The haddock failure you're seeing is caused by using relative symlinks with lndir. See #7879. Thanks, that's one problem less :) I applied the patch and I'm running the validation to see if that works.
Perhaps you have different library versions checked out in different trees; you could check whether or not this is the case by looking at the various trees' fingerprints. According to fingerprints the trees are identical. I diffed both directories and there are some minor differences, but I think these come from the fact that ghc-working never contained build artifacts, whereas ghc-validate was cleaned with 'make maintainer-clean', which might have been inaccurate.
Personally, I think the current mix of git submodules and non-git submodule dependencies is horrible. It makes it impossible to check out a consistent tree without a full fingerprint while also incurring all the disadvantages of submodules. I admit I don't understand motivation for using both git submodules and non-git submodules.
What worries me the most is unreliability of the test suite. After getting 200 test failures mentioned in the previous mail, I ran the validation again and got only 3 failures. I ran one of the failed tests manually and it passed, leaving me with two known failures. Janek