Diemo Schwarz

The following gen code does not compile because of a weird error in the generated C code: the variable C is used outside the sample loop before being defined inside :

            const t_sample in1 = (*(__in1++));

            t_sample H = ((((t_sample)0.78) * in1) * ((int)360));

            t_sample C = ((t_sample)0.5);

here's the gen code (hsl to rgb, if anyone has a working more elegant code or patch snippet, I'm all takers...), also I'm not yet totally firm with gen, there is a chance I've missed something obvious.

// visible hue
dx = 0.78;
H  = dx * in1 * 360.; // in degrees
S  = 100.; // percent
L  = 50.;

// hsl to rgb
s = S / 100.;
v = L / 100.;
C = s * v;
// float X = C*(1-abs(fmod(H/60.0, 2)-1));
X = C * (1 - abs(mod(H / 60.0, 2) - 1));
m = v - C;

if (H >= 0 && H < 60)
{
  r = C; g = X; b = 0;
}
else if (H >= 60 && H < 120)
{
  r = X; g = C; b = 0;
}
else if (H >= 120 && H < 180)
{
  r = 0; g = C; b = X;
}
else if (H >= 180 && H < 240)
{
  r = 0; g = X; b = C;
}
else if (H >= 240 && H < 300)
{
  r = X; g = 0; b = C;
}
else
{
  r = C; g = 0; b = X;
}

r = (r + m); //*255;
g = (g + m); //*255;
b = (b + m); //*255

out1 = r;
out2 = g;
out3 = b;

The following gen code does not compile because of a weird error in the generated C code: the variable C is used outside the sample loop before being defined inside :

                t_sample m = (((t_sample)0.5) - C);
		// the main sample loop;
		while ((__n--)) {
			const t_sample in1 = (*(__in1++));
			t_sample H = ((((t_sample)0.78) * in1) * ((int)360));
			t_sample C = ((t_sample)0.5);
                        ...
here's the gen code (hsl to rgb, if anyone has a working more elegant code or patch snippet, I'm all takers...), also I'm not yet totally firm with gen, there is a chance I've missed something obvious.
Cheers!

The following gen code does not compile because of a weird error in the generated C code: the variable C is used outside the sample loop bef

gen-code-generation-error

Thanks for the report. It looks like there is a bug in the optimization pass which tries to lift constant and once per vector expressions outside of the main calculation loop (gen uses the same code model as gen~ just with one sample vectors).

We'll look at solving this issue. In the meantime, you can "trick" gen into compiling by placing an assigment of 

 or similar inside any of the if blocks to push the calculation of m on a per sample rather than a per vector basis.

Thanks for the report. It looks like there is a bug in the optimization pass which tries to lift constant and once per vector expressions outside of the main calculation loop (gen uses the same code model as gen~ just with one sample vectors). 

We'll look at solving this issue. In the meantime, you can "trick" gen into compiling by placing an assigment of m = m; or similar inside any of the if blocks to push the calculation of m on a per sample rather than a per vector basis.

Hi Joshua, thanks for the confirmation and workaround! 

In the meantime, I found jit.hsl2rgb, so now instead of getting to know gen better, I'll have to learn jitter from scratch =-)

Hi Joshua, thanks for the confirmation and workaround! 
In the meantime, I found jit.hsl2rgb, so now instead of getting to know gen better, I'll have to learn jitter from scratch  =-)

gen code generation error