The structures of my Git repositories look like this:
A-B-C-D-E-F # master branch in separate repo1
A-B-C-D-E-G-H # master branch in separate repo2
I think Jim is right. Also bear in mind that if two commits do exactly the same patch/diff against the code, they will have exactly the same SHA1 hash. So A through E should have the same SHA1 in both repos, so it shouldn't be a problem to merge from one to the other, and keep the merged repo as the sole repo to move forward with.
You can setup on repo1 a tracking branch for repo2, and then keep them distinct that way, and delay the merge to whenever you want.