Mockingboard or Phasor cards can be inserted into any slot.
Change Config->Sound to that slots 4+5 to be individually selected for the 3 soundcard types.
Add MockingboardCardManager class to manage multiple cards and mix the sound buffers.
Reason: calling SetVolume() in Play() is too early, and the previous phoneme won't have completed.
Getting the timing right will be difficult, so better to just manipulate the samples as they get written to the playback buffer.
Fix for rare race-condition, where a phoneme could complete immediately before SSI263's interrupt is enabled:
- resulting in A2 code failing to detect SSI263 or speech phrase never completing.
Only ever possible when debugger-stepping or in full-speed mode.
- Simplify common StrFormat(), especially in Debugger (changes upcoming)
- Add helpers StrAppendByteAsHex() and StrAppendWordAsHex()
- Add helpers StrBufferAppendByteAsHex() and StrBufferAppendWordAsHex() for plain string buffer
And these Debugger related:
. Improve FormatAddress() and GetSymbol().
. GetSymbol(), FindSymbolFromAddress(), FormatAddress() are changed to use std::string instead.
. Remove static variable (not nice) in FormatAddress().
. GetSymbol() returns std::string reference instead of pointer.
New class for SSI263 (2 instances per MB/Phasor card)
- support 2nd SSI263 at $Cn20 connected to 6522-A (at $Cn00)
Switch from one-shot phoneme playback to ring-buffer
Use a new dedicated ring-buffer (distinct from the AY8910 ring buffer)
Gets rid of: SSI263Thread, the 2 events, and the 64 voices
Support the 2-bit Duration b7:6 of phoneme register for both SC01 & SSI263 (to fix#929)
Updated save-state
Updated IRQ logging (output cycle & 6522 source)
