[cairo] PDF memory usage?

[Ian Britten]

Hi all,
I'm researching some possible problems/solutions trying to
export our data to PDF via Cairo, and wondered if anyone could
fill in some background information for me?

- Does the PDF surface keep all data in memory until the final
   finish/flush is called?

- If yes, is there any way to progressively write out part/all
   of the data, mostly to free up the memory?

We produce large maps, which typically contain a mixture of
both vector data, and high-resolution image data.  I'm trying
to get this data written out to a PDF, while preserving the
vector data as vector in the PDF.  However, I'm concerned that
Cairo is going to try to carry the large image(s) in memory
while the PDF is being created, and may exhaust the memory.

For example, a 32-bit image that completely fills an A0 sheet
at 1200dpi would need over 8G of memory.  Now, worst case,
imagine several overlapping/transparent images, plus a slew
of vector data - Phew!
[ No, I have no idea what the users/customers expect to do
with files like these - I'm just trying to create them... ]

The typical approach for dealing with large output like this
seems to be to try and chunk/tile the data.  However, with
the target being PDF, I'm not sure if this is possible (And
whether the Cairo interfaces support it).  My basic PS
knowledges makes me think it could/should be possible to just
keep writing to a PDF file and let the viewer worry about
rendering the final data, but I'm not certain...
So, I figured I'd ask some of the clever people here!  :)

Many thanks for any info, suggestions, etc!
Ian