🚀 go-pugleaf

From article <426@laas.laas.fr>, by ralph@nastassia.laas.fr (Ralph P. Sobek):
" 
" Oh my god!!  Another version?!!  ...

Yes, well, my version is ready in case anyone is interested.
Here is a list of ways in which it _may_ differ from other
extant versions:

1. It uses pk or pxl font files, or a combination of the
two.  (Not gf files, though.)

2. It will use pk fonts with up to 256 characters (as one
finds in Silvio Levy's Greek fonts and JTeX, though I
haven't tested dvi2ps with JTeX yet).

3. PS fonts native to the printer are ok, too.  This dvi2ps
does not keep track of what fonts may or may not be
available inside the printer nor of character widths
in those fonts.  It just passes the info it finds in
the dvi file about names and positioning right through
to its output ps file.

4. The size of the output is somewhat smaller and PS
interpretation time somewhat faster than for the other
versions of dvi2ps that I've seen.  As suggested in Glenn
Reid's PS Program Language Design, characters are
collected into long strings with space characters and
submitted to widthshow.  (dvips 4.0 does a nice job
at optimizing PS output, but it doesn't do this particular
optimization.)

5. In calling dvi2ps, flags and flag arguments work in
a more standard way (e.g. `dvi2ps -hs -n10 some' is an
acceptable way to ask for 10 copies, no header, and
statistics, as well as `dvi2ps -h -s -n 10 some').

6. \special texts with no keywords are sent literally to the LW
(instead of being interpreted as a list of files to include).

7. There is no longer a header .ps file to be kept in a
library directory somewhere.  The header code is compiled
into dvi2ps.  The source for the header code, tex.ps,
may contain #ifdefs.

		------------
The thing is in 4 shar files, 161k.  Email me if you
want a copy.

			Greg, lee@uhccux.uhcc.hawaii.edu