Knut Anders Hatlen (JIRA) wrote:
[ http://issues.apache.org/jira/browse/DERBY-989?page=all ]
Knut Anders Hatlen updated DERBY-989:
-------------------------------------
Attachment: derby-989-testfix.diff
derby-989-testfix.stat
derby-989-testfix.diff is an alternative solution. This patch only changes what
the test expects, not the behaviour of Derby. With this patch, the test allows
the client to be serviced once after it has called unsubscribe().
Derbyall runs cleanly with the patch, and I'm not able to reproduce the failure
in daemonService.unit.
Comments/opinions on which solution to choose would be greatly appreciated.
Thanks!
I am not familar with the daemon code implementation, but have written
user code of the interface. I have always assumed that "spurious" calls
could be made,
from your description it sounds like the test is not exactly matching
the interface. Changing the test seems reasonable. Given your insight
it may be reasonable to expand the documentation of the interface to
include the case you have tracked down.
