Tag Archives: ios

Swift Hate

I’m seeing a surprising amount of vitriol aimed at Swift, Apple’s new programming language for iOS and Mac development. I understand that there can be reasoned debate around the features (or lack thereof), syntax and even the necessity of it but there can be little doubt about the outcome: if you want to consider yourself an iOS developer, it’s a language that you will need to learn.

The only variable I can think of is when you learn it.

I think it’s perfectly reasonable to delay learning it as you have code to deliver now and because Swift is, well, very beta currently.

But asking what the point of Swift is not constructive. Asking what problems can be solved exclusively by Swift makes no sense at all you can do almost anything in most programming languages. Just because Intercal is Turing complete doesn’t mean that you’d want to use it for any real work. What varies between languages is what’s easy and what’s hard.

Objective-C undoubtedly makes some things easier than Swift. It’s a more dynamic language and its C foundations open up a lot of low-level optimisations that probably are not there in higher level languages.

But that same flexibility comes with a price: segmentation faults and memory leaks; pointers; easy-to-get-wrong switch statement; a lack of bounds checking. It also inherits a lot of ambiguity from original C language specification which makes certain automatic optimisations impossible.

How many applications require low-level optimisations more than safety? (And that’s ignoring that the biggest optimisations are usually found in designing a better algorithm or architecture.) How often is it better to have five lines of code instead of one? Every line is a liability, something that can go wrong, something that needs to be understood, tested and maintained.

Whatever its failings, it’s already clear that Swift is more concise and safer than Objective C.

There are absolutely applications where Objective C, C or C++ would be a better choice than Swift. It’s the old 80-20 rule applied to a programming language. And, for those resistant to learning a new language, the 80% is not on “your” side.

Right now, some of this requires a bit of a leap of faith. Swift clearly isn’t finished. You can either learn it now and potentially have some say on what the “final” version looks like, or you can learn it afterwards and just have to accept what’s there. But, either way, you’ll probably using it next year. Get used to it.

Learning Swift

Swift is a new programming language designed by Apple for development on OS X and iOS. I thought that I should try to learn it a little so I decided to convert a non-trivial collection of classes from one of my apps (www.cut) into Swift. I always find it better to work on a real project rather than just to play around with things aimlessly. Also, by re-working an old project, I knew that all the problems I would find would be language related rather than anything to do with the architecture.

The classes are also data related rather than being UI, so it is mostly a test of the language itself rather than how it interfaces with Objective C.

First impressions are good. Swift is mostly nice and consistent, which although sounding like damning with faint praise, is actually a compliment. I read a little of the language guide and dove straight in. A lot of the attempts to get the following right were me just typing stuff, guessing the syntax rather than looking it up.

Quite by accident I think my code sample inadvertently shows an area of strength for Objective-C and weakness for Swift.

The idea of the code is that it reads a Plist, instantiates a class based on that configuration and fills in a number of properties.

The first half of the code looks like this:

        NSError* error = nil;
        NSString *plistPath = [[NSBundle mainBundle] pathForResource:@"XXX" ofType:@"plist"];
        NSData *plistXML = [[NSFileManager defaultManager] contentsAtPath:plistPath];
        NSDictionary *temp = (NSDictionary *)[NSPropertyListSerialization propertyListWithData:plistXML
                                                                                       options:NSPropertyListMutableContainersAndLeaves
                                                                                        format:NULL
                                                                                         error:&error];
        NSArray* values = [temp objectForKey:@"Entries"];

This was pretty straight forward to convert into Swift, though the type system gave me issues:

    let plistPath =  NSBundle.mainBundle().pathForResource("XXX", ofType: "plist")
    let plistXML = NSFileManager.defaultManager().contentsAtPath(plistPath)
    var error:NSError? = nil
    var format:CMutablePointer<NSPropertyListFormat>? = nil
    let immutable:NSPropertyListReadOptions = 0
    var pList = NSPropertyListSerialization.propertyListWithData(plistXML,
        options:NSPropertyListReadOptions(NSPropertyListMutabilityOptions.Immutable.toRaw()),
        format:format!,
        error: &error) as NSDictionary

Getting the options property seems very clumsy; I’m sure that there must be a better way of doing it.

I had a real problem with the in/out parameters format and error. Not only was the documentation confusing but the Swift Playground kept crashing making it difficult to distinguish between what I was doing wrong and where the compiler itself was messing up. It’s also a bit odd that, though both are in/out parameters, that they both need different methods to extract the values.

