On Sat, Dec 5, 2009 at 7:32 AM, Tomek Kaczanowski <
[email protected]> wrote:

> Please forgive me my weak understanding of Gradle, but would it be
> good if both TestNG support and JUnit support was somewhat external,
> and not in core ? Isn't it really a good example of plugins (TestNG
> plugin and JUnit plugin) ?
>
>
I think I'm the ignorant one here, which is partly why I asked the
question.  I'm not sure that having JUnit in Gradle's libs directory
restricts the version of JUnit used by Gradle.  I have no idea, one way or
the other.  I'm just concerned it might.  BTW, the JUnit and TestNG support
are both in the "gradle-plugins" module.  That's not quite the same as
having them in separately distributed plugins, but it's a move in the right
direction.  Actually, until Gradle gets a solid plugin ecosystem, including
automatically fetching needed/updated plugins from a standard repository I
think keeping the "core" plugins bundled is a good solution.



> If JUnit or TestNG are in lib directory, than how can I upgrade them ?
> By changing that file to the latest version of testing library, I
> suppose. This is kind of "hack". I would prefer, to use the same
> Gradle version, but to download a latest testing plugin that would
> include the letest version of JUnit/TestNG.
>
>
There are a couple of possibilities.  I agree that changing the version in
the lib directory would be a bad option.  It might even cause Gradle to
break (if the library update changes API signatures that Gradle is using,
for example).  I suspect Hans or Adam will come in here and set us straight
soon enough.


> My apologies if what I'm saying does not strictly fit into Gradle
> architecture, but I'm trying to provide a general idea, not a
> ready-to-use solution.
>
>
As I said above, I agree that a separately distributed plugin model makes
sense, but not until better support for it has come online.  If Hans/Adam
were to focus on it, this probably would not take long.  However, they are
focused on other things that I think are good as well, so I'm not
complaining.  :)


-- 
John Murph
Automated Logic Research Team

Reply via email to