Closed kennykerr closed 5 years ago
Going to leave this for now. The fast iterators are great for optimizing and simplifying range-based for loops for WinRT collections, but if you're going to do any serious algorithm-based computation you're probably better off getting your data into a standard container to avoid the ABI overhead. If we come across a compelling scenario we can always revisit.
The fast iterators were originally designed to simplify and optimize use of
IVector
andIVectorView
in range-based for loops and thus only support forward iteration. These containers are however capable of supporting random access. The fast iterator should be updated to provide complete random access iterator semantics and the reverse fast iterator can then probably just rely onstd::make_reverse_iterator
.