Subject | CS incompatability |
---|---|
Author | david_beutel |
Post date | 2002-11-15T08:48:17Z |
Hi All,
I just started using Firebird yesterday. When I run the CVS head
client-java unit tests with FirebirdCS-1.0,
TestFBDriver.testWarnings() throws NullPointerException from line 150
(the printStackTrace()):
SQLWarning warning = dialect1Connection.getWarnings();
warning.printStackTrace();
assertTrue("Connection should have at least one warning.",
warning != null);
I don't see this error with FirebirdSS-1.0. The rest of the tests
pass on both CS and SS. But, the TJDO tests have a lot more problems
with CS. I'm not sure whether the root cause is an incompatability
between CS and SS, or just with my CS configuration. I used the RPMs
on RedHat7.2 with jdk1.3.1_02, jdk1.3.1_06, and j2sdk1.4.1, as well as
ant-1.5.1, and last night's JBoss snapshot.
By the way, I'd like to suggest moving the not-null assertion line
above the printStackTrace(), so the test will fail gracefully, and
using assertNotNull(). Also, why print the stack trace? Is somebody
watching the output? Should there be another assert about the
contents of the stack trace?
Is there anyone using CS without this problem? Is anyone running a
ContinuousIntegration tool (e.g., CruiseControl) on the client-java
CVS updates?
Thanks,
11011011
I just started using Firebird yesterday. When I run the CVS head
client-java unit tests with FirebirdCS-1.0,
TestFBDriver.testWarnings() throws NullPointerException from line 150
(the printStackTrace()):
SQLWarning warning = dialect1Connection.getWarnings();
warning.printStackTrace();
assertTrue("Connection should have at least one warning.",
warning != null);
I don't see this error with FirebirdSS-1.0. The rest of the tests
pass on both CS and SS. But, the TJDO tests have a lot more problems
with CS. I'm not sure whether the root cause is an incompatability
between CS and SS, or just with my CS configuration. I used the RPMs
on RedHat7.2 with jdk1.3.1_02, jdk1.3.1_06, and j2sdk1.4.1, as well as
ant-1.5.1, and last night's JBoss snapshot.
By the way, I'd like to suggest moving the not-null assertion line
above the printStackTrace(), so the test will fail gracefully, and
using assertNotNull(). Also, why print the stack trace? Is somebody
watching the output? Should there be another assert about the
contents of the stack trace?
Is there anyone using CS without this problem? Is anyone running a
ContinuousIntegration tool (e.g., CruiseControl) on the client-java
CVS updates?
Thanks,
11011011