Well, while this release cycle is supposed to be focused on pangoization, we've not really done much work on that yet, so I've been thinking about where to start...
At this point I think the plan looks like this:
1. C++ify libnrtype (completely)
2. Introduce a pango libnrtype backend
3. get the pango backend debugged on all platforms and switch to using it exclusively
4. refactor libnrtype away, bringing pango "to the surface"
5. Investigate using Pango in other areas, like <text> layout
One motivation for the C++ification is because, while it's a significant project in its own right, it is much easier to create and refactor C++ classes than GObject/NRObject ones. Another advantage is that it'll get everyone involved more familiar with the code.
So, anyway ... Nathan's already done some of the preliminary work on libnrtype. I think the next step switching from aggregation to C++ inheritance.
(which simply means that this:
struct NRSomething { NRSomethingBase base; ... };
becomes this:
struct NRSomething : public NRSomethingBase { ... };
...and no other changes should be necessary at that step)
Any volunteers? I'll start hitting this myself once I get through with the SPSelection cleanup.
-mental