On Tue, 30 Aug 2005, David Berry wrote:
> On Mon, 29 Aug 2005, Tim Jenness wrote:
>
> > A quick look in figaro3/arc.f indicates the reason why it doesn't work. It
> > doesn't actually use AGP at all - it just calls PGBEGIN directly with the
> > supplied device name. No attempt is made to translate from Starlink device
> > to PGPLOT device name - it relies on GKS PGPLOT simply understanding the
> > device names.
> >
> > Mystery solved.
> >
> > Is it possible to call the AGP translation routine without doing a lot of
> > unnecessary AGP calls to set things up?
>
> Not really. If you want to avoid using AGI/P for device control, you
> could maybe just call agp_asoc to open the deivce (assuming it has an
> associated ADAM parameter), then call pgqinf to inquire the name of the
> open device, then call agp_deact to shut down pgplot.
But it looks like a lot of work to sort the graphics in FIGARO out. A
quick deep grep reveals a lot of PGBEGIN calls. Not surprised given the
mixed heritage and sticky plaster approach to its recent development.
|