Re: [Modeling-users] slooow delete

[Sebastien B. <sbi...@us...>]

John Lenton <jl...@gm...> wrote:
> Thank you very much for your reply. Just one small doubt
> remains:
> 
> 
> On Fri, 18 Mar 2005 12:18:57 +0100, Sebastien Bigaret
> <sbi...@us...> wrote:
> > 
> > And, if you do this: just make sure that
> > 
> > - either task.getMeal().isFault() == true
> > 
> > - or that, in case it is not, that
> >     task.getMeal().getTasks().isFault()==true
> > 
> >   In this last case, if it is false, then you can do the
> >   meal.addToTasks(task) without any overhead, since this
> indicate
> > that
> >   the fetch has already be done.
> 
> is there any case where the last line would return false,
> instead of
> raising an AttributeError?

Good question ;)  Here are the details

First situation: you have a newly created object ; in this case,
meal.getTasks() will be a real python list (actually a list or a
tuple), assigned by the constructor, and you'll get an
AttributeError.

Second situation: the object has been fetched by the framework. 
In this case, any attribute corresponding to a tomany
relationship is initialized with the so-called <<to-many fault>>
object, responsible for lazily initializing the list when
appropriate.  If you're curious, you'll have a look at
FaultHandler.AccessArrayFaultHandler, which is the corresponding
class proxying the underlying list.  So, in this case, you'll
get a true or false answer when asking
meal.getTasks().isFault().

  As a conclusion, you're right, you have to check for
AttributeError as well and your code will handle all possible
situations.

-- Sébastien.


-- Sébastien.