[WIP] Correct fusion algorithm in RZ

[Yin-YinjianZhao]

This PR mimics the one in PR #2510 to correct the fusion algorithm in RZ.

[RemiLehe]

@NeilZaim I have added you as a reviewer here, but we could discuss in more detail why this rotation is needed, during the next developer meeting (next Tuesday).
Also: I think that we could discuss whether the second part of the fusion algorithm (the creation of particles with the correct momenta, as implemented in #2330) is already valid for RZ.

[NeilZaim]

Hello. Yes, this change sounds reasonable to me if we have a perfect axial symmetry (order 0 RZ). I don't know if modifying the momenta to machine precision is an issue (I guess not) but if we want to avoid that then it's easy to do.

Also, yes we would need to do something similar for the initialization of the product particles (rotate the momentum of one of the colliding particles before and rotate back the momentum of half the produced particles after). I can do that in a follow-up to #2330 if needed.

[RemiLehe]

This is superseded by #3255