30 Mar
2017
30 Mar
'17
1:13 p.m.
On Thu, Mar 30, 2017 at 8:51 AM, Lee Clagett via Boost
How would it notice that a bucket was partially overwritten though? Wouldn't it have to _always_ inspect the entire key file?
One way to fix this is to prefix each bucket in the log file with a cryptographic digest, making it easy to verify the integrity. This means an extra pass through the log file, which isn't too bad at all (much better than a pass through the entire data file).