Postscript printer question

[Tim Writer]

lsorense-1wCw9BSqJbv44Nm34jS7GywD8/FfD2ys at public.gmane.org (Lennart Sorensen) writes:

> On Mon, Mar 13, 2006 at 11:40:39AM -0500, Madison Kelly wrote:
> > Hi all,
> > 
> >   I've got a bit of a problem... I've got a perl script that looks at a 
> > PS file that I print from OpenOffice2.
> > 
> >   The program (which I can't modify) reads the postscript and looks for 
> > '~' (tilde) and substitutes in values where they are found. I have tried 
> > creating a printer (the recommended printer is the Apple LaserWriter NT) 
> > and told to print in PostScript level 1. Then I print a given document 
> > to a file using this (dummy) printer.
> > 
> >   The problem is that when I look at the resulting PS file instead of 
> > getting '~' where appropriate I am getting (the ASCII?) '<7E>' instead. 
> > Has anyone run into this before? If so, any idea how to tell it to print 
> > the tilde instead of the ASCII?
> > 
> >   As an aside, when the same printer in PS level 1 is used on MS' 
> > Office the result is the tilde and I need to mimic this behavior.
> 
> Well after a quick look, if I did something like:
> 
> echo "abc ~ 123" | a2ps -o a2pstemp.ps
> 
> I get something which contains the code:
> x0 y0 moveto
> (abc ~ 123) p n
> 
> Which seems reasonable.
> 
> Do the same from open office with print to file and you get:
> 162 131 moveto
> 0 setgray
> (TimesNewRomanPSMTHGSet1) cvn findfont 50 -50 matrix scale makefont
> setfont
> <616263207E20313233>
> 
> Where the stuff in <...> is the hex codes for the stuff that was ascii
> in (...) above.
> 
> Of course this probably means it is reasonable to look for <...> lines
> in the output from OOo parsing 2 characters at a time, looking for 7E
> and replacing it with the hex encoding of the desired string, which
> before you just had to look for ~ in (...) lines and replace that.

The PostScript Red Book has this to say:

    A hexadecimal string consists of a sequence of hexadecimal
    digits (09 and either AF or af) enclosed within < and >. Each
    pair of hexadecimal digits defines one character of the
    string. White-space characters are ignored. If a hexadecimal
    string contains characters outside the allowed character set,
    a syntaxerror occurs.  Hexadecimal strings are useful for
    including arbitrary binary data as literal text.

So, while this is legal behaviour on the part of OOo, I would argue that it
departs from the intent as your example uses plain text.

In case you decide to implement a scanner, be aware of this:

    If the final digit of a given hexadecimal string is missing in
    other words, if there is an odd number of digits, the final
    digit is assumed to be 0. For example, <901fa3> is a
    3-character string containing the characters whose hexadecimal
    codes are 90, 1f, and a3, but <901fa> is a 3-character string
    containing the characters whose hexadecimal codes are 90, 1f,
    and a0.

There's no reason for OOo to escape ~ as, according to the Red Book, only
(, ), <, >, [, ], {, }, /, and % are special. Having said that, PostScript
uses <~ and ~> as delimiters for base 85 strings.

> Or you can try and figure out how to make OOo not use the hex encoding,
> although I suspect it has to do with unicode and other multilanguage
> support and the fonts it uses, where perhaps on windows you are getting
> a simpler ascii only version.  Even setting OOo to postscript level 1
> did nothing to change the output.

It could be as simple as playing with your LANG or LOCALE environment
variables.

-- 
tim writer <tim-s/rLXaiAEBtBDgjK7y7TUQ at public.gmane.org>                                  starnix inc.
647.722.5301                                      toronto, ontario, canada
http://www.starnix.com              professional linux services & products
--
The Toronto Linux Users Group.      Meetings: http://tlug.ss.org
TLUG requests: Linux topics, No HTML, wrap text below 80 columns
How to UNSUBSCRIBE: http://tlug.ss.org/subscribe.shtml