Re: Model redesign?

Samuele Pedroni <[email protected]>:
[snip]
> the latter by the way is what more or less Dallas's code already do.
> So I agree with Magnus that is far from true that it's wasted effort
> ... What changes is that the model observes the compenent instead of
> having the componenet take care of directly updating the model.

Yes, I guess that is true.

> It should be noted that under this approach unless one knows that no
> model is attached to comp, is not guaranteed that this works:
> 
> L=['a',...]
> comp.items=L
> 
> # later
> L.insert(0,)
> comp.push('items')
> 
> unless set({items = }) for a list change the a preexisting list
> in-place...

You mean that the risk is that comp.items no longer refers to L? That
is true.

And, as I said, if we choose to go with this sort of decoupled
approach (without direct in-place modification etc) we might also
consider making all attributes work like the aliases. Thus, the above
would be guaranteed *not* to work.

> Otherwise the above idiom to be model-agnostic under this approach
> should be:
> 
> comp.items.insert('items',...)
> comp.push('items')

Again, this would be guaranteed not to work with copy semantics in the
attributes. You'd have to do

  comp.items = 'foobar' + comp.items

or something.

This is an area where it seems we may potentially see some changes
from the initial 0.2 design, but I think it's important that we get
this right, so...

> regards.

-- 
Magnus Lie Hetland        Practical Python          The Anygui Project
http://hetland.org        http://ppython.com        http://anygui.org


-------------------------------------------------------
This sf.net email is sponsored by:ThinkGeek
Welcome to geek heaven.
http://thinkgeek.com/sf