thereby decoupling the two systems (for updates, anyway)
That means, consumers of Post need to update_document themselves!
Post (in update) now only talks to the database)
Tests for this will now be failing, as they haven't been touched yet
This function does what was previously done in main:
create a search index, and try to recover for errors
if possible.
This commit also fixes plume-cli to use the new Searcher
API (that depends on a DbPool, not on a single Connection).
this way, we don't need to pass along a conn into the function.
This should make splitting PlumeRocket up into its components a little
easier. This PR is another attempt at fixing #799.
