[cairo] Is Cairo an easy optimisation for WinCE?

[Elliott Bignell]

Hi, all,

I have been investigating Cairo in relation to some anti-aliased drawing code
that we require for Windows CE, desktop and conceivably Linux later on. It seems
very impressive. We currently render using the Windows GDI and API calls into
vanilla HDCs wrapped in out own DC class. Performance is always an issue, and I
was wondering if there is an argument for converting the entire class to a
wrapper for Cairo on the basis of performance. One post here, for instance,
referred to creating a surface in video memory. Does Cairo offer an performance
improvement following a 1:1 replacement of GDI code by accessing the video
memory, or is it constrained to the performance of the GDI calls that seem to be
the basis of the Windows port?

Thanks in advance for any guidance...
Elliott Bignell