Yesterday I ran into a frustrating situation while I was writing a unit test for some new code that I was working on. I had a test that was expecting a mock to have one of its methods called a specific number of times.
Only I was trying to avoid having too many static values defined all over my test case so I was calculating the number of times the method would be called.
$callCount = ceil(self::TOTAL \ self::LIMIT); $myMock->expects($this->exactly($callCount))->method('myMethod'); $myObject->doSomething();
The issue I was having was that my test kept failing but the reason it was failing didn’t seem to make any sense.
Method was expected to be called 3 times, actually called 3 times.
My google searches for what this might be turned up nothing. So I then
tried simplifying the test. That is when I noticed that if I changed
the calculated count to a integer everything worked fine. What I didn’t
take into account is that in PHP the
ceil() method returns a float
and PHPUnit apparently does an exact comparison (
===) of the number
of times a method is expected to be called an the number of times it is
actually called. But when the failure is output the float of value 3
prints exactly the same as the integer value of 3, thus my confusion.
In the end what I did was to cast the result of the
ceil() to an
integer since I know that the number will not be large enough to be an
$callCount = (int) ceil(self::TOTAL \ self::LIMIT); $myMock->expects($this->exactly($callCount))->method('myMethod'); $myObject->doSomething();