It was always changing CR1 (starting at CR bit 4) instead of the CR selected by crfD.
Also, it was clearing all but the FL,FG,FE,FU bits of FPRF of FPSCR.
Adds a "Grab Mouse" command to the "Window" menu which toggles
SDL_SetRelativeMouseMode. That "traps" the mouse to the current window
(and hides it) which makes it less annoying to deal with the separate
acceleration curves of the host and guest OS.
This needs to be done via AppKit code since SDL does not have a
cross-platform way to add menu commands.
I also looked into the extended ADB mouse protocol but it does not
appear to do that we want. If we use it to report a device class of
classAbsolute (0) the acceleration curve changes to a flat response
(we end up using 'accl' 0 from 9dd3c4bef8/base/SuperMarioProj.1994-02-09/Resources/MiscROMRsrcs.r (L64-L76))
but the coordinates are still interpreted as being deltas. If I use a
device type 4 (Absolute pointing devices) then the mouse position is
never queried. Per https://68kmla.org/bb/index.php?threads/anyone-have-an-adb-graphics-tablet-or-joystick.39128/
all tablet devices end up using custom drivers.
The `SubOpcode31Grabber[1024] = { ppc_illegalop }` initializer only
populates the first entry with ppc_illegalop (at least on some compilers),
switch to explicitly initializing the entire array with std::fill_n.
Also fix a couple of sign and overflow issues flagged by the Xcode
undefined behavior sanitizer.
Cherrypicks a small piece of joevt/dingusppc@117ca1e449
so that booting from the 10.2 CD gets past it trying to change the video
mode to 15bpp.
Co-authored-by: joevt <joevt@shaw.ca>
It's used in the main emulator loop (ppc_exec_inner), and the function
call overhead adds up.
By inlining it, time to boot to the Finder using a 7.1.2 install CD
and a 6100 ROM goes from ~6700ms to ~6400ms (with clang 14 on a
M2 Max)
