What's so bad about the idea of using the database itself to guarantee
atomicity?

On Fri, 30 Jul 2004 05:33:17 +0900, Lennon Day-Reynolds
<rcoder / gmail.com> wrote:
> I appreciate everyone being so interested in this, but I think we're
> wander far afield of the whole point here, which is implementing a
> simple ORM layer atop a standard RDBMS backend. Implementing a
> concurrent atomic sequence manager external to that database sounds a
> lot like reinventing the wheel, and is certainly way outside the scope
> of what I was trying to do with Active Record.
> 
> At this point, I think I'm going to have to go one of two ways on
> this. The first option would be to simply write a SQL Server-specific
> (rather than generic ODBC) adapter for Active Record, using the
> built-in identity management functionlity. The simpler option for my
> purposes may actually be to transfer the data I need (which is mostly
> periodic reporting from a sales transaction database) from SQL Server
> to something easier to work with (like SQLite) via straight ODBC/SQL,
> do my Active Record-based manipulations on that local snapshot, and
> then drop the local database.
> 
> The general issue of supporting arbitrary ODBC backends through Active
> Record will probably take some more time to solve.
> 
> Lennon
> 
>