zjat02@apctrc.trc.amoco.com (Jon A. Tankersley) (04/27/89)
I ran into some problems with dvi2ps (March '89 and the previous version).
Seems that certain documents (proc.sty is all I've seen so far) can cause
some PostScript errors. I think that the real bug is in dvi2ps, but the
fix seems to not interfere with other documents. Case occurred when
a string of text started in the upper left-hand corner of the page.
A
-41 r(moveto) (string) s(how) was failing.
Doing a
-41 0 p[moveto] (string) s(how) worked.
@bop1 definition does an initgraphics. This sets the currentpoint to
unknown. Under certain circumstances, dvi2ps outputs rmoveto commands
immediately after the bop1 call. I modified tex.ps to do a 0 0 moveto
after the initgraphics. Anybody else seen this? Anybody see a problem
with this?
*** tex.ps Wed Apr 26 17:15:43 1989
--- tex.ps.old Wed Apr 26 17:16:03 1989
***************
*** 9,17 ****
% <holtz%cascade.carleton.cdn@ubc.csnet>
% June, 1985
% Last Modified: Aug 25/85
- % jat 26-Apr-1989
- % Modified @bop1 operator to do a 0 0 moveto to fix rmoveto bug with
- % some document styles. Dvi2ps may need to have the real fix.
% oystr 12-Feb-1986
% Changed @dc macro to check for a badly formed bits in character
% definitions. Can get a <> bit map if a character is not actually
--- 9,14 ----
***************
*** 173,179 ****
erasepage initgraphics
Mtrx setmatrix
/SaveImage save def() pop
- 0 0 moveto % JAT make sure we have a currentpoint!
} def
%-- tjh sept. 87: if this page has a mac drawing on it, we have to
--- 170,175 ----
-tank-
#include <std/disclaimer.h> /* nobody knows the trouble I .... */
tank@apctrc.trc.amoco.com ..!uunet!apctrc!tankpiet@cs.ruu.nl (Piet van Oostrum) (04/28/89)
In article <818@apctrc.UUCP>, zjat02@apctrc (Jon A. Tankersley) writes:
`I ran into some problems with dvi2ps (March '89 and the previous version).
`Seems that certain documents (proc.sty is all I've seen so far) can cause
`some PostScript errors. I think that the real bug is in dvi2ps, but the
`fix seems to not interfere with other documents. Case occurred when
`a string of text started in the upper left-hand corner of the page.
`A
` -41 r(moveto) (string) s(how) was failing.
`Doing a
` -41 0 p[moveto] (string) s(how) worked.
`
My version of dvi2ps initializes the position at the beginning of a page to
some impossible value, so it always starts with an absolute movement.
--
Piet van Oostrum, Dept of Computer Science, University of Utrecht
Padualaan 14, P.O. Box 80.089, 3508 TB Utrecht, The Netherlands
Telephone: +31-30-531806. piet@cs.ruu.nl (mcvax!hp4nl!ruuinf!piet)