[Rt-devel] Re: RT::Attribute::Name Unimplemented in RT::Attributes error

[Jesse Vincent]

On Wed, Dec 21, 2005 at 05:46:21PM -0800, Joby Walker wrote:
> Thanks,
> 
> Hadn't seen that yet.  This is pretty easy to override though by having
> RT::Record->Attributes always return a fresh RT::Attributes object.  You
> lose the caching of all attributes in the parent object, but I don't see
> that as a disadvantage.  All the attributes that get regularly called we
> actively save in the user's session, and loading all of a user's (either
> directly or via a group) SavedSearches, Reports, and Schedules any time
> I want one of the user's attributes is not going to be efficient.

*nod* That's why we're doing the caching in the first place. I wonder
what's _actually_ wrong with the caching layer.