[P026][Sysvars] Add Internal temperature sensor value for ESP32

[tonhuisman]

Resolves #4817

Features:

• P026; Add Internal temperature (ESP32) as a value for ESP32 processors
• System variables: Add %inttemp% (Internal temperature) for ESP32
• Show correct branch name in Actions run builds
• [Build] Stop Documentation-only commits on existing branches from triggering an Actions run

TODO:

• Testing
  • ESP32 Classic (Currently works using IDF 4.4.5, but rumors are that it might be dropped from support)
  • ESP32 S2
  • ESP32 C3
  • ESP32 S3
• Update documentation
• Test: Documentation commit. Result: Not working as expected, (but still 'by design' 🤔 as described here

[TD-er]

• System variables: Add %internaltemperature% (Internal temperature) for ESP32

Couldn't you come up with an even longer name? I think this one can even be used on nearly all display plugins in their line strings. ;)

[tonhuisman]

Couldn't you come up with an even longer name?

Shortened to %inttemp%

[iz8mbw]

Testing %inttemp% (HEAD_1f9e4a7) on ESP32, Temperature values alternates between "sense" values (likes 65 degrees) to -273,15 (no sense value):

[TD-er]

I think the sensor needs some 'time' to stabilize as there is a check to see if the value returned is 128 and if so it will return the error value.

[TD-er]

By the way, it looks like the temperature sensor on the ESP32 (classic) is just (barely) capable of giving proper readings.
So you should only use it to track trends.

I think it just tracks the offset in frequency of the internal 150 kHz oscillator (its accuracy is highly temperature dependent) compared to the crystal frequency.

The ESP32-S2/S3/C3/C2/C6 seem to have a more reliable temperature sensor, or at least have a factory calibration.

I think that's why Espressif removed the "temperature sensor" from the features summary of the ESP32 classic.

[TD-er]

I just upload this PR into the 'latest' web flasher: https://td-er.nl/ESPEasy/latest/

[iz8mbw]

I think that's why Espressif removed the "temperature sensor" from the features summary of the ESP32 classic.

So if on the classic ESP32 we cannot obtain acceptable values (-273.15 is unacceptable) I suggest to remove this feature for the ESP32 board.

[TD-er]

Like I said, I think the temperature sensor may need some more polling as it will only return this extreme value when the sensor returns this specific value (128 if I'm not mistaken)

Still the temperature reading on ESP32 classic should not be used as a true temperature value. Only as a trend.

[iz8mbw]

Just tested the last commit of this PR (via Action run https://github.com/letscontrolit/ESPEasy/actions/runs/6305109576) and I can confirm now all is OK and I don't have more any -273.15 value.
Thanks!