Non-blocking incrementation with single button pressure

Question

I am looking for a way to increment a variable by one each time a button is pressed. Current approach is the following :

 if (redButton.isPressed())
 {
 a++;
 delay(100);
 }

isPressed() just return true if the pin state is HIGH.

delay(100) is used to compensate the small amount of time the button is being pressed. Otherwise a would be incremented multiple times in a single press. But it's blocking behavior, my program can't do anything else during that delay.

In short I'm looking to add a method in my Button class to detect if the button "has just been released".

• VE7JRO: Have you seen the Bounce2 library? github.com/thomasfredericks/Bounce2

Answer 1

In your approach a button press will be detected every 100ms when you just keep the button pressed. So you have to set a flag when the button is pressed. When the button is not pressed while your flag is set, that's the event you're looking for. In this way you can detect both a button press and a release respectively. As @JRobert mentioned, there still has to be an amount of time between two calls to such a routine to provide some button debouncing.

bool pressed_flag=false;
bool button_timestamp=millis();
uint16_t button_period=100;
if(millis()-button_timestamp>button_period){
 button_timestamp=millis();
 if(redButton.isPressed() && !pressed_flag){
 // button has just been pressed
 // you can do sth. here
 pressed_flag=true;
 }
 else if(!redButton.isPressed() && pressed_flag){
 // button has just been released
 // you can do sth. here
 pressed_flag=false;
 }
}

• JRobert: Also use either a timer library, or hand-code with millis(), to come back and check the button after a time. That interval defends against contact chatter or bounce (multiple inputs for a single press). Using the flag prevents you "detecting" one long button press multiple times. And your code will still not get blocked.
• Sim Son: @JRobert theoratically my code is allready non-blocking. But I aggree: it doesn't include any debouncing and my routine should only be called in a specific period.
• JRobert: Agreed - it doesn't block. I just intended to suggest the addition of chatter removal to your already appropriate solution. (And I got caught accidentally submitting a partially-edited comment! ):
• Sim Son: @JRobert you're absolutely right, I added what you suggested!