(To be fair, this is a beta and it is the first version of a whole language and compiler. I mention the crashes not because they’re unexpected or even especially bad, just as an honest description of the difficulty I had.)

The next section, using the data in the plist, was much more problematic. The code looks like this, but I’ve trimmed a lot so what we have here isn’t terribly useful now!

        proxy = nil;
        for (NSDictionary* i in values) {
            if ([thing isEqualToString:[i objectForKey:@"Class"]]) {
                dynamicClass = NSClassFromString([i objectForKey:@"BaseClass"]);
                proxy = [[dynamicClass alloc] init];
            }
        }

I didn’t do the straight conversion. A few more years of Cocoa programming allowed me to notice an optimisation:

    let valueList = values.filteredArrayUsingPredicate(NSPredicate(format: "Class = %@", value))
    let valueData = valueList[0] as NSDictionary

This same approach would work in Objective-C.

Next I tried:

  let dynamicClass = NSClassFromString(valueData["BaseClass"])
  let proxy = dynamicClass()

The first line works as expected. The second line doesn’t compile.

Is there anything else that we can do with dynamicClass? Let’s see. It’s an AnyClass which is a type alias for AnyObject.Type. Which doesn’t really help.

I tried casting it to a base class but no matter what I tried I couldn’t alloc/init it (in Objective C terms).

Josh Smith figured out how to do it by creating a factory class in Objective-C.

I tried (and failed) to get it to work by calling some of the Objective-C runtime directly:

var bytes:Byte[] = [0,0,0,0]
let b = objc_constructInstance(a, &bytes)

But the second line doesn’t work when using ARC. (To be fair, Xcode struck through the definition so I didn’t have much confidence that it would work!)

So that leaves Josh’s call out to Objective-C to be the best method that I’m aware of.

In the end I just used a switch statement to select between the relatively limited number of options that I had to choose between. Not as clever, but maybe that’s a good thing?

Not so smart phones

The flood of new so-called smart watches continues. Some people seem to love theirs, others remain to be convinced.

Count me in with the unconvinced, though only because the current ones seem to be poorly conceived.

Marco Arment says:

Portability is critical to modern device usefulness, and there are only two classes that matter anymore: always with you, and not… Smartphones dominate always with you.

I think this gets to the heart of why the current range of devices — both those for sale and also those just announced at CES — just are not very compelling.

Let’s ignore for the moment the fact that most of them look awful.

Actually, no. Let’s not. You can’t sell a device for hundreds of pounds, one designed to sit on your wrist, replacing the only jewellery that many men wear, and make it look like a digital watch from 1981. I half expect the next smart watch to have a calculator keyboard on like the old Casios.

(For what it’s worth, I think new Pebbles are a big step forward over the original version. Unfortunately I like my super-thin Skagen so I’d still consider it a long way from acceptable.)

But yes, let’s assume the form-factor was more pleasing. Then it still doesn’t work. They’re not replacing anything. They’re accessories for an already expensive and always-with-you device. Sure, looking at your wrist is easier than getting your phone out of your pocket, but is it really that much easier? Is it several hundred pounds better? Is it worth the hit on you phones battery life and the inconvenience of yet another device to charge? I think most people will conclude, no.

So in summary, I think that smart watches have two main problems:

  1. Aesthetics
  2. They’re companion devices

The first is solvable with advancing technology and a degree of taste. The latter means that not every manufacturer will solve it but once the components become small enough putting them in an attractive case becomes possible.

Moore’s Law can partly solve the second point, too, but it’s not enough on its own. You’d also need changes in the way you interact with the device if it’s to be a full replacement for a smartphone.

I don’t think the answer to that is speech. Sure, Siri will get better, but there are places where you can’t or wouldn’t want to talk to your devices. And it would be hard to do an Android and get bigger and bigger screens — at least until we evolve to have bigger fore-arms.

Instead, I wonder if smart-watches are really a bit of a technological dead-end. If over time components tend smaller and smaller, then why stop at wrist watch size?

The other side of the equation is the smart phone. Is it really the best form-factor that we can possibly imagine? Do we use smart phones because they are the best place to put small computers, radios and piles of sensors?

Or put another way: if you could have the same functionality that’s currently in your smart phone in the form of a watch, would you take it? If you could take all that functionality and not even have to wear a watch, would you take it?

The smart phone is a great compromise. It’s small and with you most of the time. But you still have to carry it. You can still easily lose or drop it and break it.

