FWIW, I added code in Turbine 3 for this sort thing.
Daniel Rall <[EMAIL PROTECTED]> writes:
> Byron Foster <[EMAIL PROTECTED]> writes:
>
> > Kurt Schrader wrote:
> >
> > > On Saturday, August 11, 2001, at 01:37 PM, Daniel Rall wrote:
> > >
> > >> So I finally reviewed this patch. A more flexible implementation
> > >> would involve a one-method name conversion interface to be implemented
> > >> by the application programmer. I would involve setting a
> > >> "nameConversion" attribute on the schema, table, or column one would
> > >> like their name conversion implementation to be applied against. The
> > >> nameConversion attribute would be overridden if the javaName attribute
> > >> was set for any fields. Some simplicitly is lost with my proposal
> > >> (compared against the patch below). Thoughts on this?
> > >
> > >
> > > I would prefer it this way also, as long as there is a default
> > > implementation
> >
> > > that would be executed if the property wasn't specified. I've never
> > > found a need to deviate from the default mapping and this is something
> > > that I don't want have to worry about.
> > >
> > >> If it's not liked as much as the patch below, I'll check in with
> > >> changes to the names of the alogorithm types supplied by Byron and a
> > >> modification to the DTD to take this into account.
> > >
> > >
> > > I would be -0 to Byron's patch as I don't want to see this happening
> > > whenever someone comes up with a new way to change the
> > > capitalization of columns.
> > >
> > Actually the patch preserves the current default behavior. If the
> > user does not specify a method for conversion then torque will apply
> > the current algorithm.
>
> Yes, it does.
>
> > I think given the current trend such as "javaName" and "basePeer"
> > attributes it would be appropriate to add a "nameConversion" attribute
> > to the database schema. I would be willing to do this if it is the
> > consensus.
>
> I would prefer this (I hope you don't feel jerked-around :). Unless
> someone else speaks up soon, let's go with this, since as you point
> out it's much like the existing attributes.
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
