Re: NSView bounds transformations

[Richard Frith-Macdonald]

On 12 Nov 2008, at 08:10, Fred Kiefer wrote:

> As I wrote I am currently working on that area again and this time I
> think I understood the whole idea behind all these frame and bounds
> transformations (OK, not for bounds size, but all the rest).
> Before I go on and make some more changes I would like to discuss the
> next concept here. What I want to change is the frame transformation,
> these should no longer try to adjust the _boundsMatrix when this is
> present. We should keep on changing the _bounds directly, as long as
> there isn't a _boundsMatrix, but when that is present,w e should just
> compute the new bounds with the inverse transform and leave the
> transform itself unchanged. Does this sound correct to you?

I can't have much input here as I never do this sort of graphical  
programming.  What you say sounds reasonable, but I think the real  
answer is to figure out what the possibilities are, then devise a test  
program to to show what is happening and run it on MacOS-X to get a  
definitive answer.

> One other point, I would like to extend the test suit with more
> transformation test (mixed ones). For this I will need the correct
> expected values first and this would be fairly easy if I were able to
> run the test on an Apple machine. So now the question, how would I
> achieve that? What is needed to run our test suite on an Apple machine
> against their frameworks?

I have built and run the testsuite for the base library on MacOS-X  
using the Apple Foundation, so it's certainly possible to do that  
(though obviously some tests for gnustep specific features need to be  
omitted by putting the correct #ifdef in the test programs etc)  I  
don't recall anything very special being needed - just gnustep-make  
configured for apple-apple-apple.