Smart watches and Google Glass try to solve these problems but, as Marco says, they do so with some pretty serious draw-backs. The smartphone is better for most people right now but that won’t always be the case.

What you forgot from your Computer Science Degree

Last night I did a short presentation about my WSLHTMLEntities open source project at the London iOS Developer Group meeting. You can see the slides here:

Since last time I did a talk there people snickered because I built the slides using PowerPoint, this time I decided to use the latest Apple technology: Keynote in iCloud. Unfortunately this was a bit too new for the Mac Pro they use in the Apple Store, so we ended up downloading a copy in PowerPoint format and loading that into the local copy of Keynote. Nothing is ever simple.

One question I got at the end that I was unable to answer is how well it performs compared to other solutions.

This afternoon I ran a few quick tests.

Firstly I added two further methods of performing the HTML entity replacement:

The last option, while built-in and easy to implement, is not something you’d want to consider if you had any performance requirements. First, it has to run on the main thread. Second, it’s memory footprint is way higher than either of the other two solutions. And, finally, it’s slow. In my tests I did a loop of 10,000 iterations. Because of the memory problem I only ran 1000 iterations of the NSAttributedString solution, but it was still about 15 times slower than my version when doing the full 10,000 records (i.e., over 150 times slower overall).

So I’ve not included it in the charts below.

HTMLEntityChart

I ran eight simple tests with each of the parsers.

There is no clear winner.

As I speculated, WSLHTMLEntities is much more consistent than the other two. Google’s varies a lot depending on where in the mapping list the “solution” lies. Replacing &loz; (near the end of the list) takes 50% longer than &amp; (near the beginning) for example.

Still, you’d have to be pretty obsessive to pick one of these solutions for performance reasons alone. Interesting to find out, though!

CameraGPS debrief

As happy as I am with the way that my new app, CameraGPS, a GPS logger application for people who want to geotag their photographs, came out I can’t say that it’s exactly as I envisioned it at the start of the process.

The idea was something like this: many of the GPS logger apps in the App Store require you to either use iTunes file sharing (who connects their iPhone’s to iTunes any more?) or mail yourself the exported document or sign up to some third party fitness or trekking website. Mailing yourself stuff just didn’t feel very slick and I didn’t want to record my trails for fitness purposes.

I felt that there must be a better way. The better way, I thought, was to use iCloud to share the files between iOS and a Mac. This would require writing a recorder app for iPhone and an “exporter” app for the Mac.

It didn’t work out that way in the end. Here’s why.

You see all kinds of horror stories from people who have had to deal with iCloud. Most of those stories come from people who used the Core Data syncing, while there are a few apps that successfully use the document syncing model, not least some of Apple’s own. Using the document storage I figured that, worst case, if something gets screwed up it would only affect a single trail. Using Core Data syncing, an error could mean corrupting the whole database.

So this became the plan: record each GPS trail into a separate document. Because I wanted to “random access” parts of each trail and because I didn’t want to have to load the whole trail into memory all the time, I decided to use Core Data to manage the document. On iOS this is called UIManagedDocument.

The first and most obvious fail here is with UIManagedDocument’s cross platform capabilities. Or rather the lack of them. Unless you count iPhone to iPad as cross platform it just doesn’t work. There is a class on the Mac called NSPersistentDocument that is also a Core Data store and looks on the surface to do exactly the same thing… but it’s in a different format. On iOS it’s a “package” which includes a Core Data store. On the Mac it’s just the store. Mike Abdullah has written an open source component that attempts to be bridge the gap but — as we’ll see shorty — I didn’t get to the point where I was able to test it in anger.

But let’s take a step backwards. Let’s assume that all we need is to move data between iPhone and iPad. Then what?

Well, technically it works — in the sense that documents transfer between machines — just not in a way that would make sense to ship.

To make sense of why it doesn’t work let’s take a quick diversion to look at the format of a UIManagedDocument. As hinted above, it’s not a file. Rather it’s a “package” which a number of files within.

An important file is called DocumentMetadata.plist. This tells you about the rest of the package and is created when the document is. There’s little in here that needs to be changed after document creation. The interesting stuff is all stored in a SQLite file, which is stored in a folder.

When files are stored in iCloud you can’t just use NSFileManager to poke around and see what’s there. Instead you create a query using NSMetadataQuery — which supports both a predicate and a sort descriptor — and listen for updates. Updates come in two parts, the initial state and then changes to that state.

