OASIS Mailing List ArchivesView the OASIS mailing list archive below
or browse/search using MarkMail.

 


Help: OASIS Mailing Lists Help | MarkMail Help

wsbpel-spec-edit message

[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]


Subject: undefined behavior



My recollection is that when we closed 144 we said that no undefined behaviors had been identified after the restructure and that any that were found should have issues opened.  However, if these are trivial, an action item should be ok.  As the first of their class it might be best to review  them with the TC.  

Mark, if you think action items are ok, please resend this to the TC list with the note that we will open action items for the spec editing team unless there is an objection.  (I've copied the spec editing team - they may want to weigh in beforehand).

Regards, Diane
IBM  Emerging Internet Software Standards
drj@us.ibm.com
(919)254-7221 or 8-444-7221, Mobile: 919-624-5123, Fax 845-491-5709



"Peter Furniss" <peter.furniss@choreology.com>

03/08/2006 02:24 AM

To
"'Mark Ford'" <mark.ford@active-endpoints.com>, Diane Jordan/Raleigh/IBM@IBMUS
cc
Subject
RE: new issue





Whoops - I found this in my inbox. But I seem to remember some message (from Diane ?) leading me to think there was some doubt about whether it should go in.
 
I'm happy to put it in, and apologise if I've just been asleep on this. Mark - do you want it submitted as an issue (and if so, can you give me a title please). Or is it just a candidate for the action items ?  
 
(Hope to be on the call tomorrow, but don't expect to be able to update the issue list until later in the evening, or sometime on Thursday).
 
Peter
-----Original Message-----
From:
Mark Ford [mailto:mark.ford@active-endpoints.com]
Sent:
28 February 2006 15:47
To:
'Peter Furniss'; 'Diane Jordan'
Subject:
new issue

In response to closing Issue 144, there are still a few places in the current working draft that mention undefined behavior. I'm not sure if these are "issue worthy" or not since they are pretty straight forward.

#1 undefined changes to bpws:missingReply

5.5 The Lifecycle of a Business Process

(Paragraph 8)

A receive activity for an inbound request/response operation is said to be open if that activity has been performed and no corresponding reply activity has been performed.  If the process instance reaches the end of its behavior, and one or more receive activities remain open, then the status of the instance becomes undefined. This condition indicates a modeling error that was not detected by static analysis.

Change to:

A receive activity for an inbound request/response operation is said to be open if that activity has been performed and no corresponding reply activity has been performed.  If the process instance reaches the end of its behavior, and one or more receive activities remain open, then the process faults with a bpws:missingReply. This condition indicates a modeling error that was not detected by static analysis.

#2 undefined changes to bpws:selectionFailure

Section 8.3 Expressions

(2nd to last paragraph)

This function extracts global property values from variables. The first argument names the source variable for the data and the second is the qualified name (QName) of the global property to select from that variable (see Variable Properties). If the referenced property is not defined or if there does not exist a propertyAlias to associate the property with the referenced variable then the semantics of the process is undefined.  

This sentence in bold was introduced with the acceptance of Issue 145. The line should be changed to the following:

If the referenced property is not defined, if there does not exist a propertyAlias to associate the property with the referenced variable or if the given property definition selects a node set of a size other than one, then the standard fault bpws:selectionFailure MUST be thrown by a compliant implementation.

#3 undefined changes to bpws:missingRequest

10.4 Providing Web Service Operations

(Paragraph beginning with "The correlation between a request and the corresponding reply…)

…For the purposes of this constraint, an onMessage clause in a pick is equivalent to a receive (see Pick). Moreover, a reply activity must always be preceded by a receive activity for the same partner link, portType and (request/response) operation, such that no reply has been sent for that receive activity. The semantics of a process in which this constraint is violated is undefined.

Strike the last two sentences since the bpws:missingRequest fault is detailed a few paragraphs later.



[Date Prev] | [Thread Prev] | [Thread Next] | [Date Next] -- [Date Index] | [Thread Index] | [List Home]