> > I was wonder what the thinking was behind removing the preSave method
> > and if there are plans on implementing alternatives?  The preSave method
> > offered a nice way for the base class to provide a more managed and
> > safer environment for a derived object to prepare itself to be saved.
> > For example a method such as preSave could do the following:
> > 
> > o Guarantee that the object has been modified, and really will be saved.
> > o preSave would be called only once during a save operation.
> > o The Base class could prepare itself for operations within the preSave
> > that may change what is saved and offer protection against unexpected
> > state changes.  
> > o The Base class could manage when the preSave is called in relation to
> > other objects being saved and other preSave method calls.
> > o Prevent infinite loops because of OM object method calls from within
> > save operations.
> > 
> > I lean toward making the save methods final,  And only allowing the
> > derived OM object to override preSave, or some such method.  I think
> > this allows more flexibility in future changes to torque without hosing
> > current applications of it.
> > 
> > Thoughts?
> 
> I think all the things you mention above should be handled transparently.
> The methods available to application objects should be as small as possible,
> your basic CRUD and that's about it. I don't see much point in preSave().
> 

 
I agree that these issues should be handled transparently, but I think
that preSave is the way to do it.  I also agree in simplicity.  The
preSave method allows the Base class OM an opportunity to handle these
issues transparently.  I know this is vague but I'll send some concrete
examples.

Thanks,
Byron





--
To unsubscribe, e-mail:   <mailto:[EMAIL PROTECTED]>
For additional commands, e-mail: <mailto:[EMAIL PROTECTED]>

Reply via email to