[tex-live] [slc at topo.math.u-psud.fr: probable bug]

[Thomas Esser]

A suggested fix for the config file of dvipdfm.

Thomas

----- Forwarded message from Laurent SIEBENMANN <slc at topo.math.u-psud.fr> -----

To: slc at topo.math.u-psud.fr, te at dbs.uni-hannover.de
Subject: probable bug
From: Laurent SIEBENMANN <slc at topo.math.u-psud.fr>
Date: Wed, 17 May 2006 03:42:49 +0200
X-UIDL: fF&#!![E"!c3g"!F$2"!
X-Spam-Status: No, hits=0.000000 required=0.900000



Hi Tom,

Here is a probable bug of your dvipdfm distro.

Cheers

LS



| Newsgroups: comp.text.tex
| From: "Laurent S." <postel at topo.math.u-psud.fr>
| Date: Wed, 17 May 2006 03:38:01 +0200
| Subject: Re: dvipdfm graphics problem
| Organization: Guest of ProXad - France
| Message-ID: <postel-1705060338010001 at mar92-5-82-225-147-180.fbx.proxad.net>
| References: <r6r72uejoa.fsf at skye.ra.phy.cam.ac.uk>
| 
| 
| 
|    *** Re: dvipdfm graphics problem
| 
| Hi dvipdfm users,
| 
|      The bug I have been hunting is a system-specific dvipdfm 
| implementation bug -- of teTeX in my case -- rather than one 
| that concerns dvipdfm itself. I suspects it bites teTeX users of 
| dvipdfm for all sufficiently big eps objects.
|  
|      Of the 3 suggested fixes the replacement of "-sPAPERSIZE=a0"
| by "-dEPSCrop" is the only right one I believe.
| Here is my revised "config" file for dvipdfm.
| 
| %%% ----------cut
| % "config" file for dvipdfm, prepared for teTeX by Thomas Esser.
| % For more info, read the dvipdfm manual (texdoc dvipdfm) and
| % the original file which is provided as README.config.
| 
| % Distiller config. Uses ghostscript and works on compressed and
| % uncompressed files.
| % D "zcat -f %i | gs -q -sPAPERSIZE=a0 -sDEVICE=pdfwrite -dCompatibilityLevel=1.2 -dUseFlateCompression=true -dSAFER -sOutputFile=%o - -c quit"
| %%%%%%%%%%%%%% begin bug comment
| % Laurent S. & Heiko Oberdiek bugfix of May 2006 on comp.text.tex:
| % Any fixed papersize as in above line will cause unwanted clipping of very big objects, 
| % so -- following a (version 8.50)  gs log suggestion spotted by Heiko Oberdiek 
| % -- use  "-dEPSCrop" instead of  "-sPAPERSIZE=a0".  
| % The other gs suggestion "-dEPSFitPage" causes unwanted downscaling. Here is the revised line:
| %%%%%%%%%%%%%% end bug comment
| D "zcat -f %i | gs -q -sDEVICE=pdfwrite -dEPSCrop -dCompatibilityLevel=1.2 -dUseFlateCompression=true -dSAFER -sOutputFile=%o - -c quit"
| 
| % Set default paper size here
| p letter
| 
| % The "g" option specifies the amount to "grow" annotations by
| % Many TeX macro pages set the annotation bounding box equal
| % to the TeX box that encloses the material.  That's not always
| % what you want
| g 1
| 
| % PDF version stamp to use in output file and max version of files
| % allowed to be included (2 for 1.2, 3 for 1.3)
| V 2
| 
| % Font map files; maintained by updmap(1)
| f dvipdfm.map
| %%% ----------cut
| 
| On my institutional linux machine I can simply place the above 
| patched config file in the same directory as the dvi files for 
| processing by dvipdfm.
| 
| Here is my current TeX test file.
| 
|   %%%%------cut
|   %%% fig2-ab_ptest.tex, needs fig2-ab.eps
|   %%% use Plain TeX; then process the ".dvi" output with dvipdfm
|    %% or compatible extension (such as dvipdfmx ?)
|   \output{\shipout\box255}
|   \noindent\null
|   \vfill
|   \hrule
|   \vskip 3pt
|   \hrule
|   \vbox to 407.6bp
|     {
|     \vfill
|     \hrule width \hsize
|     \vskip 3pt
|     \hrule
|     \vfill
|     }
|   \special
|      {PSfile=fig2-ab.eps llx=0 lly=0 urx=3609 ury=4076
|           rhi=4076}
|   \hrule
|   \vskip 3pt
|   \hrule
|   \vfill
|   \eject
|   \end
|   %%%%------cut
|   
| The complete example is posted at:
| 
|       http://topo.math.u-psud.fr/~slc/dvipdfm_puzzle
|   
|      My sincereest thanks to Heiko Oberdiek and Sanjoy Mahajan 
| for the key suggestions. 
| 
|      Which implementations of dvipdfm are affected by this bug?
| 
|      Cheers,
| 
|            Laurent S.
| 

----- End forwarded message -----