Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!husc6!bbn!oberon!ll-xn!ames!sdcsvax!ucbvax!ATHENA.MIT.EDU!swick
From: swick@ATHENA.MIT.EDU (Ralph R. Swick)
Newsgroups: comp.windows.x
Subject: Re: Configure/Expose event ordering.
Message-ID: <8710221356.AA27610@LYRE.MIT.EDU>
Date: Thu, 22-Oct-87 09:56:32 EDT
Article-I.D.: LYRE.8710221356.AA27610
Posted: Thu Oct 22 09:56:32 1987
Date-Received: Sun, 25-Oct-87 01:47:38 EDT
References: <2259@super.upenn.edu>
Sender: daemon@ucbvax.BERKELEY.EDU
Organization: The ARPA Internet
Lines: 19


    Date: 21 Oct 87 22:23:49 GMT
    From: super.upenn.edu!dsl.cis.upenn.edu!bradley@rutgers.edu  (John Bradley)

                   I've set the event mask to "ButtonPressMask | ExposureMask
       | StructureNotifyMask".  In my main loop I simply check for those three
    events.  Everything works well until I re-size the window (via uwm).  When I
    do, I get an ExposeEvent, followed by a Configure and another Expose.  This
    causes the program to redraw the contents of the window twice, the first time
    scaled wrong.

The first Expose comes from the un-mapping of uwm's resize pop-up window.
You could discover this by examining the bounds of the region in the event.
For purposes of uwm, this should be fixed as soon as save-unders are
implemented.  However, for general simple-minded clients that ignore the
region information in exposure events (and repaint their entire window),
it may be desireable to look ahead in the event queue for more exposures
(with XCheck*Event) before processing the current one.  Such clients can,
of course, ignore all exposure events with count>0.