Ulrich Drepper <drep...@gmail.com> ha scritto: >Yes, and this is already trivial to do with the operator() interface. > >The fill() interface is needed for performance, everything else is >taken care by the operator() interface.
Understood, but you do *not* loose performance by having those fill functions templates, you add genericity and your interface becomes much more consistent with the rest of the library. If, for a different new feature, you really have to restrict the arguments to be plain pointers, so be it, but earlier today - quickly, admittedly - I went through your first patch and it appear to only use basic operations on pointers, which any forward iterator supports, so... Let me add Daniel too in CC, anyway, he may be interested, if these features will be submitted for standardization at some point. Paolo
