Re: Bitmap code - 3rd alpha version

[Pete French]

> Sadly my first test with the nsimagecell.app in examples\gui\Old failed,
> which may not so much be a problem of the new code. My current

Havent look at this one - will do so tomorrow.

> impression is that a lot of code in the gui will need some rewrite to
> catch up with this improvement. So far I found that the NSImage does not
> propagate its size change to the NSImageReps and also does not
> invalidate the cached reps, as it now should (when this flag is set).

*pause* - O.k., obviously there is a lot of stuff going on above
the level I normally use images at that I hadnt considered. Can you
give me a specific example of what you mean ? (i.e. this that setSize in
NSImage doesnt propagate down - I never used anything above NSImageRep in
my code over all these years I am ashamed to say :_0 )

> Another problem that results from a recent change is that NSDrawBitmap()
> sets just the current point, which is no longer used when drawing. So it
> should now set (and later reset) the transformation, just as [NSImageRep
> drawInRect:] does.

Not quite sre what you mean by the current point it, but I am also comming
to the conclusion that preserving the atrix in the same way that drawInrect
does is probably a very good idea - this addresses the colourspace concerns
as well.

> There also a few minor concerns with the new code. Having to build up
> the dictionary and than decompose it again for the NSDrawBitmap() call
> looks like a lot of overhead. Why not write a better image draw method

I didnt change that - it always worked that way, so if things have slowed
down that isnt the reason. i too was surprised that it was done that way, but
I gather that its implemented like that to be more like the actual Display
Post Script call ?

-bat.