Multiple selections don't work correctly in parallel
This issue was created automatically from an original Mantis Issue. Further discussion may take place here.
Multiple selections don't work correctly in parallel. I suspect that this is the tip of an iceberg. I have seen single point selections also be weird - but this finally allows us to replicate it.
Utkarsh took a few tries, but was able to replicate this.
- 8 remote pvservers, Linux, 5.0.1. (I don't know if it replicates the same with fewer pvservers.)
- Open g1s1. Ask Utkarsh or me for this dataset.
- Run ProcessIdScalars, Apply. Not necessary, but helps me describe where to select.
- Go to last timestep. Back one timestep.
- Select points on.
- Not necessary, but you may want surface with edges.
- Now, you want to rubber band select the lower tip of the medium blue material (white in vtkBlockColors). You want to rubber band select about 1/2" by 1/2", including a bit of the orange, the beige, and the medium blue. (You may need to cycle through this a few times.)
<> Notice that many points are selected that are inside the dataset. I think they are on processor boundaries, but cannot tell. This is a bug.