I recently required a function to return an input range over some base element type E. I was under the impression that this is exactly what the interfaces in std.range.interface are for but found that these are more restrictive than the checks from std.range.primitives.

In particular, isInputRange(R) only requires R to implement front, empty and popFront, while InputRange!E requires moveFront and opApply in addition. To me this was unexpected and seems inconsistent, but maybe there's a good reason for that.

Can someone explain whether (and why) this is intentional.