Archive for the ‘Development’ Category

Circles within circles: custom types and extensions

So the other day Paul and I were shooting the breeze about spirographs. He was saying I needed to get back to blogging now that I have the time. Of course, the second you set out to try to write a post, so many ideas flow forth at once, it actually is hard to pick just one and focus on it. Since he had enjoyed my drawing work, I thought I’d start out with a little push at circles.

That quickly went out of control because I ended up with enough material for a few chapters rather than a single blog post. (And, no, my chapter for his book is about property wrappers, which I love, rather than drawing)

To give a little context, before I decided to try my hand at full time work, I had had a contract with Pragmatic to update my Drawing book. I had to cancel that because commuting full time to another city left me with zero free time for my self, my family, let alone writing. (Now that I am back at my own desk, I may see about returning to that project because geometry is a core passion.)

Anyway, let me start at the end of all this and show you some pictures and then I’ll (pardon me) circle back and start telling the story of the code and how all this came to pass.

I started my day by opening a playground, deciding to just flex my coding and see where it would take me. I wanted to work with circles, so I plugged a range into a map to get me there.

// Get some points around a circle
let points = (0 ..< nPoints).map({ idx -> CGPoint in
  let theta = CGFloat(idx) * (CGFloat.pi * 2 / CGFloat(nPoints))
  return CGPoint(x: r * cos(theta), y: r * sin(theta))
})

// Create a path
let path = UIBezierPath()

// Get the last point before the circle starts again
let lastIndex = points.index(before: points.endIndex)

// Draw something interesting
for (point, index) in zip(points, points.indices) {
  let p0 =  index == lastIndex
    ? points.startIndex
    : points.index(after: index)
  let p1 = index == points.startIndex
    ? lastIndex
    : points.index(before: index)
  path.move(to: .zero)
  path.addCurve(to: point, controlPoint1: points[p0], controlPoint2: points[p1])
  
}

It’s not the worst code in the world, but it’s not great. I was rewarded with a pretty path, so whee. Still, the code was screaming out that it wanted to be better.

I started by attacking the index work. Why not have an array that supported circular indexing by offering the next and previous items? So I built this:

public extension Array {
  /// Return the next element in a circular array
  subscript(progressing idx: Index) -> Element {
    idx == indices.index(before: indices.endIndex)
      ? self[indices.startIndex]
      : self[indices.index(after: idx)]
  }
  
  /// Return the previous element in a circular array
  subscript(regressing idx: Index) -> Element {
    idx == indices.startIndex
      ? self[indices.index(before: indices.endIndex)]
      : self[indices.index(before: idx)]
  }
}

Before you start yelling at me, I quickly realized that this was pretty pointless. Arrays use Int indexing and I had already written wrapping indices a billion times before. Just because I has the concept of a circle in my head didn’t mean that my array needed to. I pulled back and rewrote this to a variation of my wrap:

public extension Array {
  /// Return an offset element in a circular array
  subscript(offsetting idx: Index, by offset: Int) -> Element {
    return self[(((idx + offset) % count) + count) % count]
  }
}

Much shorter, a bit wordy, but you can offset in either direction for as far as you need to go. I did run into a few out of bounds errors before remembering that modulo can return negative values. I decided not to add in any assertions about whether idx was a valid index as arrays already trap on that, so a precondition or assertion was just overkill.

Next, I decided to design a PointCircle, a type made up of points in a circle. Although I initially assumed I’d want a struct, I soon realized that I preferred to be able to tweak a circle’s radius or center, so I moved it to a class instead:

/// A circle of points with a known center and radius
public class PointCircle {
  
  /// Points representing the circle
  public private(set) var points: [CGPoint] = []
  
  /// The number of points along the edge of the circle
  public var count: Int {
    didSet { setPoints() }
  }
  
  /// The circle radius
  public var radius: CGFloat {
    didSet { setPoints() }
  }
  
