[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]
Subject: Re: [ws-caf] [Bug 129] New: SOAP mustUnderstand should not be determined by ws-context
I agree. I was only pointing out that the current schema is broken anyway :-) Mark. ----- Original Message ----- From: "Furniss, Peter" <Peter.Furniss@choreology.com> To: "Mark Little" <mark.little@arjuna.com>; "Doug Bunting" <Doug.Bunting@Sun.COM>; <ws-caf@lists.oasis-open.org> Sent: Friday, May 28, 2004 2:08 PM Subject: RE: [ws-caf] [Bug 129] New: SOAP mustUnderstand should not be determined by ws-context > ok. > > But: > > If mustUnderstand is on context, then, since context is usually a SOAP > header element, it collides with the SOAP:mustUnderstand attribute. > Having a ws-context defined mustUnderstand on the xsd:any - or perhaps > on any children of a complex type there, would seem possibly, especially > for the sort of use Doug suggested. > > We don't seem to have a definition of what "mustPropagate" means. I > believe it only applies if (soap) mustUnderstand=0. If the receiver > understands the context, they will no what they want to do with it, and > whether it should be propagated as received or after modification. It's > only if the receiver doesn't understand the particular context, but the > original sender wants it to be passed on that mustPropagate comes into > play. (this would provide a utility in ws-context - headers that can be > propagated mindlessly - though one could argue that it should be > promoted to soap level) > > And as Doug points out, we need to say, or have some way of saying where > mustPropagate is to apply to. Synchronous responses only (rather useless > - the carrier does that) ? WSDL-defined out's to in's, regardless of > carrier ? Anything off the same thread (what if there are unknown > invocations on ws that don't understand ws-context) ? > > Peter > > > > -----Original Message----- > > From: Mark Little [mailto:mark.little@arjuna.com] > > Sent: 28 May 2004 13:38 > > To: Furniss, Peter; Doug Bunting; ws-caf@lists.oasis-open.org > > Subject: Re: [ws-caf] [Bug 129] New: SOAP mustUnderstand > > should not be determined by ws-context > > > > > > Peter, the mustUnderstand/mustPropagate are in the wrong > > place for the schema in the currently available spec. I'm > > sure I've mentioned this before on the list (though could be > > wrong), but it may not be an issue because it's something > > that I think was editorial. They should be on the context itself. > > > > Mark. > > > > ----- Original Message ----- > > From: "Furniss, Peter" <Peter.Furniss@choreology.com> > > To: "Doug Bunting" <Doug.Bunting@Sun.COM>; > > <ws-caf@lists.oasis-open.org> > > Sent: Friday, May 28, 2004 12:52 PM > > Subject: RE: [ws-caf] [Bug 129] New: SOAP mustUnderstand > > should not be determined by ws-context > > > > > > > I strongly agree with your point about general extensibility - just > > > about any "any" needs to have an indication of whether it can be > > > safely ignored by the ignorant, or whether would corrupt > > the sender's > > > purposes. > > > > > > But you end up saying that we can have > > soap:mustUnderstand="0" if and > > > only if we require neither understanding or propagation > > further in. I > > > agree. But the implication of that is that we don't, at > > specification > > > level, need soap:mustUnderstand="1" everywhere. Just the > > sender needs > > > to think carefully. > > > > > > But looking at our schema, what are the indicators doing on > > the list > > > of participating services - they surely belong on the any ? > > (and, for > > > that matter, how does the participating services list get modified) > > > > > > Peter > > > > > > > -----Original Message----- > > > > From: Doug Bunting [mailto:Doug.Bunting@Sun.COM] > > > > Sent: 27 May 2004 20:14 > > > > To: ws-caf@lists.oasis-open.org > > > > Subject: Re: [ws-caf] [Bug 129] New: SOAP mustUnderstand > > should not > > > > be determined by ws-context > > > > > > > > > > > > We may need soap:mustUnderstand="1" everywhere. The > > receiver must > > > > understand WS-Context enough to know a set of "related > > messages" (as > > > > mentioned in my previous email[1]) must include the same > > structure. > > > > More generally, the receiver must implement the correct > > > > propagation rules. This > > > > receiver does not need to understand the entire contents of > > > > the context. > > > > > > > > We are hitting on a general point about extensibility in SOAP > > > > messages (or, even more generally, XML instances). > > Wherever "any" > > > > content is allowed, > > > > understanding may or may not be required of added content -- > > > > depending upon > > > > the semantics of the container and whatever additional > > > > indicators that > > > > container provides (and how they are set in a particular > > > > message). The > > > > soap:Header element has an indicator (soap:mustUnderstand) > > > > usable in every > > > > child element for this purpose. For ctx:context, the SOAP > > > > indicator must > > > > be true[2] so that our indicators (ctx:mustUnderstand and > > > > ctx:mustPropagate) are used correctly. The only case in which > > > > soap:mustUnderstand may be false occurs when ctx:context > > contains no > > > > ctx:activity-list or both indicators are false on the > > > > ctx:context/ctx:activity-list element. (No, I am not sure > > > > the new schema > > > > will contain the same indicators or place them in the same way.) > > > > > > > > thanx, > > > > doug > > > > > > > > [1] http://lists.oasis-open.org/ws-caf/200405/msg00142.html > > > > > > > > [2] Please note my use of values from the value space[3] of XML > > > > Schema part 2. This may confuse an issue originally[4] about the > > > > lexical space[5]. > > > > That original issue ("1" versus "true") should also appear in > > > > the issues list. > > > > > > > > [3] http://www.w3.org/TR/xmlschema-2/#value-space and > > > > http://www.w3.org/TR/xmlschema-2/#boolean > > > > > > > > [4] http://lists.oasis-open.org/ws-caf/200405/msg00086.html > > > > > > > > [5] http://www.w3.org/TR/xmlschema-2/#lexical-space and > > > > http://www.w3.org/TR/xmlschema-2/#boolean-lexical-representation > > > > > > > > On 26-May-04 04:30, bugzilla-daemon@arjuna.com wrote: > > > > > > > > > http://services.arjuna.com/wscaf-issues/show_bug.cgi?id=129 > > > > > > > > > > Summary: SOAP mustUnderstand should not be > > > > determined by ws- > > > > > context > > > > > Product: WS-Context > > > > > Version: 1.0 > > > > > Platform: PC > > > > > OS/Version: Windows 2000 > > > > > Status: NEW > > > > > Severity: normal > > > > > Priority: P2 > > > > > Component: Implementation and interoperability > > > > > AssignedTo: ws-caf@lists.oasis-open.org > > > > > ReportedBy: peter.furniss@choreology.com > > > > > QAContact: mark.little@arjuna.com > > > > > > > > > > > > > > > Draft 0.2, section 2.1 says that SOAP:mustUnderstand must > > > > be true for > > > > > a context > > > > > sent with an application message. WS-Context should not > > > > specify this - it > > > > > should be left to a higher level. There will certainly be > > > > some cases where a > > > > > context is propagated in hope, but not certainty, that it > > > > will be recognised. A > > > > > referencing specification could mandate a value, or leave > > > > it to the particular > > > > > use. > > > > > > > > > > It should be stated that SOAP:mustUnderstand=1 on a > > > > ws-context means > > > > > the > > > > > receiver is required to understand the particular > > > > context-type, not just that > > > > > it is a ws-context. > > > > > > > > > > > > > > > > > > > > ------- You are receiving this mail because: ------- > > > > > You are the assignee for the bug, or are watching the assignee. > > > > > > > > > > > > > > > >
[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]