"John H. Embretsen (JIRA)" <[EMAIL PROTECTED]> writes:

>     [ 
> https://issues.apache.org/jira/browse/DERBY-2274?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel#action_12503871
>  ] 
>
> John H. Embretsen commented on DERBY-2274:
> ------------------------------------------
>
> I'm afraid I'm not able to say whether or not you've done the right thing 
> wrt. the JUnit framework stuff (it looks OK to me), but I do have one comment 
> about the test itself (SSLTest.java).
>
> In the public void sslTest() method, there is the following code
>
> +        try {
> +            // Should fail
> +            Connection c2 = DriverManager.getConnection(plain);
> +            c2.close();
> +        } catch (SQLException e) {
> +            if (e.getCause() != null && 
> +                e.getCause() instanceof 
> org.apache.derby.client.am.DisconnectException) {
> +                // Assumes protocol fail due to SSL on the other end
> +            } else {
> +                throw e;
> +            }
> +        }

Two more nits about this code fragment:

1) (e.getCause() != null) is redundant, since (e.getCause() instanceof
DisconnectException) is false if e.getCause() returns null.

2) Would it be better to replace the contents of the catch block with a
call to assertSQLState()? Or is the SQL state non-deterministic in this
test?

-- 
Knut Anders

Reply via email to