This paper reports a systematic comparison of the performance of alternative methods of including relativistic effects on the nature of metal-metal bonding in the Re2Cl8(2-) anion. The comparison involved the description using a scalar relativistic Douglas-Kroll-Hess (DKH2) Hamiltonian with all-electron basis sets and the relativistic effective core potential (ECP) basis sets. The impact of the above methods on the picture of the bonding was analyzed using the so-called domain averaged Fermi holes (DAFH). Besides comparing the impact on the picture of the bonding of the two above methods, the focus was also put on the systematic comparison of the "exact" AIM generalized form of DAFH analysis with the approximate Mulliken-like approach used in an earlier DAFH study of ReRe bonding. It has been shown that in contrast to descriptions using ECP basis sets where the differences in the picture of the bonding emerging from the approximate and "exact" DAFH analysis are only marginal, the approximate DAFH approach has been found to dramatically fail in the case of all-electron basis sets required for the description in terms of the Douglas-Kroll-Hess (DKH2) Hamiltonian.