* Better memory stats (CLI)
* Bitmap buffer overflow check added
* Lua bitmap improvements
* Proper handling of symbolic links in simulator
* S6R Lua script fixes:
* added progress screen while loading bitmaps
* memory for bitmaps is now freed when leaving the script
* S6R Lua script: improved bitmap loading
* * Lua Themes and Widgets moved to separate file and separate Lua state
* Stand-alone, mixer and function scripts now behave the same as on Taranis (restarted after the stand-alone script was run)
* both Lua states are independent: Themes and Widgets is initialized only at the start, the other one is initialized before and after the stand-alone script is run
* Better leak test and report in BitmapBuffer
* Re #3318: Lua compiler enabled in simu and Companion simulator. Usage:
* any time <xxx>.lua file is about to be loaded and if file <xxx>.lua.src exists:
* load contents of <xxx>.lua.src
* compile Lua code
* save compiled code into <xxx>.lua (effectively overwrites existing file)
* immediately following the compilation the real file loading is done from <xxx>.lua (which by now contains compiled Lua bytecode)
When crossfireTelemetryPush() or sportTelemetryPush() is called without arguments, it returns a boolean telling whether or not the output buffer is currently available.
This simplifies the LUA in case SPORT or Crossfire is used as a transport for bigger requests (like MSP/SPORT).
* Add constants for LUA telemetry
This commit added 896 bytes to flash size which is a significant size. An alternative would to just document the values in the lua telemetry function.
Shortening some of the UNIT names (e.g. UNIT_MW instead of UNIT_milliwatts) or shortening UNIT_ to U_ could save a few bytes but the
* #if 0 ... #endif constants
* horus has 2MB flash
* Fix CUSTOM_COLOR, also change to way setColor work to a more natural : lcd.setColor(CUSTOM_COLOR, RED)
* Remove _INDEX from TBS main to match the recent change to setColor
I did a refactoring for the latitude and longitude decoding which should be tested carefully for FrSky S.PORT and D telemetry
There was a double conversion in S.PORT protocol, there are less maths now, and perhaps precision will be better
