this emphasizes the transposition issue.

initial condition the same, but trajectories themselves are different.

both NavierStokes and NSVorticityEquation fail with segfaults, but the
test code is minimally sane

Merge branch 'feature/new-particles' of gitlab.mpcdf.mpg.de:clalescu/bfps into feature/new-particles

Fill field data with zero after allocation

See merge request !7

Debug when some processes may have 0 particles to write to file (output hdf5 file)

See merge request !6

Add showHtml method to scope timer

See merge request !5

I think this does the job. I don't get any errors from overwriting, so
things seem ok.

Debug for a tiny number of particles -- interval builder was incorrect when number of intervals greater than number of items

New fluid solver NSVorticityEquation should be stable from now on.
Old fluid solver NavierStokes should work as before.