Glad You're Ready. Let's Get Started!

Let us know how we can contact you.

Thank you!

We'll respond shortly.

Wrapping Objective-C Delegates with Blocks, Part 1

Since adding blocks to Objective-C, Apple has designed more and more of their APIs to use them as the preferred callback mechanism over delegation—and for good reason. Blocks are (relatively) easy to use, lightweight, they play well with GCD, and most importantly for this post, they put the callback code right where it has the most context, thus making your code easier to read and reason about.

Read more »

Test Driven iPhone Development with Cedar, Part II

Co Author: Andy Pliszka

In Test Driven iPhone Development with Cedar, Part I, we created a new Xcode project called Recipes and set it up to use Cedar for test-driven development. In this post, we’ll test-drive our first piece of actual functionality.

Read more »

Test Driven iPhone Development with Cedar

Co Author: Andy Pliszka

Cedar is an open source BDD testing framework from Pivotal Labs that makes test driven iPhone development quick and easy. The framework provides a large library of matchers so you can start testing right away on a large collection of objects.

Read more »

Cedar target templates for Xcode 4.3.x

Announcing Cedar target templates for Xcode 4.3.x: Getting started with Cedar has never been easier!

These templates let you quickly add Cedar targets to your iOS or Mac OS X project. It's super simple to build and install from source now:

git clone --recursive && cd cedar ; rake install

If you don't like building from source, you can also install a build of the Cedar templates from the downloads page - just untar to your home directory.

Using Cedar Templates

Restart Xcode after installing, open your project and add a new target. You'll find targets for both iOS and Mac OS X. There are 2 types of targets: Suites and Bundles. A Suite target will build your tests as a separate application which you then run to execute your tests. Bundle targets work just like OCUnit testing bundles, providing closer integration with Xcode. If you do create a bundle target, make sure to edit the scheme of your application and add the new target to the list of tests.

Every template comes with a simple example which should run out of the box to help you get running.

ARC: A word of caution

Automatic Reference Counting setting: Don't use it for your spec targets. Your application is free to use ARC, just don't use it for any Cedar targets. This is because there is a flaw in the compiler that ships with Xcode (see here for the details, and if you care please feel free to use this information to file a bug with apple).

Have fun speccing

We really hope these changes help you to get started writing specs and spending less time setting up Cedar. If you have any feedback, whether it's about these templates or something else about Cedar, please feel welcome stop by the discussion group and share your thoughts.

Cedar and OCUnit

tl;dr Latest version of Cedar can nicely integrate with Xcode just like OCUnit.


When creating new Xcode project you might have noticed that New Project dialog has 'Include Unit Tests' checkbox at the bottom. When that option is selected in addition to creating an app target Xcode will also create a separate target that is used for testing your application. This target most likely will be named [app]Tests. Convenient, huh! Since Xcode is such a nice IDE[1] it will also create an example test file that uses Xcode's built-in unit testing framework -- OCUnit (also known as SenTestingKit). If you never heard of it below is a screenshot of how that example OCUnit test file looks.

Because Apple ships OCUnit with Xcode it is tightly integrated into the IDE and that gives it somewhat of an advantage over other testing frameworks. Example of such unfair treatment is relatively nice looking red error markers (shown above) that show up next to failed test cases when application tests are run. If you have ever used Cedar before you would know that when Cedar specs fail we do not get those markers but rather we are presented with boring gray console output (we have dots though!). Unfortunately there is no documented way for custom testing frameworks to mimic OCUnit behavior without building on top of it. But even if do decide to use OCUnit as a base for your custom testing framework you will run into several problems. One of those problems, at least a year ago, was that you could not reliably use debugger with OCUnit tests. That was one of the main reasons why Cedar was not build on top of it. Recently, however, Xcode 4 was released with a lot of enhancements including better support for running OCUnit tests (with or without debugger) or so it seems. Red markers being a driving force, once again it was worth to investigate on how take advantage of OCUnit goodies from Cedar. Latest version of Cedar succeeds at that. What follows are the steps on how to setup Cedar for maximum Xcode integration.

Setup steps

Check 'Include Unit Tests' checkbox in New Project dialog. If you want to add application tests to an existing project here is a very nice tutorial:

After Xcode is done creating a new project file go to Build Settings for application tests target (in this case ExampleTests) and find Other Linker Flags settings. Add -ObjC, -all_load, and -lstdc++ to that setting and click Done.

Next step is to add Cedar static library to application tests target.

Now you are ready to create a Cedar spec file. Create a new empty file and end the filename with ''. 'mm' file extension is needed to be able to use built-in Cedar matcher library. Also since this file should only be run as part of application tests and not the actual application make sure that you only include it in application tests target.

Add a test or two (see above for example) to be able to test that Cedar is working correctly. Example above tries to assert that 2+2 equals 5 and that should hopefully fail. Notice that here I'm using recently added Cedar matchers should syntax. If you do not like that you can always rewrite that expectation as expect(2+2).to(equal(5));

Larger image. And finally to run specs you just wrote select scheme for your app (in this case Example) from the drop down in the upper left corner and press 'Cmd+U'. 'Cmd+U' in Xcode by default runs tests associated with selected scheme. That should first build your app (because application tests target lists the app as a target dependency), then build application tests target and then run the tests. Since 2+2 does not equal 5 you will see that Xcode will highlight that line in red and show that as an error in the errors panel on the left side. Fixing that expectation should make the test pass.

