''moving' default SDA pin for hardware I2C(1)

[robhamerling]

Currently I'm using: v1.26.0-preview.148.g49f81d504 with my ESP32-SPIRAM. With this version the default SDA-pin for I2C(1) is apparently pin 27, which was also the case with some other preview versions. However in other preview versions it was pin 26, which confirms with the documentation.
Strange that the default for the SDA pin seems to be 'moving' between preview versions.

[robert-hh]

That is indeed strange, because in the code the default is set to 26, and there is not setting in the codebase for SDA to be at Pin 27. Code snippet from machine_i2c.c. Maybe there is a different setting on the machine which creates the preview files.

#ifndef MICROPY_HW_I2C1_SCL
#if CONFIG_IDF_TARGET_ESP32
#define MICROPY_HW_I2C1_SCL (GPIO_NUM_25)
#define MICROPY_HW_I2C1_SDA (GPIO_NUM_26)
#else
#define MICROPY_HW_I2C1_SCL (GPIO_NUM_9)
#define MICROPY_HW_I2C1_SDA (GPIO_NUM_8)
#endif
#endif

[robhamerling]

Just tried to find out which of some previous preview versions have which default SDA pin for I2C(1). To my surprise all defaults - even of version 148 as in my opening message - report pin 26! That is the good news.
Leaves a riddle why I saw pin 27.....

[robhamerling]

Found out: when explicitly specifying a pin for I2C (like sda=27) in one script it becomes the default for next scripts. And it survives soft resets. Maybe that was the cause of my experiences/observations.

[ned-pcs]

I would recommend not depending on the default pin assignments but instead specifying explicit sda and scl pins.