Quoting Niclas Hedhman <[email protected]>:
On 6/4/10, Stanislav Muhametsin <[email protected]> wrote:
* Declared concerns and side-effects are executed for public methods
IF AND ONLY IF the call originates through @This -injected variable.
OR from outside the Composite.
True. I realized that only *after* I sent mail (typical situation),
but decided to fix that only if I would need to reply here again, hehe.
And what about when the call is cross-mixin but not via @This, i.e.
abstract ones sharing the implementation burden, which may or may not
be aware of each other??
Yeah, that was exactly why the private method aliases might not always
work for bypassing the full invocation chain of mixin. It is true that
it might get complicated if they are completely unaware of each other.
However, there always could be built another mechanism for such
situations, it's all about balance between complexness and limiting
usage-scenarios.
After thinking about it some more I think that Rickard's consistency
argument is quite important, like @Optional.
Well, I just hope that scenarios where it is specifically important to
bypass the mixin invocation chain will always be handlable by private
method aliases, or some other way which doesn't pop onto my mind yet.
I for one have had some "close-by" situations where I certainly would
have wanted to bypass the mixin invocation chain, but since the
performance loss was not critical, I didn't put much weight on it.
_______________________________________________
qi4j-dev mailing list
[email protected]
http://lists.ops4j.org/mailman/listinfo/qi4j-dev
