Swift 2: Change Is in the Air
Like this article? We recommend
Like this article? We recommend
Like this article? We recommend
Since Apple introduced Swift in 2014, it has catapulted into a leading position, becoming one of the "must learn" languages for software developers. This is due in no small part to its leading role as star of iOS and Mac OS X, in addition to platforms such as watchOS and tvOS. Swift is also a language whose syntax favors readability and an ease of use that has heretofore been the domain of scripting languages. With Apple's promise of making an open source version of the Swift compiler available by the end of 2015, the language is expected to gain even more followers and popularity, as well as additional mobile and desktop platforms.
It's not surprising that in Swift's mere 18-month existence as a public language, it has undergone some hefty revisions. Apple has never feared change; and tinkering with the language has affected source code compatibility from Swift 1 to Swift 2. Now the language is at version 2.1, and while the changes haven't been quite as drastic of late, there are bound to be more in the future.
In this article, I'll show some examples of how Swift has evolved from its first major release to its second. I'll also go over a few of the newer features of the Swift 2 language, demonstrating how you can use those features to make your own code better.
The Keyword Is Evolution
Playgrounds have been around since the beginning of Swift and Xcode 6. They're a great feature for trying Swift code without committing to creating a full project. Go ahead and create a new playground from the File > New > Playground menu. Type a name for your new playground document, and then follow along with the example in this article by entering the code snippets in the following sections.
Goodbye, println!
Printing information to the screen is a common operation in any computer language. Swift 1 introduced two functions to perform this activity: print and println. Both functions take an object (usually a String) as a parameter and send it to an output stream, such as a log file. print sends the output verbatim, and println performs the same operation but appends a newline at the end.
In Swift 2, the println keyword has been removed and its functionality merged into the print command In fact, the default behavior of print is to behave like println and append the newline automatically. Its original behavior of not adding a newline is preserved when the terminator parameter is used with an empty string, as shown in Figure 1.
)
Figure 1 The new behavior of print.
A New Loop
The do-while loop is a common feature in any programming language. It allows developers to create an iterative construct in which code can be executed repeatedly before an "escape clause" terminates the loop.
Swift 1 used the familiar do-while keyword pair, but the do keyword has been repurposed in Swift 2 for its new error-handling feature. The new keyword pair is repeat-while, shown in action in lines 12–15 of Figure 2.
)
Figure 2 The repeat-while loop in action.
Here, a simple loop has been created to show the value of the variable x, which is incremented until it reaches the value 10. At that point, the loop exits.
Watch Your Language
The moral of the story in these two examples is that keywords in Swift already morphed between versions 1 and 2. Such changes can break existing code, rendering it useless until the keywords are fixed. Keeping up with versions of Xcode will reveal these types of compatibility issues, and you should be ready to update your code as needed to address the changes to the language.
do Gets a Redo
As I alluded to earlier, the do keyword wasn't thrown away in Swift 2, but has been repurposed. do is now part of a robust error-handling mechanism that's designed to make your code more impervious to unexpected errors. If you're familiar with the try-catch feature in Objective-C, you'll immediately feel at home with Swift's new do-catch construct.
Enter the following code into your playground file (as shown in Figure 3):
enum NumberError : ErrorType {
case Negative
case Zero
}
func mustPassPositiveNumber(value : Int) throws {
if value == 0 {
throw NumberError.Zero
} else if value < 0 {
throw NumberError.Negative
}
}
do {
try mustPassPositiveNumber (33)
} catch {
print(error)
}
do {
try mustPassPositiveNumber (0)
} catch {
print(error)
}
do {
try mustPassPositiveNumber (-5)
} catch {
print(error)
}
)
Figure 3 Examples using do-catch to trap errors.
