Folks -

We're ironing out some wrinkles involving 3D, endiannesss and BitBlt.  Please stand by (or jump in and join the fun).

	- Dan
--------------------------
4038geeMailOffset-raa -- Bob Arning -- 17 May 2001
Fix a problem where morphs dropped in a GeeMail were sometimes moving to the right unexpectedly"

4039RecompileUpdateCS-AFi -- Alain Fischer -- 17 May 2001
Changing definition of a class force a compilation of the class
and the metaclass methods through Behavior>>recompile:from:
With the old method, the current ChangeSet is not updated which
could let some obsolete reference in the system causing problems
when you try to save a project.
The new method update the current ChangeSet in a way similar to
the one used when you accept change to a method in a code pane."

4040FixBitBltEndianess -- Andreas Raab -- 17 May 2001
This change set fixes the determination of endianess from within BitBlt. Rather than assuming that all non-Bitmap bits are little endian pixels (which leads to problems when Form>>hackBits: is used to perform pixel hacking) we now check for NEGATIVE depth values of forms. This has a variety of advantages, such as giving *explicit* information about the format of the form, allowing for transparent pixel access through a Form (depending on the endianess the shifts are either left or right aligned so the sign of the depth tells us which way to go) and also allows extending various support primitives (such as #supportsDisplayDepth:) to deal with little endian formats more easily.
Note that this CS deals *exclusively* with the BitBlt related side of the issue - support for the image side hasn't been written yet."

4041LSBForms-ar -- Andreas Raab -- 17 May 2001
Display
Make forms capable of dealing with little endian pixels. Whenever we ask a form about its depth we will return the 'absolute' depth but internally we may use negative depths to indicate little endian pixels."