Re: KeyValueCoding compatibility issues

[David Ayers]

[Markus asked me to post his replies to the list]

Marcus Müller schrieb:
> Am 29.11.2007 um 21:55 schrieb David Ayers <address@hidden>:
> 
>> Hello Markus,
>>
>>> Attached, find a patch that fixes all these problems. As an added
>>> bonus, all compatibility code is prepared for exclusion from
>>> compilation, making it very easy to actually migrate to the new-style
>>> KVC completely
>>
>>
>> I don't necessarily like #ifdef's as you mention there are some apps
>> which expect set of semantics while others expect the other set, but
>> there will generally be only one -base installed (per library combo).
> 
> 
> Agreed. The official stance seems to be mimicing Cocoa which deprecated
> this API, so preparing it to be excluded in a later release looks ok  to
> me.
> 
>> (This is why GDL2 replaces the implementations during +load)
> 
> 
> But that's a different story - GDL2 alters the behavior of the very 
> same API (AFAIK).
> 
>> I'm also interested into which concrete version of KVC this is  supposed
>> to be compatible with.
> 
> 
> The previous gnustep-base version. To the best of my knowledge, this  is
> also the way it is done in Apple's Foundation.
> 
>>  The reason I ask, is that if this matches WO45,
>> I may consider implementing these semantics as categories in GDL2.
>>
>> Maybe we should create a micro KVCvX Framework/Library/Bundle that
>> implements these categories when loaded without burdening -base.  And
>> I'd have GDL2 depend on that (or dynamically load that bundle though
>> that might be too late for some cases).
> 
> 
> Not sure if this is necessary, but it sounds more flexible.
> 
> Cheers,
> 
>   Marcus
>