SofaROM bug report thread

Latest observations here: the turboR mapper just seems unreliable in Z80 mode. Even with MEGAMU, some bytes are corrupted in RAM, but not at the same locations as with SofaROM (MEGAMU is switching the CPU to Z80 DRAM mode before launching the game, when SofaROM does not change it).

The corrupted addresses are always at a #?C00 location. In the case of #EC00, that's causing a direct GFX corruption that can be seen in MoG or USAS, for other locations, that's corrupting the game data, problems are less likely to be observed.

I might just force R800 mode in SofaROM if that's confirmed, like in TRLOAD.COM. The Z80 mode in turboR mapper has little interest anyway, it's quite slow and you'd better use the Memory Mapper device... Unless we manage to find a ROM mode using some undocumented registers...

Currently thinking about another approach for switching pages: first setting the 9th bits register to 0, so that the next write is always done on a ROM area, hoping that this write in 7FF8 is not conflictual... do not believe too much in it but have to try!

The PMs on MRC do not work too well these days (have not received anything yet)...

Try a direct email to: lhalter gorzlub free blotz fr (gorzlub->@, blotz->.). You're never too cautious with those damned email fishers...).

So there was indeed a regression in the Game Master 2 support. You can try the fix in the same ZIP file as above: sofarun31new.zip

a mapper page is false GDX have you the add circuit for decode D5 in your Turbo-R