@DataPoints-related fixes

[MatrixFrog]

This fixes both issue 125 and issue 137. There's probably a better exception to throw other than just RuntimeException but I don't know what that would be.

[jschneider]

I am quite sure, this patch will not be accepted (unfortunately), since it changes the current behavior. While I prefer your version, there might exist some use cases, where silent failure is preferred (I can't think of one, but I have read that some people prefer it that way...).

Your change in line 79 should probably be modified to stay backwards compatible with the current behavior.
Maybe the RuntimeException should one be thrown, if a (newly created) parameter is set at DataPoints.

What about the "DataPoint" annotation? I think the same changes should be applied for that anntotation, too.

[MatrixFrog]

I can't think of one either, so I would love to hear one. :)

You're right about @DataPoint. If it looks like a change along these lines is going to be accepted, I'll definitely make the equivalent change on @DataPoint as well...

[jschneider]

I have added some checks checking for void methods:

https://github.com/jschneider/junit/commit/2760cf967da1bfb400f07fec818b064411022fde

[dsaff]

Sorry for the delay on this. Indeed, changes that cause currently passing tests to fail (even if they should fail) are handled with kid gloves. However, I think this one does make sense, in some form. I'd like a quick way for people to opt out of the change, however. Perhaps:

@DataPoints(ignored={SomeExceptionWeShouldSilentlyIgnore.class, AnotherOne.class})

?

[dsaff]

@MatrixFrog, are you still available to work on this?

[dsaff]

I'm guessing @MatrixFrog has moved on. Is anyone else on the thread available to fix this up?

[dsaff]

Closing for now, can re-open later