[gs-bugs] [Bug 691424] Issue with multiple ghostscript processes running at a time

bugzilla-daemon at ghostscript.com bugzilla-daemon at ghostscript.com
Tue Jul 6 15:59:14 UTC 2010


--- Comment #6 from Ray Johnston <ray.johnston at artifex.com> 2010-07-06 15:59:06 UTC ---
The customer wrote (in email on 7/2):

I think I have finally found the crucial point.
Now I am able to reproduce an AV with the *original* gs 8.71 sources and
the original fonts without any private customer changes.

I unpacked the GS archive into d:\v\gs8.71org and the font archive into
The *only* change is in psi\msvc32.mak, line 660:
I changed
Without this change no error occurs.

I have created a debug build with Microsoft Visual C++ 2008.

Here are the command lines:
gswin32c.exe -P- -dBATCH -dNOPAUSE -Z: -r50x50 -sOUTPUTFILE=$.tif
-sDEVICE=tiff24nc -dDEVICEWIDTH=1408 -dDEVICEHEIGHT=1170 -dFirstPage=1
-dLastPage=1 -dNOCIE -dNOPSICC -dUseCropBox -d.IgnoreNumCopies=true -c
(56957.pdf) run

Here is the output (copied from the console):
% Init started, instance 0x003F4CA8, with args: -dDisplayFormat=198788
Resolution=96 -P- -dBATCH -dNOPAUSE -Z: -r50x50 -sOUTPUTFILE=$.tif
24nc -dDEVICEWIDTH=1408 -dDEVICEHEIGHT=1170 -dFirstPage=1 -dLastPage=1
dNOPSICC -dUseCropBox -d.IgnoreNumCopies=true -c (56957.pdf) run
Artifex Ghostscript 8.71 (2010-02-10)
Copyright (C) 2010 Artifex Software, Inc.  All rights reserved.
This software comes with NO WARRANTY: see the file PUBLIC for details.
% Start time = 0.281, memory allocated = 2522312, used = 2347563
Processing pages 1 through 1.
Page 1
[:]width=1407, band_width=1407, band_height=134, nbands=9
[:]width=1407, band_width=1407, band_height=134, nbands=9
[:]Beginning compression
% Outputpage start time = 156.125, memory allocated = 5739628, used =
[:]clist_end_page at cfile=-1476326238, bfile=32544
[:]Number of raw buffers allocated=65
And the program crashed with an AV, displaying a box which asked whether I
should send message this to Microsoft.

I could reproduce this error on my colleague's PC.
This error occurs for every run.

The division by zero mentioned in other messages could have been only a
symptom of this error.


The problem here was that the gxclmem.c that implements the memory based
clist was never made to handle logical (uncompressed) file sizes over 2Gb.

I will attach the patch for that and the file that generated 2.8Gb of clist
that the customer provided.

Configure bugmail: http://bugs.ghostscript.com/userprefs.cgi?tab=email
------- You are receiving this mail because: -------
You are the QA contact for the bug.

More information about the gs-bugs mailing list