PR #2995 didn't take into account the data model storage actually does change when you change a variable name.
Boxer yaml generation was missed in earlier PR.
* chore: Run `generate-yaml.sh`
* chore(cpn): carryTrim -> trimSource
* fix(build): Boxer missing from `generate-yaml`
* dodgy attempt to read old carryTrim
* fix: Don't break the byte boundary + reorder
* fix(yaml): indentation for attributes that should not be output
* chore: Regenerate YAML after rebase
---------
Co-authored-by: raphaelcoeffic <raphael.coeffic@frafos.com>
These invocations will all cause bugs when pwd returns something that
can be interpreted by bash, the most common case being a folder
with whitespaces in its name. In other words these can be triggered by
calling the scripts from any folder that contains whitespaces.
More info on this specific finding by shellcheck:
https://github.com/koalaman/shellcheck/wiki/SC2046
Added special UART4 init to sample if FlySky sticks are present or not.
Added boolean globalData.flyskygimbals to save the detection result.
Removed FLYSKY_HALL_STICKS and FLYSKY_HALL_STICKS_REVERSE CMake build options.
Tested on two RM TX16S radios, one with FlySky digital hall sticks and second with stock analog sticks.