  /// The circle's center point
  public var center: CGPoint {
    didSet { setPoints() }
  }
  
  public init(count: Int, radius: CGFloat, center: CGPoint = .zero) {
    (self.count, self.radius, self.center) = (count, radius, center)
    setPoints()
  }

  /// Calculate the points based on the center, radius, and point count
  private func setPoints() {
    points = (0 ..< count).map({ idx -> CGPoint in
      let theta = CGFloat(idx) * (2 * CGFloat.pi / CGFloat(count))
      return CGPoint(x: center.x + radius * cos(theta), y: center.y + radius * sin(theta))
    })
  }
  
}

I added observers to each of the tweakable properties (the radius, point count, and center), so the circle points would update whenever these were changed. By the way, this is a great example of when not to use property wrappers. Wrappers establish behavioral declarations, which this is not. My circle uses the observers instead to maintain coherence of its internal state, not to modify, limit, or expand type side effects for any of these properties.

Now I could easily create a circle and play with its points:

let circle = PointCircle(count: 20, radius: 100)
let path = UIBezierPath()
for (point, index) in zip(circle.points, circle.points.indices) {
  path.move(to: circle.center)
  path.addCurve(to: point,
                controlPoint1: circle.points[offsetting: index, by: 1],
                controlPoint2: circle.points[offsetting: index, by: -1])
}

I decided not to add some kind of API for passing a 2-argument (point, index) closure to my circle instance. It doesn’t really add anything or make much sense to do that here. It wouldn’t produce much beyond a simple for-loop, and the point here is to build a Bezier path, not to “process” the circle points.

My goal was to build enough support to allow me to iterate through the points and reference adjacent (or further offset) points to build curves. This does the job quite nicely.

It didn’t feel quite finished though. I wanted to add a little extra because one of the things I often do with this kind of drawing is create stars and points and loops using offsets from the main radius, often interpolated between the main points on the circle’s edge. Because I had built such a simple class, adding an extension for this was a snap:

extension PointCircle {
  /// Returns an interpolated point after a given index.
  ///
  /// - Note: Cannot pass the existing radius as a default, which is why the greatest finite magnitude is used
  public func interpolatedPoint(after idx: Int, offsetBy offset: CGFloat = 0.5, radius r: CGFloat = .greatestFiniteMagnitude) -> CGPoint {
    let r = r == .greatestFiniteMagnitude ? self.radius : r
    let theta = (CGFloat(idx) + offset) * (2 * CGFloat.pi / CGFloat(count))
    return CGPoint(x: center.x + r * cos(theta), y: center.y + r * sin(theta))
  }
}

The most interesting thing about this is that you cannot use a type member to default a method argument, which is why I ended up defaulting it to `greatestFiniteMagnitude`. It seems to me that it would be a natural fit for the Swift language to allow that kind of defaulting. In this case, it would say: “If the caller doesn’t specify a radius, just use the one already defined by the instance.” What do you think? Is that too weird an ask?

To wrap up, here’s an example using the interpolation:

let path2 = UIBezierPath()
for (point, index) in zip(circle.points, circle.points.indices) {
  path2.move(to: point)
  path2.addCurve(to: circle.points[offsetting: index, by: 1],
                 controlPoint1: circle.interpolatedPoint(after: index, offsetBy: 0.33, radius: circle.radius * -2),
                 controlPoint2: circle.interpolatedPoint(after: index, offsetBy: 0.67, radius: circle.radius * -2))
}

All the shapes at the start of this post are created from this and the preceding loop by tweaking the curve parameters. What pretty results will you come up with? If you build something particularly lovely, please share. I’d love to see them!

Reading stdin on the fly

Shai Mishali asks, “Is there something like readLine that lets me read stdin as the user types instead of waiting for an entire line?” The readLine function waits until the user presses enter or otherwise recognizes EOF to process input.

A little web search led me to this SO post, which is  based on this example, using termios, the Unix API for terminal I/O. After looking through both posts, I built a RawMode type to support direct input based on that code because nothing makes me happier than messing with solutions to make them a little swiftier.

