[CDBI] ANNOUNCE: Ima::DBI 0.35 released

[Perrin Harkins]

On 6/10/07, Matt S Trout <dbix-class at trout.me.uk> wrote:
> There's always -somebody- who uses the worker MPM on !Win32.

That person should contact me so I can tell them why it's a bad idea.
And if any Windows users want to use Class::DBI with threads, I'd be
happy to take a patch.

> Well, that would make prepare_cached not dangerous where your DB supports
> it appropriately

Isn't it always safe with a value of 3?  That's the impression I got
from the DBI docs.

> But a global flag to say "look, just don't use prepare_cached" would also
> be nice - the problem is that the choice is set at ->set_sql time so if your
> superclass (whether CDBI or otherwise) says "yes" you have absolutely no
> way to change your mind in a subclass, which is a complete pain.

Class::DBI already overrides set_sql, so that seems like the
appropriate place to put a configuration option like this.  Ima::DBI
tries to be agnostic about using prepare_cached or not and just does
what it's told.

- Perrin