I added a changeset which fixes the original problem, mind someone might want to do a Sunit to confirm all this, I won't even say this works on intel. Thus taking the short integer of 16 bits with a bit pattern 0xFFFF and asking for it's unsigned results in 65535, not a signed value of -1 which any novice can see would be wrong.
Can't say I'll comment on where the primitive should be be, or called, or invoked etc... Or even if the users of this primitive asking for unsigned short won't now fall over and die... Mind I'll note that unsigned byte and unsigned long do the right thing and only an MacRGBColor, an FFI Type
On 29-Mar-06, at 5:06 PM, tim Rowledge wrote:
On 28-Mar-06, at 4:10 PM, tim Rowledge wrote:
This bug has been sitting around a while and should be killed. Putting aside the debate it lead to about where certain prims should go, what is the opinion on the original problem? Contributions to the actual mantis report please, rather than email.
If there are no substantive suggestions by the end of the week I intend to close this report and leave things as they are.
tim
tim Rowledge; tim@rowledge.org; http://www.rowledge.org/tim Klingon Code Warrior:- 1) "Behold, the keyboard of Kalis! The greatest Klingon code warrior that ever lived!"
-- ======================================================================== === John M. McIntosh johnmci@smalltalkconsulting.com 1-800-477-2659 Corporate Smalltalk Consulting Ltd. http://www.smalltalkconsulting.com ======================================================================== ===