- From: Gilbert Pilz <gilbert.pilz@oracle.com>
- Date: Tue, 15 Dec 2009 17:15:50 -0800
- To: Ram Jeyaraman <Ram.Jeyaraman@microsoft.com>
- CC: "johannes.echterhoff@igsi.eu" <johannes.echterhoff@igsi.eu>, "public-ws-resource-access@w3.org" <public-ws-resource-access@w3.org>
- Message-ID: <4B2834C6.2060403@oracle.com>
I think Johannes is really looking for some example event data and some likely XPath expressions that do and do not result in the transmission of a notification of that event. - gp On 12/7/2009 2:35 PM, Ram Jeyaraman wrote: > > This is issue 8323 http://www.w3.org/Bugs/Public/show_bug.cgi?id=8323. > > > > A possible way to address this issue is to add the highlighted > sentence in [1] to WS-Eventing specification. How does this sound to you? > > > > Thanks. > > > > [1] > > > > *[Body] /wse:Subscribe/wse:Filter/@Dialect * > > Implied value is "http://www.w3.org/2009/09/ws-evt/Dialects/XPath10". > > XPath expressions provided for the event filtering should take into > account the notification format (wrapped or unwrapped). While an XPath > predicate expression provides great flexibility and power, alternate > filter dialects MAY be defined. For instance, a simpler, less powerful > dialect might be defined for resource-constrained implementations, or > a new dialect might be defined to support filtering based on data not > included in the notification message itself. If desired, a filter > dialect could allow the definition of a composite filter that > contained multiple filters from other dialects. New dialect > definitions MUST include sufficient information for proper > application. For example, it would need to include the context (which > data) over which the filter operates > > > > -----Original Message----- > From: public-ws-resource-access-request@w3.org > [mailto:public-ws-resource-access-request@w3.org] On Behalf Of > Johannes Echterhoff > Sent: Tuesday, December 01, 2009 10:48 AM > To: public-ws-resource-access@w3.org > Subject: AW: context node for XPath filter dialect ambiguous? > > > > Hi again. > > > > It seems like it is the responsibility of the subscriber to know which > type or types of event XML it is subscribing for. As WS-Eventing does > not add a wrapper around notifications by itself, it can only be the > events that the client subscribes to which come in multiple formats - > which the subscriber should know and then write the Xpath expression > accordingly (or rather, avoid different event XML types in one > subscription altogether if Xpath [1.0] based filtering should be used). > > > > So that would solve the issue with the wrapper format for me (second > half of my previous email). > > > > Now the only thing that would be good to clarify in the WS-Eventing > specification is how an Xpath expression should be created if the > context node is the root node of the event XML - this would best be > done with an example and some expressions which do or do not result in > a match. > > > > Best regards, > > Johannes Echterhoff > > > > > > > -----Ursprüngliche Nachricht----- > > > Von: public-ws-resource-access-request@w3.org > > > [mailto:public-ws-resource- access-request@w3.org] Im Auftrag von > > > Johannes Echterhoff > > > Gesendet: Freitag, 13. November 2009 10:54 > > > An: public-ws-resource-access@w3.org > > > Betreff: context node for XPath filter dialect ambiguous? > > > > > > Hi. > > > > > > Reading the latest draft of WS-Eventing, I was wondering what the > > > context node for an Xpath 1.0 filter really is. The specification says > > > that the context node is "the root of the event XML". Maybe its just > > > me but does > > this > > > mean the root node of the document where the event XML is contained or > > > the top element of the event XML (which one could also call 'root of > > > the event XML'). > > > > > > Suppose I have got the following event XML: > > > > > > <swes:SWESEvent xmlns:swes="http://www.opengis.net/swes/1.0" > > > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" > > > xmlns:wsa="http://www.w3.org/2005/08/addressing"> > > > <swes:identifier > > > codeSpace="http://www.example.org">a87s9c76s</swes:identifier> > > > <swes:code>CAPABILITIES_CHANGED</swes:code> > > > <swes:service> > > > <wsa:Address>http://my.swe-service.com/path</wsa:Address > <http://my.swe-service.com/path%3c/wsa:Address>> > > > </swes:service> > > > </swes:SWESEvent> > > > > > > Now I want to filter via Xpath to get all events with codeSpace > > > "http://www.opengis.net/swes/1.0#EventCode" and value > > "CAPBILITIES_CHANGED". > > > > > > Let us consider the following Xpath expressions: > > > > > > 1. swes:SWESEvent/swes:code='CAPABILITIES_CHANGED' > > > 2. /swes:SWESEvent/swes:code='CAPABILITIES_CHANGED' > > > 3. swes:code='CAPABILITIES_CHANGED' > > > > > > I evaluated the expressions with different choice of context node: > > > - with context node root node: 1+2 are true ... 3 is false > > > - with context node swes:SWESEvent: 1 is false ... 2+3 are true > > > > > > ----------- > > > > > > What would happen if the event XML is somehow wrapped already before > > > the Xpath is applied - or is that not allowed / considered? > > > > > > Example: > > > > > > <abc:Wrapper xmlns:swes="http://www.opengis.net/swes/1.0" > > > xmlns:abc="http://www.example.org" > > > xmlns:xsi="http://www.w3.org/2001/XMLSchema-instance" > > > xmlns:wsa="http://www.w3.org/2005/08/addressing"> > > > <abc:someProperty> > > > <swes:SWESEvent> > > > <swes:identifier > > > codeSpace="http://www.example.org">a87s9c76s</swes:identifier> > > > <swes:code>CAPABILITIES_CHANGED</swes:code> > > > <swes:service> > > > <wsa:Address>http://my.swe-service.com/path</wsa:Address > <http://my.swe-service.com/path%3c/wsa:Address>> > > > </swes:service> > > > </swes:SWESEvent> > > > </abc:someProperty> > > > </abc:Wrapper> > > > > > > Again, I evaluated the expressions with different choice of context > node: > > > - with context node root node: 1+2+3 are false > > > - with context node abc:someProperty: 1 is true ... 2+3 are false > > > - with context node swes:SWESEvent: 1+2 are false ... 3 is true > > > > > > What do you think? > > > > > > Best regards, > > > Johannes Echterhoff > > > > > > > > >
Received on Wednesday, 16 December 2009 01:17:41 UTC