Thank you for your attention ant! You encouraged me reading osoa's specification again, then I found your point. OK, now let me speak about my question to some further: I found in WebService_Binding 2.3, wsdl should be generated for an binding.ws, and there follows a generation rule in the spec, OK now we know Tuscany' s binding.ws works noly if we provide a wsdl(not generate it). I don't think it a big problem, but since I found a <binding.ws uri="[url]"> just override the <soap:address location="[url]">, and it seems without this binding.ws's uri, the url of the web service would be some where random(not according to the soap:address's).
ant elder <[EMAIL PROTECTED]> wrote: On 7/18/07, shaoguang geng wrote: > > When I work on the svn code, I found that the service address of a < > binding.ws> depends on it's uri attribute, not the inside > the wsdl file. If the is some thing different from the < > binding.ws>'s uri, or it does not exists absolutely, the client will get a > confusion form it http://[host]:[port]/[servicename]?wsdl, > > If I don't give a , I will see a warn, but without the < > binding.ws>' uri, Tuscany runs without any message. The WS service address is calculated based on section 2.1.1 of the WS binding spec and section 1.7.2.1 of the assembly spec (see [1]), and there's a bit about it in the Tuscany doc at [2]. From that, Tuscany should be using the from the WSDL if you reference the WSDL port from the < binding.ws wsdlElement= ...>, the uri attribute is only used if you don't reference the wsdl port or if it is a relative url. Not sure if that answers you question though? ...ant [1] http://osoa.org/display/Main/Service+Component+Architecture+Specifications [2] http://incubator.apache.org/tuscany/sca-java-bindingws.html --------------------------------- Fussy? Opinionated? Impossible to please? Perfect. Join Yahoo!'s user panel and lay it on us.
