On Mon, Mar 10, 2003 at 11:59:05PM +0000, Keith Whitwell wrote: > > Unless C++ can figure out at compile time *exactly* which class is being > invoked, I think... It's hard to know.
It can, with templates. > Anyway, I don't think Jose really wanted a virtual member here, as he > definitely seems to want the only usage of these functions to be expanded > inline, right? If so, the virtual keyword is unneccessary -- isn't it? Yes, when using templates they become sort of unecessary. These are the two main C++ paradigms for generic programming: - virtualization: generic algorithms operate on abstract classes using their virtual methods. (no inlining on the virtual methods is possible because the compiler doesn't know which function is calling until run-time) - templates: generic algorithms are instantiated for each classes, using their methods (inlining of the methods is possible, becuase the compiler knows the function in compile time) So far I've been following more the first paradigm (hence the virtual keyword in everything), which matches more closely the current C code and OOP principles in general, but it's becoming clearer that for the 3D driver, it will be necessary the second paradigm to avoid the C++ overheads. Still, there is no harm in leaving those virtual keywords and give/make an choice of paradigms later. These two paradigms don't exactly match, but you can interchange/mix them to get a good compromise. José Fonseca __________________________________________________ Do You Yahoo!? Everything you'll ever need on one web page from News and Sport to Email and Music Charts http://uk.my.yahoo.com ------------------------------------------------------- This sf.net email is sponsored by:ThinkGeek Welcome to geek heaven. http://thinkgeek.com/sf _______________________________________________ Dri-devel mailing list [EMAIL PROTECTED] https://lists.sourceforge.net/lists/listinfo/dri-devel
