Well, maybe, but I fail to understand your position here, first you tell me that singularities are meaningless and that this is the reason GR is wrong, and now you answer that the universe must be geodesically incomplete and therefore inhomogeneous (or the other way around?). But you do realize that the only reason to consider it geodesically incomplete is the fact that is supposed to have singularities, right? If there were no singularities, as you (and according to you also wikipedia) claim, then it would be geodesically complete.
(...)
The problem is that the FRW metric is based in the mathematical use of homogeneity, and it surely admits perturbative approximations that may vary the scale of homogeneity, but can never admit singularities, because if a not (mathematically)homogeneous 3-space is used in the metric (and there are only three choices :euclidean, elliptic and hyperbolic), then expansion(or contraction) is not mathematically feasible, and in addition it wouldn't fulfill the cosmological principle.