Christopher Barker wrote:
> Paul McNett wrote:
>> So, my patch has bracketed code based on platform. I humbly attach it 
>> for review.
> 
> I've finally gotten around to applying your patch. Thanks.

Cool!


> I don't have any non-24 bit systems to test on, so it would be nice if 
> you could test it -- it's in current SVN (the wxPython 3rd party one...)

I tested it before submitting on:

Linux: 8, 16, 24, 32 bit
Mac: 32 bit
Windows: 8, 16, 24, 32 bit

Obviously the lower the bit rate the fewer objects can be uniquely identified...


> A couple thoughts:
> 
> 1) I wonder if we should bypass the colormatch code on 24bit systems -- 
> it's a bit intensive to be run all the time.

Do you think it matters? It isn't necessary to create that pixel and then read 
in 
that pixel for 24 or 32 bit, as the next rgb value the generator yields will 
work, 
however I doubt it is slowing things down enough to worry about.


> 2) I think we should re-factor this into a class -- when I see "global" 
> I get itchy! It uses a generator now because when generators where new I 
> was trying to find a use for it, and it was simpler then anyway.

+1 on the class idea. I don't like globals either, although it is pretty 
explicit as 
written.

Paul

_______________________________________________
FloatCanvas mailing list
[email protected]
http://mail.mithis.com/cgi-bin/mailman/listinfo/floatcanvas

Reply via email to