I also didn't think we were going to backport all the junit stuff
to 10.2. As Myrna points out some of those changes involve decisions
not valid for 10.2, a released product.
It is unfortunate that some bug tests won't backport cleanly, but
that could be the case with any backport, and even if we fix 10.2
still is a problem for 10.1. Maybe it might make
sense to have new bug fixes in the trunk go into "new" junit tests
rather than existing files - would this make the backport easier?
Myrna van Lunteren wrote:
On 11/20/06, Daniel John Debrunner <[EMAIL PROTECTED]> wrote:
Kristian Waagan wrote:
> Rick Hillegas wrote:
>> Hi Kristian,
>>
>> I agree that we should stabilize the 10.2 tests soon since we're
>> planning a 10.2.2 release in early December. If we don't merge the
>> JUnit infrastructure into 10.2, then porting a bug fix to the 10.2
>> branch may involve re-writing the test case which verifies the fix.
>
> Rick,
>
> I see your point, but had the impression this was not going to be done.
> Maybe I have missed a lot of commits, but I thought most of the JUnit
> work on trunk has not been merged to 10.2. Am I wrong?
I thought the tests in 10.2 were stable. 10.2 uses derbyall only and not
the Junit top-level suites. I didn't even know the top-level JUnit
suites were in 10.2. They were merged up from a point in time when they
were known not to be working.
http://svn.apache.org/viewvc/db/derby/code/branches/10.2/java/testing/org/apache/derbyTesting/functionTests/suites/All.java?view=log&pathrev=476832
Dan.
I feel apprehesive about backporting all changes to junit framework in
trunk back to 10.2.
1. 10.2 is stable...which should mean changes get applied more selectively.
2. The changes in the trunk have been made with happy disregard of jcc
and jdk spec 1.3.1 based jvms.
I can understand the wish to just put a specific test back to 10.2
without any work, but I am worried that backporting all the current
changes to 10.2 is going to be more painful than meets the eye.
Of course, I have no actual proof there is going to be trouble, nor
that there isn't...
Myrna
