From Rom's email quoted in the changelog thread:

We are feature complete for 6.2, from here on out 6.2 will just be bug fixes. Any of the remaining new features that didn’t make it are going to be punted to the next release. Windows 9x should work again.

...

This is a good time to bring up any issues that you might have raised in the past, but was tabled for one reason or another, if you think there are major bugs that haven’t been addressed. I’m not promising any radical changes or anything like that, but I would like to make sure that anything that hasn’t been fixed is at least logged as a bug in Trac.

Why don't we take him at his word, and gather together here any bugs which have been reported, but haven't (yet) made it into the v6.2.1 pre-release?

I'll open the batting. Trac [trac]#588[/trac] (indefinite suspension of computing when changing system clock) is still present in v6.2.1, as is the other problem discussed in the related Benchmarking bug - indefinite suspension of computing thread (<active_frac> set very low and doesn't self-correct).

In fact, the 'indefinite suspension' bug is even worse than I described it at the time: BOINC Manager shows the tasks as still 'running, high priority', but no computing is being done (CPU time not increasing), and a remote viewer like BOINCview shows the 'suspended' status.