|
From: | David Ayers |
Subject: | Re: [RFC] Header organization of -base & -gui |
Date: | Thu, 03 Jul 2003 12:31:58 +0200 |
User-agent: | Mozilla/5.0 (Windows; U; Windows NT 5.0; en-US; rv:1.4b) Gecko/20030507 |
Hello Richard,Thanks for jumping in!
Richard Frith-Macdonald wrote:
On Thursday, July 3, 2003, at 10:23 am, David Ayers wrote:and the following files should be part of -baseadd: preface.h GSFileHandle.(h/m)Do we really want to expose these? I'm not sure how feasible it is. Certainly we would need quite different versions of the code to work with the MacOS-X foundation.
Hmm... GSFileHandle does have some useful features, that don't seem to be part of the superclass' interface. If it's meant to be part of a class cluster, then fine. But looking at the header it seemed more than that. It also inherits from NSFileHandle which doesn't declare any ivars that could be gnustep specific. So I assumed, it would at least compile on OS X. I didn't expect it to use any features available on GNU/Linux, BSD and Darwin but not on OS X. But you know better what it does, so it's your call.
I wasn't sure about preface.h, but are you sure it shouldn't be public? Not that one should compile Tools/defaults on OS X, but this tool seems to use it (along with some headers on Source/Testing) If you really think those declarations should be private, then fine. If some belong in GSObjCRuntime maybe, then lets move them and then make it private.
GSLocal.(h/m)I don't know about this ... is it worthwhile and portable to MacOS-X?
Adam, could you say something about this? I had the impression that these where meant to be public. I wouldn't have a problem with them just logging error messages on OS X, if it's a hastle to make them portable.
GSUnion.h GSIArray.h GSIMap.hYes, these were always intended to be public.
Thanks for the verification. Cheers, David
[Prev in Thread] | Current Thread | [Next in Thread] |