|
From: | Richard Frith-Macdonald |
Subject: | Re: KeyValueCoding compatibility issues |
Date: | Fri, 30 Nov 2007 07:08:41 +0000 |
User-agent: | GNUMail (Version 1.2.0) |
On 2007-11-30 06:39:42 +0000 David Ayers <address@hidden> wrote:
David Ayers schrieb:But even if the define is set, I'm unsure that the runtime checks will lead to the expected results. Just because the method in NSObject isn'toverriden by a subclass doesn't necessarily mean it expects the newbehavior. Most EO (and other KVC traversal) classes I deal with do notoverride the KVC primatives.So I understand the intention but I don't think we are reaching the result.
You may be right (I can't claim to be sure what happens in all cases)... but I think this code is an improvement over the previous version (which also tried to be backward compatible, but missed the cases added by Marcus of course). My impression is that mostly the new KVC behavior is compatible with the old behavior anyway, so what needs looking at are any cases where it isn't.
Ideally we would update other frameworks to work with the newer API, but where things haven't been updated, we want to maximisae backward compatibility, and this is an improvement to that.
[Prev in Thread] | Current Thread | [Next in Thread] |