Re: How to avoid $class->new($value) with inflate

[prev] [thread] [next] [Date index for 2005/03/03]

On Thu, Mar 03, 2005 at 05:10:25PM +0000, William Ross wrote:
> On 3 Mar 2005, at 16:50, William McKee wrote:
> 
> >On Thu, Mar 03, 2005 at 11:10:27AM -0500, Charles Bailey wrote:
> >>I don't know that there's a conceptual benefit (though the same could 
> >>be
> >>said of any Relationship, I guess).  I'm wondering, though, if 
> >>wrapping
> >>into a module the four-or-so triggers necessary to manage this sort of
> >>transformation would have a practical benefit for new users (easier to
> >>understand) and more experienced users (code reuse).  I'm willing to 
> >>be
> >>convinced either way, though.
> >
> >OK, I'm convinced that the new relationship sounds like it'd be a good
> >thing.
> 
> It does sound useful, but it doesn't sound like a relationship. It's 
> just a transformation applied to a column (as is a relationship, I 
> suppose, but that doesn't make this a relationship :).
> 
> CDBI::Column::Transform?
> 
> or CDBI::Column::Rule, if you wanted to be more ambitious with it in 
> future.

Currently I can find all relationship modules by searching CPAN for
"Class::DBI::Relationship". I don't think I want to give that up, even
if the name ends up being imperfect. Also the relationship mechanism
allows easy addition of methods and triggers and is probably the sensible
place to start - so while the resultant module might not be a relationship
as such it *will* probably be a CDBI::Relationship in an ISA sense.

        -- 
            Matt S Trout            Brag sheet:    http://trout.me.uk/services.html
LAMP, Infrastructure        Contact:       services@xxxxx.xx.xx
   and Automation
     specialist                                       Do it once. Do it right.