DERBY-4463, it was found there was a bug with the IBM jvms in the handling of interrupt() and waitFor().
This is fixed with IBM 1.6 SR9 (and is planned for IBM 1.5 SR13).
As SR9 is now GA, I thought to adjust the skipping of this test to only with ibm 1.5.
However, I now notice that the test drops java core dumps with ibm 1.6 SR6. It does not do this with Sun's 1.6 jvm.
I also believe that it didn't do this a couple of weeks ago when I first tried this with ibm 1.6 SR9, so I think something has changed in derby, but it is possible I was not paying proper attention.