Top-notch video tutorials for Swift developers

Now that we have some streaming output from the ffmpeg process, we can take this and create a custom progress bar. Since this could useful in other utilities we can create it as a separate Swift package and import it using a local file reference.

When dealing with long running tasks it would be nice to be able to gather output as the task runs and not hang until the entire process is done. In this episode we will extract some useful information out of ffprobe, so we can get the total number of frames in a video file, and then kick off ffmpeg to encode the video. We'll use the Subprocess package to provide a simpler interface over gathering output from a pipe as it is sent, rather than waiting for the end.

I always like it when command line programs offer colored output, which helps to discern between various stages of output, highlight errors and other important information. This is done with ANSI color codes, which are a somewhat archaic technique, but we can implement a nice wrapper to make it easy to work with in Swift.

In this episode we will launch other programs as sub processes and show how to capture its output.

Working with files and folders with FileManager is cumbersome. Instead we'll lean on the excellent Files package from John Sundell.

I use a collection of command line scripts that help me in the production of NSScreencast episodes. Most are written in Ruby, which is a language I know and love, and some are written in bash, which is... a language. I sometimes wonder what these scripts would look like if I were to write them in Swift. So in this new series I'm going to explore rewriting a script that I use to encode videos using Swift. In this episode we'll bring in the excellent Swift Argument Parser library and use it to give us a clean and consistent command line interface to start with.

Pigeon is an interesting and powerful library for SwiftUI and UIKit for fetching and caching data. In this episode we will explore the library and how we can use it to build a simple Star Wars Universe explorer app using the Star Wars Dev API.

Foundation has a wonderfully useful support for dealing with measurements like distances, durations, temperatures and more in Swift. The Measurement class combines with a Unit type to constrain the measurements into their own domains. With this you can get functionality like conversion and addition. With MeasurementFormatter we can format these values for display in our UI based on the user's locale. In this episode we will explore how to use this API, how to extend the existing dimensions to add new measurement values, and how to create your own domain of measurements. For this we will model Throughput of Items in the game Factorio.

One of the benefits we can get by programming to protocols and leaving our concrete implementations isolated, is that we enable swapping out these dependencies for different ones for use during tests. Doing so will require us to restructure our app setup a bit so we can have a completely separate AppDelegate and SceneDelegate for use in testing. With this setup we can control the environment completely. We'll setup mock dependencies that we can verify that the interaction with these dependencies is as expected.

Managing dependencies can be a pain. Unless you're using shared instances accessible everywhere, you end up having to pass along dependencies from class to class. One strategy for decoupling your types from external dependencies is to create protocols and "inject" the actual concrete implementations at runtime. This enables you to test them easily and also isolates library-specific APIs from the rest of your application. This can make switching between libraries more feasible. In this episode we'll take a look at a library called Resolver that does this. Technically it implements Service Location rather than Dependency Injection, but the result is similar.

Using Item Anchors we can position decoration views anchored to our layout items. This is could be used to outline or underline items, or in our case to add a little unread indicator badge to our items.

Not every layout will be appropriate for every screen size. In this episode you'll see how you can use the layout environment when constructing your compositional layout to provide a more suitable layout for iPad.

One of the things you'd get for free with UITableView in the grouped style is a nice rounded rect background around your sections. With UICollectionView, you can implement these with background decoration views. We'll see how to set these up in our compositional layout to give the series section a different feel.

Our Diffable Datasource and snapshots are generic over the type of data that we pass to the cells. So how can we make sections with completely different data? In this episode we'll cover one approach which involves defining an enum with associated data for each of the sections. We'll use this to add a Series strip of data mixed in with our collection view.

In this episode we'll make a custom header view to give some of our sections a title. The approach we use here with compositional layout is more flexible than with UITableView. We'll start with a UICollectionReusableView implementation for our header, add the desired item to our layout, and then vend the desired view using the datasource's supplementaryViewProvider.

Diffable datasources provides a great API for driving your collection view updates in a transactional, state-driven way. We no longer have to manually call insert/delete/move rows when the data is changed. Instead, we apply a new snapshot and the changes are made for us, including animations.

First introduced in iOS 13, UICollectionViewCompositionalLayout is an amazing and powerful addition that gives you lots of flexibility when describing layouts. There are a few new types to get used to (namely sections, groups, and items) but they all work together allowing you to keep layout separate from your views and your data.

With UITableView no longer being encouraged for use, we need to replace this behavior with UICollectionView. This is where UICollectionViewListLayout comes into play. Using this layout we can get the familiar table view appearance in plain and grouped styles (as well as additional styles to support sidebars on iPad and macOS). This includes support for sticky headers and footers, swipe actions, and other UITableView behaviors that we've come to rely on.

In this episode we migrate our collection view to use the new cell registration API. Using this API we no longer need to cast dequeued cell types to our custom types. Instead, we set up the registration object with the cell type and the data we'll be passing to each cell. This further reduces the code we have to write in our datasource implementation and gives us more flexibility on how and where cells are configured.

In this episode we review the basic example app and start setting up our collection view in code. We start with the basic flow layout which is most common. Later we'll refactor this to use the newer style, but this episode introduces the series and sets up the foundation we'll build upon.

If you can describe your animation with a small number of parameters that interpolate over the animation, animation is pretty easy. But once you want to combine an arbitrary number of animatable data values, for instance an array of Doubles representing our points, then you have to resort to a custom VectorArithmetic implementation. In this episode we will create an AnimatableVector type that is capable of animating between arrays of Double values. We then use this type to animate between 2 sets of points using our waveform algorithm.

Animating in SwiftUI can seem like magic, until you want to animate something custom. In this episode we'll add animation to our frequency, amplitude, and phase parameters for our waveform by leveraging SwiftUI's AnimatableModifier protocol. We'll see how to implement the animatableData property with one, two, and ultimately all three parameters. We'll cover implicit and explicit animations, as well as the behavior when attempting to mix animations.

I've been working on rendering waveforms using mathematical functions and have found the experience to be both fun and enlightening. In this episode we will develop a method to render arbitrary functions using a Shape, then explore some mathematical concepts that can help us render a nice looking waveform that could be use to indicate activity in sound, speech, or other effects.

Simple animations in SwiftUI are refreshingly easy. However, some animations are deceptively tricky, as we don't have access to completion handlers. Animation is entirely state-based, so if we have an animation where an item needs to move and return to its original state, we need to take a different approach. In this episode we will model a bounce animation using a simple sine function, feeding in the time value t using SwiftUI's linear animation interpolation.