im not sure what you mean

to brake it down into r,g,b,a and then have it rebuilt in the method

you can do:

image.imageData.set(x, y, 0xFF00FFFF, true)

or

image.imageData.set(x, y, 0xFF0000)

as I said you breaking it up and then putting it back together

and how are you proposing?

taken from cornerContour/color/ColorHelper

inline
function toHexInt( c: Float ): Int
    return Math.round( c * 255 );
inline
function rgbInt( c: Int ): Int
    return ( c << 8 ) >> 8;
    
inline
function getAlpha( c: Float ): Float
    return ((Std.int(c) >> 24) & 255 )/255;

inline
function getColor( c: Float ): Int
    return rgbInt( Std.int( c ) );
        
// throws aways alpha on c and uses the new a value.
inline
function colorAlpha( color: Int, alpha: Float ): Int
    return ( toHexInt( alpha ) << 24 ) | rgbInt( color );

inline
function colorIntAlpha( color: Int, alpha: Int ): Int
    return ( alpha << 24 ) | rgbInt( color );

I think it's ideal to provide

0xFF0000 ?1.

and

0xFFFF0000

since the methods are inlined you can use some of them to manipulate rgb, argb, and alpha to your needs.

generally I find color in most targets is used as 0xFFFF0000 and only broken up perhaps for a shader.

for canvas it's sometimes 0xFF0000

but the methods I linked to allow you to use lineStyle like flash

so 0xFF0000, 1.

normally you want to define a color with just Int32

i dont really understand how any of that is relevant tbh... the pixel data needs discrete r, g, b, values (for seperate indexes) the a is a totally different part of memory (different pointer)...

it would be nice to not have the "useAlpha" param, granted

well I am saying supply both argb int and rgb int with a alpha

since that's what users need the methods above help extract the alpha from argb int

right, so "set(x, y, 0xFF0000, .5)" and "set(x, y, 0xFF0000)" ?

yer I think that usually less operations in total and closer to what users will have.

because you can extract the alpha.

it only matters because over a large image likely takeing apart all channels and putting together again is going to slow it down compared to just separating rgb and alpha but not made a test.

id be very surprised if it made any serious difference tbh

if you try to do this every frame...

there is no "every frame" in wx

but how are you proposing to set the image data anyway if not taking it apart into r,g,b?

i guess you could just set

p[i] = 0xFF0000; i += 3

inline function getColor( c: Float ): Int return rgbInt( Std.int( c ) );

inline function getAlpha( c: Float ): Float return ((Std.int(c) >> 24) & 255 )/255;

right, i get that, but im talking about, now you have, say "0x123456"

oh you have to set each channel in there

yes, ive been saying that for ages 🤦‍♂️