Ok. My point is that a single-cross Punnet’s square, and simple, single gene inheritance can be used to predict progeny (for what can be treated as a single locus) for the cross in question. Codominance doesn’t affect the genotype ratios of a Punnet square, even if codominance violates a law of Mendle.
I honestly forgot Mendel didn’t understand codominance. I tend to think of nonmendelian as quantitative, extranuclear inheritance, and epigenetic type examples. Regardless you can pretend the “B-locus” is a single, codominant gene and that this “single locus” predominantly determines chemotype. Unless you are specifically searching for rare, intermediate recombinants, you can ignore them here.