I already worked around TUSCANY-1558 by editing the generated
WSDL and using my updated version. TUSCANY-1558 only applies
when using exceptions that are in a different package from the
business interface. There's another report of the same problem
in TUSCANY-1664.
Fixing the WSDL allows the service to start up. Now try actually
throwing a business exception and the results aren't pretty at all :-(
I independently discovered and hacked around the issue described by
Scott. This got me on to the next problem which was that the
exception was turned into the dreaded AxisFault "unknown" rather
than being marshalled and unmarshalled according to the correct Axis2
pattern for business exceptions (helpfully documented in the Axis2
samples).
I'll write a JIRA for the runtime faliure very soon and attach a
simple test case.
Simon
Luciano Resende wrote:
Are you talking about the same issue described in TUSCANY-1558 [1] ?
If so, looks like the issue is fixed in Axis 2.1.3.
[1] https://issues.apache.org/jira/browse/TUSCANY-1558
On 9/11/07, Scott Kurz <[EMAIL PROTECTED]> wrote:
There are some particular issues relating to dealing with, alternately the
exception DataType and the fault DataType, and getting these through the
Tuscany databinding framework.
Some of these I've worked around already and am working to post a more
detailed update myself (for example I modified DataTransformationInterceptor
to match element names of XMLType fault logicals rather than using an
equals() comparison).
Scott
On 9/11/07, Simon Nash <[EMAIL PROTECTED]> wrote:
Yes, I think this should work. I'm trying to get a simple case
through the Web Service binding and the databinding transformers at the
moment and I'm running into various issues. I'll post a more detailed
update later.
Simon
Scott Kurz wrote:
Say I wanted to have a remotable Java interface with a method like:
int myMethod() throws java.sql.SQLException;
Should I be able to throw this exception, say, across the web service
binding?
I don't see why not. JAX-WS Sec 3.7 describes how to build a fault
bean
out of an exception like SQLException which doesn't conform to the
pattern
in JAX-WS Sec 2.5. A tool like wsgen should be able to generate a
WSDL
with a corresponding fault element, typed by the default mapping
obtained by
viewing SQLException as a JavaBean (Sec 3.7) per JAXB.
(For a complicated data type the default JAXB mapping isn't enough
without
annotations, but for simpler examples it might be OK.)
Do we view this as a valid remote interface? I'm not asking whether it
works today.. just whether it seems like it should work.
Scott
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]