Running from the command line

If you want to run your application tests from the command line you can use rake ocunit:application to do that. Before doing so make sure you have updated APP_NAME and OCUNIT_APPLICATION_SPECS_TARGET_NAME in the Rakefile. Also you'll need to use Debug for CONFIGURATION so that your tests are able to include classes from your application. What's nice about this rake task is that it is ready to be used on your CI machine since it returns exit code 1 when any tests fail and 0 if they all succeed. It also prints in color.

How it actually works

If you are wondering how Cedar pretends to be OCUnit here is a short summary:

  • When Cedar is loaded it overrides runTests class method on SenTestProbe class (related file to look at is CDROTestIPhoneRunner.m)
  • SenTestProbe +runTests gets called by otest -- an executable that Xcode uses to run tests after it injected the application tests bundle into TEST_HOST app
  • Since SenTestProbe +runTests was overridden with Cedar's run method Cedar executes all of its specs following all the usual rules: focused, pended, etc.
  • When specs are running CDROTestReporter (source) gets notified which tests passed, failed, got skipped, etc. and prints out appropriate information which you see in console log
  • CDROTestReporter prints out failed tests in a way that is recognized by Xcode (OCUnit spits out same output for failed examples) e.g.:
Test Case '-[ExampleTests testExample]' started.
/Users/work/workspace/Example/ExampleTests/ExampleTests.m:16: error: -[ExampleTests testExample] : '4' should be equal to '5': 2+2 != 5
Test Case '-[ExampleTests testExample]' failed (0.000 seconds).
  • Xcode notices that pattern in the console output, parses it, gets filename and line number and marks that file location with a red marker to indicate a test failure

Drawbacks/Not implemented features

  • Debugger support is still a bit flaky but this is a problem with Xcode/OCUnit integration
  • I have not tried running it on the device but it should work
  • If you are used to specifying -SenTest option to run specific test case it will be ignored; however, you can use Cedar's fit, fdescribe, and fcontext to focus on specific tests
  • Some third party OCUnit add-ons might not work (let us know about them)

Useful links

[1] It's not.

Too many parentheses

Cedar comes with its own built-in matcher library and here is Adam Milligan's blog post that explains on how to use it. An example of expectation written using Cedar matcher library:


Even though this is much better than OCHamcrest it is still not satisfying (in my opinion) because there are so many parentheses involved. Since Cedar matcher library is written in C++ we can do some magic here. The result is you can write your expectations like this:

controller.button should_not be_nil;

So let's dig into how Cedar's should syntax is implemented.

  controller.button   should_not   be_nil
    actual value       C macro     matcher

As you might have guessed should and should_not are C macros. Besides making possible for each expectation to know its filename and line number (with __FILE__ and __LINE__ preprocessor variables), they also hide the unnecessary complexity which makes this syntax possible. Here is how it looks expanded:

  controller.button   ,(ActualValueMarker){__FILE__, __LINE__},true,   be_nil
    actual value                        mumbo jumbo                    matcher

So how does this mumbo jumbo work? Since Cedar matcher library was written in C++ we can overload the comma operator. Besides being a weird operator to overload, it has a useful property that we can exploit here. Comma operator has the lowest precedence among all other operators and that allows us to write 2+2 should equal(4) instead of (2+2) should equal(4). Now let's break mumbo jumbo into pieces:

(1) <actual value>
    (ActualValueMarker){__FILE__, __LINE__} -> (2) ActualValue
                                                   true/false  -> (3) ActualValueProxy

(1) This step captures actual value (can be anything since it is templated), filename and line number. It is implemented like this:

template<typename T>
const ActualValue<T> operator,(const T & actualValue, const ActualValueMarker & marker) {
    return ActualValue<T>(marker.fileName, marker.lineNumber, actualValue);

(2) Now we overload comma operator to match ActualValue (returned from step 1) and a bool that indicates whether matcher result should be negated.

template<typename T>
const ActualValueMatchProxy<T> operator,(const ActualValue<T> & actualValue, bool negate) {
    return negate ? actualValue.to_not :;

(3) And finally we overload comma operator again so that it matches ActualValueMatchProxy (returned from step 2) and anything else. Second argument will be considered to be a matcher.

template<typename T, typename MatcherType>
void operator,(const ActualValueMatchProxy<T> & matchProxy, const MatcherType & matcher) {

That's pretty much all. Full implementation of should syntax (which is just several more lines more than what is shown above) can be found in ShouldSyntax.h (and corresponding spec file that makes sure syntax works). Everything above allows us to write expectations as follows:

controller.button should_not be_nil();

Since the goal is to get rid of as many parentheses as possible we still have some work to do. Nil matcher that comes with Cedar is implemented with BeNil class. The way it worked before was be_nil was a function defined in Cedar::Matchers namespace. Calling it each time would return a new instance of BeNil class. To make parentheses go away be_nil was turned into a static variable that is initialized only once to an instance of BeNil class. (be_nil matcher still supports being used with two parentheses for backwards compatibility by implementing call operator on BeNil class.) Bam.

controller.button should_not be_nil;

Usually it is not recommended to use operator overloading for non-operator related purposes (e.g. don't use + operator for subtraction); however, in this case I think it's fine to bend the rules just a bit to escape all those parentheses and provide clean expectation DSL. You can use should syntax with the latest Cedar master branch.

