[cairo] Meta surface proposal

[Keith Packard]

Around 17 o'clock on Feb 6, Owen Taylor wrote:

> Why we conceivably want to drag the PDF rendering model into the
> equation for rendering to a printer on Windows? That sounds like an
> incredible headache for no gain compared to sticking with something
> simpler.

The question isn't whether we use the "PDF rendering model" here, it's
whether the PDF generated by cairo holds an accurate and concise
representation of the rendering operations performed on the surface -- we
need only understand a tiny fraction of PDF, and only that which cairo
itself generates.

I think it's worth a tiny amount of exploration here to see how feasible 
this is; with luck, we'll need only produce a cairo-PDF to printer GDI 
conversion function and have complete vector output for GDI devices 
wherever possible.

With your wrapper routines, we could bail on PDF, if necessary, and create a 
custom meta file format in the future.

> Note there is no "file" in the meta surface, it's just a list of
> primitives with extra information (need an image fallback). You only
> need one page at a time, so I don't think memory size will be a big
> issue.

You need to preserve the state of source images at the point of use, which 
may well represent significant storage...

I'd like to see an API designed which hides this implementation detail 
from the upper level driver.

-keith


-------------- next part --------------
A non-text attachment was scrubbed...
Name: not available
Type: application/pgp-signature
Size: 228 bytes
Desc: not available
Url : http://lists.freedesktop.org/archives/cairo/attachments/20050206/999d8946/attachment.pgp