These simple examples demonstrate how the do-catch combination ensures that code works properly. Here's a breakdown of the code:
enum NumberError : ErrorType {
case Negative
case Zero
}
The enumeration inherits from the ErrorType protocol, and it's used to define appropriate error cases. In this instance, there are two cases: Negative and Zero. These cases are used in the following function:
func mustPassPositiveNumber(value: Int) throws {
if value == 0 {
throw NumberError.Zero
} else if value < 0 {
throw NumberError.Negative
}
}
This function insists on taking a parameter that's a positive number (greater than zero). The passed value is checked against 0 and < 0. In the first case, the Zero ErrorType is thrown. In the second case, it's the Negative ErrorType.
The throw keyword is used by a function to alert the program that an error condition has occurred. Since errors are "thrown," it makes sense that they should be "caught." The act of catching an error is encapsulated in the do-catch, implemented in the subsequent lines of code:
do {
try mustPassPositiveNumber (33)
} catch {
print(error)
}
do {
try mustPassPositiveNumber (0)
} catch {
print(error)
}
do {
try mustPassPositiveNumber (-5)
} catch {
print(error)
}
Each section tests the mustPassPositiveNumber method with a positive number, a negative number, and zero. The sidebar in the playground window shows that for the second and third examples, the error is printed in the catch block.
Code inside the braces following do is intended to be executed under the careful watch of the throw handler. Any code within that block (or any code called from within that block) that throws an error will cause the code in the accompanying catch block to execute. This mechanism gives the code the ability to handle anomalous situations gracefully, and to communicate specific error conditions so that the calling code can handle the situation appropriately.
The takeaway with do is that Apple has demonstrated its willingness to change the meaning of a language keyword, recycling it to be used in another part of the language. Bold changes like this have big implications on existing source code, however; you need to be ready for potentially drastic code changes as you carry your Swift projects to newer versions of the language.
Swift 2 Functions, Be on Your Guard
Succinct code that expresses your intent in a readable but optimal way is a worthy goal of any programming language. Often, however, readability must be sacrificed for functionality, especially when it comes to anticipating and reacting to problems. Consider the following example, in which two integers are added, but only if they're positive numbers:
func addPositiveNumbers(value1: Int, value2: Int) -> Int {
if value1 >= 0 && value2 >= 0 {
return value1 + value2
}
return -1
}
The two input values, value1 and value2, are checked to see whether they're positive numbers. If they are, the sum of the two numbers is returned; otherwise, -1 is returned, indicating invalid input.
When writing a function, it's good practice to handle the failure cases at the very beginning and do any error handling and early returns there, leaving the rest of the body for the work that the function needs to do. In the example above, the success case is checked first, leaving the failure case at the end.
To test the success case much earlier in the function and act on it, Swift 2 introduces the guard keyword. It has the following syntax:
guard <success_case> else <failure_action>
The code that tests for success immediately follows the guard keyword. If it evaluates to true, success is verified and execution continues past the else clause. Otherwise, the code following else (representing the failure case) is called. We can modify the string-length verification function as follows:
func addPositiveNumbers2(value1: Int, value2: Int) -> Int {
guard value1 >= 0 && value2 >= 0 else {
return -1
}
return value1 + value2
}
This early test allows the function to determine quickly whether there's a problem, allowing the code past the else clause to execute in the nominal success case (see Figure 4).
)
Figure 4 Demonstrating the guard feature of Swift 2.
The new guard keyword is one example of how Apple has made an attempt to improve code readability by introducing a totally new construct to the language. This change doesn't break source code compatibility, and it gives you the opportunity to refactor your code to take advantage of good practices.
Swift Can Change Swiftly
In the short period between Swift's two major releases, Apple clearly has made bold changes and additions to the language. This amounts to accelerated evolution of a computer language at an ultra-fast, technology-driven pace—something that other languages have been leery of doing at such breakneck speed. While it's challenging to keep up with all of these rapid changes, you should always be ready and willing to modify your code to take advantage of new features and best practices that the language introduces. The rapid pace of language modifications may abate for a while, but you never know when it will take off again. After all, this language does carry the name Swift!