Improve TextureProgressBar.set_radial_initial_angle() by removing loops

[arkology]

Replace two while loops with if and fposmodp.
Document radial_initial_angle wrapping.
Infinite numbers will cause error and will not change radial_initial_angle value.
Add testcase for set_radial_initial_angle() argument wrapping and skipping non-finite values.

Should fix #60338.

[AThousandShips]

See above, does not solve the issue linked (only non-finite inputs)

[arkology]

Thanks for review, looks like I opened it too early.

does not solve

Issue was Godot freeze, so technically it solves issue.
I'll add what you suggest.

[AThousandShips]

It normally does but it'll cause errors, it also doesn't behave correctly with negative input

[arkology]

Yes, for now it doesn't work well with negative input and I will fix it.

[AThousandShips]

A check with !Math::is_finite and then checking with:

if (p_angle < -360.0 || p_angle > 360.0) {
	p_angle = Math::fmod(p_angle, 360);
}
if (p_angle < 0.0) {
	p_angle += 360.0;
}

Should be sufficient

You could also simply use:

if (p_angle < 0.0 || p_angle > 360.0) {
	p_angle = Math::fposmod(p_angle, 360);
}

And I think some of the checks in that will be optimized out

Concerns resolved

[AThousandShips]

Will test this code when I am able!

[arkology]

@tool
extends EditorScript

func _run() -> void:
	var temp_variable452 = TextureProgressBar.new()
	
	var values : Array[float] = [360*1e4 + 10, - (360*1e4 + 10), INF, -INF]
	
	for val : float in values:
		temp_variable452.set_radial_initial_angle(val)
		print(val)
		print(temp_variable452.radial_initial_angle)
		print("---")

Output:

  Angle is non-finite.
  Angle is non-finite.
3600010.0
10.0
---
-3600010.0
350.0
---

This matches with how it worked before (except for infinite numbers).
Changing value from editor also works as before.

[arkology]

Replaced fposmod with fposmodp, should work the same.

[Mickeon]

I suppose it's not strictly required but adding a test about this would be nice.

[arkology]

I suppose it's not strictly required but adding a test about this would be nice.

Sure! I didn't touched Godot unit tests before but it will be definitely interesting to try.
I suppose this should be done in separate PR, right?

[Mickeon]

I suppose, yeah. The PR as is looks like an objective improvement.

[arkology]

@Mickeon test added - #98930

[akien-mga]

Sorry for the conflicting directives, but I think that adding the test in this PR would be best, so it can validate the change from this PR being correct.

[arkology]

Good, I'll do it and close separate PR with test.

[Mickeon]

I apologise for my own mistake!

[arkology]

@Mickeon don't worry, it didn't take much time to merge changes into single branch :)
Several minutes to verify that everything works as expected and I'll push changes.

[AThousandShips]

Would be good to add test cases for non-finiteness as well

[arkology]

Would be good to add test cases for non-finiteness as well

Done.
UPD: Now for real :) Not only for INF, but also for NaN.

[arkology]

I applied suggestions and rebased branch.

[Repiteo]

Thanks!