Potential crash on I2C communication failure due to ESP_ERROR_CHECK

[devprofile98]

Hi, Thanks for the great library.

I am using i2cdev and esp-idf for an ESP32 project to connect to MPU6050, everything is ok in debug build, but in production, I think the ESP_ERROR_CHECK is misused, for example, the i2c_master_cmd_begin function can fail because the bus is busy (ESP_ERR_TIMEOUT), In debug environments this result in calling aborrt(), which could be ok in many cases, but in production the ESP_ERROR_CHECK macro became empty, and the writeBytes function will return true. This behavior is not suitable for a robust production application, considering I2C is a sensitive protocol.

Isn't it better to leverage a more appropriate error handling using if/else statements or just by returning the esp_err_t from the called function? What do you think?

[jrowberg]

This is a good observation, but I feel like I don't have much to offer directly because I've never used the ESP IDF development approach--only the Arduino platform implementation. Someone else contributed the IDF variant of the I2Cdevlib core. Could you fork, modify, and submit a tested pull request that improves things to add the functionality you recommend?

[devprofile98]

Of course, I will try to give the function correct return types and replace ESP_ERROR_CHECKs with robust error handling. Thank you.