[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: NSWindowController should be a NSResponder (part 2)
From: |
Fred Kiefer |
Subject: |
Re: NSWindowController should be a NSResponder (part 2) |
Date: |
Sat, 01 Nov 2008 02:03:10 +0100 |
User-agent: |
Thunderbird 2.0.0.17 (X11/20080922) |
Wolfgang Lux wrote:
> Fred Kiefer wrote:
>
>> Thank you for those follow up patches. I applied them with slight
>> changes. Please check whether I made any mistake here.
>
> Unfortunately, you did. The method -encodeWithCoder: always encodes
> the super class fields in the archive, even if the class version is
> set to 0 (this is the ``trick'' used by Gorm for creating backward
> compatible archives).
>
I did not know that we do such tricks. I am also not aware of other
places in the gui library where we do so. But if it helps for Gorm it is
certainly worth doing.
> W.r.t. the changes in NSApplication.m, I appreciate the introduction
> of an auxiliary method to check the responder chain of a window.
> However, I'd prefer the semantics of my patch, which checks the
> window's delegate immediately after the window itself and not at the
> end of the responder chain. This does not only comply with Apple's
> documentation -- cf.
> http://developer.apple.com/documentation/Cocoa/Conceptual/EventOverview/EventArchitecture/chapter_2_section_6.html#
>
> -- but also looks more sensible to me since the delegate will act
> on behalf of the window and is not a responder in its own right.
>
I just now looked at pictures on the Apple web side and you are right.
The delegate comes before the document and the window controller.
Feel free to apply both of these changes.