Relay-Version: version B 2.10 5/3/83; site utzoo.UUCP
Path: utzoo!mnetor!uunet!seismo!rutgers!sri-spam!ames!amdcad!sun!imagen!atari!apratt
From: apratt@atari.UUCP (Allan Pratt)
Newsgroups: comp.sys.atari.st
Subject: Re: desktop bug
Message-ID: <822@atari.UUCP>
Date: Thu, 27-Aug-87 13:17:57 EDT
Article-I.D.: atari.822
Posted: Thu Aug 27 13:17:57 1987
Date-Received: Sun, 30-Aug-87 05:08:24 EDT
References: <386@nikhefh.UUCP>
Organization: Atari Corp., Sunnyvale CA
Lines: 31

in article <386@nikhefh.UUCP>, t19@nikhefh.UUCP (Geert J v Oldenborgh) says:
> 
> 
> [...] do a copy from one window to another one, preferably to 
> or from floppy.  Impatiently start clicking the closebox of the active 
> window repeatedly.  When the copy is completed BOTH windows will 
> disappear...
>                                          Geert Jan van Oldenborgh.

GEM is notoriously bad at mouse-ahead (analogous to type-ahead).  In this
case, it buffers up two close messges, and assumes they are for the topmost
window each time.  Since after the first one is processed the inactive 
window is now the top window, it closes, too.

The moral of the story is this: GEM has 0 to 1 event's worth of mouse-ahead.
More than that and unpredictable things happen.

I have a worse problem for you: it's most obvious when your double-click
timer is at its longest setting.  Click someplace, then move the mouse.
The click is registered at the END of the double-click timeout, and it
is registered at the place where the mouse is at the timeout, not where
the mouse was when the click actually took place.

Considerations like this are what make GEM a drag compared to a real
windowing system like the Xerox Development Environment (XDE).  My two
summers at Xerox spoiled me.

/----------------------------------------------\
| Opinions expressed above do not necessarily  |  -- Allan Pratt, Atari Corp.
| reflect those of Atari Corp. or anyone else. |     ...lll-lcc!atari!apratt
\----------------------------------------------/  "Yow! Am I interfaced yet?"