Cedar Expectations

As I wrote here we've used OCHamcrest matchers for some time for writing expectations in Cedar, but have found them unsatisfying. We wanted convenient matchers, like the ones Jasmine provides for JavaScript, but for Objective C. To that end, we added Cedar-specific expectation functions and matchers that specifically solve the problems we had with OCHamcrest.

To use Cedar's expectations you need to make a couple small changes to your spec files:

1) Cedar matchers use C++ templates. Tell the compiler to expect some C++ code by changing the file extension on your spec files from .m to .mm.

2) Cedar matchers live in a C++ namespace. At the top of your spec file, after the includes, add this line:

using namespace Cedar::Matchers;

The Trouble With Expectations in Objective C

At Pivotal we write a lot of tests, or specs if you prefer. We TDD nearly everything we write, in every language we write in and on every platform we write for, so we actively work to improve every aspect of our testing tools. Personally, as I've written tests in Objective C I've found that the syntax of expectations has left much to be desired.

Like many people who write Objective C, I've spent a fair bit of time with languages like Ruby and JavaScript. When I write specs I often yearn for the simplicity of the expectation syntax in those languages. For example, some simple expectations in JavaScript, using Jasmine:


In comparison, the same expectations in Objective C, using OCHamcrest:

assertThat(, equalTo(@"Ludwig"));
assertThatUnsignedInt([composer.symphonies count], equalToUnsignedInt(9u));
assertThat(composer.symphonies, hasItem(@"Eroica"));
assertThat(composer.symphonies, isNot(hasItem(@"Appassionata"));

Meanwhile, over in Objective-C land

Two new interesting open-source Objective-C frameworks bookended our week this week:

First, Peter Kim sent a message to the Cedar-discuss mailing list to introduce us to his new matcher framework Expecta. Now, instead of OCHamcrest's assertThatBool(maybe, equalToBool(NO)) shenanigans one can simply expect(2+2).toEqual(4); There's also a nice DSL for implementing custom matchers for objects.

Excited about the possibilities, Kurtis and I spiked on implementing a set of custom matchers for commonly-used Cocoa structs such as CGPoint, something we have found lacking in current tools and testing patterns. Our entire project team looks forward to writing more tests using Expecta.

Also, Twitter Engineering just announced TwUI, a new Core Animation-based UI framework for OS X. Fun times.

Cedar vs. Xcode 4 (round one: the command line)

I've finally found a bit of time to update Cedar to work with Xcode 4, and I hope to have it working smoothly some time in the next few days. However, I've already come across my first significant issue with the Xcode 4 changes: the location of build products.

Not unexpectedly, the problem has to do with command line builds using xcodebuild. By default, Xcode 4 now puts build products into a project-specific directory in the "Derived Data" folder; this looks something like /Users/pivotal/Library/Developer/Xcode/DerivedData/Cedar-somegianthashstring/Build/Products/Debug-iphonesimulator/libCedar-StaticLib.a. This isn't a problem, generally, because the BUILD_DIR compiler variable contains the build directory, should you need to find this location during the build process.

Sadly, when you build from the command line, using the xcodebuild command, the build products still go into the old Xcode 3 build location, but the BUILD_DIR compiler variable contains the new Xcode 4 build directory. This means any script that looks for the build results in the directory specified by BUILD_DIR won't find anything.

The build target for Cedar's static framework is simply a script that uses xcodebuild to build the static library for both the simulator and the device, and then uses lipo to make a fat binary from the results. Because it can't find the build results at the location specified by BUILD_DIR it now fails messily.

The easiest workaround I've found is to change where build products go using the Locations setting in the Xcode 4 preferences (details below). Unfortunately, this isn't a project-specific setting, so you'll have to change your preferences similarly to make it work. I haven't found any problems with changing the location of the build products, but this does mean the Cedar static framework (as well as the related static frameworks for OCHamcrest and OCMock) won't build with the default settings. Unsatisfying.

The longer term solution is for Apple to act on the bug I filed. We'll see how that goes.

UPDATE: Thanks to Christian Niles for pointing out the SYMROOT environment variable in a pull request. Setting this for command line builds forces Xcode to use the specified location for all build products, and updates the BUILD_DIR compiler variable.

Steps for changing the build location in Xcode 4:

  • Open Xcode preferences (Command-,)
  • Select the "Locations" tab
  • Change the "Build Location" drop down from "Place build products in derived data location" to "Place build products in locations specified by targets."