It's no secret there's a camp staunchly opposed to enabling C2 in EAC. And, for obvious reasons, if you're seeking 'flawless' rips, you're completely dependent upon the ability of your drive to accurately use C2 to catch errors. Not all drives are have it implemented well (so I'm told), and from time to time, would let an error slip through the cracks undetected.

So suppose a disc is ripped with Test & Copy, where the tracks are extracted two times. Would the CRCs be mismatched if one extraction let an error slip through?

Granted there would be a slight possibility of the drive failing to catch an error in the same exact position two times in a row and the CRCs would match (even with an undetected error present), but the probability of that happening is almost impossible, no?

This is true. If one is concerned about C2 error correction it would be in there best interest to see what drives support it. I really don't see how it's a major issue myself, but I guess some people loose sleep over it at night.

Yes, absolutely.
I've seen it happen.

Anyway, it depends on the disc and the drive. The same holds true regarding the "slight possibility" aspect of your statement. I don't think it's a good idea to make simple statistical judgements about how a CD may be flawed and how it will behave in different drives and with different settings in EAC. Furthermore, EAC has a horrible way of masking the true nature of a rip. For instance, despite drives that have accurate stream and that setting being checked, EAC still performs synchronization, but just not as often.

Here are a couple of links to look at:
http://www.digital-inn.de/exact-audio-copy...-detection.html
http://www.hydrogenaudio.org/forums/index....mp;#entry441313
http://www.hydrogenaudio.org/forums/index....st&p=411848