[pstricks] pst-geo

[Alan Ristow]

On Fri, 16 Jul 2004 15:24:06 +0200, Herbert Voss
<Herbert.Voss at alumni.TU-Berlin.DE> wrote:

>Manfred Braun wrote:
>> I have been following up the discussion on pst-geo, having the same 
>> problem as Alan Ristow.  In my case, however, the problem isn’t solved 
>> by inserting the right (absolute) path for /data. I still get the error 
>> message from GSview (version 4.3)
>> 
>> Error: /invalidfileaccess in --file--
>> 
>> Operand stack:
>> 
>>    (C:/Programme/TeXLive/texmf-local/pst-geo/data/river.dat)   (r) 
>> 
>> when displaying page 4 of doc-geo2d, although the path of river.dat is 
>> provided correctly.

I actually had the same problem and just used a relative path name for
purposes of compiling the documentation.

>seems to be a problem with the GSView. Create a PostScript
>file and view it with gv. Then it should work ...

AFAIK, there is no Windows version of gv. If I have time later today
I'll give it a try on a Solaris box, but I'm headed out of town and
probably won't be able to do it today. Then I'm Windows-only for the
next week....

I tried opening the file with the absolute path name using two different
versions of GSView, two different versions of Ghostscript, and the
commercial PS file viewer RoPS and had the same error from all. (Well,
RoPS only told me I had a stack underflow and provided no additional
diagnostic information, but the error occurred on the same page.)

AFAIK, the way Manfred and I specified our absolute path names is
perfectly legal Postscript, but something definitely seems amiss. RoPS
is stand-alone and does not make use of gs -- well, now that I try it,
files with relative path names fail under RoPS, too, so forget about
RoPS. Anyway, is there some trick to making Postscript recognize that
C:/ is a device? I'm not the world's greatest Postscript programmer, but
that's the first thing to occur to me -- that maybe the device name is
not being parsed correctly.

Alan