Re: [wsbpel] Issue 157 - Rania's latest revision of RE desc

[Rania Khalaf <rkhalaf@watson.ibm.com>]

Hi Alex,

the semantics is the same. However, the description is different.

Perhaps a middle of the way solution would be not to talk about a 
sequence of steps (first, then, extra step) but just about what the 
desired result is shown below. This is proposed to replace the entire RE 
section of Charlton's document:

RE: Replace the entire element at the destination with a copy of the 
entire element at the source. An attribute named 'keepSrcElementName' is 
provided for flexiblity on the name of the copied element:
	- Its default value is 'no', in which case the name [ie. [namespace 
name] and [local name] of the original destination element is  used as 
the name of the copied element.

	- When the keepSrcElementName attribute is set ("yes"|"no"), the 
selection results of the from-spec and to-spec MUST be elements. A BPEL 
processor MAY enforce this checking through static analysis of the 
expression/query language. If a violation is detected during runtime, a 
bpws:selectionFailure fault MUST be thrown.

Examples:

consider the following variables and copy statements. Note that in both 
cases, var1 stays the same
:
   var1: <tns:a z='0'>123</tns:a>
   var2: <tns:b x='1'>456</tns:b>

<copy keepSrcElementName='yes'>
    <from>$var1/tns:a</from>
    <to>$var2/tns:b</to>
</copy>

Would result in var2: <tns:a z='0'>123</tns:a>

And

<copy>
    <from>$var1/tns:a</from>
    <to>$var2/tns:b</to>
</copy>

Would result in var2: <tns:b z='0'>123</tns:b>




Alex Yiu wrote:
> 
> Hi Charlton,
> 
> I am OK with the changes in general, except:
> 
> "*      If the keepSrcElementName attribute is set to *"yes"*, an extra 
> replacement step MUST be performed to  replace the destination 
> element-name with the source element-name  (i.e. [namespace name] and 
> [local name] properties)."
> 
> In WS-BPEL, the convention is to use "yes"-vs-"no", NOT, "true"-vs-"false".
> 
> *The default is still kept as "no"*. The semantics of 
> "keepSrcElementName" completely equivalent to 
> "replaceDestinationElementName". This should be just a pure syntactic 
> change.
> 
> I hope that is OK with everyone.
> 
> 
> 
> Regards,
> Alex Yiu
> 
> 
> Charlton Barreto wrote:
> 
>> Hi Rania,
>>
>> On further review I don't see any issue with using  
>> "keepSrcElementName" (or "keepSourceElementName").
>>
>> With your amendment, do you intend to replace:
>>
>> "*      If the replaceDestinationElementName attribute is set to  
>> "yes", an extra replacement step MUST be performed to replace the  
>> destination element-name with the source element-name (i.e.  
>> [namespace name] and [local name] properties)."
>>
>> with
>>
>> "*      If the keepSrcElementName attribute is set to  "false" 
>> (default), an extra replacement step MUST be performed to  replace the 
>> destination element-name with the source element-name  (i.e. 
>> [namespace name] and [local name] properties)."
>>
>> and to replace:
>>
>> "RE (Replace-Element-properties): First, remove all [children] and  
>> [attribute] properties of the destination EII. Then, copy and add all  
>> [children] and [attribute] properties of the source EII to the  
>> destination EII. An OPTIONAL replaceDestinationElementName attribute  
>> is provided to further refine the behavior:"
>>
>> with
>>
>> "RE (Replace-Element-properties): Replace the element at the  
>> destination with a copy of the element at the source - by first  
>> removing all [children] and [attribute] properties of the destination  
>> EII, then copying and adding all [children] and [attribute]  
>> properties of the source EII to the destination EII. An OPTIONAL  
>> keepSrcElementName attribute is provided to further refine the  
>> behavior:"
>>
>> This is how I see the text reading with your amendment.
>>
>> I would use the example provided for EII-to-EII copy with  
>> keepSrcElementName to illustrate it.
>>
>> I also wanted to pin down whether you sought to have the examples  
>> incorporated into the spec itself or a part of a primer. Let me know  
>> what your stance is on this....
>>
>> -Charlton.
>>
>> On 10/08/2005, at 08:30, Rania Khalaf wrote:
>>
>>> I strongly  prefer  'keepSrcElementName' instead of  
>>> 'replaceDestinationElementName' for the sakes of simplicity and XML ;)
>>>
>>> That would also change the description of RE (Replace element  
>>> properties) to say: "replace the element at the destination with a  
>>> copy of the element at the source. If the 'keepElementName'  
>>> attribute is set to false (default), then an additional step is  
>>> required where the name of the copied element is replaced with the  
>>> name of the destination element (add an example)."
>>>
>>> This is instead of the current description:
>>> •    RE (Replace-Element-properties): First, remove all [children]  
>>> and [attribute] properties of the destination EII. Then, copy and  
>>> add all [children] and [attribute] properties of the source EII to  
>>> the destination EII. An OPTIONAL replaceDestinationElementName  
>>> attribute is provided to further refine the behavior:
>>>
>>> Charlton Barreto wrote:
>>>
>>>> Attached is the latest revision of the Issue 157 proposal,  
>>>> reflecting  the comments Alex and I have received since last week,  
>>>> including  examples of each of the copy operations in the  
>>>> proposal's matrix.
>>>> -Charlton.
>>>> -- 
>>>> Charlton Barreto
>>>> P 1.408.536.4496
>>>> cbarreto@adobe.com
>>>> www.adobe.com
>>>>
>>>
>>>
>>>
>>
>> Charlton Barreto
>> P 1.408.536.4496
>> cbarreto@adobe.com
>> www.adobe.com
>>
>>
>