Re: [ePiX-users] bounding box problems on linux amd64

[P. T. Kornman]

On Wed, 2007-09-26 at 11:58 -0400, Andrew D. Hwang wrote:
> On Tue, 25 Sep 2007, P. T. Kornman wrote:
> 
> > This probably an os problem, but I thought I would ask the group. I
> > recently started using Debian testing amd64 on an intel core2 duo
> > notebook. I have noticed that the picture sizes are now page sized
> > rather than the requested smaller size. An example would be Andy's
> > calculus.xp in sample subdirectory of the distribution. This problem has
> > occured in 1.0.24, 1.1.18 and the latest 1.2.0 release. This problem
> > does not appear on the 32 bit Debian Testing. Any help/hints/suggestions
> > would be appreciated.
> >
> Internally, a file undergoes the conversions:
> 
> xp -[epix]-> eepic
>     -[latex]-> dvi
>     -[dvips]-> ps
>     -[ps2epsi]-> epsi
>     -[sed]-> eps
> 
> Here are some ideas to isolate the problem.
> 
> * Compare calculus.eepic compiled on amd64 with that compiled on a 32 bit.
>    They should only differ in the time stamp (first line of the comment
>    block); if that's the case, epix isn't at fault. :)
> 
> * On each platform, compile the eepic file with "elaps -ps" and compare
>    the BoundingBox of the PostScript file with that in the eps file from
>    "elaps". They should be different. If they're not, ps2epsi looks guilty.
>    (elaps' use of sed doesn't touch the BoundingBox.)
> 
> * Compare your ps2epsi scripts on the two platforms (by diff-ing). For
>    some reason, one-byte errors near the bottom of the file seem to creep
>    in with regularity.
> 
>    If the scripts differ, try copying the 32 bit machine's ps2epsi to the
>    amd64 and compile the eps file manually from the PostScript file:
> 
>    ./ps2epsi calculus.ps calc_test.epsi
> 
>    Compare the BoundingBox with that in calculus.eps from elaps. If they're
>    the same, ps2epsi seems not be the problem; I'd then suspect dvips.
> 
>    Repeat, copying the amd64's ps2epsi script to the 32 bit machine.
> 
> Hope that's helpful,
> 
> Andy
> 
> Andrew D. Hwang                       address@hidden
> 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)
> 
> 
> _______________________________________________
> ePiX-users mailing list
> address@hidden
> http://lists.nongnu.org/mailman/listinfo/epix-users
> 
Andy you are good. You nailed it - the diff against the 32 bit ps2epsi
showed the problem. The simple replacement you suggested fixed my
problem. Many thanks for the help and a trully outstanding and
indispensible piece of software.
ptk