Although you can do your testing from the main function of a program, it is generally easier to use a testing framework. A framework is a code library that generally imposes a structured way of performing some task. In this case, it is a structured way of writing tests.
One popular testing framework for C++ is doctest. Doctest is not a standard library, you have to download it to your computer as a single file called doctest.h, place it with your code, and include it in your file using #include "doctest.h". Doctest is already on the server that runs your code in this book. So anytime we use the include statement, it will be available.
// Bring in unit testing code and tell it to build a main function
#define DOCTEST_CONFIG_IMPLEMENT_WITH_MAIN
#include "doctest.h"
// ...function goes here...
// The tests
TEST_CASE("heronsFormula") {
CHECK(heronsFormula(3, 4, 5) == doctest::Approx(6));
CHECK(heronsFormula(2.5, 4.1, 6.25) == doctest::Approx(3.20322));
CHECK(heronsFormula(0, 3, 3) == doctest::Approx(0));
}
Don’t worry about memorizing all the details. For now you will be using tests that are already written. Later we will worry about learning the syntax well enough to write tests. Here is what we are looking at:
Line 2 says "doctest, create a main function that automatically runs the tests". That #define must come before the include.
Line 8 is our TEST_CASE for the heronsFormula function. The name in quotes is just to help us read the output. Usually each function we want to test will get its own TEST_CASE and have a name that matches the function it tests.
Lines 9-12 are our CHECKs that make up the test. Each CHECK compares two values with == to see if they are the same. (We’ll learn more about == in Section 7.2). For example, the line:
compares the result of heronsFormula(3, 4, 5) to the value doctest::Approx(6). The doctest::Approx part says that the value does not have to be precisely 6, it just needs to be “close enough to” 6. (Recall that double values are only approximations, so it is important to allow for some leeway when testing them. See Section 4.2).
Below is that same testing code, with the same buggy version of the heronsFormula function. Before you fix it, run it to see the output. You should see a section that looks like:
===============================================================================
Test messages:
test.cpp:27:
TEST CASE: heronsFormula
test.cpp:28: ERROR: CHECK( heronsFormula(3, 4, 5) == doctest::Approx(6) ) is NOT correct!
values: CHECK( 2.44949 == Approx( 6 ) )
Lines 3 and 4 show that we are looking at the TEST_CASE name heronsFormula on line 27 of the program. If we had more function, each would have its own TEST_CASE, so this would help identify which test was failing. Then on line 6 we are shown a CHECK that is failing. Immediately after that on line 7, we are shown the actual values that were used. Here, I can see that heronsFormula(3, 4, 5) evaluated to 2.44949 and that value is not Approx( 6 ).
