In a unit test, is it generally a good idea to test returned values according to the string their toString() returns?
For example, doing the following to make sure the expected list is returned:
assertEquals(someExpression.toString() ,"[a, b, c]");
It seems to me that the considerations are as follows:
Pros: Saving time (constructing the actual expected value requires longer code).
Cons: The test depends on toString(), which is not formally defined in the doc, and thus can change in any future version.
Aucun commentaire:
Enregistrer un commentaire