After a few false starts I ended up using KVO to populate a table view showing the current list of documents. This worked nicely, showing the full list, updating automatically when changes arrived over the network.

But, and it’s a big but, what if you want to sort the list. Pretty basic requirement, no?

One of the fun things about NSMetadataQuery is that it can only look for files. Folders are “invisible.”

Most of Apple’s documentation — where there is any — suggests searching for the DocumentMetadata.plist. Which makes sense but with two significant drawbacks: every file that you’re looking for now has exactly the same name; and the modification date is the creation date of the document rather than the time of the last update.

This means that you can’t sort the documents alphabetically or by modification date using only the metadata predicate. To get the names you need to listen for updates, copy them elsewhere and then butcher the file name, that is removing the last component and possibly the file extension.

To get the last modification date you probably need to open the document and do a Core Data query. This sounds pretty straight-forward but, again, there are difficulties. Remember that NSMetadataQuery tells us what’s available in iCloud rather than what has been downloaded onto the current device. Just because the query tells us the file is in iCloud doesn’t mean that you can open it immediately.

UIDocument helps us a bit here: if you ask to open a document it will download it if necessary. However there’s a big difference between opening a local file and downloading one over the network and then opening it. Also remember, at this point we’re not opening it as the result of a user request. We’re just trying to find the last modification date so that we can sort the list of documents conveniently.

None of this, of course, is impossible. It’s just a lot of faff for something that pretty much every app that wants to use UIDocument must have to go through.

Okay, so we get all that done. We’re good to go now, right? Not quite.

Part of the idea of using a UIDocument was that it would appear as a, well, document. But that whole thing about what NSMetatadataQuery “sees” comes back to haunt us.

The short, weird version is that if you configure the application to look for DocumentMetadata.plist, then that’s what is seen in the Settings app. Not very helpful as all the documents will all have the same name.

On the other hand you can configure the app to “know” about the document type. This means that the Settings app understands that they’re document and displays them with the correct “file” name. Unfortunately then the app can’t see the documents properly. Specifically, you can search for the document by name — which is nice — but you can’t actually open it. It’s odd and I’m not entirely sure what goes wrong, but when you call openWithCompletionHandler: it never actually calls the completion handler and the document never becomes available for use. (I initially thought that I just wasn’t being patient enough. I later accidentally left the app running for over two hours and the completion handler still never ran. That convinced me that it wasn’t a patience thing.)

I convinced myself that this couldn’t possibly be how it was supposed to work. My question on Stack Overflow didn’t get any answers and a reply to another, similar question said that the behaviour was “by design.”

I think it’s just broken and that no one is really using it. Either that or there’s a secret handshake that you have to give to the right person at WWDC.

So what’s shipping? Pretty much none of the above. I read about the new iOS 7 methods of Core Data – iCloud syncing and liked what I saw. Additionally iOS 7 comes with a few more APIs with better error handling hooks. This made me happier shipping with Core Data syncing. Just using “raw” Core Data meant that I could simplify a lot of code, use NSFetchedResultController and remove a bunch of the more fragile iCloud code (mostly NSMetadataQuery). While I was quite pleased with the KVO code I wrote, having far less code overall is a much bigger win.

And, in case you were wondering, I also de-scoped the Mac app and integrated Dropbox support instead.

But the bottom line is that this is all a lot harder than it should be. My concerns originally were around error conditions — something that has been very weak in iCloud since it’s introduction in iOS 5 — but it quickly morphed into questions about how it’s all supposed to work at all.

I once quipped on Twitter that you can tell which APIs are the ones Apple uses themselves as they’re the well designed ones. I think it’s pretty clear that Apple don’t use UIManagedDocument.

NSFetchedResultsController and iCloud

This took me a while to figure out so I thought it was worth blogging about. The short version: I’m using Core Data with iCloud syncing and it works… mostly. When starting up for the first time — when there is already data in iCloud —  none of the data appears in a table view, but restarting the app correctly displays it.

I know what you’re thinking: you’re not merging the updates into the right managed object context. Nope. Sorry. Thinking that was the problem is probably why it took me quite so long to track the real problem down!

So, what does the problem look like?

