I'm assigning this to me because I keep forgetting the number. If someone else feels interest, be welcome to take it away.
In the mean time, I have tried to find out what is wrong with this and I've been unsuccessful so far.
DERBY-855 Dan stated that he had found a problem with the security manager and had reported this to IBM. I checked on the status of that problem; it is fixed in the follow-up release of wctme5.7_foundation (called weme6.1, which is based on jdkj2ME specification 1.1). I tested with that product and indeed the problem Dan saw has been fixed...
However, I still see the access permission problems with junit with that version.
I will continue to investigate. I think this is an important detail - we have said we're supporting JSR169, yet it appears the only check we have are our test runs with "j9" foundation. If someone else is running JSR169 tests, I hope they'll announce it.
The problem needs to be better identified, and possibly a fix needs to be arranged.
As a workaround there are 2 approaches:
- ignore the failing tests. This basically means skipping every junit test. We can do this for each test as it gets morphed into a junit test, which means a ridiculous maintenance task, or in the test harness. But this would hollow out the testing for JSR169 support.
- make the junit tests not run with security manager if the jvm is j9. This seems fairly ugly
(third option I thought maybe a separate policy file could be made, but judging from the error, it isn't having the permissions to read the policy file, so I'm not sure that will work.)