When people say that Google has "done a crawl", they're speaking rather loosely, since Google obviously can't start at one end of the web and keep going until they reach the "other end".  Their crawlers are constantly visiting various websites, and with different frequencies for different websites, so the fact that they've visited this forum thread isn't that significant.  You might check your usage logs to see if they've recently re-encountered the area of your site that contains the pages you don't want crawled.  If they crawled the other pages from that area, but not the ones you wanted excluded, then maybe robots.txt is working.  Even so, it's only been about a week since your original post, and I would think it might take several weeks to two months for your pages to drop out, assuming that it takes about the same amount of time for them to drop out as to get in in the first place.  Maybe even longer, since obviously they would be worried about too quickly dropping pages that really do still exist, due only to a server error at the time they crawled. 

The answer I've seen elsewhere, at least regarding Google, is that they are very reluctant to purge URLs that they know once existed. Ok, make that very, very reluctant. It could also have to do with their making backup archives and failing to purge those. If they ever have to revert to the archive for some reason, your URLs are resurrected. In a previous post I mentioned "several weeks to two months". Since then I saw a complaint somewhere that URLs that had been dead for 2 years were still being crawled.