I set up my Core Data stack in the app delegate. Part of that includes connecting to iCloud:

    NSURL *storeURL = [[self applicationDocumentsDirectory] URLByAppendingPathComponent:@"XXX.sqlite"];
    NSDictionary *options = @{
                              NSMigratePersistentStoresAutomaticallyOption: @YES,
                              NSInferMappingModelAutomaticallyOption: @YES,
                              NSPersistentStoreUbiquitousContentNameKey : @"XXX"
                              };
    NSError *error = nil;
    _persistentStoreCoordinator = [[NSPersistentStoreCoordinator alloc] initWithManagedObjectModel:[self managedObjectModel]];
    if (![_persistentStoreCoordinator addPersistentStoreWithType:NSSQLiteStoreType
                                                   configuration:nil
                                                             URL:storeURL
                                                         options:options
                                                           error:&error]) {

The app delegate passes the managed object context through to the main view controller which then uses it to create a NSFetchedResultsController:

    NSFetchRequest *fetchRequest = [[NSFetchRequest alloc] init];
    // Edit the entity name as appropriate.
    NSEntityDescription *entity = [NSEntityDescription entityForName:@"XXX" inManagedObjectContext:self.managedObjectContext];
    [fetchRequest setEntity:entity];
    // Set the batch size to a suitable number.
    [fetchRequest setFetchBatchSize:20];
    // Edit the sort key as appropriate.
    NSSortDescriptor *sortDescriptor = [[NSSortDescriptor alloc] initWithKey:@"someTimestamp" ascending:NO];
    NSArray *sortDescriptors = @[sortDescriptor];
    [fetchRequest setSortDescriptors:sortDescriptors];
    // Edit the section name key path and cache name if appropriate.
    // nil for section name key path means "no sections".
    NSFetchedResultsController *aFetchedResultsController = [[NSFetchedResultsController alloc] initWithFetchRequest:fetchRequest
                                                                                                managedObjectContext:self.managedObjectContext
                                                                                                  sectionNameKeyPath:nil
                                                                                                           cacheName:@"Master"];
    aFetchedResultsController.delegate = self;
    self.fetchedResultsController = aFetchedResultsController;
	NSError *error = nil;
	if (![self.fetchedResultsController performFetch:&error]) {
        // Replace this implementation with code to handle the error appropriately.
        // abort() causes the application to generate a crash log and terminate. You should not use this function in a shipping application, although it may be useful during development.
	    NSLog(@"Unresolved error %@, %@", error, [error userInfo]);
	    abort();
	}

(Again, this is pretty much Apple-standard boilerplate code —  nothing clever going on here.)

But what I found was that when there was already data in iCloud, the records did not automatically appear in the main view controller.

Eventually I found that this was not a threading issue — merging the changes into one context and trying to read them from another — by adding an explicit fetch request on a “debug” button. Doing this I could see the new data, even though the fetched result controller could not.

In my app delegate I listened for a number of notifications: NSPersistentStoreDidImportUbiquitousContentChangesNotification and NSPersistentStoreCoordinatorStoresWillChangeNotification. My expectation was that NSPersistentStoreCoordinatorStoresWillChangeNotification would fire before switching from the fallback store to the “real” one and NSPersistentStoreDidImportUbiquitousContentChangesNotification would fire when the new data was available.

I was half right. NSPersistentStoreCoordinatorStoresWillChangeNotification fired on a background thread, so I used GCD to ping it onto the main thread and reset the context.

But NSPersistentStoreDidImportUbiquitousContentChangesNotification didn’t fire at all. I guess the objects didn’t change as such, they just became available, which feels like a slightly false distinction to me.

So my next guess was to fire the fetch request again. The data was visible in the main threads context so surely this would find the data?

Nope. (And don’t call me Shirley.)

I was getting pretty lost at this point so I ended up just semi-randomly stopping the code and looking around in the debugger.

And, long story short, I realised that fetch requests have a reference to the persistent store in them — check out the affectedStores method. This meant that the NSFetchedResultsController was happily, and correctly, reporting on the empty and no longer used fallback store and completely ignoring the new and fully populated iCloud store.

The simple solution was to listen for the NSPersistentStoreCoordinatorStoresDidChangeNotification and create a completely new fetch request.

- (void)storesDidChange:(NSNotification*)notification {
    self.fetchedResultsController = nil;
    [NSFetchedResultsController deleteCacheWithName:@"Master"];
    [self.tableView reloadData];
}

(I did think of just adding the new persistent store to the old fetch request but I wasn’t sure that this would create the refresh anyway and, given the frequency with which this is likely to happen, I thought it would be cleaner just to start from scratch.)

These weird problems almost always boil down to just a couple of lines of code. This time was no exception.