[cff] Optimize the blend vector computations.

* src/cff/cffload.c (cff_blend_build_vector): Use FT_MulDiv and skip
multiplying by 1.
1 file changed