[cairo] PS backend

Carl Worth cworth at east.isi.edu
Mon May 10 19:14:51 PDT 2004

On Wed, 05 May 2004 10:15:25 -0700, Keith Packard wrote:
> Around 17 o'clock on May 5, "Rob Taylor" wrote:
> > i'm going to have to hook into the cairo API at a much higher level than
> > current backends (the same will go for a decent PDF backend to, i
> > suppose).
> PS doesn't support the whole cairo drawing model, and it has some serious 
> precision problems in drawing objects, so we want to hook lower than you 
> might at first think.  For geomtery, we really want cairo to compute 
> trapezoids and send those to the output device instead of sending the high 
> level paths.

That's true, and it's definitely the route we want to take for a
high-quality output-oriented PDF backend.

But I can imagine other uses that will require a higher-level
interface. One compelling scenario is SVG output. For this, users will
expect to get editable SVG out, (eg. a path must still be a
path). They'll probably even request that arcs still remain arcs, which
we could support without too much reworking of the internals.

So maybe we'll end up with two different routes for generating PDF?

I agree that the trapezoid based implementation is the priority for now.


More information about the cairo mailing list