Having read the paper, it seemed very intuetive to add new constraints to the framework. Looking at the avaiable code its very limited to the set parameters and solver.
I found that you could add callback functions in the constraint loop. However the predicted positions aren’t present in the FlexSolverCallbackParams. The “original” predicted positions could be calculated with the velocity, but not the ones achieved from the stabilization loop if those aren’t updated to the velocities instead.
Morever I can’t see how you could fit my new constraint in the established constraint groups by this approach, which will be very limiting.
Is it possible to get the source code of the flex lib? Or do anyone have the solution to my problem?
The idea is to integrate cohesion, thermodynamics and compression to granular materials to simulate snow.
//
Christian Markowicz (“Jawsarn”)