MantisBT - ParaView
View Issue Details
0008964ParaView(No Category)public2009-05-05 12:592009-06-03 20:35
David Doria 
Utkarsh Ayachit 
normalminoralways
closedfixed 
 
3.6 
0008964: Bad default glyph size
If I open a small model (bounding box edge length about .1) and add vector glyphs of the normals, the "scale factor" changes to 397794.090337098 and hence the arrows are ENORMOUS and it makes the model look like a single point. If I check "Edit", change the scale factor to .1, and hide then show the model (to reset the zoom), it looks correct.
No tags attached.
related to 0005659closed Dan Lipsa Glyph filter scale factor starts wrong 
? Scene.vtp (34,727) 2009-05-05 12:59
https://www.vtk.org/Bug/file/7207/Scene.vtp
Issue History
2009-05-05 12:59David DoriaNew Issue
2009-05-05 12:59David DoriaFile Added: Scene.vtp
2009-05-29 11:08Utkarsh AyachitNote Added: 0016591
2009-05-29 11:08Utkarsh AyachitStatusbacklog => @80@
2009-05-29 11:08Utkarsh AyachitFixed in Version => 3.6
2009-05-29 11:08Utkarsh AyachitResolutionopen => fixed
2009-05-29 11:08Utkarsh AyachitAssigned To => Utkarsh Ayachit
2009-06-03 20:18Alan ScottNote Added: 0016624
2009-06-03 20:18Alan ScottStatus@80@ => closed
2009-06-03 20:18Utkarsh AyachitNote Added: 0016625
2009-06-03 20:33Alan ScottNote Added: 0016626
2009-06-03 20:33Alan ScottStatusclosed => tabled
2009-06-03 20:34Alan ScottRelationship addedrelated to 0005659
2009-06-03 20:35Alan ScottNote Added: 0016627
2009-06-03 20:35Alan ScottStatustabled => closed
2011-06-16 13:10Zack GalbreathCategory => (No Category)

Notes
(0016591)
Utkarsh Ayachit   
2009-05-29 11:08   
Why in the world were we taking the different of
magnitudes when computing scale factor? We just needed to consider the maximum
value. This was resulting in humungous scale factors when differences between
vector values were small. Fixed that.

/cvsroot/ParaView3/ParaView3/Qt/Components/pqGlyphPanel.cxx,v <-- Qt/Components/pqGlyphPanel.cxx
new revision: 1.9; previous revision: 1.8

* Pending commit to 3.6 branch
(0016624)
Alan Scott   
2009-06-03 20:18   
I think I am going to fix this, since it is MUCH better with this fix. However, we have still not fixed the underlying problem (using time 0 for ranges, as opposed to taking something better like the last time).

This fix will only work if time step 0 has reasonable numbers in it such as velocity. It won't work if time step 0 has initial conditions, such as displacement.

This bigger problem is documented in bug number 5659.

Tested local client, using can.exo.
(0016625)
Utkarsh Ayachit   
2009-06-03 20:18   
Committed to 3.6 on May 29.
(0016626)
Alan Scott   
2009-06-03 20:33   
Trying to form relationships. This is a test.
(0016627)
Alan Scott   
2009-06-03 20:35   
Getting a relationship to work. Closing again.