To use this, you fetch bytes until you catch an EOF. It’s interesting to test both direct keyboard entries as well as pasting complex emojis. As I was working on this on both 10.14 and 10.15, I was mildly startled when my initial test code stopped working when I moved to 10.14. UnicodeScalar does not offer a utf8 view until 10.15:

extension Unicode.Scalar {
  @available(macOS 10.15, iOS 13, tvOS 13, watchOS 6, *)
  @frozen
  public struct UTF8View {
    @inlinable
    internal init(value: Unicode.Scalar) {
      self.value = value
    }
    @usableFromInline
    internal var value: Unicode.Scalar
  }

  @available(macOS 10.15, iOS 13, tvOS 13, watchOS 6, *)
  @inlinable
  public var utf8: UTF8View { return UTF8View(value: self) }
}

So thank you, Swift team, for giving us that. It’s a nice addition to the language!

Anyway, if you have a different approach or you can see ways to improve my take further, drop a comment and let me know.

Whisky tango foxtrot: Xcode allows ObjC switch unindenting

This happened.

You might think I’m about to go off on some Swift rant (and trust me there is a Swift rant inside me waiting to emerge) but it’s the second checkbox that made my mind explode.

From:

To:

Who thought this was a good idea? I’ve never been a fan of left-aligned case in Swift although I embrace it as the standard. But in Objective-Freaking-C? As a standard Apple-blessed toggle in Xcode? No! Thrice no! The option enabling the choice is bad for Swift and worse for Objective-C.

Why is this option in there and why is it available for both languages? It would be best to, as Joe Groff put it, “let sloth naturally lead everyone to pick the default” given that the feature has been expressed in Xcode. Or better yet, file some bug reports for the broken feature.

Each language default reflects years (and decades) of language style consensus:

  • Swift: keyword-aligned.
  • Objective-C: “scope”-aligned.

This new choice in preferences is madness.

Talk me down from here, friends.

The knee-jerk else-clause

I’ve gotten into several heated discussions recently (I wouldn’t call them arguments or fights as no blows were exchanged and we departed as friends) about whether an if-statement should always include a balancing else clause in languages where it is not mandated by a ternary if form.

My feeling is a firm “no”.

Let me attempt to represent the opposing viewpoint. As far as I can tell, the “yes” faction believes that as proper diligence, an else clause should always be added to address both truth scenarios, even if the clause is populated with nothing other than a comment saying “this clause intentionally left blank”. It promotes a consistent inspection of all possible cases, on par with a switch statement requiring full coverage.

I firmly disagree. In my opinion, if-statements boil down to three scenarios:

  • Special CasingAction that applies solely to certain logic conditions, where the absence of those conditions has no effect in the logic. In such case, I recommend to skip the else. There’s no point to it as the flow of execution continues past the additional steps introduced by the if-statement. Adding an else statement removes the clarity of the step-by-step logic and reduces the weight of the special casing in understanding the code.
  • Either-Or: Separate actions of approximately equal weight that apply to both outcomes. If statements are naturally biased towards the positive clause. If both statements are weighted the same, consider refactoring to switch. Even if the switch is ordered (as it must be in some fasion), a switch-statement reduces the emphasis given to either outcome. If both cases are quite large, consider breaking out functionality to dedicated methods. Further, if the logic overlaps between both cases, consider localizing the conditions closer to the differences, even if the test may be applied several times.
  • One case complex, one case simple: Actions that are of disparate impact or weight, where one clause greatly exceeds the complexity of the other, I always place the more complex clause first when used in an if-statement to emphasize the positive path. At the same time, if the simple case is non-trivial, this is an opportunity to consider whether the logic should be broken up or refactored in some way. If an if-clause is so significantly big that it takes up the majority of a method implementation, perhaps you should be using early return for the lighter condition and promoting the more complex logic to the main method body. To me, a heavy complex if-clause usually signifies an opportunity to refactor.

