bandit

Human friendly unit testing for C++11

Reference

Entry point

This is the entry point for a bandit application:

#include <bandit/bandit.h>

int main(int argc, char* argv[])
{
  return bandit::run(argc, argv);
}

Assertions

Bandit uses the snowhouse assertion library. Go there for references on how to write assertions.

go_bandit()

The go_bandit() construct registers a function that contains tests. There can be one go_bandit() construct per .cpp file.

The boilerplate for each .cpp file containing tests should look like this:

#include <bandit/bandit.h>
using namespace bandit;

go_bandit([](){
  describe("a thing", [&](){
    it("should do something", [&](){
      AssertThat(something, IsTrue());
    });
  });
});

describe()

This function is setting up the context for a test. That is, declaring variables for holding the objects that we need for our tests. You can have several levels of nested describe() where each level specifies specific cases of its parent level.

This function is similar to the concept of a fixture in xUnit based frameworks.

describe("a calculator", [&](){
  calculator calc;

  describe("in scientific mode", [&](){
    // Additional setup and tests.

    describe("in hex mode", [&](){
      // Additional setup and tests.
    });
  });

  describe("in basic mode", [&](){
    // Additional setup and tests.
  });
});

xdescribe(), describe_skip()

This tells bandit to skip running all it() functions in this describe() and all its nested describes. Bandit will report the number of skipped it() functions after the test run has completed.

it()

This function describes something that should hold true for the component we’re describing.

it() is similar to a test method in xUnit based frameworks.

xit(), it_skip()

This causes the function to be skipped during a test run. The number of skipped functions will be reported by the bandit executable after a test run.

before_each()

This function is used to set up all prerequisites needed before we can start testing. before_each() is called before each it() during a test run. This way you know that each it() function has access to a fresh new context that hasn’t been contaminated by other tests.

You can have several before_each() methods in a describe() function. They will be called in the order they are declared before each it() function.

describe("a calculator", [&](){
  calculator_ptr calc;

  before_each([&](){
    // Make sure each 'it()' gets a fresh new
    // calculator to work with.
    calc = calculator_ptr(new calculator());
  });

  it("can add", [&](){
    AssertThat(calculator->add(3,2), Equals(5));
  });

  // More tests...
});

When you have nested describe() functions, bandit will call all before_each() for the outermost context first, then all for its child contexts until each reaches the context containing the current it().

NOTE: All before_each() must be declared before any it() or nested describe(). Otherwise bandit cannot guarantee that everything is set up correctly and will complain with an error message.

after_each()

Sometimes (more rarely than you might think) you need to clean up after each test. after_each() gets called after each it() function.

NOTE: All after_each() must be declared before any it() or nested describe(). Otherwise bandit cannot guarantee that everything is set up correctly and will complain with an error message.