earlier today owen pointed me at this post to the xorg list by keith.
the comments in the post greatly clarify the intended semantics of the effect of window redirection on drawing and expose event delivery. the patch in the post also fixes a bug in x.org that has been bugging me for a little while.
bonus points on top of all of this: the fix allows me to remove a timer abuse from my code! (before the fix, my application wasn’t receiving events so i had to manually schedule refreshes of the window content).
keith: today you rock my world.
now all that is left is to (*hint*) get someone to look at the corresponding bug in gdk and to convince the ubuntu guys to include keith’s patch for feisty.