[Paraview] Re: Problems with State File

[Kent Eschenberg]

A few additions to my list of problems with state files:

1) The reason the file gets larger is that more and more duplicates of the 
"trackball manipulators" are added.

5) The origin of the slice plane is not restored. The values are correct in the 
pvsm file but after loading they are at their default values.

One of the most promising features of ParaView was the way it could be used to 
work with very large datasets:

    - on your workstation interactively create a visualization
      on a small version of the dataset;

    - save state; then

    - use the state file in batch with the full dataset.

The unreliability of the state files makes this impossible. Would others use 
this feature? If so fixing state files might be important.

Kent
Pittsburgh Supercomputing Center


Kent Eschenberg wrote:
> Using 3.2.1 on Linux it is impossible to exactly restore the display by 
> recalling the state file. This is particularly important at facilities 
> using pvbatch to render very large data sets, and using the state file 
> to define the image.
> 
> 1) The file gets larger at every load/save cycle. This doesn't look right.
> 
> 2) The camera zoom is never restored. It seems to "zoom out" as if going 
> to some default location.
> 
> 3) In the XDMF reader I turn off 80% of the variables to reduce memory 
> demands. These are always turned on again after loading the state.
> 
> 4) A simple box is included (its useful for positioning the camera) but 
> turned off before the state is saved. Upon reloading the state its 
> always on.