- fixed indenting on existing code in a few places for consistency
- introduced IR_TIMER_USE_ESP32 for ifdefs within the code as per
request
- added comments explaining what's missing for irsend support on ESP32
- IRrecvDemo.ino gets a warning before and after interrupt is enabled in
case it causes a crash
TESTED=IoTuz ESP32 board and original 328p arduino to make sure current
code did not break.
- disable a lot of defines not relevant to ESP32, set them to 1 (no-op)
- change default IR pin to 35 from 11
- changed serial speed to 115200 (9600 is too slow to keep up with IR input)
- irSend disables code that will not compile on ESP32. It won't work,
but it won't break compilation either.
parameters changed from int to unsigned int to allow longer mark/space durations and signal length.hz changed to allow for potential future use of 455kHz carrier frequency. (Ther may be existing modes to the library, using this frequency)
removed "asm" workaround for compiler, because it was not need ed on my system.
Original autor should verify this again. It could be alternatice compiler optimization settings?
Alternatively, place the volatile keyword before the variables in the function to avoid the "optimization out"
Use micros() to delay based on "real-time" instead of approximation with
delay() or delayMicroseconds()
Changed name to _usec to correspond to MicroSeconds.
_ms is MilliSeconds.
Improve output for recvDumpV2
Added my name to the contributors list (not 'cos I really care for the credit <whatever> but so people know who to "blame" [non-pejorative])
Moved the decode() function to the top of the source as it is likely to be edited the most
Utterly failed to reduce the MARK_?? functions back down to MACROs - every time I try, the decoders start failing ...However, I have found a considerable number of bugs in the toolchain, so I'm starting to wonder if the fault is not mine.
