Fix FourDivergence (and ThreeDivergence) to handle derivatives in different dimensions

[vsht]

Currently FourDivergence claims that a derivative of a 4-dim vector w.r.t to a D-dim vector is zero. This is not really consistent as one can see in the following example

FourDivergence[FV[p, mu], FVD[p, nu]]
FourDivergence[FVD[p, al] MT[al, mu], FVD[p, nu]]

FourDivergence[FVD[p, mu], FV[p, nu]]
FourDivergence[FVD[p, mu], FVD[p, al]] MT[al, nu] // Contract

Likewise for Dirac matrices, propagators and tensors. This should be fixed.

[vsht]

@duartefontes Could you please have a look and test if it works for you? I think that as of now FeynMaster is the only tool that needs/uses four derivative in the t'Hooft-Veltman scheme.

FCSetDiracGammaScheme["BMHV"];
FourDivergence[FVD[p, mu], FV[p, nu]]

[duartefontes]

Hi, Vlad, I see that it no longer gives warnings; yet, FeynMaster uses by default NDR. I have played around a bit with other schemes, but I have almost no experience.

[vsht]

I see. I just noticed that FeynMaster has a dedicated derivative routine for such cases, since FourDivergence could not handle those properly. So I thought that it was needed for some calculations done by FeynMaster.

[HBelusca]

Oh nice! A feature that was necessary!