Hi Stephan,
I had to poke around for hours to come to that solution. In my case there are variants where leaving the points where they are is actually the correct answer (if all strain components happen to be zero, which I didn't exclude in the randomization).
The purpose of the workaround is actually to generate specific feedback even if nothing has been done except pressing "check".
I'd expect some method like touch() to exist, which you can call in order to trigger an update. It is interesting that the offset must be big enough to actually shift the point to a different grid snap position. Otherwise the engine is smart enough to check that actually nothing has changed.
Martin
I had to poke around for hours to come to that solution. In my case there are variants where leaving the points where they are is actually the correct answer (if all strain components happen to be zero, which I didn't exclude in the randomization).
The purpose of the workaround is actually to generate specific feedback even if nothing has been done except pressing "check".
I'd expect some method like touch() to exist, which you can call in order to trigger an update. It is interesting that the offset must be big enough to actually shift the point to a different grid snap position. Otherwise the engine is smart enough to check that actually nothing has changed.
Martin