Here's a follow-up two months later to this thread.....
We seem to need a small additional fix in WSDLOperation$Wrapper.getChildElements(). In this line we're again wrongly enforcing the requirement that the wrapper element can't have a reference on to the child element, which I think should be allowed to be a reference. if (childElement.getName() == null || childElement.getRefName() != null) { return null; } Agreed? Scott On 1/3/07, Scott Kurz <[EMAIL PROTECTED]> wrote:
OK, I see you used r491901 to fix this so I won't open a JIRA. On 1/2/07, Scott Kurz <[EMAIL PROTECTED] > wrote: > > Thanks Raymond, > Should I open a JIRA? > > > > On 1/2/07, Raymond Feng <[EMAIL PROTECTED] > wrote: > > > > Hi, Frank. > > > > I think you're right after reading the paragraph again. > > > > Scott, I'll adjust the code to fix the rule. > > > > Thanks, > > Raymond > > > > ----- Original Message ----- > > From: "Frank Budinsky" <[EMAIL PROTECTED] > > > To: <tuscany-dev@ws.apache.org > > > Sent: Tuesday, January 02, 2007 9:31 AM > > Subject: Re: bug in determing WSDL-> Java mapping style? > > > > > > > Hi Raymond, > > > > > > Just my 2-cents worth, but my reading of the quoted section 2.3.1.2is > > > that "they" in "furthermore, they must not be nillable" is referring > > to > > > the "wrapper elements", not the "child elements". > > > > > > Frank. > > > > > > "Raymond Feng" <[EMAIL PROTECTED]> wrote on 01/02/2007 12:12:00 > > PM: > > > > > >> Hi, Scott. > > >> > > >> JAX-WS spec 2.0 section 2.3.1.2 actually says: > > >> > > >> "The wrapper elements only contain child elements, they must not > > contain > > > > > >> other structures such as wildcards (element or attribute), > > xsd:choice, > > >> substitution groups (element references are not permitted) or > > > attributes; > > >> furthermore, they must not be nillable." > > >> > > >> Please note the last phrase requires that the child elements must > > not be > > > > > >> nillable. > > >> > > >> Thanks, > > >> Raymond > > >> > > >> ----- Original Message ----- > > >> From: "Scott Kurz" < [EMAIL PROTECTED] > > > >> To: <tuscany-dev@ws.apache.org> > > >> Sent: Tuesday, January 02, 2007 8:34 AM > > >> Subject: bug in determing WSDL-> Java mapping style? > > >> > > >> > > >> > Before opening a JIRA I thought I'd throw this out there to > > ensure > > > it's > > >> > really a bug. > > >> > > > >> > In code such as org.apache.tuscany.idl.wsdl.WSDLOperation we try > > to > > >> > enforce > > >> > the JAX-WS criteria for using "wrapper-style" mapping mentioned > > in > > > JAX-WS > > >> > 2.3.1.2. > > >> > > > >> > The WSDLOperation code has the spec statement as a comment: > > >> > ..... > > >> > * (v) The wrapper elements only contain child elements, they > > must > > > not > > >> > contain other structures such as > > >> > * wildcards (element or attribute), xsd:choice, substitution > > > groups > > >> > (element references are not permitted) or > > >> > * attributes; furthermore, they must not be nillable. > > >> > > > >> > Our code, however, seems to go further and require that the child > > > > > elements > > >> > themselves are non-nillable in order to use wrapper-style > > mapping. > > >> > > > >> > So if I'd modified the example in Figure 2.1 of JAX-WS to look > > like: > > >> > > > >> > <xsd:element name="setLastTradePrice"> > > >> > <xsd:complexType> > > >> > <xsd:sequence> > > >> > <xsd:element name="tickerSymbol" nillable="true" > > >> > type="xsd:string" /> > > >> > <xsd:element name="lastTradePrice" nillable="true" > > >> > type="xsd:float" /> > > >> > </xsd:sequence> > > >> > </xsd:complexType> > > >> > </xsd:element> > > >> > > > >> > then the Tuscany runtime would assume I was working with a > > >> > non-wrapped-style > > >> > interface like: > > >> > > > >> > SetLastTradePriceResponse setLastTradePrice(SetLastTradePrice > > >> > setLastTradePrice); > > >> > > > >> > Doesn't this seem incorrect? > > >> > > > >> > When I used the JAX-WS wsimport tool against a doc-lit-wrapped > > style > > > WSDL, > > >> > I > > >> > had no problem with nillable="true" on the child elements. I > > still > > > got > > >> > wrapped-style Java. > > >> > > > >> > (A related but separate issue that I will open a JIRA for is the > > fact > > > that > > >> > there is no option in Tuscany's WSDL2Java tooling to generate > > > non-wrapped > > >> > Java from doc-literal-wrapped WSDL.) > > >> > > > >> > Thanks, > > >> > 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] > > > > > > > > > --------------------------------------------------------------------- > > To unsubscribe, e-mail: [EMAIL PROTECTED] > > For additional commands, e-mail: [EMAIL PROTECTED] > > > > >