Given these styles, I don’t see any reason, whether from code-reading clarity or a strict adherence to potential future expansion, to automatically include else-clauses as a mandatory part of a coding style. They add heaviness without real functionality and speak to a philosophy of form weighed over functionality. Code should be light, clear, and direct. Mandatory-else is none of those things.

Am I wrong? Go ahead, convince me otherwise. I’m listening.

My enemy the Minimap

The minimap is one of Xcode 11’s starring features. I know many people were excited for it at its debut but after months of exposure, I now just disable it in the Adjust Editor Options menu (the five horizontal lines of unequal width indicating the contents of the primary editor) and grab back that room on my screen:

I’ve spent some time recently considering exactly why it is I hate the minimap so much. It’s not just that it takes up valuable horizontal space (although to be honest, it’s mostly that). If it did a better job to help me navigate or conceptualize my code, I wouldn’t resent that space. Rather, I get little out of it and it’s a blurred colorful distracting mess for me.

In theory, the minimap is a specialized scrollbar for code. It highlights where are you with respect to the file. You hover the cursor over it to view tooltip details. Upon finding a section,  you can quickly jump to it. Hold the command key down and all the tooltips appear at once for a quick and actionable  overview. All contextual bookmarks are larger sized so technically “legible” although I can barely read them with my bad eyes.

I don’t care where I am physically inside my file, so the scope highlights don’t offer much. I care where I am conceptually. My jump bar gives me the same overview as the minimap (using the command key) without stealing screen space or distracting me with extra pointless blurry information.

What I am at a loss to figure out is how the minimap does that better than my beloved jump bar or why people prefer having the big jumble on-screen. As far as I can figure, the overview (whether jump bar or minimap command key) provides the most powerful tool because it relies on recognition of the details you’re looking for. But most of the minimap is designed around recall: where you are in the file, what the shape of your code is like, etc, which is a far weaker place for cognitive support.

So what am I missing? Why should I love the minimap? Please help me figure out what I am not understanding here and help me turn the minimap from an enemy to a friend.

Xcode: Basics of the four-block wonder

The official name is “Navigate to Related Items” but to me, it’s the four block wonder, a menu button that sits at the top left of the Xcode editor. With this menu, you can hop between file counterparts (for example, .m/.h, or .swift/generated interface, which is what a lot of people use it for). But there’s so much more.

Set the cursor on a type, and you can view or navigate its superclass, subclasses, or siblings, as well as the protocols, extensions, and categories it connects to. A single click navigates you to the interface or implementation in question:

Use the cursor to establish the context for the menu, otherwise you’ll only see a smaller subset menu options, such as recent files. The callers option shows you your clients, and callees the items your code is calling — all specific to the current cursor context.

One of my favorite tools in the four-block menu is Generated Interfaces, which allows you to view an item’s Swift interface or see how it translates to Objective-C. For example, if you use an obvious preposition label, the ObjC generated interface subsumes it into its selector:

With this, you don’t have to wonder if your selector is specified right and you don’t have to override the selector with objc. You just look up the definition and you’re good to call.

In addition to contextual helpers, the four-block lets you select recently viewed files and, when using version control, recent files that have been locally modified and not yet committed (basically the ones showing the “M”-for-modified).

The four little squares may be tiny but it is a powerful tool in your Xcode arsenal.

Update: Lilly reminds me that you can bind the related items menu. I have mine bound to ^1 but I don’t remember if that’s something I did or something that is default. If you want to add or change, visit Preferences > Key Bindings and search for “related”.

Fun with Xcode Search Domains: Excluding match text

Most Xcode users quickly become familiar with the basics of the Find Navigator panel.

With it, you can find text, regular expressions, and perform search-and-replace, whether matching or ignoring case. But that’s just scratching the surface of the Find Navigator.

I thought I’d drop a few words today about search scopes. Controlled from the bottom left,  under the search field, you can create narrowed searches. This enables you to, for example, search only in Swift files or exclude files containing the word Test.

