I think a main source of confusion is the distinction between what empirical claims virtue ethicists make, imply, or are committed to, and what their normative claims are. Here’s my interpretation of at least part of the story:
Empirical Claim: People ordinarily have various states of character, which regulate their behavior (and emotions) across time and across situations. Some of these states of character are virtuous (full honesty when called for), some are vicious (dishonesty), and some are in between (partial honesty when called for).
Normative Claim: People ought to cultivate all and only those states of character that are virtuous (because that fulfills their human function and allows them to flourish, blah blah blah).
Situationists aren’t really attacking the normative claim, or the rarity of people meeting its demands. Rather, they are using various bits of empirical evidence to challenge that people have states of character at all. Virtues are only a subset of the possible states of character a person can have. But, if people don’t have states of character in the virtue ethicists’ sense at all, then they cannot by extension have virtues. So the situationists only indirectly make trouble for the normative claim.
Virtue ethicists who are sensitive to the empirical nature of their presuppositions must explain what they mean not just by “virtue,” but more importantly by “states of character” such that it is consistent with the situationists’ data. And this is where things get tricky, and people in the literature start talking past each other, and everyone sounds correct. With that, I’ll quit beating this dead situationist horse for a while :-)