The driver files known to have broken cassette implementations are: cgenie, intv, jupiter, nascom1, osi, ti990_10, trs80, tx0, kim1. Reports of broken cassette implementations are welcome. These things should hopefully be fixed before the next release. This change also made it clear that some driver say they support cassettes but they actually simply copy the data into RAM or do some other 'evil' things. In other news: The cassette code has been changed into a mame device, this may be causing some regressions. I first need to get rid of the ugly DMA hack used by the pc_hdc code. The DMA locally is triggered by timer 1 to execute the RAM refreshes. I also have quite some updates for the DMA controller locally. The EGA POST tests use some of the timers which are causing a timer irq some time later and that is causing the AT BIOS timer tests to break somehow.
The EGA is locally passing all of it's POST tests. I'm sure VGA will get sorted in time though I don't think Judge has finished EGA yet.
