@kareltucek Side note: I always felt like the GUI would just create some kind of macro commands anyway. But yes, I would have compiled macros on the Agent side and submitted binary execution trees (p-code …) to the firmware side. Different design choices.
Well, if I was there at the time, I would have done a number of things differently…
but specifically parsing the command text on the firmware side has turned out to be very practical. Like this, new features can be added very easily, with no changes to the Agent. Compiling stuff Agent side would make the process of adding new features significantly more laborous.