Bill.Chiles@WB1.CS.CMU.EDU (10/24/88)
Would someone please post a theory behind graphics context's being associated with a drawables. I don't really see an efficiency issue here, but I don't know much about putting monitors or servers together. I would also like to know what is expected of a client when he creates graphics contexts. Are you supposed to have one GC per drawable? Can you make all your GC's on the root for convenience, even though you apply them with your windows? Can you have one GC for N windows because they all act the same? I recently discovered that if you are putting an image on a pixmap for the megapel display, then you must have a 1:1 mapping with the GC's, or you get some error I don't remember. I was doing this to make a cursor. Before, I was using one GC for both the cursor and mask pixmaps. It seems to me that a graphics context manifests all the information about how you like to do some graphics operation, modulo how you affect a request by passing specific arguments to some programmer's interface for that operation. If this is solely true, then there is no reason a graphics context should ever need a pointer to some drawable. It should be applicable with any drawable. If this is documented anywhere, I'd be glad to only receive a pointer. In case it is pertinent, I always deal with X through CLX. Bill