opacity binding

[19jrushlow]

The binding works, but the value displayed in the cheat sheet does not update until some other binding has been called by the user. I suspect the issue lies in docDblOpt, but I cannot figure it out and would like someone to review the problem.

[snoyer]

@19jrushlow

The binding works, but the value displayed in the cheat sheet does not update until some other binding has been called by the user. I suspect the issue lies in docDblOpt, but I cannot figure it out and would like someone to review the problem.

I don't think it's your code, the cheat sheet is only updated if it has been invalidated, and it doesn't look like it is (yet) when changing the opacity, see

f3d/vtkext/private/module/vtkF3DRenderer.cxx

Lines 1154 to 1162 in dc202ae

	void vtkF3DRenderer::SetLightIntensity(const double intensityFactor)
	{
	if (this->LightIntensity != intensityFactor)
	{
	this->LightIntensity = intensityFactor;
	this->LightIntensitiesConfigured = false;
	this->CheatSheetConfigured = false;
	}
	}

vs

f3d/vtkext/private/module/vtkF3DRenderer.cxx

Lines 1678 to 1685 in dc202ae

	void vtkF3DRenderer::SetOpacity(const std::optional<double>& opacity)
	{
	if (this->Opacity != opacity)
	{
	this->Opacity = opacity;
	this->ActorsPropertiesConfigured = false;
	}
	}

---

@mwestphal this // clang-format off looks a bit early

f3d/library/src/interactor_impl.cxx

Line 890 in dc202ae

// clang-format off

should it go just before all the this->addBindings instead?

also should the bindings added by this PR go on the P key instead of T to match the existing "Toggle Translucency" binding? And should increasing/decreasing opacity turn translucency support on if it was off?

[snoyer]

ternary operator as a replacement for if else without assigning the value looks confusing, would be clearer as:

Suggested change

	val.has_value() ? (valStream << val.value()) : (valStream << 1.0);
	valStream << (val.has_value() ? val.value() : 1.0);

and then std::optional already has a shortcut for this pattern

Suggested change

	val.has_value() ? (valStream << val.value()) : (valStream << 1.0);
	valStream << val.value_or(1.0);

[19jrushlow]

fixed, though I just realized I can do the same in IncreaseOpacity(). I will address it in the next commit.

[19jrushlow]

Actually, I am also now realizing that it may not be best to default to 1 automatically, in case docDblOpt needs to be used for a different binding in the future. Should I pass the default value as a parameter to the lambda instead?

[mwestphal]

it doesn't look like it is (yet) when changing the opacity, see

@snoyer is right, please add this->CheatSheetConfigured = false; in the method he pointed.

[mwestphal]

should clang-format go just before all the this->addBindings instead?

Agreed

[mwestphal]

should the bindings added by this PR go on the P key instead of T to match the existing "Toggle Translucency" binding?

Good point

And should increasing/decreasing opacity turn translucency support on if it was off?

I think that make sense but I would like @Meakk input on this

[19jrushlow]

The bug appears to be fixed after doing the cheat sheet invalidation thing, and I have switched the binding to P. It seems like all that should be left is:

1. Add unit tests
2. Possibly turn on translucency support when calling the binding, pending input from Meakk

Am I missing anything? Here is the original issue

[mwestphal]

Hum, I dont think thats correct, it should default to "N/A" to be coherent with docTglOpt imo.

[19jrushlow]

Is that not misleading? The user may think that the opacity binding is incompatible with their model if they see that, and you said it can be forced to 1 if it is not set. In fact, I think using "N/A" for docTglOpt might also be a bit misleading, since when I initially saw the values displayed as "N/A" I assumed that I would be unable to change those values even though that is not the case.

[mwestphal]

"1" is incorrect, but maybe N/A is not the right choice. The idea to convey is that it is not defined. I dont have a better idea than N/A.

In any case, lets be coherent at least :)

[19jrushlow]

Maybe one of these: [*], [_], [—], [?] could be used instead? Or if you are okay with being a bit more verbose, [Unset] might be a bit more descriptive.

[mwestphal]

Lets discuss that on discord