* Reduced PROGMEM / DRAM usage slightly (Using defaults w/HW5: Before -> 214,668 bytes / 5,757 bytes; After -> 213,414 bytes / 5,751 bytes)
* Optimized some menu options and calls
* Added more string constants and implemented them where found.
* Fixed some stylization
* Constants should always be uppercase.
* Features should use the `ENABLE_<feature name>` format.
* Options for features should use the `OPTION_<feature name>_<option name>` format.
* Added ENUMs for more clarity and better type checking.
* Moved some defines over to `constexpr` and `const` types. These are preferred over preprocessor constants when not intended for use with `#if` and other preprocessor language.
The `setVoltage()` function should be called even when `ENABLE_VSELECT` is disabled because `ENABLE_3V3FIX` also uses it. There is no resource cost to do this as when both options are disabled the compiler will optimize this function out. This just "future proofs" the code so if that function ever does more it doesn't need updated everywhere. This applies to `setup_FlashVoltage()` as well.
The changes to OSCR.cpp are just for code formatting and additional comments to clarify this.
Lots of changes/additions.
Added:
* Firmware Updater support: Supports the Firmware Updater app (release to follow soon). Enabled by default, can be disabled in the config.
* 3.3V Fix (3V3FIX): Enable if you have stability issues when using 3.3V, works best with VSELECT. Disabled by default, can be enabled in the config.
* `DynamicClockSerial`: Class that extends and modifies HardwareSerial to be compatible with a dynamically changing clock speed. Used through the `ClockedSerial` object/variable.
* `OSCR.cpp` & `OSCR.h`: New files for storing globals. Only contains these new additions for now. More code cleanup to come.
Changed:
* Moved configuration flags to `Config.h` and documented them better.
* Removed `vselect()` function. Now uses `setVoltage()` with the params `VOLTS_SET_3V3` and `VOLTS_SET_5V`.
Known Issues:
* Rarely the LCD backlight turns white when using 3V3FIX. Resetting fixes it. Doesn't affect functionality/usability; it's just weird.
This code was removed as part of
a555f21 ("All: Resolve all compiler and linker warnings")
because it was filling a 1kB buffer which was never read from.
While I expect reading values from PINC should be invisible from the
cartridge's perspective, writing to PORTL, PORTF and PORTK is definitely
visible.
I have currently no way of testing if this code is required, so in doubt
restore the parts which have effects visible from the cartridge.
I am not sure why I changed some variables from 16 bit types to 8 bit types
when the warnings being resolved were only about signedness.
Fixes: a555f21 ("All: Resolve all compiler and linker warnings")
Also, make print_FatalError(*) independent from print_Error(*).
Also, factorise common parts of print_Error(*).
Frees 2336 bytes of program space and 5 bytes of global ram space.
