|
From: | Fred Kiefer |
Subject: | Re: Patch to fix view rotation and hitTest: |
Date: | Sun, 13 Mar 2005 17:40:40 +0100 |
User-agent: | Mozilla/5.0 (X11; U; Linux i686; en-US; rv:1.7.2) Gecko/20040906 |
Banlu Kemiyatorn wrote:
This patch will remove translation information from _frameMatrix and will only create _frameMatrix when the frame is rotated and manually prepending the translation and prepend rotation only if _frameMatrix isn't nil (rotated). I've tested this patch with gsimageapp's diagramkit (my local tree) and it looks OK though, I am not sure.. actually.. I'll never be. Please revise it for me OK?
I looked at that patch, but I am surely no expert in that area. What I did not quite get is the intention of this patch. It seems to be fine, as far as I can tell and in many cases we save us the creation of yet another matrix. Is this the whole point behind the patch? NSAffineTransform is in many cases the class we create the most objects. (only notifications seem to be worse) So saving some should be worthwhile, but somehow I can only see this as a half hearted attempt on this. We really should try to find out where all this matrixes are coming from.
But then, I may have totally missed the point behind the patch.
BTW. I don't like that we load a few methods to NSAffineTransform. Is it a bad idea to subclass it for NSView?
We could get rid of some of them, because they are no longer used, or could be replaced easyly. But where do you see the problem? This is internal to the GUI framework, any additional categories shouldn't do much harm.
[Prev in Thread] | Current Thread | [Next in Thread] |