DrawCylinder Color.FromRGBA

[ptcsoderlund]

Issue description:
Seems like bytes are in order of ARGB and not RGBA as function indicates.
Hence, Color.FromRGBA should be named Color.FromARGB. Or its just when drawing cylinder?
I dont know.

Steps to reproduce:

public override void OnDebugDraw()
{
    base.OnDebugDraw();
    const float radius = 100f;
    const float height = 200f;
    //This gives a semi transparent red color, indicating its actually ARGB.
    DebugDraw.DrawCylinder(Actor.Position, Quaternion.Identity,radius,height, Color.FromRGBA(0x55FF0000u));
}

Flax version:
1.8.1

[cNori]

letts see

internal code:

R = ((rgb >> 16) & 0xff) / 255.0f;
G = ((rgb >> 8) & 0xff)/ 255.0f;
B = ((rgb >> 0) & 0xff)/ 255.0f;
A = ((rgb >> 24) & 0xff)/ 255.0f;

where rgb is 0x55FF0000u with is in bits (01010101111111110000000000000000)
set do it manualy step by step
bit shift

R = 00000000000000000101010111111111
G = 00000000010101011111111100000000
B = 01010101111111110000000000000000
A = 00000000000000000000000001010101

mask the bit shift with 0xff

R = 00000000000000000000000000000001
G = 00000000000000000000000000000001
B = 00000000000000000000000000000001
A = 00000000000000000000000000000001

the 00000000000000000000000000000001 is just 1 so

R = 1 / 255.0f
G = 1 / 255.0f
B = 1 / 255.0f
A = 1 / 255.0f

output color

R = 0,003921568627451
G = 0,003921568627451
B = 0,003921568627451
A = 0,003921568627451

Conclusion funcions is broken

data masked out in the process

So, Lets fix it

Assuming the function is supposed to convert decode bytes inside the uint32

Input

rgba = 0x55FF0000u

Chanel	Bits
R	01010101
G	11111111
B	00000000
A	00000000

version where top bits are R chanel

uint MaskR = 0b11111111_00000000_00000000_00000000;
uint MaskG = 0b00000000_11111111_00000000_00000000;
uint MaskB = 0b00000000_00000000_11111111_00000000;
uint MaskA = 0b00000000_00000000_00000000_11111111;
R = (rgba & MaskR) >> 24
G = (rgba & MaskG) >> 16
B = (rgba & MaskB)  >> 8
A = (rgba & MaskA) >> 0

with results in following byte value output

Chanel	Byte Value	Bits	Float Value
R	85	00000000000000000000000001010101	0,3333333333333333
G	255	00000000000000000000000011111111	1
B	0	00000000000000000000000000000000	0
A	0	00000000000000000000000000000000	0

version where buttom bits are R chanel

uint MaskR = 0b00000000_00000000_00000000_11111111;
uint MaskG = 0b00000000_00000000_11111111_00000000;
uint MaskB = 0b00000000_11111111_00000000_00000000;
uint MaskA = 0b11111111_00000000_00000000_00000000;

R = (rgba & MaskR) >> 0
G = (rgba & MaskG) >> 8
B = (rgba & MaskB)  >> 16
A = (rgba & MaskA) >> 24

with results in following byte value output

Chanel	byte Value	Bits	Float Value
R	0	00000000000000000000000000000000	0
G	0	00000000000000000000000000000000	0
B	255	00000000000000000000000011111111	1
A	85	00000000000000000000000001010101	0,3333333333333333

[mafiesto4]

Thanks for your suggestions and for spotting this issue. I refactored color processing with those bits, added proper info in doc comments to clarify which channel is used in the bottom bits, and added unit text to ensure this won't be broken again.

2529312