fix i2c scanning eror for Arduino

[martgras]

revert to the old scan implementation.

What does this implement/fix?

getting correct i2c scan results

Quick description and explanation of changes

Types of changes

• Bugfix (non-breaking change which fixes an issue)
• New feature (non-breaking change which adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Other

Related issue or feature (if applicable): fixes esphome/issues#2453

Pull request in esphome-docs with documentation (if applicable): esphome/esphome-docs#

Test Environment

• ESP32
• ESP8266

Example entry for config.yaml:

substitutions:
  updates: 60s
  unique_id: bme280

esphome:
  name: ${unique_id}

esp32:
  board: TinyPICO
  framework:
    type: arduino

logger:
  level: VERBOSE
i2c:
  sda: 21
  scl: 22
  scan: True
  id: bus_a

sensor:
  - platform: bme280
    temperature:
      name: "Temperatur (Huette)"
      id: hut_temperature
    humidity:
      name: "Luftfeuchtigkeit (Huette)"
      id: hut_humidity
    update_interval: ${updates}
    address: 0x76

Checklist:

• The code change is tested and works locally.
• Tests have been added to verify that the new code works (under tests/ folder).

[probot-esphome]

Hey there @esphome/core, mind taking a look at this pull request as it has been labeled with an integration (i2c) you are listed as a code owner for? Thanks!
_{^{(message by CodeOwnersMention)}}

[mmakaay]

Why revert the code? I think the roll forward option might be to use:

auto err = writev(address, nullptr, 0);

That should work too IMO, and would be closer to the intended target code.

[OttoWinter]

ERROR_OK is of type i2c::ErrorCode; but endTransmission returns type int - it may just happen that the i2c error codes match up with what endTransmission returns, but we can't assume that in general.

[martgras]

updated the PR to use writev.
Is there a reason why the scan ends at address 119 and not 127?

[OttoWinter]

iirc those addresses are reserved for special uses and so can't have i2d devices connected

looks like this explains it: https://electronics.stackexchange.com/a/7553

[martgras]

ok changed the range to 8 - 119 to exclude the 2 reserved ranges

[martgras]

actually for esp-idf readv works as well but switching to writev for consistency.

[martgras]

@mmakaay is right auto err = writev(address, nullptr, 0); works as well for me and we don't have to use the underlying wire_ object here.

[OttoWinter]

@mmakaay is right auto err = writev(address, nullptr, 0); works as well for me and we don't have to use the underlying wire_ object here.

The same applies to the esp_idf implementation, could you also update the _esp_idf.cpp file? Thanks!

[martgras]

BTW: The CI check doesn't seem to pull the correct files https://github.com/esphome/esphome/actions/runs/1261652210 shows files that are not part of this commit

[oxan]

BTW: The CI check doesn't seem to pull the correct files https://github.com/esphome/esphome/actions/runs/1261652210 shows files that are not part of this commit

The CI runs on a simulated merge between dev and your PR, but you unfortunately just hit the moment where I broke CI on dev. I've scheduled a rerun, it should be fine now.

[OttoWinter]

Thanks!