[DBIx-Class-Devel] [ANNOUNCE] A much faster DBIx::Class public trial - v0.08240

[Peter Rabbitson]

On Thu, Feb 14, 2013 at 02:49:42PM +0100, Alexander Hartmaier wrote:
> On Thu, Feb 14, 2013 at 1:24 PM, Peter Rabbitson <rabbit+dbic at rabbit.us>wrote:
> 
> > On Thu, Feb 14, 2013 at 01:09:39PM +0100, Alexander Hartmaier wrote:
> > > My ResultClass of one of the Cat apps triggered [1] but doesn't seem to
> > > cause harm.
> > >
> > > [1] DBIx::Class::ResultSet::all(): ResultClass
> > > NAC::Web::NAC::Model::DBIC::ResultClass::API does not inherit from
> > > DBIx::Class::Row and therefore its inflate_result() will receive the full
> > > prefetched data tree, without any branch definedness checks. This is a
> > > compatibility measure which will eventually disappear entirely. Please
> > > refer to t/resultset/inflate_result_api.t for an exhaustive description
> > of
> > > the upcoming changes at
> > >
> > /home/ahartmai/perl5/perlbrew/perls/perl-5.16.2/lib/site_perl/5.16.2/Catalyst/Controller/DBIC/API.pm
> > > line 292
> >
> > Right, I wasn't certain how to word this correctly, help to imrpove the
> > warning would be much appreciated. It boils down to what do we feed to
> > inflate_result() - the new code uses the same format, but omits "NULL
> > branches" of data, so that the (HashRefInflator|Row)::inflate_result
> > have less work to do. However when a custom result class is used (as it
> > is in our case) - I can't know if omitting stuff will break anything, so
> > I give you everything the way it used to be, with a warning about it.
> >
> > Tell me more about the class in question - what does it do, what is in its
> > inflate_result() etc.
> >
> >
> You might remember when I wrote it,

Of course I do not ;)

Anyhow - this was actually helpful, it highlighted a new way to solve 
this that did not occur to me before. Please do not change your code for 
the time being, a solution will most likely materialize that does not 
require any changes.

Cheers