[cairo] Surprising cairo_win32_printing_surface Behavior

[Brent Fulgham]

Just a followup on my printing surface problem:

On Wed, Jan 6, 2010 at 10:53 PM, Brent Fulgham <bfulgham at gmail.com> wrote:
> I have been attempting to get printing of web plugins to work under a Cairo-backed version of WebKit
> that runs on Windows.  The basic problem is that regions of the screen drawn by various plugins seem
> to be partially clipped, even though the other elements on the page (text, graphics, etc.) are not overlapping.
> Cairo is responsible for nearly all elements that appear on the page, but hands a windows device context
> and a rectangular region for the plugin to use for its drawing.
>
> I have simplified the test case down to a single program, which draws a set of graphics (using some of
> the examples from the cairographics.org site), and finally draws a blue rectangle on a region of the screen
> that overlaps with some of the Cairo drawings.  I have placed a ZIP file of the example here:
> "http://files.me.com/bfulgham/aztjcp".
>
> I built using Cairo 1.8.8 (and pixman 0.17.1), and would be happy to provide my binaries if anyone needs them.

If I use the postscript printing surface, I get pretty much the same
behavior.  It's actually slightly worse, since the GDI drawn rectangle
is not shown at all in the ps output, rather than simply being clipped
as in the win32_printing_surface.

-Brent