[swift-evolution] [swift-evolution-announce] [Review] SE-0067: Enhanced Floating Point Protocols

Chris Lattner clattner at apple.com
Wed Apr 27 14:58:36 CDT 2016


> On Apr 27, 2016, at 9:10 AM, Jordan Rose <jordan_rose at apple.com> wrote:
> 
> 
>> On Apr 26, 2016, at 11:42, Chris Lattner via swift-evolution <swift-evolution at swift.org> wrote:
>> 
>> On Apr 26, 2016, at 8:47 AM, Tony Allevato via swift-evolution <swift-evolution at swift.org> wrote:
>>> That seems like a purely syntactic concern that could potentially be addressed in other ways, though. I'm not sure the choice of "duplicate all operators using verbosely-named methods" is the best one for the reasons I mentioned above, and the question of "how do we cleanly unify operators with other protocol requirements?" seems out-of-scope and orthogonal to this proposal.
>> 
>> There is a strong motivation for this approach though: we want the type checker to be scalable.  John recently wrote an epic piece about why having tons of overloads is a really bad idea:
>> https://lists.swift.org/pipermail/swift-dev/Week-of-Mon-20160404/001650.html
>> 
>> It is *much* better for type checker performance to have (e.g.):
>> 
>> func +<T : FloatingPoint>(lhs : T, rhs : T) -> T { return lhs.add(rhs) }
>> func +<T : Integer>(lhs : T, rhs : T) -> T { return lhs.add(rhs) }
>> 
>> Rather than overloads for 4 floating point types, and 8+ integer types.   We really need to eliminate all the “expression too complex” classes of issues, and this is an important cause of them.
> 
> That’s a reason to pull operators in as members, not push them out as second-class free functions.

There are a ton of open questions that would have to be resolved, but I agree that in theory that could resolve the issue as well.  

The practical problem is that those issues won’t get resolved in the Swift 3 cycle, and yet we still want improved numeric protocols, compile time, and decent overload failure diagnostics.

-Chris


More information about the swift-evolution mailing list