[swift-evolution] [Review] SE-0050: Decoupling Floating Point Strides from Generic Implementations

[Nicola Salmoria]

> * What is your evaluation of the proposal?

+1. It might not be pretty, but it is an improvement.

The Proposed Solution doesn't compile with development snapshot 05-09
(SE-0067 hasn't landed yet, IIRC). I hacked it to get it to compile, and
tested it with two problematic cases:

Test #1:
let expected = Array(stride(from: 10, through: 20, by: 1).map { Float($0) *
0.1 })
// [1.0, 1.10000002, 1.20000005, 1.30000007, 1.39999998, 1.5, 1.60000002,
1.70000005, 1.80000007, 1.89999998, 2.0]

let actual = Array(stride(from: Float(1.0), through: 2.0, by: 0.1))
// old result: one iteration less than expected
// [1.0, 1.10000002, 1.20000005, 1.30000007, 1.4000001, 1.50000012,
1.60000014, 1.70000017, 1.80000019, 1.90000021]
// new result: correct
// [1.0, 1.10000002, 1.20000005, 1.29999995, 1.39999998, 1.5, 1.60000002,
1.70000005, 1.79999995, 1.9000001, 2.0]

Test #2:
let expected = Array(stride(from: 10, to: 100, by: 9).map { Float($0) * 0.1 })
// [1.0, 1.89999998, 2.79999995, 3.70000005, 4.5999999, 5.5, 6.4000001,
7.30000019, 8.19999981, 9.10000038]

let actual = Array(stride(from: Float(1.0), to: 10.0, by: 0.9))
// old result: one iteration more than expected
// [1.0, 1.89999998, 2.79999995, 3.69999981, 4.5999999, 5.5, 6.4000001,
7.30000019, 8.19999981, 9.09999943, 9.99999905]
// new result: correct
// [1.0, 1.89999998, 2.79999995, 3.69999981, 4.5999999, 5.5, 6.39999962,
7.29999971, 8.19999981, 9.09999943]

So the proposed solution works as advertised.

I would suggest to include two cases like the above in the test suite.


> * Is the problem being addressed significant enough to warrant a change to
Swift?

Yes. This is a higher level abstraction, so it's important that it
guarantees the best possible accuracy. The current generic implementation
gives unacceptably surprising results.

> * Does this proposal fit well with the feel and direction of Swift?

Yes. It nicely follows the principle of least astonishment.

> * If you have used other languages or libraries with a similar feature,
how do you feel that this proposal compares to those?

I always flat-out avoided floating point loops in other languages, fearing
that they would be unreliable. It's good that with this change it'll be
possible to use them with more confidence.

> * How much effort did you put into your review? A glance, a quick reading,
or an in-depth study?

careful reading, tested the proposed solution, and read and contributed to
the relevant threads.

--
Nicola