I really like the IncompleteRange/RangeExpression idea!
I don’t think that IncompleteRange/RangeExpression should, by themselves,
conform to Sequence. It seems like necessary information is missing there.
Instead, there needs to be a conditional conformance to Sequence based on
another protocol that provides the natural bounds for the Bound type.
For example, what if we have another protocol:
protocol FiniteComparable : Comparable { //Any finite set which is
comparable will have a lowest value and a highest value...
static var lowestValue:Self {get}
static var highestValue:Self {get}
}
Something like UInt would have a lowestValue of 0 and highestValue of (UInt.max
-1). Then you could conditionally conform a RangeExpression where the Bounds
are FiniteComparable to Sequence.
Now the behavior is consistent. In the case of ‘array[0…]’ the array is
providing the missing upper bound. In the case of ‘for n in 0…’ the
conformance to FiniteComparable is providing the bound (and it doesn’t trap,
because it is enumerating all values IN that type above the lower bound).
for n in UInt8(0)… {/* Will get called for every possible value of
UInt8 */}
I agree that trapping when an infinite sequence of integers goes past the max
value is the only reasonable thing to do in that situation... but since we get
to define the bounds (and we have defined them in other cases to be the largest
usable value), why not define them the same way here (i.e. not infinite). In
this case, I don’t see the added value in making the sequence infinite instead
of just bounded by what the type can represent. The only thing it seems it
adds is the trapping behavior. With the natural bound, you can use things like
filter on partially defined ranges (which would trap if they are defined as
infinite):
let odd:[UInt8] = (0…).filter({$0 & 1 != 0}) //returns an array of all
the odd UInt8
In cases where the bound type doesn’t conform to FiniteComparable, it could
still be used as a RangeExpression, but not as a sequence.
Thanks,
Jon
> On Feb 2, 2017, at 4:07 AM, Brent Royal-Gordon via swift-evolution
> <swift-evolution@swift.org> wrote:
>
>> On Feb 2, 2017, at 3:06 AM, Jaden Geller via swift-evolution
>> <swift-evolution@swift.org> wrote:
>>
>> It's not infinite (else subscript would trap)
>
> I'm not necessarily a fan of the idea of allowing you to iterate over an
> `IncompleteRange`, but I have to ask: What do you imagine an infinite
> sequence of integers would do when it tried to go past the `max` value? As
> far as I can tell, trapping is the *only* sensible possibility.
>
> (If you used a `BigInt` type, the sequence could of course then be infinite,
> or as infinite as memory allows.)
>
> --
> Brent Royal-Gordon
> Architechies
>
> _______________________________________________
> swift-evolution mailing list
> swift-evolution@swift.org
> https://lists.swift.org/mailman/listinfo/swift-evolution
_______________________________________________
swift-evolution mailing list
swift-evolution@swift.org
https://lists.swift.org/mailman/listinfo/swift-evolution
