Open RattlesnakeLodge opened 1 year ago
Undo behavior of control activities is something that I never really have designed consciously. It just happened or not happened.
I think that different use cases ask for different undo behavior. So maybe that should be configurable. I'll try to solve this in a target-agnostic way one day. For now, I think that writing no undo point is at least in the spirit of ReaLearn as a performance tool.
I agree with wanting to at least be able to set undo points!
I totally get that the initial purpose for this tool might have been a different one than mapping controllers to volume faders for mixing. But I think this tool is so good and essential, that it can be used in a ton of different ways!
I think it should be an option in the target section as the way how to implement the undo (without too much resource usage) is target-dependent. And undo also doesn't make sense for all targets. And for some mappings the user might wish to enable undo, for others not.
One challenge is finding the right delay after which an undo point is created. If it would be done after each target value change, there would be way too many undo points.
Greetings!
Looks like the issue from #787 is fixed, but maybe a little TOO fixed. ;-)
On the most recent ReaLearn release (2.14.2) adjusting volume (source is fader, target is Track: Set Volume) does not generate any undo points. Same is confirmed with Track: Set Pan. I have not looked into other parameters yet.
It is much better to have ZERO undo points than an extreme amount per adjustment, as is the case since issue #787. However, one undo point per adjustment would be ideal.
Thanks in advance!