"Insert point" unavailable

[oprcmr00]

In the attached screen-cast, I am trying to split a face with an edge by point insertion. It can be split by insert edge, but I want to split it with point insertion. It doesn't work, except in diagonal orientation.

→→ I copy paste the body into a whole new document, and point insertion works again as it should.

Note:

• It doesn't work if I save, close and reopen
• It doesn't work if I export / the entire file
• It doesn't work if I export the component to which the body belongs, or any parent component
• It doesn't work if I copy-paste the component to which the body belongs, or any parent component
• It only works if I copy-paste the body itself.

By being forced to do so I lose the structure of the document.

I have seen other (similar) glitches with Pull, where instead of pulling the geometry on the closest object, it does it on a seemingly random point of an unrelated object far away, even if target mode is set to manual.

Pasting the bodies (t-spline model and mesh target) into a whole new document solves the problem.

Are they known issues / have they been documented before?

Thank you.

[oprcmr00]

Update: the glitch goes away by copy-pasting the body within the same document.

This fully addresses the effects of the problem as it is possible to preserve document structure.

I would delete the thread at this point but I can't.

[glenn-chun]

@oprcmr00 Looks like the glitch was caused by how the model was transformed in different components.

• Root component: Main Model (copy)
• Main Model (source): Main Model (source)
• Experiments 1: Main Model (copy), Body1
  
  
  

[TrippyLighting]

---

@glenn-chun wrote:

@oprcmr00 Looks like the glitch was caused by how the model was transformed in different components.

•  

---

Maybe in this case, but I've seen this often enough in other models. I will record it and share the model if I come across it again.  

[glenn-chun]

@oprcmr00 Thanks for reporting this issue. We logged this as a defect (FUS-155253).

As a workaround, if you remove the transformations made for Experiments 1 component (i.e., move the component back to the original position. See screenshot), then you will see Insert Point operation working again.

BTW, you moved Experiments 1 component by the following distances from the original position.

X: 541.306 mm
Y: -65.409 mm
Z: -15.971 mm

 

[oprcmr00]

Hi, it seems that the bodies are fine, only the component is glitched.

Interestingly any new form that is created within the glitched component will inherit defectiveness.

I tried to move it from child O to root O and by your numbers but it didn't fix it. I suspect that while moving it you created a copy outside the problem component and tested insert point on that copy...

Pull is also not working:

[glenn-chun]

@oprcmr00 Since you originally moved Experiments 1 by (541.306 mm, -65.409 mm, -15.971 mm), you need to move it by (-541.306 mm, 65.409 mm, 15.971 mm) in order to move  back to the original position.

[oprcmr00]

It works... yes I inputed the wrong numbers...

How is it possible that by just translating a child component from the O of the parent component, it breaks?

[glenn-chun · Accepted solution]

---

@oprcmr00 wrote:

It works... yes I inputed the wrong numbers...

How is it possible that by just translating a child component from the O of the parent component, it breaks?

---

It's a defect on the Fusion side (FUS-155253). Here's an example of what happened behind the scene. When the midpoint of an edge is selected in the Insert Point command and the edge interval ranges from 0 and 1, Fusion code is supposed to pass 0.5 to T-spline library. The defect is that Fusion sometimes doesn't take into account the component transform in geometry calculations. Due to the defect, the parameter for the edge midpoint is computed as a number, say 12, outside of the [0, 1] range, and then 12 is clamped to 1. Fusion passes the selected edge and its parameter 1 (instead of 0.5) to T-spline library that does a no-op because a vertex already exists at parameter 1.

Thank you again for reporting the issue.
Glenn