PDF to SVG: Image compositing before SVG backend output

Hello!

I'm trying to convert PDF to SVG using poppler and cairo. Some of the PDFs
has mosaic ornament combined with several images repeated many times with
clipping.

SVG backend writes all tiles with clipping to the output. That's why output
becomes very huge.

I'm trying to find a way to make raster images compositing before surface
will reach SVG backend

Could anybody helps me with this task?
How to find out cairo processing pipeline documentation?
How to find out cairo internals documentation?

Regards,
Vladimir

-- 
cairo mailing list
[email protected]
http://lists.cairographics.org/mailman/listinfo/cairo