analog pin paramter for analogRead not zero based

[jnsbyr]

Basic Infos

Hardware

Hardware: ESP8266 - all boards
Core Version: 2.1.0

Description

The analog pin parameter for analogRead() is currently not zero based but absolute (A0 = 17) in contrast to the standard Arduino library implementation (see https://www.arduino.cc/en/Reference/AnalogRead). This makes ports of existing projects (e.g. https://github.com/firmata/arduino) harder.

Possible Solution

• change in all variants/.../pins_arduino.h:

...
static const uint8_t A0 = 0;
...

• change in core_esp8266_wiring_analog.c:

...
extern int __analogRead(uint8_t pin) {
  if(pin == 0){
    return system_adc_read();
  }
  return digitalRead(pin) * 1023;
}
...

[igrr]

It's not zero based for at least two variants of the AVR core:
https://github.com/arduino/Arduino/blob/master/hardware/arduino/avr/variants/standard/pins_arduino.h#L47-L54
https://github.com/arduino/Arduino/blob/master/hardware/arduino/avr/variants/leonardo/pins_arduino.h#L114-L125

Edit:
the actual implementation of analogRead allows passing in both A0 (i.e. number of pin) and 0 (number of the analog channel). I think we can adjust analogRead implementation to handle both values, like AVR analogRead does. I don't think that changing A0 back to 0 is a good idea though, as that would imply that analog input is on pin 0, which isn't the case.

[jnsbyr]

Thx for considering the suggestion. Keeping A0 at 17 will not be a problem if analogRead works with pin=0.

[jnsbyr]

An analysis has shown that the wifio board has some extras that need to be considered, especially A1 to A7 in addition to A0. Best I can think of would be to do it like we agreed for A0 and support access both absolute and zero based. There are no overlaps between absolute pin numbers and zero based pin numbers even with wifio so this should be possible. This issue blocks the integration of the ESP8266 into firmata/arduino (see firmata/arduino#278 (comment)).

[igrr]

@jnsbyr this is now fixed in git version.

[jnsbyr]

Thanks, will let you know how it works.

[jnsbyr]

With the new version analogRead() now works for A0 as expected.

There are similar problems with the macros digitalPinToInterrupt(p) and digitalPinHasPWM(p) from pins_arduino.h. E.g. the second macro returns true for pins 1 to 17 with an Adafruit board. But it should be 0 to 16 or even better 0 to 5 and 12 to 16, because you should not do PWM on your flash IOs. The reason is that the current macro uses NOT_A_PIN but that resolves to -1 and that is true for 17 and it returns the pin number and that is false for 0.

My suggestions:

• minimum change

#define digitalPinToInterrupt(p)    (((p) < EXTERNAL_NUM_INTERRUPTS)?1:0)
#define digitalPinHasPWM(p)         (((p) < NUM_DIGITAL_PINS)?1:0)

• saveguard approach

#define digitalPinToInterrupt(p)    (((p) < EXTERNAL_NUM_INTERRUPTS && ((p) < 6 || (p) > 11))?1:0)
#define digitalPinHasPWM(p)         (((p) < NUM_DIGITAL_PINS && ((p) < 6 || (p) > 11))?1:0)

EDIT:
Problem with macros raised as issue #1831

[jnsbyr]

@igrr
From my point of view this issue is resolved and can be closed.

[igrr]

@jnsbyr Thanks! It will get closed (automatically) once we deploy the new release (2.2.0).