VisualARQ API wall missing from automated viewport captures in Rhino 9

VisualARQ
1

Hello, I was testing some integration between VisualArq and the new Rhino 9 AiPanel and I was unable to use it,

I have this file with a house badly designed and I wanted to replace everything with VA definitions but the RhinoAI was unable to do it, so I asked for it to file a report explaining the details of what happened. Since I am late for work, I didn’t try myself without the RhinoAI intervention.

Find his report below:

This report was written by ChatGPT/Codex, the AI assistant that performed the tests through the RhinoAI integration. The user is sharing my findings; these are not observations independently verified by the user.

I repeated a minimal test and reproduced the same symptom: a wall created through the VisualARQ API exists in the document and reports valid dimensions and volume, but does not appear in the viewport image returned by the integration.

Test environment

  • Rhino: 9.0.26300.9401
  • VisualARQ: 3.12.2.14590
  • Runtime: .NET 10.0.12
  • Windows build: 26200
  • Document units: millimeters
  • Absolute tolerance: 0.01
  • Viewport: Perspective, Shaded
  • Execution: Python through RhinoAI, using VisualARQ.Script and the document handle injected as __rhino_doc__

Steps performed

  1. Created a new wall style with a 120 mm layer and a height of 2,800 mm.
  2. Created a wall using AddWall, from (20000, 0, 0) to (23000, 0, 0).
  3. Explicitly set its height to 2,800 mm and alignment to Center.
  4. Created a regular Rhino Brep box with identical dimensions, offset by 1,000 mm along Y as a visual control.
  5. Captured the viewport with both objects inside the frame.
  6. Selected the VisualARQ wall and ran _vaUpdate _Enter.
  7. Queried the wall again, hid the control box, and captured the same viewport again.

Expected result

Both the VisualARQ wall and the regular Rhino box should be visible.

Observed result

Only the regular Rhino box appeared in the capture. After hiding that box, neither test object was visible, although the VisualARQ wall remained unhidden on a visible layer.

The wall returned:

IsWall: True
Found in document: True
Height: 2800.0 mm
Thickness: 120.0 mm
Volume: 1007999999.9999986 mm³ ≈ 1.008 m³
Hidden: False
Layer visible: True

Bounding box:
Min: (20000, -60, 0)
Max: (23000, 60, 2800)

InstanceDefinition.Name: *EmptyDefinition
InstanceDefinition.GetObjects(): 0 objects

After vaUpdate, the wall still returned IsWall: True, the same volume, and an empty instance definition. It remained absent from the capture.

Scope and limitations

I reproduced this through the automation integration in the existing project document. I have not independently verified the user’s live Rhino display, execution directly in Rhino’s Python editor, or behavior in a fresh document.

Consequently, I cannot yet attribute this to VisualARQ itself. An empty instance definition alone may not establish a fault if the plugin uses a custom display mechanism.

Questions for the developers

  • Is *EmptyDefinition with zero definition objects expected for a VisualARQ wall created through the API?
  • Is an additional initialization or regeneration step required in this execution context?
  • Could automated viewport capture omit VisualARQ’s custom drawing even when the live viewport displays it correctly?
  • What diagnostic would best distinguish an API creation issue from an integration or capture issue?

All temporary test objects, the test style, and the test layer were removed afterward. I verified that the document’s pre-existing object IDs were preserved.