Re: 3.4 release

Ok, good, lets regress the change to the mpeg source, it adds a bit of  
complexity to the mpeg
code that's not really needed. So I'll attempt to back out the changes  
tomorrow if I figure out how to
do that (pointers are welcome).

On Thursday, February 6, 2003, at 03:31  PM, Andreas Raab wrote:

> John,
>
>> Well I'm afraid this change set was forgotten when we were building
>> 3.4. Mostly I was looking for feedback from the Croquet folks if
>> this actually works for them as intended.
>
> Just to reconfirm: This change was the one which made the alpha values
> appear correctly, right?! If so, we've actually solved this problem
> differently in the meantime because the MPEG player wasn't the only  
> place
> where the alpha channel was killed - it turns out that this happens
> regularly when you blt 16bpp -> 32 bpp and considering both overdraw  
> issues
> and the generally higher requirements in terms of computing power in  
> Croquet
> I decided that it's worthwhile to add an extra pass where it is needed  
> (done
> by BitBlt's new fixAlpha:with: rule). In other words, if it's  
> expensive, it
> is not necessary to keep it in the MPEG plugin for the sole sake of  
> Croquet.
> We could simply #ifdef it with an appropriate comment saying what this  
> is
> doing and what is needed if you want to use it.
>
> Cheers,
>   - Andreas
>
> _______________________________________________
> Squeakfoundation mailing list
> [email protected]
> http://lists.squeakfoundation.org/listinfo/squeakfoundation
>
>
--
======================================================================== 
===
John M. McIntosh <[email protected]> 1-800-477-2659
Corporate Smalltalk Consulting Ltd.  http://www.smalltalkconsulting.com
======================================================================== 
===