[Info-vax] Ghostscript update for VMS
JOUKJ
joukj at hrem.nano.tudelft.nl
Fri Jul 8 04:00:23 EDT 2011
Mark Berryman wrote:
[snip]
>
> Do you have a version of GhostView that works with this version of
> Ghostscript? The one I could find, GV v3.5.8, does not work at all,
> either with the current GhostScript (it can't give GS a simple file
> name) or with VMS V8.4 (the file selection window does not paint
> anything until you mouse over where the files are supposed to be listed).
Seems I have the same version (I only have Alpha machines). It is "old"
and I compiled it more than 10 years ago. I'm running VMS v8.4 and when
I select "open" the file dialog seems to be normal to me.
Also "gv file.ps" works. Be sure that the symbol gs is defined as the
command to run ghostscript.
If you are interested I can zip-up the directory containing Ghostview
and put it on my web-site.
[snip]
> I'll be including a backup save set of the source code when I release
> the PCSI kit. The code changes would probably be easy enough to get
Thanks, please keep me informed.
> into the official release. However, I had to make several changes to
> the .mak files, in some cases completely rewriting them. The changes
> are not compatible with the Unix versions.
>
> For example, all of the .mak files use $(D) as the path separator,
> defined as / for unix and \ windows and it is null for VMS. That is
> because you cannot simply concatenate things together to build a path on
> VMS. As long as the reference in question is simply $(DIR)$(D)filename,
> this is not a problem. However, when the reference needs to go more
> than one level deep, i.e. $(DIR1)$(D)$(DIR2)$(D)$(DIR3)($D)filename,
> everything breaks. There are a number of such instances in the build
> files. I simply edited them for VMS when doing the build. We'd need to
> come up with a way to fix the issue that doesn't break the build file
> for other platforms in order to submit it.
>
I know. In the past I made several changes to them. I would have
preferred "separate" makefile for OpenVMS, but this was the way the
GhostScript developers liked it.
The problem of $(D) should be solved in the same way as some other
macros i.e. $(I_) and $(_I) which were introduced for the specific
naming in OpenVMS. So $(D) should be split into 2 or three macros
depending it is the first, middle of end occurrence (i.e. $(_D) $(D) and
$(D_)), which are different for OpenVMS but the same for other OS's.
Jouk
More information about the Info-vax
mailing list