I have discovered that these seem to be the two main ways of testing for exceptions:
Assert.Throws<Exception>(()=>MethodThatThrows());
[ExpectedException(typeof(Exception))]
Which of these would be best? Does one offer advantages over the other? Or is it simply a matter of personal preference?
The first allows you to test for more than one exception, with multiple calls:
Assert.Throws(()=>MethodThatThrows());
Assert.Throws(()=>Method2ThatThrows());
The second only allows you to test for one exception per test function.
The main difference is:
ExpectedException()
attribute makes test passed if exception occurs in any place in the test method.
The usage of Assert.Throws()
allows to specify exact
place of the code where exception is expected.
NUnit 3.0 drops official support for ExpectedException
altogether.
So, I definitely prefer to use Assert.Throws()
method rather than ExpectedException()
attribute.
I prefer assert.throws since it allows me to verify and assert other conditions after the exception is thrown.
[Test]
[Category("Slow")]
public void IsValidLogFileName_nullFileName_ThrowsExcpetion()
{
// the exception we expect thrown from the IsValidFileName method
var ex = Assert.Throws<ArgumentNullException>(() => a.IsValidLogFileName(""));
// now we can test the exception itself
Assert.That(ex.Message == "Blah");
}
You may also strong type the error you're expecting (like the old attrib version).
Assert.Throws<System.InvalidOperationException>(() => breakingAction())
If you are using older version(<=2.0) of NUnit
then you need to use ExpectedException
.
If you are using 2.5 or later version then you can use Assert.Throw()
https://github.com/nunit/docs/wiki/Breaking-Changes
How to use:
https://www.nunit.org/index.php?p=exceptionAsserts&r=2.5