The answer to What common sites are known to be troublesome for Tor use? is a "community wiki [that] will list commonly used websites known to be troublesome for Tor users." I have encountered such problems over the years with many websites, and I've seen many such complaints in the tor-talk mailing list, and elsewhere.
And so I've decided to test how well the top 50 websites work for Tor users, and to determine the pervasiveness of any blocking. I'm using a script that runs "torsocks midori -s http://foo.bar/" to snapshot sites, while using exit_used.py to determine exit relays used. Midori is configured send a Firefox user agent string.
An example for http://craigslist.org/ from preliminary results is here. It's possible that, given enough such information, someone might determine what server(s) I'm using. But that would arguably require collusion with entry guards. It wouldn't be problematic if the server(s) were tied to my mirimir persona.
Is there concern that unblocked exit relays will be blocked as a result? I don't think so. All of the exit relays in my example are obviously in Atlas, and they are all in the "DAN TOR" and "DAN TOREXIT" blacklists. But maybe I'm missing something.
What are other key issues around publishing which websites block particular exit relays?