See Ubuntu bug report https://bugs.launchpad.net/bugs/160203 ----------------------------------------------------------------------------- Ghostscript in Gutsy (and Hardy) gives wrong results for the Bounding Box for most EPS files I've tried. It used to work fine with gs-gpl in Feisty. One simple example of the failure is the following (using the example.ps file attached): gs -sDEVICE=bbox -dNOPAUSE -dBATCH example.ps the result is wrong. Reverting back to a previous version is hard, since the package changed its name in Gutsy. Example.ps file's content: ---------------------------------------------------- %!PS-Adobe-2.0 %%Title: example.ps %%Date: Sat Nov 3 20:58:13 WET 2007 %%BoundingBox: 52 97 267 163 %%Page: 1 1 0.8 0 0 setrgbcolor 1 setlinewidth 100 15 160 {100 exch moveto 120 0 rlineto stroke} for 0 setgray 0.3 setlinewidth newpath 140 130 3 0 360 arc gsave 0.7 0.6 0 setrgbcolor fill grestore stroke /Times-Italic findfont 12 scalefont setfont 136 138 moveto (q) show 55 127 moveto (V ) show currentpoint 225 127 moveto (V ) show /Times-Roman findfont 12 scalefont setfont (menor) show moveto (maior) show showpage Here is another PostScript file that shows better the bug. The command "gs -sDEVICE=bbox -dNOPAUSE -dBATCH example2.ps" gives an empty bounding box, which is wrong: GPL Ghostscript SVN PRE-RELEASE 8.61 (2007-08-02) Copyright (C) 2007 Artifex Software, Inc. All rights reserved. This software comes with NO WARRANTY: see the file PUBLIC for details. %%BoundingBox: 0 0 0 0 %%HiResBoundingBox: 0.000000 0.000000 0.000000 0.000000 The gs-gpl package in Edgy used to give the correct result. (file example2.ps attached to original bug report) -----------------------------------------------------------------------------
This is likely a duplicate of Bug 689548 and has been fixed in the current head (r8363). We were just discussing yesterday how unfortunate it was that the Ghostscript version number didn't give revision information (PRE-RELEASE 8.61, isn't definitive, since that is what the r8165 through r8363 report). In any case, I'm working on finding a patch for this.
*** Bug 689548 has been marked as a duplicate of this bug. ***
Till, do you know that date or revision of the svn snapshot that was us to build the Ghostcript that is generating the bad bounding boxes? I'm having trouble finding it.
Ubuntu ships revision 8187, so that one has definitely the problem. See bug 689548 for whether the original poster observed the bug also in other revisions.
I'm not able to duplicate the bounding box problem with r8187, or any of the other various revisions that I've tested. I thought perhaps that the problem related to the fact that under Ubuntu 7.10 /usr/bin/gs is just a front end for /usr/lib/libgs.so.8.61, but I've tried building Ghostscript r8187 as a shared object ('make so', after adding "CFLAGS_SO=-fPIC" to the Makefile) and then running sobin/gsc, but that works as well. I suspect it's related to the build process.
The bounding box problem has turned out to be caused by the CJK support patches applied to the Ghostscript in Debian and Ubuntu. After rebuilding without the patches I get correct results. The patches will be dropped in the Debian and Ubuntu packages from Ghostscript 8.64 on, as Ghostscript 8.64 provides the CJK support natively. Ghostscript 8.64 will get released on February 1.