUAE CPU core options are used, although they
  work fine with oldUAE CPU core, both with
  "--compatible" option enabled and disabled,
  and with TOS4 and EmuTOS.

  Could you check that it's not some long term
  regression, but e.g. due to cycle-accuracy
  reasons?


* these code comments seem out of date:
--------------------------------------
$ git grep -i -e olduae -e "old uae"
....
src/cycles.c:                           AddCycles -= 0; 
       /* NOTE [NP] : this is used by old uae cpu core but does not 
happen */
src/cycles.c: 
       /* FIXME : fix old uae cpu to remove long accesses to memory for 
68000 ? */
src/includes/configuration.h:  bool bAddressSpace24;           /* Always 
set to true with old UAE cpu */
src/includes/m68000.h:/* Values for IACK sequence when running in 
normal/prefetch mode or when using old UAE CPU */
src/spec512.c:/*                        not possible on a 68000 STF, 
this is a bug in old UAE CPU core.         */
--------------------------------------

Could you check / update them where appropriate?


	- Eero