empty function optimizations

[Joern Rennecke]

> I personally believe that the right fix is in the compiler, which
> should statically determine that instantiations don't depend on the
> some template parameters, and mangle the function names with a
> wildcard. Different instantiations would then still get the same
> assembler name. The compiler would avoid emitting duplicate
> instantiations of functionally-equivalent code, and the linker would
> remove duplicates as it does today.

I thought we were going towards a standarized ABI and mangling scheme, I.e.
we'll have no freedom to change the mangling?