converting float[9] to three vec3's

[marmots]

Is the following a valid conversion(?):




uniform float aNiceFloat[9];

void main()
{
vec3 vA=vec3(aNiceFloat[0]);
vec3 vB=vec3(aNiceFloat[3]);
vec3 vC=vec3(aNiceFloat[6]);
// Do some relevant operations stuff
}


I have not tried using this yet, but I was afraid it might work, but is not allowed somewhere in the spec. I tried searching for an example that shows that this is valid but I did not come across one.

[Ilian Dinev]

uniform float aNiceFloat[9];

void main()
{
vec3 vA=vec3(aNiceFloat[0],aNiceFloat[1],aNiceFloat[2]);
vec3 vB=vec3(aNiceFloat[3],aNiceFloat[4],aNiceFloat[5]);
vec3 vC=vec3(aNiceFloat[6],aNiceFloat[7],aNiceFloat[8]);
// Do some relevant operations stuff


What you're currently doing is



uniform float aNiceFloat[9];

void main()
{
vec3 vA=vec3(aNiceFloat[0],aNiceFloat[0],aNiceFloat[0]);
vec3 vB=vec3(aNiceFloat[3],aNiceFloat[3],aNiceFloat[3]);
vec3 vC=vec3(aNiceFloat[6],aNiceFloat[6],aNiceFloat[6]);
// Do some relevant operations stuff

[kyle_]

It is valid code. It should compile and do what its specified to do.

Note however, that vec3(aNiceFloat[0]) creates a vec3 with all three components set to aNiceFloat[0]. This probably isnt what you want.

[marmots]

Thanks for the answer - that makes perfect sense I don't know why I didn't see the error of what I was doing (and the solution of setting the vec3 as vec3(f[0],f[1],f[2]) is pretty obvious).