I'm tired of Gumball

[ShynnSup]

Another case:

When selecting a line with Gumball set to Object mode it works fine.

But when adding points to the selection, the you loose the line direction.

[theoutside (Kyle Houchens)]

double click the rotation handle then use osnap and snappy dragging to align it however you want.

[Gijs (Gijs de Zwart)]

I can see the logic for that specific case, but I don’t think it is going to be very feasible to keep analyzing the users’ selection to see if the objects could be treated as a single entity or not.
If @Joshua_Kennedy thinks differently, I’d be happy to be proven wrong though.

[Eugen]

Thing is, when multiple objects are selected, the gumball is re-calculated to the bbox of those objects, obviously.
What else should it do? Take the gumball alignment of the first selected object, maybe… that could work. How about it? Could be optional.

What’s also bothersome, as mentioned, is that an extrusion, since it is a new object, gets it’s gumball re-calculated to something new (What exactly? Looks like the midpoint of the edge extruded from the curve’s first point).
I would prefer if the new extrusion object just copied the gumball from the curve!

To generalize: every time a new object is created from input objects (sweeps, booleans, …), it’s gumball is not ‘estimated’, but simply taken from the (first?) input object.

Does that make sense?

[Joshua_Kennedy]

@Eugen

To generalize: every time a new object is created from input objects (sweeps, booleans, …), it’s gumball is not ‘estimated’, but simply taken from the (first?) input object.

Does that make sense?

I think that would make sense in about half the time and drive you crazy in other half. For a short time the gumball used to do that since the saved orientations were copied over with the input objects attributes. In the case of exploding or joining curves you’d get get some really funky results. I think we could do a lot better job giving you a default orientation for multiple selections and I have an open issue about it.

[eobet]

What I’m most tired of the Gumball about is that it has no memory. In fact, I’d argue that it has an acute case of alzheimers. If I set the Gumball on an object in a certain way, and then use it once, it resets back to whatever was before.

It’s like Rhino is actively fighting me.

[Eugen]

@eobet

If I set the Gumball on an object in a certain way, and then use it once, it resets back to whatever was before.

Even when in 'align to object" mode? It shouldn’t do that. Doesn’t here.

[Charles]

@eobet

the Gumball about is that it has no memory

Yes.
Also see here:

Gumball forgets orientation on control points

When I edit a control point using the gumball, I want to orient the gumball into a direction of my choice. No problem so far. But when I select another control point, and select the former control point then, the orientation is forgotten. Not nice.

And when it comes to groups, the gumball orientation is not stored as well.

[eobet]

@Eugen

Even when in 'align to object" mode?

You forgot the key words “I set”. In other words, custom orientation.

@Charles

Also see here:

And here and here.

[Eugen]

Strange. Works here.

The ‘align to object’ mode shows whatever the gumball was set to, manually. Right?

[eobet]

@Eugen

Strange. Works here.

Did you try

@eobet

use it once

[Eugen]

You might have ‘AutoReset Gumball’ on.

[Joshua_Kennedy]

@eobet

If I set the Gumball on an object in a certain way, and then use it once, it resets back to whatever was before.

Does it reset as soon as you are done dragging it? You might have GumballAutoReset enabled

@Charles

And when it comes to groups, the gumball orientation is not stored as well.

Fixed in the Rhino WIP

[ShynnSup]

@Gijs

I don’t think it is going to be very feasible to keep analyzing the users’ selection to see if the objects could be treated as a single entity or not.

@Eugen

To generalize: every time a new object is created from input objects (sweeps, booleans, …), it’s gumball is not ‘estimated’, but simply taken from the (first?) input object.

@Joshua_Kennedy

I think that would make sense in about half the time and drive you crazy in other half.

Guys, it is not so complicated. The Gumball just needs to calculate the average Gumball when multiple objects are selected and the Gumball is set to OBJECT.

In the specific case I mentioned, points have no direction, thus the direction of the Lines should be maintained.

When two parallel lines are selected, the Gumball might appear in the middle given the average point, but the direction will be identical to just selecting one line.

[Charles]

@Joshua_Kennedy

Fixed in the Rhino WIP

When you say “fixed”, I understand it’s an error in V8.
So this error will be fixed in V8?
Or is the fix an addition to V9?

[Joshua_Kennedy]

@Charles

Or is the fix an addition to V9?

This one

[eobet]

@Eugen

You might have ‘AutoReset Gumball’ on.

@Joshua_Kennedy

You might have GumballAutoReset enabled

Literally

@eobet

And here

(Although, I must admit that now that I’m trying it, at the moment I can’t reproduce it either.)

EDIT: One tangential issue I found is that custom gumball orientation for sub-objects reset when you perform an Undo.

EDIT: Another tangential issue is that if you change the gumball align from a custom one to CPlane it doesn’t actually just change the alignment, it also resets its position.

EDIT: Removing control points from a control point only selection with a custom gumball orientation also resets its position.

EDIT: Oh, just executing the Drag Strength menu option also resets the position!

EDIT: Heck, just performing Unisolate resets the custom gumball when control points are selected.

Yeah, very tired!

[eobet]

One more.

First, I set the custom orientation I want, then I choose the reLOCATE option and only choose a LOCATION and what happens? The alignment resets.

(EDIT: Yes, ok so a nice popup told me that I could double click the origin of the gumball to move it and retain the custom orientation, but still. Also, I’d rather this method work because double clicking is the path to RSI.)

[taraskydon]

I have been trying for over a year to understand Rhino Gumball in order to achieve my needs. the video clearly shows the problem in defining and calculating the plan of orientation of Gumball in the middle of the object. I developed С# my own Gumble for my own purposes. and I’m also developing a method of orientation calculating the plan. I think that for such simple types of objects, you can select the largest and planar surface and get torientation data from it to define the vectors Gumball in the center of the object.

[taraskydon]