Andrew Borley wrote:
I tested the patch briefly on windows/VC7 - everything builds fine and
the
Calculator sample is happy.
Cheers
Andy
On 10/5/06, Pete Robbins <[EMAIL PROTECTED]> wrote:
I attached a patch of my proposed fix to the Jira
On 05/10/06, Pete Robbins <[EMAIL PROTECTED]> wrote:
>
> I think it does not. I raised TUSCANY-798 as I was getting a windows
> compile warning when Composite constructor was passing it's this
pointer
to
> the ComponentType constructor. ComponentType saves away the pointer
to a
> Composite but never uses it, nor is the getComposite() method ever
called.
> My patch attached to the Jira removes the Composite* form the
ComponentType
> constructor and all is well.
>
> Looking further into the code there is a method
> ImplementationExtension::getImplementation(Composite* composite,
> DataObjectPtr scdlImplementation). The only place that the value
composite
> is ever used is to construct the ComponentType (and this is not
needed).
>
> I propose to change the extension method signature to
>
> ComponentType* getImplementation(DataObjectPtr scdlImplementation)
>
> and obviously change the impls of this method in the extensions.
>
> Any objections/comments?
>
> Cheers,
>
> --
> Pete
>
--
Pete
I thought that a ComponentType would need to know the composite it lives
in, to reflect interfaces, or use complex data types for properties,
which are currently packaged with a composite.
However if we are not using this Composite* pointer yet, then I'm happy
to see it removed. We can always re-introduce it when we actually come
across an actual use case, instead of over-engineering things for
imaginary use cases :)
So I am fine with the change.
--
Jean-Sebastien
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
