Looking for your thoughts on a bit of a conundrum.
It is working pretty well to migrate the client code to throw
SQLException directly rather than org.apache.derby.client.am.SqlException.
There are some diagnostic features, however, that can't migrate, because
they rely on extra information stored in SqlException that is not
available in SQLException.
The alternative is to continue using SqlException internally, passing it
up the chain to the top-level classes, and then convert it into
SQLException. Ths problem with this is in JDK 1.3 the stack trace is
completely lost. However, in JDK 1.4 and greater I can use the
initCause() method and the stack trace is retained.
My preference is to actually to continue using SqlException and then
convert at the top level before throwing the exception to the user (this
is actually a common practice). This allows us to keep our diagnostic
info and to be able to increase it if we want to.
I can use reflection to see if the initCause() method is available, and
call it if it's there. For JDK 1.3 vms, I can ensure that the stack
trace is logged prior to converting the exception.
This seems pretty reasonable to me. Any comments?
David
begin:vcard
fn:David W Van Couvering
n:Van Couvering;David W
org:Sun Microsystems, Inc.;Database Technology Group
email;internet:[EMAIL PROTECTED]
title:Senior Staff Software Engineer
tel;work:510-550-6819
tel;cell:510-684-7281
x-mozilla-html:TRUE
version:2.1
end:vcard
