On Mon, Aug 1, 2011 at 4:30 PM, Mike Edwards <mike.edwards.inglen...@gmail.com> wrote: > Folks, > > Simple answer to Scott's questions: > > On 26/07/2011 16:08, Scott Kurz wrote: >> >> Trying to understand myself as the spec doesn't seem crystal clear... >> >> I assume we'd agree also that the answer should be the same for this >> variation: >> >> composite B with component C2 >> >> <service name="S1Promoted" promote="C2"> >> <binding.ws/> <!-- Binding now specified on composite service >> --> >> </service> >> <component name="C2"> >> <implementation.java class="..."/> >> <service name="S1"> >> <binding.ws/> >> </service> >> </component> >> >> In contrast, suppose we that composite A with component C1 did not >> explicitly specify >> <binding.sca>, e.g.: >> >> <component name="C1"> >> <implementation.composite name="B"/> >> <service name="S1Promoted"/> >> </component> >> >> Should this be<binding.ws> then? > > YES > >> Does it matter whether B's >> composite-level service has the binding or not? > > NO - if the component service has one or more bindings specified. > YES - if the component service has zero bindings specified. > >> >> Scott >> > > To quote chapter & verse from the OASIS Assembly spec (line numbers for > CD06): > > 770 ... If no > 771 binding elements are specified for the service, then the bindings > specified for the equivalent service in > 772 the componentType of the implementation MUST be used, but if the > componentType also has no > 773 bindings specified, then <binding.sca/> MUST be used as the binding. If > binding elements are > 774 specified for the service, then those bindings MUST be used and they > override any bindings specified > 775 for the equivalent service in the componentType of the implementation. > [ASM50005] > > > Yours, Mike. >
To close this off I applied Greg's patch and get a clean build so it's in svn now. Simon -- Apache Tuscany committer: tuscany.apache.org Co-author of a book about Tuscany and SCA: tuscanyinaction.com
