We're going to try this ASAP.  THANK YOU SO MUCH!!!  I *really* appreciate
your digging into this!  Swiki's have taken off here at Tech much faster
than I expected -- we have *FOUR* classes using them, with three other
classes requesting them, plus four more being used by various research
groups.  (Why? Ward's Wiki is just a marvelous idea, and Swiki is cheap and
easy to run.) I'm running out of server capacity!!!  Moving to NTs will
help alot.

Many thanks!
  Mark



>Mark,
>
>> But I also replicated the problem that Noel Rappin and Georg Gollman had
>> reported to me.  It served once, then died.  I stopped the server and
>> restarted.  It worked. Once.  After a few times, it stopped working even
>> after restart.  I restarted on a new port. It worked. Once.  I kept moving
>> around the port numbers trying to figure out what was going on, when I
>> stopped being able to access any files at all -- even the File List just
>> reported something like "I can't access that file right now" for *ANY*
>> file.  I quit Squeak, restarted Squeak, and then, suddenly, it worked.
>> Time after time.  No server restart.
>
>I've looked into your problem and I _think_ that I've found the problem.
>It seems that Socket>>primSocketSendDone: has to return false if the
>socket is no longer connected (see Socket>>waitForSendDoneUntil:).
>
>This leads to a problem in Socket>>sendData: since if the connection is
>closed by the client the process will hang there. You can try this very
>easily by starting the PWS as foreground process and press the reload
>button of your web browser twice. If you interrupt the PWS then you should
>end up in an endless loop in Socket>>sendData:
>
>For now, the easiest fix seems to be changing the return statement of
>Socket>>waitForSendDoneUntil: into
>
>  ^sendDone and:[self isConnected]
>
>I'll change it in the next VM version. Please let me know if this solves
>your problem.
>
>Hope this helps,
>  Andreas
>
>PS. I've tried Swiki on various Windows versions (Win95, NT3.51, NT4) and
>except from the above it worked quite nicely.
>
>--
>Linear algebra is your friend - Trigonometry is your enemy.
>+===== Andreas Raab ============= (raab at isg.cs.uni-magdeburg.de) =====+
>I Department of Simulation and Graphics      Phone: +49 391 671 8065  I
>I University of Magdeburg, Germany           Fax:   +49 391 671 1164  I
>+=============< http://isgwww.cs.uni-magdeburg.de/~raab >=============+

--------------------------
Mark Guzdial : Georgia Tech : College of Computing : Atlanta, GA 30332-0280
(404) 894-5618 : Fax (404) 894-0673 : guzdial at cc.gatech.edu
http://www.cc.gatech.edu/gvu/people/Faculty/Mark.Guzdial.html