- division vector was not normalized after orthogonal projection onto orientation-defined plane
    - this would lead to inconsistent displacements
    - e.g. in 2D, a rand_vec of (0,0,1) would lead to 0 displacement
- the daughter cell was displaced by the full radius and the parent cell by half the radius in the opposite direction
    - this led to a center of mass shift (in the rand_vec direction) of radius/4
    - now, both are displaced by half the radius in opposite directions, preserving center of mass
    - by using 1/2 the radius, the cells will remain overlapping, which the mechanics module will resolve over time

Fix division displacement my pc

- bottom line: at each mechanics time step, it will suffice to push the cell in a cardinal direction to increase the shortest distance to the boundary
- this is precisely what the current code currently does the vast majority of the time.
- now, it will do so for the previously defined edge cases
- in 2D sims, they looked for cells on line y=x
  - then used signs of x and y positions to determine which direction to push
  - no check for cells in upper left or lower right corners
  - the sign of the positions does not accurately convey which way to push (if all corners are negative, they would all push the same way)
- in 3D sims, they looked for cells on line x=y=z
  - similar issues but missing more cases (x=\pm y = \pm z)

Fix dist to edge edge cases my pc

always check sbml_filename exists and set properly

Bump

1.14.2-drbergman-2.3.1