Implement a learning mode about rc-switch HOT 3 OPEN

Sounds cool. From my experience it was tricky to detect unknown signal patterns as there is a lot of noise and hard to tell where a signal starts/ends plus arduino has pretty less memory to analyse long samples.

from rc-switch.

FWIW the example ReceiveDemo_simple is what i used to debug/clone one of my transmitters.
The only piece missing was the pulse length. I had to add
Serial.print("Delay: "); Serial.println( mySwitch.getReceivedDelay() );
to the end of the current debug output (line 31) in order to get that but then I was able to successfully duplicate/clone the incoming rf codes with the mySwitch.send function
#include <RCSwitch.h>
RCSwitch mySwitch = RCSwitch();
void setup() {
// Transmitter is connected to Arduino Pin #10
mySwitch.enableTransmit(10);
// Optional set pulse length.
mySwitch.setPulseLength(192);
}
void loop() {
mySwitch.send(<received int value>, 24);
delay(1000)
}

from rc-switch.

I'll voice my support for the first half of this feature. I had to modify rc-switch to do the first part (printing a recorded protocol's timings) while adding a protocol so having this available by default would have been useful. If there's interest I could post a portion of that code here, though it's probably pretty obvious how that would be accomplished.

As for the play back of a learned code, here's a guy who implemented that with a RF ceiling fan remote: http://arduinobasics.blogspot.com.au/2014/07/433-mhz-rf-module-with-arduino-tutorial_30.html . There's some weird stuff going on like hooking the receiver up to an analog input (I have the same receiver and it's digital...) but the idea is to skip the statistical analysis, just store the raw timing values. I'm not saying that's necessarily the way to go, but here's someone who's done something similar to the second part of what your suggesting.

from rc-switch.