Re: [CDBI] Re: ANNOUNCE: Ima::DBI 0.35 released

On 6/11/07, Edward J. Sabol <sabol-2oVx0MVsMgiP/[email protected]> wrote:
> And there's always some kind of wacko ;) who wants to use
> Ima::DBI without Class::DBI, so why should Ima::DBI's set_sql() cache
> statements in manner different from Class::DBI's set_sql()?

Said wacko can get the desired effect pretty simply by subclassing
Ima::DBI (which you're supposed to do anyway) and overriding set_sql.
That's what Class::DBI is doing actually, for some of the parameters
to set_sql.

> What do you think? If someone were to submit a bug for Ima::DBI with a patch
> like that (along with docs and tests)...

I still think it makes more sense to put it in Class::DBI, but I don't
see it doing any real harm to put it here, so if you want to make a
patch for it, go ahead.  I'll change the if_active flag to 3 as well.

Would you guys mind explaining how this is causing problems for you?
I've used prepare_cached on Oracle, Postgres, and MySQL, and never had
any unexpected behavior from it.  I use it practically everywhere. And
I don't even set the if_active flag.

- Perrin