[PATCH 4/4] database: Handle error while deleting uncompacted database

[David Bremner]

Ben Gamari <bgamari.foss at gmail.com> writes:

> We never checked to ensure that the rmtree() of the old database
> succeeded.
>

Is there some reason these patches (and the existing compact code in the
lib) is using fprintf and not the status callback function? Most places
we don't (yet) have logging callbacks, but here we do.

d