Analyzing the "gs regression report - 2005-06-16", I run HEAD on peeves (using the binaries created by the nightly regression test). Got a segmentation faut. See console log below. Can't reproduce on Windows. A suspicious patch is http://ghostscript.com/pipermail/gs-cvs/2005-June/005567.html. login as: igor Password: Linux peeves 2.4.26 #4 SMP Thu Apr 29 14:37:16 PDT 2004 i686 GNU/Linux Last login: Tue Jun 14 08:25:51 2005 from mlc.fbc.spb.ru igor@peeves:~$ /home/regression/HEAD/bin/gs -I/home/regression/HEAD/lib -r300 - Z? -dNOPAUSE -dNOOUTERSAVE -dBATCH -sDEVICE=pkmraw -sOutputFile=out.ppm - dLastPage=1 /home/regression/comparefiles/244-01.ps AFPL Ghostscript 8.52 (2005-04-19) Copyright (C) 2005 artofcode LLC, Benicia, CA. All rights reserved. This software comes with NO WARRANTY: see the file PUBLIC for details. StartTestModule. Loading NimbusSanL-Bold font from /home/regression/gshead/share/ghostscript/fonts/n019004l.pfb... 2205032 812467 1538036 242797 0 done. Loading NimbusSanL-ReguItal font from /home/regression/gshead/share/ghostscript/fonts/n019023l.pfb... 2205032 856202 1538036 243401 0 done. Loading NimbusSanL-Regu font from /home/regression/gshead/share/ghostscript/fonts/n019003l.pfb... 2205032 896232 1538036 243998 0 done. StartOneTest. Segmentation fault igor@peeves:~$
Now I've got an error message on Windows : ***FILE*** H:\AuxFiles\comparefiles\244-01.ps AFPL Ghostscript CVS PRE-RELEASE 8.52: f:\afpl\gs-hd\src\gsalloc.c(770): gs_image_cleanup_and_free_enum: freeing 0xc28a38, not owned by memory 0x3e4b40! Will analyze so far.
Patch to HEAD : gsimage.c rev 1.14 gsimage.h rev 1.7
Oops, the comment #2 appears to mention a wrong revision number, and later we committed another patch. So here is a corrected summary : Patch 1 to HEAD : gsimage.c rev 1.14 gsimage.h rev 1.8 Patch 2 to HEAD : gsimage.c rev 1.15 gsimage.h rev 1.9