So, in this case, when using Tuscany models/processors to read a composite that has <service name="service" requires="">, then writing back <service name="service"> would be considered ok and expected behavior ?
On Wed, Jan 7, 2009 at 3:59 PM, Raymond Feng <[email protected]> wrote: > Our our java models and processors are strongly typed and checked for the > values. I don't think they are designed to round trip the composite XML that > contains invalid values. It can only ensure that valid content can be > written back in a semantically equivalent way. For example, we cannot > differentiate the case between @multiplicity="1..1" or non-presence of the > @multiplicity. > > Thanks, > Raymond > -------------------------------------------------- > From: "Luciano Resende" <[email protected]> > Sent: Wednesday, January 07, 2009 3:37 PM > To: <[email protected]> > Subject: Expected behavior while parsing empty value policy intent (e.g > requires="") > >> I came across a interesting scenario with our composite processors. >> When a user has defined a service that has a policy intent "requires" >> attribute with empty value (e.g <service name="service" requires="">) >> the policy processors are ignoring this require and when we try to >> write back the composite, the attribute is gone. What should be the >> expected behavior of the processors in this case ? Should it generate >> a warning and store the invalid policy information to enable fidelity >> during write phase ? >> >> -- >> Luciano Resende >> Apache Tuscany, Apache PhotArk >> http://people.apache.org/~lresende >> http://lresende.blogspot.com/ > > -- Luciano Resende Apache Tuscany, Apache PhotArk http://people.apache.org/~lresende http://lresende.blogspot.com/
