Star tool right-angle visualization should be drawn parallel to star corner edges

[0HyperCube]

When using the star mode of the polygon tool, you can edit the star by dragging the corner whilst still in the star tool. This causes some very strange functionality as seen below:

Apparently this is meant to be a right angle. However it is not.

[Keavon]

We can draw it more accurately like this, where the annotated blue lines here are parallel to the sides so the square mark is skewed in the same manner as the whole star shape:

[0HyperCube]

It would be better if it actually snapped to being a right angle.

[Annonnymmousss]

I guess it is resolved.

[0HyperCube]

A standard in geometry is to use a small square in the corner between two lines to show a right angle (90° or π/2 rad). Other angles are typically given an arc.

This program will mistakenly put the small square for angles that are not 90°:

still_not_right_angle.mp4

Therefore I would not conclude that this issue is resolved.

[devviktoria]

Actually I was able to reproduce what was depicted in the video. What I had noticed that when a star is created, then in the node graph there is a star node that is followed by a transform node. The problem is that the angle indicator ignores the transform node: it indicates the angle as if the transformation has uniform scale and 0 skew angles.

IMHO this is a problem because the users are expecting to indicate the actual visible angle not the non transformed angle.

If its OK I would like to take a look at this.

[Keavon]

The expected behavior is that the gizmos are defined to work how they work for an untransformed star. Whatever occurs to its transform downstream is not a concern to the gizmo, other than projecting its overlays from viewport space back into the original local transform space of the star object. I just checked and confirmed that it is doing that in the current build, so I think we are good with the behavior working as expected now.