- * XXX Other than clearing the Redisplay flag or not,
- * XXX we may as well rely on the INVOK_WCB() doing this
- * XXX pointer-check.
- * XXX
- * XXX If we do not invoke the display because the pointer
- * XXX is not defined (should never happen, really), then
- * XXX we may enter an infinite busy-loop trying to update
- * XXX the window. Similarly, if we skip because the window
- * XXX is not visible. However, if the window becomes visible
- * XXX at a later time, the window should get its callback
- * XXX invoked. I would recommend removing the first check,
- * XXX and making the second check only affect whether the
- * XXX callback is invoked---but always clear the flag, if
- * XXX the {window} pointer is defined.
+ * XXX REALLY shouldn't be done. GLUT docs state that this
+ * XXX isn't even processed immediately, but rather waits
+ * XXX for return to the mainloop. "This allows multiple
+ * XXX glutReshapeWindow, glutPositionWindow, and glutFullScreen
+ * XXX requests to the same window to be coalesced." (This is
+ * XXX having some deleterious effect on a sample program of mine.)
+ * XXX Not only does GLUT not flush at this point, GLUT doesn't even
+ * XXX *do* the reshape at this point! We should probably rip this
+ * XXX out and do what GLUT promises. It would be more efficient, and
+ * XXX might be more compatible.