Wow: “xrandr doesn't work anymore on xorg-git” “I do not think this should be specifically on you, it is not unreasonable to expect that the author of a change tries their change before even submitting it upstream.” does not give a warm fuzzy feeling about the author of the at-fault patch leading a fork.
I think it's not the one to blame who broke this but those who implemented everything all the time without adding any tests. Xorg has quite a large codebase but almost no automated tests.
So we agree that the maintainer is at fault: he wanted to change things and not have to thoroughly test his changes by doing the boring work of adding test coverage to the modified area.
