the problem was more likely a patching error on his part, or even possible something he did to the ROM after patching (he did evaluate the binary under a tool, at one point, which may have inadvertantly changed it).
For what it's worth, he did specifically state that he verified the ROM with the MD5 listed. If he knows enough to do that, the chances that he accidentally corrupted the ROM (and
had it 99.99% functional) are pretty slim. Maybe it was an issue that only shows up on certain emulators or on real hardware, maybe he did in fact corrupt the ROM, who really knows? But he did have an issue that he believes in good faith is attributable to the hack, and while it's not the ideal approach, addressing the issue in a review isn't wrong
This is the reason that version numbers are now included in the review. If you release an update, you can mention in the description that a bug has been fixed and people can see that the review is for an old version of the hack. From a policy standpoint, letting hack authors request removal of reviews is a terrible idea. And while you can contact the reviewer to try to rectify the issue that way, you should be careful not to drum up unnecessary drama.
You do have a bit of a point in that reviews for old versions can become somewhat irrelevant. To me a reasonable solution would be to include some sort of warning at the top of a review if it is for a previous version of the hack, but I can only offer that as a suggestion to the powers that be.