Blog

Swift Semantics
Oct 16, 2020 swift

In this post, we’ll look at various ways to improve readability and writeability of Swift code by introducing extensions to Swift’s native types.

TLDR;

If you find this post too long, I have added most of the extensions in this post to SwiftKit and SwiftUIKit. I have excluded UIKit-specific code, since I focus on SwiftUI and don’t have UIKit code in these libraries.

Background

There will always be parts of a language that you may think are lacking or don’t fit your personal coding style. Some things are just strange (e.g. why does UIKit only have an isHidden view property, but no calculated isVisible) while some may be specific to your own needs.

Swift aims to be very focused, which means that it will intentionally omit parts that you may think should be there, which may explain missing properties like isVisible. You can easily create such properties yourself with extensions, but doing so can be repetetive and tiresome. I have therefore started gathering reusable extensions in the libraries mentioned above, to make my code more straightforward.

I will split this post into different sections for specific focus areas. If you have suggestions or comments, don’t hesitate to discuss in the comment section or send me PRs.

Inverse semantics

I personally think that using ! in control flow makes code harder to read, since it adds logic operators to code that otherwise pretty much reads like English.

In many cases, you can just replace a guard with an if and vice versa to reverse the control flow. For instance, instead of the double negation nature of this guard:

guard !view.isHidden else { return }

I think the code below is easier to read:

if view.isHidden { return }

Sometimes you can’t do this, though. For instance, here:

if !view.isHidden { /* do something with the view */ }

you can’t replace the if with a guard, since guards have to return.

In these cases, inverse properties can help. The inverse of isHidden would be isVisible:

if view.isVisible { /* do something with the view */ }

I find that it’s easier to read code that expresses intent instead of inversed intent, so I try to reduce negations as much as possible.

Syntax hiding semantics

For an overall readable language like Swift, I think that == and != reduces readability (not always true, though, read on).

Consider if you want to check that an optional has a value:

if myOptional != nil { ... }
if myOptional == nil { ... }

I think the following is easier to read:

if myOptional.isSet { ... }
if myOptional.isNil { ... }

The same goes for comparisons. Consider comparing two dates:

if date1 > date2 { ... } 
if date1 < date2 { ... } 
if date1 == date2 { ... } 

I think the following is easier to read:

if date1.isAfter(date2) { ... }
if date1.isBefore(date2) { ... }
if date1.isSame(as: date2) { ... }

This is a common feature in test libraries that strive for readable code, with comparison functions like 5.isGreaterThan(4). I think this is a great practice in other cases as well.

However, I think the syntax above may reduce readability in some cases, e.g. when comparing numeric values. I don’t think that if 5.isGreater(than: 4) is much more readable than if 5 > 4.

Chaining semantics

In many cases, chaining operators can improve readability by removing guards or if lets when using optional values. Consider this case, where you cast an optional SwiftUI View to AnyView:

var body: some View {
    content
}

var content: AnyView {
    if let form = loginForm {
        return AnyView(form)
    } else {
        // Logged in content
    }
}

var loginForm: LoginForm? {
    isLoggedIn ? nil : LoginForm()
}

In the code above, you have to perform an if let to get the login form before you can cast it. To simplify things, you could add the following extension:

public extension View {
    
    func any() -> AnyView {
        AnyView(self)
    }
}

then make the code more compact by not having to if let the login form:

var content: AnyView {
    loginForm?.any() ?? // Logged in content
}

The same can be used to simplify code for converting optional numeric values. Instead of:

var myInt: Int?
...
guard let value = myInt else { return nil }
return Double(myInt)

you can add an Int extension:

extension Int {
    
    func toDouble() -> Double { Double(self) }
}

that lets you chain the code and make it more readable:

return myInt?.toDouble()

Chaining operators are as great here as they are when e.g. chaining maps, filters, SwiftUI modifiers etc.

Conclusion

This post discusses things that I think improves the overall readability of the code. One could argue that it’s just a matter of becoming familiar with the syntax of the language, but I would argue that readability trumps keeping to the basics of the language, although what is readable is highly subjective.

Below is a list of semantics that I use in my projects:

Collection hasContent instead of !isEmpty
Date isAfter, isBefore and isSame instead of >/, < and ==
Decimal doubleValue instead of Double(truncating...)
Numeric toX() e.g. Int toDouble() instead of Double(self)
Optional isNil instead of == nil
Optional isSet instead of != nil
String hasContent instead of !isEmpty
View any instead of AnyView(self)
View hidden(if:) instead of if x { view.hidden() } else { view }
View visible(if:) instead of if !view.hidden(if:)

Source code

I have added most of the extensions in this post to SwiftKit and SwiftUIKit. Feel free to try them out and let me know what you think.

Swift Semantics Oct 16, 2020 swift