After developing the initial prototype of Cryptonite in Costa Rica, I continued working on it independently. After a much needed cleanup of the code (prototype development had been hectic and had left not much time to refactor or test the code), I worked on a number of Perl modules and components that would be needed next, to make the jump from a simple prototype to a scalable product. These included Crypt::GPG (with an interface almost identical to that of Crypt::PGP5, so that switching to GnuPG for the crypto operations in Cryptonite involved little more than a single-line change to the code), and Persistence::Database::SQL and Persistence::Object::Postgres (which provide object persistence in a Postgres database, with a similar interface to Persistence::Object::Simple, making the backend database switch quite seamless as well).
Persistence::Object::Postgres, like Persistence::Object::Simple,
uses a blessed reference to a hash container to store key-value pairs, which can be
committed to the database with a
commit method call. It also uses Perl's
Tie mechanism to tie Postgres' large objects (BLOBs) to filehandles, enabling natural filehandle-based access to large binary objects in the data-base. One of the major benefits of Persistence::Database::SQL over Persistence::Object:: Simple, of course, is that it enables proper queries into a real database. For example, with Persistence::Object::Simple, there's no clean way to quickly search for a particular ...