> > Well, that was one of the things that first puzzled me - Mike reported > this with DerbyNet, and none of the DerbyNet tests actually run with > security manager. > However, the network server seems to *always* start with security > manager, even when the actual test has security manager switched off. > I'll have a look at that also.
That's correct. The network server when started separately always runs under the security manager. The noSecurityManager flag only applies to the JVM running the test, the "client" jvm, as in "JDBC client", embedded or network. That was the original setup, when only the network server started separately was run under the security manager. Since all the tests passed in that mode it seemed to me like that model should continue. > (pf). > In the mean time, would it be acceptable to add the permissions so the > nightlies pass on? Yes, as long as it is commented with DERBY-616 bug number so that when the bug is fixed the policy file will get cleaned up. You will also need to comment DERBY-616 that the permissions you add will need be removed (I think) in order to see the issue. Currently just by removing the noSecurityManager=true property in the test's _app.properties file (eg. for access.sql) the tests will fail. Adding these properties will make the tests pass with a security manager and we just don't want someone to think that the bug was already fixed. Dan.
