In order to generate all of the chromaticity coordinates on the chromaticity diagram one just needs two sources, pass each of those sources through a monochromator, and mix the resulting two monochromatic sources in an integrating sphere. Any point on the chromaticity diagram can be reproduced by using the two wavelengths which are along lines that pass through the chromaticity coordinate in question and points on the spectrum locus where the line is perpetual to the spectrum locus.

One issue would be the luminance of the light at each chromaticity. Without a pair of equal energy sources the luminance of each chromaticity coordinate would be different.

It’s worth noting colored chromaticity diagrams are mostly color for sake of the aesthetics of the diagram only. Any chromaticity coordinate could appear as any color depending on the viewers state of adaptation. For instance, if a viewer were adapted to 550nm (unlikely but theoretically possible), “white” would be on the spectrum locus at 550nm and the vast majority for the diagram would be filled with shades of blue and red.

Regardless, one could easily produce any x,y chromaticity coordinate with just the setup described above.