Recenter View questions

[biagas]

Is anyone familiar with the use-case for Recenter View of 3d plots?
I am working to fix a few issues I found when navigation mode is Flythrough.

My question is this: is there any reason to make any changes to the view during a Recenter if the limits haven't changed?
If the passed in limits match the current bounding box, does anything need to be done?

Also, I have not been able generate a use-case of manually calling Recenter View where the passed in limits don't match the bounding box.

edit:
During my testing, for the most part with Trackball mode, RecenterView doesn't change any of the View parameters unless axes3DScaleflag is on.

[markcmiller86]

Is anyone familiar with the use-case for Recenter View of 3d plots?

What about in the case of time-evolving extents? Maybe at late time, extents have changed enough that recentering the view is a useful operation? But, that kinda speaks to your other question (am assuming by limits you mean spatial extents).

[biagas]

Do you know offhand if any of VisIt's test data has time-evolving extents? I'd like to take a look at what happens with Recenter view in this case.

[markcmiller86]

I don't. Makes me think we need a set of descriptors for all the different test datasets we have that might manifest a particular feature of interest. Right now, about all we really have in the way of descriptors is database format.

I think some of the Silo data gen tools have a variety of command-line args on them to cause them to produce the same data in a slightly different way. You might wanna check sources in src/tools/data/datagen for any argc, argv blocks that offer what you are looking for. If you don't find any, it would be easy to add to say wave.C by adding an offset or scale to the mesh coords on every output.

[biagas]

@markcmiller86 I tried m_plot.mili (has extents that change in time). In ViewerWindow::RecenterView3d, if the 'limits' passed in as argument do not match the boundingBox, then I see view parameters change. Seems like we could skip some unnecessary calculations: If limits == boundingBox and axis scaling NOT applied. Thanks for the pointer towards time-evolving extents.

[markcmiller86]

@biagas did you have any closing thoughts about this?