To get started, click the icon (two lines with three squares on a line between them) and then New Scope (the plus icon). Here, you can name the scope, limit the search extent, and add criteria for exactly which files should be included or not.

The logic is straightforward. You choose where to look (the project, a folder, or through the entire SDK), and whether to include all conditions or some conditions:

Each condition is based on the file name, path, extension, UTI (the kind of file, like image which is useful for finding vector assets), Workspace location (namely groups), or source control status (handy for finding newly applied changes.)

Most of my conditions are file-name-based. And for those, you get the following matching conditions. The “ends with” is an obvious win for extensions (although you can also use UTIs for that), and “starts with” can help for projects organized in hierarchical ways.

Now, interestingly enough, this list fails to offer “does not contain” but that’s fairly easy to work around. Since Xcode supports regex matching, you can easily replicate “does not contain” with an appropriate regex:

Change the file name to a path to exclude source file directories.

You can create as many search domains as you like. At least, I haven’t found an upper bound yet. I haven’t found a way to reorder the find scopes, although if you’re really controlling about this, you can pop into  your workspace (ProjectName.xcodeproj/project.xcworkspace/xcuserdata/username.xcuserdatad), convert your UserInterfaceState.xcuserstate to xml (plutil -convert xml1), and hand-edit it the way you need.

There are lots of wonderful little Xcode tweaks like these throughout this monster of an IDE. What are some of your faves? If I have time this week, I’ll share some of mine, such as the four-square — another of my favorite tools — and a few great ways to connect your editor to the navigator.

How I got Rust working in Xcode

A while ago, I posted about how I set up Xcode to work with Python. Yesterday, I was taking a class on Rust and decided to use my friendly neighborhood (sp)IDE(rman) coding environment, namely Xcode.

I’m not going to say it was a stunning success but there was enough interest that I thought I’d share the steps so you too could embrace Rust through Xcode.

Install Rust. You start, as one does, by installing Rust. Hop over to https://www.rust-lang.org/tools/install to grab a copy of the tools. They install to ~/.cargo, for whatever reason. I put a link in to / usr/local/bin.

Create a Project. Create an external build system Xcode project by choosing File > New > Project > Cross-platform > External Build System > Next. Enter a product name (I called mine “Rust” because that’s exactly how creative I am.) and set your build tool (in my case, /usr/local/bin/rustc because of the link). Save it somewhere convenient.

Create a source file. Apparently “rs” (rust source?) is the proper extension. I went with “test” as my name. File > New > Empty > test.rs

fn main() {
    println!("hello world");
}

Don’t forget to add some code.

Compile. Edit your scheme.  Choose Run > Info > Build Executable > Other and select your compiler. Adding it to /usr/local/bin made it easier to select rustc for me. Then uncheck Debug executable because you’re not debugging the Rust compiler.

At this point you can click Run and you’ll see the standard option message because you haven’t specified what it should run.

Back in the scheme editor select Run > Arguments and add the source file and output file. Unfortunately, I could not get this to work with SRCROOT at all, so here it is in all its glory with complete paths.

The Pre-action removes any build product from a previous run:

So here we are. With luck, it compiles. If not, the errors appear in pretty horrible form in the Xcode console, where curses is what we do, not how the console interprets pretty text output.

You can get slightly less horrible feedback by adding the launch argument: –error-format=json

Yeah, it’s wordy but it’s slightly less awful.

Pick a path. Unlike python, rust is just a compiler. If you build, and then add a step execute, the execution output (unlike compiler errors) will not normally print at the Xcode console. The challenge is to get that information in some form where you can access it.

At first I went with a little post-action osascript and threw up the output in a separate window:

But I really wanted to make it work with the console So back I went to Applescripting. Instead of rustc, I changed my build tool to osascript:

I added this instead to my run scheme arguments.

Yep, I’m using osascript to run a shell script that just compiles with rust and then runs it, passing the output through back to Xcode.

I know this is bad. I know I should be ashamed. I hang my head.

