[cairo] Updated ROADMAP for cairo 1.2.0 (and beyond)

[Ralph Giles]

On Thu, Apr 27, 2006 at 08:18:02AM -0700, Carl Worth wrote:

> However, by associating the new functionality with new API, we now get
> compilation-time verification that the functionality exists, (perhaps
> link-time failure in case of a broken installation), and no need to
> introduce any run-time error for this.

So an application would be expected to #ifdef all the version stuff 
based on a declared cairo api version define, or stuff it discovered
itself in the configure script?

I would think some kind of property interface that can report failure
would be cleaner, though there is some conceptual advantage to not
having cairo_surface_set_property() return ETOOLATE. :-)

I would also like to comment as far as requirements that while a numeric 
version works fro SVG and Postscript, for PDF there are non-numerical 
compatibility profiles people will eventually be interested in 
supporting, like PDF/A and the various PDF/X. (There's even a PDF/is 
which is entirely image fallbacks!) An enum would do fine here, though.

> The earlier the error checks are, (in the sense of compile-time
> vs. run-time), the better.

I come from a much more conservative school wrt API and ABI stability.

 -r