On 28 May 2010 12:26, Tomas Puverle <tomas.puverle@morganstanley.com> wrote:
I am sorry but I can't speak intelligently about the suitability of my endian implementation for this particular purpose.
I'll try to pose it differently. How would you implement the following using your library? Magic<8,32> m; m.put(0xAA); m.put(0xBB); m.put(0xCC); m.put(0xDD); assert(m.get() == 0xAABBCCDD); What about supporting this slightly more complicated case? Magic<16,32> m; m.put(0xAAAA); m.put(0xBBBB); assert(m.get() == 0xAAAABBBB); And how about this far more challenging one? Magic<4, 16> m; m.put(0xA); m.put(0xB); m.put(0xC); m.put(0xD); assert(m.get() == 0xABCD); I think I see how I'd do the first, and probably the second, but the third may require a very different approach. ~ Scott McMurray