Il 02/07/2013 01:42, Andreas Färber ha scritto: > I happened to notice that the scripted approach you describe in the > cover letter causes some unnecessary indirection here: > In this hunk a human could've just supplied the function argument obj. > But since OBJECT() is not a dynamic cast and f is referenced on the same > line for the MemoryRegion itself, I guess there is no harm in always > using OBJECT(f) for simplicity.
The conversion actually wasn't scripted, only the review. I used OBJECT(f) because it was easier to think about, and any mistakes (opaques that are not QOM objects) would be caught during the scripted review. Paolo
