Code review comment for lp:~vanvugt/mir/run-without-entropy

I appreciate Alan's caution. However I think his concern that other common software exists that would suffer from the same /dev/random mistakes is incorrect. As Alexandros points out, only software like ssh-keygen uses /dev/random, and it is designed to do so elegantly -- expecting long indefinite delays.

As for the non-blocking concern of switching entropy sources, that's not really a new problem. We will forever have the problem of developers changing code that reduces the efficacy and coverage of existing tests. It's just something we have to maintain.