[tex-k] dvips(k) 5.95a BoundingBox

Andrew D. Hwang ahwang at mathcs.holycross.edu
Tue Aug 28 18:39:50 CEST 2007

On Tue, 28 Aug 2007, Akira Kakuto wrote:

> On Mon, 27 Aug 2007, Andrew D. Hwang wrote:

>> I'm getting an unexpected BoundingBox in dvips(k) 5.95a. The attached tex
>> file seems to be minimal. When I convert to eps via:
>> latex clip_err.tex
>> dvips -f -o clip_err.ps clip_err.dvi
>> ps2epsi clip_err.ps
>> mv clip_err.epsi clip_err
>> dvips(k) 5.95a generates %%BoundingBox: 164 608 237 609
>> dvips(k) 5.92b generates %%BoundingBox: 128 572 273 717
>> (The latter is as expected.) I haven't checked carefully, but it looks as
>> though 5.95a catches characters, rules, and black ink while 5.92 finds all
>> non-white ink.
> I think that is a feature of "ps2epsi":
> dvips 5.96dev and ps2epsi gives
> %%BoundingBox: 168 608 241 609
> while
> dvips 5.96dev and ps2eps (Ver. 1.61 by Roland Bless) gives
> %%BoundingBox: 132 572 276 717
Dear Kakuto-san,

Thank you for the speedy reply. It occurred to me that ps2epsi was the 
culprit, but the problem persists if I use a "known good" ps2epsi with 
dvips(k) 5.95a, so I assumed (perhaps incorrectly?) the problem was 
elsewhere. In case it's relevant, here's my detailed setup.

The first machine runs GNU/Linux (Mandrake 10.0); dvips(k) 5.92b (the 
original binary RPMs, compiled with gcc-3.3.2). On this machine, the 
BoundingBox is "as expected".

The second runs Mac OS X (Tiger); dvips(k) 5.95a (compiled with Apple's 
port of gcc-4.0.1 from the most recent fink package, ghostscript 8.51-1).
Here, the BoundingBox is "thin".

The GNU/Linux machine's copy of ps2epsi creates a thin BoundingBox on the 
Mac. The Mac's copy of ps2epsi runs with an error on the GNU/Linux machine 
("sed: -e expression #1, char 42: Invalid range end"), but does write the
"expected" BoundingBox.

Since I'm writing software reliant on external ps to eps conversion, I'd 
prefer to use the widely-available ps2epsi than Roland Bless' (arguably 
superior) ps2eps if at all possible. Thank you for the suggestion though, 
and again, thank you for your help.


Andrew D. Hwang			ahwang at mathcs.holycross.edu
Department of Math and CS	http://mathcs.holycross.edu/~ahwang
College of the Holy Cross	(508) 793-2458 (Office: 320 Swords)
Worcester, MA, 01610-2395	(508) 793-3530 (fax)

More information about the tex-k mailing list