As far as I understand, it get the lower bound on the amount of elements yielded by the iterator, and reserves space for
lower_bound+1 elements. I may be reading that wrong, though.
Either way, these are just details of the current implementation and are explicitly not guaranteed to remain that way. So an implementation of a fixed capacity vector cannot use any Vec methods that insert more than one element at a time.
Which is a shame, because when implemented as a crate it’ll be missing out on specialization, so if we want e.g.
extend_from_slice() to be optimized into a
memcpy() we can’t do that and will have to use a separate function until specialization is stabilized.