Nobody cares about the aggregation bug?

php.internals

Wei He

23 years ago
Hi, It seems that nobody cares about the aggregation bug. I know someone can avoid using aggregation by proxy calls or even method duplication like what I'm doing now. But what I'm warrying about is that what if it is a zend hash bug. It might affect some other places at expected moments and thus costing a lot of time of many people. If so, don't say I didn't warn you. :) Wei He

(Marcus Börger)

23 years ago
At 18:26 22.05.2003, Wei He wrote:
>Hi, > >It seems that nobody cares about the aggregation bug. I know someone can >avoid using aggregation by proxy calls or even method duplication like >what I'm doing now. But what I'm warrying about is that what if it is a >zend hash bug. It might affect some other places at expected moments and >thus costing a lot of time of many people.
aggregation is still to be considered, whether it makes it into ze2 or be dropped or fixed or replaced by other oo goodies. (So far there was no final decision from zend). marcus

Wei He

23 years ago
> > aggregation is still to be considered, whether it makes it into ze2 or be > dropped > or fixed or replaced by other oo goodies. (So far there was no final > decision from > zend).
I see. It won't be a problem if you are sure it's the bug of aggregation codes rather than zend hash ones. Wei He

Andi Gutmans

23 years ago
At 08:31 PM 5/22/2003 +0200, Marcus Börger wrote:
>At 18:26 22.05.2003, Wei He wrote: >>Hi, >> >>It seems that nobody cares about the aggregation bug. I know someone can >>avoid using aggregation by proxy calls or even method duplication like >>what I'm doing now. But what I'm warrying about is that what if it is a >>zend hash bug. It might affect some other places at expected moments and >>thus costing a lot of time of many people. > > >aggregation is still to be considered, whether it makes it into ze2 or be >dropped >or fixed or replaced by other oo goodies. (So far there was no final >decision from >zend).
I remember agreeing that interfaces is going to be the aggregation alternative at the language level. If people want to build on top of this outside of Zend, although I'm not sure it's a good idea, this can be done. Andi

(Marcus Börger)

23 years ago
At 21:00 22.05.2003, Andi Gutmans wrote:
>At 08:31 PM 5/22/2003 +0200, Marcus Börger wrote: >>At 18:26 22.05.2003, Wei He wrote: >>>Hi, >>> >>>It seems that nobody cares about the aggregation bug. I know someone can >>>avoid using aggregation by proxy calls or even method duplication like >>>what I'm doing now. But what I'm warrying about is that what if it is a >>>zend hash bug. It might affect some other places at expected moments and >>>thus costing a lot of time of many people. >> >> >>aggregation is still to be considered, whether it makes it into ze2 or be >>dropped >>or fixed or replaced by other oo goodies. (So far there was no final >>decision from >>zend). > >I remember agreeing that interfaces is going to be the aggregation >alternative at the language level. >If people want to build on top of this outside of Zend, although I'm not >sure it's a good idea, this can be done.
Really? I can only remember that aggregation should be moved to the engine. And aggregation is far away from interfaces. However there is still a debate on delegates open. They are close to aggregation when done at runtime. But i would like them as in delphi and there they are done at compile time. So actually i think we could have all three, can't we? marcus

Andi Gutmans

23 years ago
At 09:10 PM 5/22/2003 +0200, Marcus Börger wrote:
>At 21:00 22.05.2003, Andi Gutmans wrote: >>At 08:31 PM 5/22/2003 +0200, Marcus Börger wrote: >>>At 18:26 22.05.2003, Wei He wrote: >>>>Hi, >>>> >>>>It seems that nobody cares about the aggregation bug. I know someone can >>>>avoid using aggregation by proxy calls or even method duplication like >>>>what I'm doing now. But what I'm warrying about is that what if it is a >>>>zend hash bug. It might affect some other places at expected moments and >>>>thus costing a lot of time of many people. >>> >>> >>>aggregation is still to be considered, whether it makes it into ze2 or >>>be dropped >>>or fixed or replaced by other oo goodies. (So far there was no final >>>decision from >>>zend). >> >>I remember agreeing that interfaces is going to be the aggregation >>alternative at the language level. >>If people want to build on top of this outside of Zend, although I'm not >>sure it's a good idea, this can be done. > > >Really? I can only remember that aggregation should be moved to the engine. >And aggregation is far away from interfaces. However there is still a >debate on >delegates open. They are close to aggregation when done at runtime. But i >would like them as in delphi and there they are done at compile time. So >actually i think we could have all three, can't we?
I'm not quite sure what the difference between aggregation and delegation is. I usually think of delegation as aggregation with auto-proxy (although I guess it can also delegate to a non-aggregated object). I think this can all be implemented in user-land by using interfaces like in Java. I don't really want to bloat the engine anymore in this area unless it is something crucial. Andi

Zeev Suraski

23 years ago
At 22:10 22.05.2003, Marcus Börger wrote:
>>I remember agreeing that interfaces is going to be the aggregation >>alternative at the language level. > >Really?
Yes. We're not planning any built-in support for delegation, aggregation or multiple inheritance. Zeev

(Marcus Börger)

23 years ago
At 14:25 23.05.2003, Andi Gutmans wrote:
>At 09:10 PM 5/22/2003 +0200, Marcus Börger wrote: >>At 21:00 22.05.2003, Andi Gutmans wrote: >>>At 08:31 PM 5/22/2003 +0200, Marcus Börger wrote: >>>>At 18:26 22.05.2003, Wei He wrote: >>>>>Hi, >>>>> >>>>>It seems that nobody cares about the aggregation bug. I know someone can >>>>>avoid using aggregation by proxy calls or even method duplication like >>>>>what I'm doing now. But what I'm warrying about is that what if it is a >>>>>zend hash bug. It might affect some other places at expected moments and >>>>>thus costing a lot of time of many people. >>>> >>>> >>>>aggregation is still to be considered, whether it makes it into ze2 or >>>>be dropped >>>>or fixed or replaced by other oo goodies. (So far there was no final >>>>decision from >>>>zend). >>> >>>I remember agreeing that interfaces is going to be the aggregation >>>alternative at the language level. >>>If people want to build on top of this outside of Zend, although I'm not >>>sure it's a good idea, this can be done. >> >> >>Really? I can only remember that aggregation should be moved to the engine. >>And aggregation is far away from interfaces. However there is still a >>debate on >>delegates open. They are close to aggregation when done at runtime. But i >>would like them as in delphi and there they are done at compile time. So >>actually i think we could have all three, can't we? > >I'm not quite sure what the difference between aggregation and delegation >is. I usually think of delegation as aggregation with auto-proxy (although >I guess it can also delegate to a non-aggregated object). >I think this can all be implemented in user-land by using interfaces like >in Java. I don't really want to bloat the engine anymore in this area >unless it is something crucial.
From a design perspective i would appreciate this, however i see a need for aggregation for BC. It has been used by the intended audience of all those OO stuff. marcus

Sebastian Bergmann

23 years ago
Marcus B?rger wrote:
> It has been used by the intended audience of all those OO stuff.
I never used aggregation (or ext/overload) in PHP 4. Mainly because it did not work when I toyed with. I think it should be safe to remove aggregate from PHP 5.
-- Sebastian Bergmann http://sebastian-bergmann.de/ http://phpOpenTracker.de/ Did I help you? Consider a gift: http://wishlist.sebastian-bergmann.de/