I didn't report a JIRA as I didn't know what the expected behavior
was.
Kristian Waagan wrote:
Mike Matrigali skrev:
Is the printing of nested exceptions a problem with my junit test
runner, our junit test, or something else.
Hi Mike,
Haven't checked this out, but I believe Derby is still using some
"custom"/pre-1.4 error chaining. I have observed that sometimes the real
error is not printed in the stack trace, even when I iterate through the
whole chain available using the standard mechanism (outside of JUnit).
However, the real error (usually?) shows up in derby.log.
If I am using junit.swingui.TestRunner or junit.textui.TestRunner
and if there is no problem with the error chaining should I expect these
tools to print the chained exception?
The test I have been working on lately (EncryptionKeyTest) seems to
suffer from this issue, so I'll have a look at it over the weekend to
confirm if what I just described is actually the case or not.
Are there some Jiras on this issue as well?
I didn't report a JIRA as I didn't know what the expected behavior
was. The specific instance I saw was DERBY-2613, but I have checked
in a fix for the actual problem, so to reproduce using that one you
will have to get your client to before change 535466.
In that case the error was a database startup error:
XJ040.C=Failed to start database ''{0}'', see the next exception for
details.
The code that packages up that exception is in the following routine -
is that "custom" chaining?:
opensource/java/engine/org/apache/derby/impl/jdbc/EmbedConnection.java!bootDatabase()
