|
Changes:
- Changed output audio sample format back to unsigned again (seems to work
better on some systems).
- Adjusted audio functions so that Hatari's sound should be right again.
- Fixed TOS patching routine - it was only working on big endian machines
(Thanks to Tony Smolar for the hint!).
- IKBD_Cmd_ReadClock() should now work.
- Some parts of the code accessed the SR directly to read the IPL - however
the UAE CPU core only updates the SR when doing a MakeSR() first. So this
is done in the affected code parts now, too.
- The IPL wasn't raised when a MFP interrupt occured - fixed now.
- Full screen resolution for ST-Low can now be selected from the screen
setup dialog.
- The IKBD emulation does not longer duplicate joystick fire buttons when
a game (like Babarian) tries to use both, joystick and mouse => Games like
Lotus Turbo Esprit Challange are now working again!
- Improved audio timer function - the code should now be a little bit faster.
- Resynced Hatari's UAE CPU core with UAE 0.8.22 - this fixes some bugs in
68k instructions like ABCD and SBCD.
- Now only reinitializing the sound subsystem if really needed, this should
fix a problem with Linux ALSA sound driver (Thanks to Tony Smolar for
discovering this problem!).
- Added another patch for TOS 2.05 so that this TOS version should now work
with Hatari, too.
- Rewrote TOS patching routine. It is more flexible now.
- Removed 0xa0ff opcode for VDI resolutions; using GEMDOS_OPCODE now instead.
- Fixed MMU RAM size configuration bug.
- Rewrote some screen conversion functions in C (low320x8.c, low640x8.c,
med640x8.c and spec640x16.c).
- When a bus or address error occurred, the PC was often not set to the
right exception handler routine. This has been fixed now - thanks to
Philippe Gerin for finding the bug and the patch for solving it!
|