Developers

[dev42]

@Romans -

Thanks for that link! Awesome stuff there.

Now, is there still a problem? Because, to me at least, once the signed/unsigned issue is understood, it should become clear that, when talking about signed integers:
&hFFFFFFFF != 4294967295

&hFFFFFFFF == -1

still working on how the 64 bit Double gets clipped to the 32 bit Signed Integer.

Edit: Double is to Float as int is to char!
4294967295 == 0x41EFFFFFFFE00000 ==
01000001 11101111 11111111 11111111
11111111 11100000 00000000 00000000

but it is getting clipped... possibly taking the least significant 32 bits. ( chime in any time gurus! )
________________________

Wikipedia: 2's Complement

[squirreltown]

"EnTerr" wrote:
Maybe calling it always with V=100, A=255 has saved your bacon so far, IDK. But check for colors with strong R, they should all be washing to cyan (?).

Thats what I found, passing it pure red returned a really pale cyan. I already had another solution so i didnt pursue it though.

[dev42]

I haven't yet tried this with DrawObject and I'm not understanding the point of converting 2^32 to "decimal" and then use that as a color IF the HSVAtoRGBA() func is working as it should ...

The main tweak's to HSVAtoRGBA() are that it now uses Integers, Integer bit shifting which supposedly avoids the Double to Integer conversion issue, and returns an Integer. The function (latest ver) returned negative values and a spectrum from Red through all the hues back to Red... so, I was pleasantly content.

I did call out how fully I tested this. 😉

I'll dig deeper later.
__________________

"squirreltown" wrote:
Thats what I found, passing it pure red returned a really pale cyan. I already had another solution so i didn't pursue it though.

I did make mention of the "baby blue" :roll: ... and again, my changes to the HSV...() fixed it. No more "pale cyan". Could some kind soul test it out?

[Romans_I_XVI]

Sorry, the dectohex function I posted earlier is not the same as what I'm using, I forgot I removed that "As Integer" so it is not getting chopped off and I'm getting back the correct hex as a string, I just now tested it.

[Romans_I_XVI]

Ok finally tested your version of the function dev42 and it works perfect! Thank you for your magic!

I think I will replace the one in the original function with yours.

[squirreltown]

"dev42" wrote:

I did make mention of the "baby blue" :roll: ... and again, my changes to the HSV...() fixed it. No more "pale cyan". Could some kind soul test it out?

Yep it works, thank you.

[Romans_I_XVI]

It's throwing a type mismatch if I give it a saturation of 0 though. I've tried 0 with everything else and it works.

Type Mismatch. (runtime error &h18) in pkg:/source/generalUtils.brs(77)
077:    for c = 0 to rgb.count()-1 : rgb[c] = int(rgb[c] * 255) : end for

[Romans_I_XVI]

I think I found the problem.

This

   if s% = 0 then
      rgb = [v, v, v]
   else

Needed to change to this

   if s% = 0 then
      rgb = [v%/100, v%/100, v%/100]
   else

[dev42]

Good catch. Glad to be helpful.

[TheEndless]

Out of curiosity, why are you using HSV instead of RGB in the first place?