Hi,
In the present state, the information the runtime is keeping about the
databases it acesses is not organized cleanly. At the moment, Torque keeps
different Maps of database adapters, of database maps, of data source
factories etc. Also, things are hidden where they are not expected: in my
eyes it makes no sense to keep the id broker in the database map, or the
schema name in the data source factory.
I'd therefore propose to change the internal representation of the data
such that there is one top level object per database, called "Database",
which keeps all the information the runtime has about a certain database.
Specifically, it would provide access to the database map, the data source
factory, the database adapter, the id broker, the id generators, and the
schema name.
To soften the transition, the id broker, the id generators, and the
schema name would stay at their current places (database map and data
source factory), but the acessors would be deprecated. After the next
release, the references themselves would be moved to the database map and
the deprecated methods would be removed.
Please send objections, requests for clarifications etc in the next few
days, otherwise I'll simply do it.
Thomas
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
