Hello, I'm Crashing....
- To: mathgroup at smc.vnet.net
- Subject: [mg8894] Hello, I'm Crashing....
- From: Eric Maiken <maiken at src.sony.co.jp>
- Date: Thu, 2 Oct 1997 22:56:46 -0400
- Organization: SONY
- Sender: owner-wri-mathgroup at wolfram.com
Hello Group:
This is my first day on the list, but I've lurked for awhile trying to
catch up on other's experiences with v3.0.1. The rest of this post is a
long-winded plea for help, involving a problem that is likely NOT the
fault of mathematica-- it seems to be either a graphics card or windows
problem. Hopefully, someone can set me straight!
I've used v2.2.3 successfully for a few years, and thought I was making
a clean conversion to my newly purchased copy of v3--until....
I started experiencing windows crashes, of various degrees of severity,
instigated by all of the varieties of Log plots found in the
Graphics`Graphics` package (other--non Log scale--plot types and
packages seem to be fine). My frustration lies partially in the fact
that these plot calls live inside of programs written in the v3
environment, which worked just fine--until a few days ago....
The problem seems to involve the system's graphics display, and provokes
EITHER
one of the red-x windows message boxes that says:
*******
*MATHEMATICA caused a general protection fault
in module GDI.EXE at 0026:000001da.
Registers:* (a list of 32 bit addresses is given)
*Bytes at CS:EIP:* (a list)
*Stack dump:* (a list)
********
OR
a dos style screen with a message:
*A Fatal Exception ...Has Occurred....*
Occasionally, as soon as I click the short-cut icon to start the
program, I get the fatal exception crash.
The GDI.exe is windows95's "Graphics Device Interface Core Component,"
whatever that means....
I've a Diamond STealth 3D 2000 (4MB) Accelerator card, which card's
driver I've set for no GDI Acceleration, and simultaneously set my
windows display for no graphics acceleration. Still, the LogPlot
inspired crashes....
There are no fancy screen savers, games or Japanese-language-anything
(ref my sig) on my new P5-200Mhz, UltraWide SCSI, 64MB RAM on which, the
v3.0 software lives.
Of course, I've deinstalled/reinstalled win95, Mathematica, and other
industrial software, looked for device conflicts, etc, etc, etc, to no
avail.
For the record, here's a simple example (variations work just as well)
of how to cause a (documented ;> ) crash:
Immediately after firing the system and mathematica up:
*****************
In[1]:=$Packages
Out[1]={"Global`","System`"}
In[2]:=Needs["Graphics`Graphics`"]
In[3]:=$Packages
Out[3]={"Utilities`FilterOptions`","Graphics`Common`GraphicsCommon`",
"Graphics`Graphics`","Global`","System`"}
In[4]:=Context[LogPlot]
Out[4]="Graphics`Graphics`"
Clear[f]
f[x_]:= x
LogPlot[f[x],{x,1,10}]
***********************
*MATHEMATICA caused a general protection fault..., blah, blah, blah....*
The frontend shuts down, the kernel locks up (but seems to remain
resident),
Norton "CrashGuard" doesn't even blink.
If I de-install, then walk over to another win95, english dos (cf the
.sig), and run the same code off the cd, all works fine.
Help Please!,
Eric
--
Eric Maiken, PhD
SONY Corporation Research Center Phone: 81 45 353-6845
Center For Environmental Technology Fax: 81 45 353-6906
174 Fujitsuka-cho, Hodogaya-ku email: maiken at src.sony.co.jp
Yokohama 240 EricMaiken at aol.com
Japan