Thanks for the correction! It makes sense. (It also partially
addresses the other question I recently sent out.) Until the Axis bug
gets fixed I guess I should modify the WSDL file to match Axis' output? Also, could you tell me why, for example, the webservice http://ws2.serviceobjects.net/nw/namewizard.asmx?WSDL embeds a <getNameInfoResult> inside of a <getNameInfoResponse>? It seems like only one of the two tags is necessary. Is it unnecessary overhead, or is it something to make certain client tools happy? Bill Anne Thomas Manes wrote: Clarification:Bug: In document/literal mode, Axis generate a <methodNameReturn> tag when writing out the return value; it shouldn't. That tag only makes sense in wrapped/literal mode. (This is what Ann was saying.)Axis should generate a <methodNameReturn> tag only when using RPC style. It should not generate the tag when using either wrapped or document style. In both cases, the child element of the SOAP Body should be the element specified in the response message part. For example, if the response message is defined as:<wsdl:message name="methodResponse"> <wsdl:part name="parameters" element="tns:foo"/> </wsdl:message> then the response message for either wrapped or document style should be: <s:Body> <tns:foo.../> </s:Body> - Anne On Wed, 12 Jan 2005 09:38:49 +0900, Bill Keese <[EMAIL PROTECTED]> wrote:OK, I think I understand what's going on now. There is both a bug and a "missing feature". Bug: In document/literal mode, Axis generate a <methodNameReturn> tag when writing out the return value; it shouldn't. That tag only makes sense in wrapped/literal mode. (This is what Ann was saying.) Missing feature: Axis obeys the specifications in the WSDL file, but you have to manually call wsdl2java to generate the deploy.wsdd. (This is what Michael was saying.) However, I would like a feature where Axis automatically parses and obeys the WSDL file, without manually calling wsdl2java. Bill -----Message d'origine-----De : Michael Schuerig[mailto:[EMAIL PROTECTED]]Envoyé : mardi 11 janvier 2005 11:51 À :[EMAIL PROTECTED]Objet : Re: Document/Literal : bad part name in axisserver responseOn Tuesday 11 January 2005 07:26, Ephemeris Lappis wrote:I don't use wsdl2java.I had it the wrong way around and confused thejava2wsdl from youroriginal message with wsdl2java.In Jonas, using the new (not so new) ejb 2.1 service endpointfeatures, ijust provide the wsdl (that is generated from the javaendpoint interface),and the ejb and webservices descriptors (for theejb deployment). Atassembly time, Jonas produces the java stubs, thewhole web application as awar with a limited wsdd descriptor. Atdeployment time, Jonas merges thewar's wsdd file with the fullplatform's one to expose the web serviceendpoint. This is the wsddfile : Presumably the following is happening: Atassembly time Jonas useswsdl2java to generate, among other things, adeploy.wsdd. At deploymenttime, Jonas uses the Axis AdminClient to mergethis deploy.wsdd intoAxis' server-config.wsdd. Do you know if this iscorrect?<?xml version="1.0" encoding="UTF-8"?><deploymentxmlns="http://xml.apache.org/axis/wsdd/"[snip]</deployment>There are no operation elements in that WSDD. Without them,Axis won'tuse the names and types you intended in the input and outputmessages.From my point of view, Axis was expected to use the given wsdl filetohandle client requests according to the descriptor it generateditself...Well, Axis (at least as of 1.2RC) does that. Not directly, though. Youneedto process your WSDL with wsdl2java and among other things, youget adeploy.wsdd with names, types and namespaces specified foroperations. Also,the deployScope can be specified.I may be mistaken and I can't claim thatAxis is doingeverything right, but the problem in your case appears to beone of integration where theJonas installation doesn't make full use ofwhat Axis offers. I don'tknow Jonas, thus I can't say whether this is agenuine limitation ofthe respective Jonas version or a shortcoming of thedeployment.Michael -- Michael Schuerig Contests between male toads overfemales aremailto:[EMAIL PROTECTED] often settled by the depth of thecroak.http://www.schuerig.de/michael/ --John Maynard Smith |
- Re: Document/Literal : bad part name in axis server resp... Bill Keese
- Re: Document/Literal : bad part name in axis server... Anne Thomas Manes
- RE: Document/Literal : bad part name in axis se... Ephemeris Lappis
- Re: Document/Literal : bad part name in axis se... Bill Keese
- Re: Document/Literal : bad part name in axi... Michael Schuerig
- RE: Document/Literal : bad part name in... Ephemeris Lappis
- Re: Document/Literal : bad part name in... Michael Schuerig
- RE: Document/Literal : bad part name in... Ephemeris Lappis
- Re: Document/Literal : bad part name in... Bill Keese
- Re: Document/Literal : bad part name in... Anne Thomas Manes
- RE: Document/Literal : bad part name in... Bill Keese
- RE: Document/Literal : bad part name in... Ephemeris Lappis
- Re: Document/Literal : bad part name in... Anne Thomas Manes
- RE: Document/Literal : bad part name in axis server... Ephemeris Lappis