AFPL Ghostscript 8.50 for 68k and x86-AROS released today

[whoosh777]

Tthe 68k release is a 68020-nofpu AmigaOS build, this will run on the maximum number of systems. It has been tried out with success on OS4, WinUAE and Pegasos II Morphos 1.4.3.

The x86-AROS build is an i386-AROS build, and I have used it successfully on AROS on my i686 machine.

On 68k the program has full Turboprint support, as well as truecolour viewers for PS and PDF documents.

 AFPL Ghostscript 8.50 is a forwards port of my original port to 68k of AFPL Ghostscript 8.00.

 This is the current major release of AFPL Ghostscript and contains many improvements not available in any earlier versions of the program.
 It is both faster and has better rendering quality than any earlier version of the program. eg it contains some advanced new optimisations for the rendering of very low resolution fonts.

It so far seems to be a particularly good release of the series.

The 2 most popular uses of Ghostscript are for printing and viewing PS and PDF documents. A lot of people who print with it use the Turboprint facility as it produces particularly good printouts.

Currently I am continuing 68k development, but intended to be run on 68k emulators, I currently use WinUAE.

Download it at http://www.whoosh777.pwp.blueyonder.co.uk

[whoosh777]

some things to note regarding x86-AROS:

1. porting 68k code to x86-AROS is very smooth,
2. if you develop code simultaneously on x86-AROS and 68k then you guarantee that your code is
endianess independent.

For me the other reason I am developing on x86-AROS as well as 68k is that I have both systems.

Anyone wishing to try out the x86-AROS port of
Ghostscript 8.50 should install it in RAM:,
it is then literally 10 x as fast and a lot of
the ops are virtually instant.

[whoosh777]

Have you tried Cary Driscoll's script?

(its on the above URL)

I think he's the person to ask on Final Writer,

if you email me with the question I can forward
it to him,

[whoosh777]

@darkcoder,

as you point out, there is a commitment element
to a port, about tracking version numbers.

So far I have done GS8, GS813 and now GS850.
With each forwards port I have become a lot wiser.
As a result the forwards porting is becoming
ever more efficient. One thing I have learnt now
is rather than allocating a fixed slot in time
for attempting a forwards port, it is better
to allocate a fixed amount of time but not allocating when, wait till a major release happens and then attempt to port that.

Trying to track each and every single
new version is not practical.

In fact my original plan was to attempt a port
approx in the second week of November.
However when I begun on this it gradually  became
clear that it was better to wait till the
soon to be announced GS850 was centrally
released.

GS850 was centrally released on 11th December 2004, so this port is very current.

regarding a 68020-fpu port: well there have
been problems running fpu ports on Morphos and OS4. As a result I decided to make 68020-nofpu
the main port. Note that the nofpu builds are
quite possibly faster than the fpu builds as
they use different fast integer algorithms
instead of floating point algorithms.


I can certainly build a 68020-fpu version and
probably will. But note that it probably wont
run on all systems.

regarding 68060 ports: the feedback from
68060 builds of the earlier GS8#?'s was that
there was no noticeable speedup and it mainly
wasnt worth the effort.

[whoosh777]

@Cary_D,

it all makes sense now!

gs_init.ps seems like exactly the right place
to customise things, for some reason its the
one place I wasnt thinking about. I must remember
this for future problems.

I attempted once to build GS to have A4 as the
default papersize. Unfortunately this leads to
some problems. So I reverted the default to
the usual papersize.

So it seems the right way to change the default
is externally, and the way you suggest seems to
be the best external way.

another external way is from the commandline one can change papersize via eg -sPAPERSIZE=a4