|
|
|
An Interview with TestNG's Cedric Beust
|
|
Page 2 of 3
JUnit 4
News that an updated version of JUnit was under development received a lot of attention- what are your thoughts on JUnit 4?
JUnit 4 is a fairly conservative evolution of JUnit 3 that incorporates
more modern techniques (annotations) and new ideas (groups). I know by
experience that it's very hard to convince JUnit 3 users to change
their thousands of existing tests, so backward compatibility and
migration paths will be key to the success of this new version
(something we also took very seriously with TestNG, hence providing a
JUnit runner and various tools to convert your JUnit tests automatically).
Once JUnit 4 is released- what is TestNG's value proposition?
JUnit remains a unit testing framework (emphasis on "unit") and Kent
and Erich seem to be adamant on sticking to this core principle, which
is a good thing.
First of all, let me share with you my definition of a unit test: a
unit test is one or several methods that exercise the functionality of
a class in isolation of all others.
Unit testing is only a small part of the art of testing in general, so
I think that from a philosophy standpoint, the two frameworks have
marked differences that should make it easy for users to decide which
one to pick.
There are also a few features that are clearly out of scope for
JUnit 4, based on Kent's messages on the JUnit mailing-list these past
weeks, among which are JDK 1.4 support, dependent methods
and the separation between static (what your tests do) and runtime
(which tests are run) models. I feel very strongly about all these
features and I know for a fact that JDK 1.4 support provided a huge
increase in acceptance for TestNG.
|
|
