Re: .89 error

[Stan Shebs <shebs@cygnus.com>]

   From: "Anthony Veale'" <veale@casa.colorado.edu>
   Date: Tue, 18 Feb 1997 12:49:06 -0700 (MST)

   I added in the lines defining OPENWINHOME to the mh-solaris config
   file so that I could compile 7.1.89 on my solaris 2.5.1 machine.
   (Taking the hardwired X11 location out is a "good thing" but there
   should also be a mechanism for giving that information to 'configure'.
   If it's already there I couldn't find it.  I'm not using Imake.)

Cygnus configure has a special option to point at X stuff, but I'll
have to find a release engineer and ask them what it is.

   But that's not the main point.  During my first test game today, it
   was cruising along for about 40 turns, then ended with:

     Error: attempting to xform -1,-1.
     Saving the game... done.

   Any advise for trying to dig out the cause?

It might be too late for that.  You could try starting the saved game
while under the debugger with a break set at run_error, but the error
is related to map drawing, and it can be hard to replicate the combo
of button clicking and movement that caused the problem.  I've been
cleaning up various accumulated probs in the X11 code, so it may be
fixed by now.

I've been thinking about how to catch this kind of problem, and
realized that while most runtime errors are caused by game design
bogosity or exceeding memory limits, the xform message is always
indicative of an actual bug in the C code, and Xconq (at least under
Unix) ought to drop a core.  I'll change the next snapshot to do that.

   Note: I haven't installed the new libs yet.  Are there changes to the
   standard game that could be the cause?

There are changes, but I doubt they would be connected.

								Stan