Improve fallback behavior for unset device config values

[pdamianik]

Describe your PR, what does it fix/add?

This should improve the fallback process for device specific, unset config values. It replaces the touchpad bool parameter for getDevice... functions with an optional fallback configuration key that describes the global configuration key to fall back to if the device specific key is unset. The previous implementation didn't quite work for me, as it would reset natural_scroll to disabled even when just enabling/disabling my touchpad. This should also handle conflicts between the fallback configs input:touchdevice:transform and input:tablet:transform and input:touchdevice:output and input:tablet:output better as the current implementation has a hardcoded order which prefers touchdevice:* keys over tablet:* keys, even for tablet devices.

Alternative solution

An alternative solution would be to use the same key structure as the normal keys for device config keys. This could improve intuitiveness for users (the keys aren't just the last section and potentially conflicting) and it would make the extra fallback argument unnecessary.

Is there anything you want to mention? (unchecked code, possible bugs, found problems, breaking compatibility, etc.)

CConfigManager::getConfigValueSafeDevice is overloaded and spreads to CConfigManager::getDevice..., so I originally tried to use std::optional (I have very little C++ experience). Is there a better way to describe an optional string reference?

Is it ready for merging, or does it need work?

If it looks good and the alternative solution is not something to be implemented, it's ready 😃 .

[vaxerski]

oh, and conflict

[pdamianik]

Fixed, what do you think about using the same key structure for device configs as the global config?

[vaxerski]

not sure what you mean by that

[pdamianik]

Nevermind, pr is ready from my side.

[vaxerski]

I don't think this will compile xD

[pdamianik]

It did (hold on) on my machine :), but we'll see

[vaxerski]

also #3189 got merged in the meantime so if you could rebase against master and fix that up

[vaxerski]

told you it wont compile :P

header:
const std::string& param = ""
source:
const std::string& param

you can't have = "" in the source

[pdamianik]

Yeah, that's what happens when you do your first C++ coding.

[pdamianik]

Fixed, compiles, starts (tested for the latest changes this time), fixes my touchpad issue, rebased and ready for merging.

[vaxerski]

some typos.

As a next step (for the future) since we have fallbacks, we could drop the HASCONFIG and move it to CConfigManager::getConfigValueSafeDevice

[pdamianik]

some typos.

As a next step (for the future) since we have fallbacks, we could drop the HASCONFIG and move it to CConfigManager::getConfigValueSafeDevice

I'll make another branch ready.

[vaxerski]

h

[vaxerski]

ya