But you know what? It works. Stray osascript-crud and all:

I’m not sure how much this makes me a programming outcast but it was kind of fun to figure out how far I could push my beloved enemy Xcode.

SwiftUI: Modified Content, Type Erasure, and Type Debugging

When working with declarative views, you should be able to reach for a full tool suite of functional application. In a typesafe language like Swift, things can prove more difficult than you’d might first think. Consider the following code:

What is core‘s type? It isn’t Text. It’s actually an application of modified content, specifically Text passed through a rotation effect:

Just add a background color and a shadow and the type jumps to this:

You might ask: why is this a problem? After all, Swift is doing all the heavy lifting, right? In my case, the answer lies in my struggle to incorporate this core image into a multi-stage bit of text art using reduce. Paul Hudson tweeted a step-by-step approach to this and I was sure I could make it simpler and more elegant.

And that’s where I started throwing myself against what at first seemed like an impenetrable wall for a couple of hours. Between SwiftUI’s stroke-style Dysarthria error messages and the typesafe system, my attempt at creating a solution along these lines felt doomed:

[Color.red, .orange, .yellow, .green, .blue, .purple].reduce(core) { view, color in
  view.padding()
    .background(color)
    .rotationEffect(theta)
}

The code wouldn’t compile and the error messages couldn’t tell me why. The problem? Each stage created a new layer of modified content, changing the type and rendering reduce  unable to do the work. It was only with the help of some deep-dives into the docs and advice from colleagues that I was able to arrive at a solution.

Type erasure, using SwiftUI’s AnyView struct enables you to change the type of a given view, destroying the nested hierarchy. Importantly, it creates a single type, allowing a reduce operation to proceed.

At first, I used AnyView the way you’d typecast in Swift, namely:

AnyView(view.padding()
  .background(color)
  .rotationEffect(theta))

But I hated that. It sticks out as so Un-SwiftUI-y, with the parentheses spanning multiple lines and throwing off the clear logical flow. If you’re going to go fluent, just go fluent. So, eventually, I decided to create a View type extension to handle this:

extension View {
  /// Returns a type-erased version of the view.
  public var typeErased: AnyView { AnyView(self) }
}

The result looks, instead, like this:

view.padding()
  .background(color)
  .rotationEffect(Angle(radians: .pi / 6))
  .typeErased

And yes, I went with a property and not a function as I felt this was expressing a core characteristic inherent to each View. I can probably argue it the other way as well.

From there, it wasn’t much of a leap to ask “what other fluent interface tricks can I apply”, and I ended up putting together this little View extensions for inline peeks:

extension View {
    /// Passes-through the view with debugging output
  public func passthrough() -> Self {
    print("\(Self.self): \(self)")
    return self
  }
}

This prints out an instance’s type and a rendering of the instance, which will vary depending on whether there’s a custom representation, passing the actual instance through to whatever the next stage of chaining is. I don’t use it much but when I do, it’s been pretty handy at taking a peek where Xcode’s normal QuickLook features hit the edge.

In any case, I thought I’d share these in case they’re of use to anyone else. Drop me a note or a tweet or a comment if they help. Cheers!

Update: It suddenly occurred to me that I could make this a lot more general:

extension View {
  /// Passes-through the view with customizable side effects
  public func passthrough(applying closure: (_ instance: Self) -> ()) -> Self {
    closure(self)
    return self
  }
}

Isn’t that nicer? The equivalent is now:

struct MyView: View {
  var body: some View {
    [Color.red, .orange, .yellow, .green, .blue, .purple]
      .reduce(Text("👭")
        .font(.largeTitle)
        .rotationEffect(Angle(radians: .pi))
        .typeErased)
      { view, color in
        view.padding()
          .background(color)
          .rotationEffect(Angle(radians: .pi / 6))
          .passthrough { print("\(type(of: $0)), \($0)") }
          .typeErased
    }
  }
}

And I can put any behavior in from printouts to timing to any other side effect I desire. To all the functional purists out there, I sincerely apologize. 🙂