Very wrong results on a zoom.

[davidharvey1986]

Describe the bug
When running Velociraptor on a SWIFT Zoom-In hydro simulation I get extremely strange results with halos not being found and positions not being correct.

I compile it with
cmake .. -DVR_ZOOM_SIM=ON -DVR_USE_GAS=ON -DVR_USE_STAR=ON -DVR_USE_BH=ON
I run it with sample_zoomhydrocosmological_run.cfg (attached)
stf -I 2 -I snap/snapshot_0008 -o velo/veloshot_0008 -C sample_zoomhydrocosmological_run.cfg

Then I run on my SWIFT hdf5 zoom.

I expect the halos and their properties to be calculated. The attached image shows the stellar particle positions in blue with their positions all reduced by a factor of 0.8 (this was a fudge). The yellow stars are the positions of halos as found by VR. We see that VR positions are almost exactly 80% of the true positions and that it doesnt find some of the largest halos.

sample_zoomhydrocosmological_run.cfg.txt

This problems was reproduced on Linux and OSX.