Re: [sca-assembly] Problems in the application of Resolution for Assembly57 to the Specification - Proposed editorial fixes

Anish,

Thanks for taking the time to check
this over - the more eyes the better for this stuff, I think.

From your comments, I assume that you're
not suggesting any changes to the editorial changes here.

Some comments line

Yours,  Mike.

Strategist - Emerging Technologies, SCA & SDO.

Co Chair OASIS SCA Assembly TC.

IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great Britain.

Phone & FAX: +44-1962-818014    Mobile: +44-7802-467431  

Email:  

Anish Karmarkar <>
wrote on 25/11/2008 06:35:44:

> [image removed] 

> 

> Re: [sca-assembly] Problems in the application of Resolution for 

> Assembly 57 to the Specification - Proposed editorial fixes

> 

> Anish Karmarkar 

> 

> to:

> 

> OASIS Assembly

> 

> 25/11/2008 06:36

> 

> Mike Edwards wrote:

> > 

> > Folks,

> > 

> > Assembly 57 is the issue that decided that any <binding/>
on a 

> > <reference/> MUST identify a target

> > service either via @uri or via binding-specific mechanisms.

> > 

> > This essentially outlaws the use of the reference @target attribute
with 

> > any <binding/> child elements

> > on the same reference.

> > 

> 

> That means we don't allow a ref with multiplicity >1 to point to
a SCA 

> service *and* an external service. Not sure if this is a mainline

> usecase, but wanted to point it out.

<mje>

No, that actually isn't
the case.  A <binding.sca/> element can be present with its
@uri

attribute set to some target.
 This can legally be mixed with other bindings:

eg.

<component name="foo">

       
<implementation.java class="com.foo.someImpl"/>

       
<reference name="someRef">

       
        <binding.ws uri="http://www.acme.com/orders/widgets"/>

       
        <binding.sca uri="bar/service1"/>

       
</reference>

</component>

...this is perfectly OK
and shows a reference linked both "internally" and "externally"

</mje>

> 

> > It also implies that there is no choosing between multiple <binding/>

> > child elements, since they all

> > apply, once they identify a target.

> > 

> 

> Not sure what you mean.

> Do you mean that a reference can't choose a specific binding of a

> service? If so, I recall that we allowed that via the use of the @name

> attribute on service bindings.

> 

<mje>

Perhaps I should have explained
myself better.  Originally, where there was a reference with

two or more <binding/>
subelements, the runtime could choose which of the bindings to use

when connecting to some
target service (ie via the @target attribute), assuming that the target

service also supported
two or more of those bindings.

Now, all of this is done
away with.  Multiple bindings can only be present if they each identify
a

specific target.  There
is no question of any choice over the binding to use relating to the

bindings specified on the
reference.

Of course, the situation
is different for a reference with zero <binding/> elements, using
the @target

attribute to wire to a
service which has multiple <binding/> elements.  In this case,
the runtime must

select one of the bindings
on the service and is free to choose any one that satisfies the requirements

(intents. policies) on
the reference.

</mje>

> > Considering  sca-assembly-1.1-spec-cd01-rev3.doc

> > 

> > Section 5.3.1:

> > 

> > - does not reflect Issue 57 at all.

> > - the following changes are necessary:

> > 

> > 1. Remove lines 866 - 868

> > (this situation can never occur - if there is an @target attribute
then 

> > there can't be any <binding/> subelement)

> > 

> > 2. Change lines 856 - 858

> > (required to state that <binding/> subelements cannot exist
alongside an 

> > @target attribute)

> > 

> > If a reference has a value specified for one or more target services
in 

> > its @target attribute,

> > 

> > there MUST NOT be any child <binding/> elements declared
for that 

> > reference. [ASM50026]

> > 

> > 

> > 

> > Yours,  Mike.

> > 

> > Strategist - Emerging Technologies, SCA & SDO.

> > Co Chair OASIS SCA Assembly TC.

> > IBM Hursley Park, Mail Point 146, Winchester, SO21 2JN, Great
Britain.

> > Phone & FAX: +44-1962-818014    Mobile: +44-7802-467431
 

> > Email:  

> > 

> > 

> > ------------------------------------------------------------------------

> > 

> > /

> > /

> > 

> > /Unless stated otherwise above:

> > IBM United Kingdom Limited - Registered in England and Wales
with number 

> > 741598.

> > Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire
PO6 3AU/

> > 

> > 

> > 

> > 

> > 

> > 

> 

> ---------------------------------------------------------------------

> To unsubscribe from this mail list, you must leave the OASIS TC that

> generates this mail.  Follow this link to all your TCs in OASIS
at:

> https://www.oasis-open.org/apps/org/workgroup/portal/my_workgroups.php

> 

Unless stated otherwise above:

IBM United Kingdom Limited - Registered in England and Wales with number
741598. 

Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6
3AU