Raw chat log for SCA bindings TC

Mike Edwards: I'll join in a moment

Eric Johnson: Scribe: Eric

Mike Edwards: trouble dialling in

Mike Edwards: in

Eric Johnson1: Topic: agenda review

Eric Johnson1: No comments on agenda.

Eric Johnson1: Topic: Approval of the minutes of last meeting

Eric Johnson1: No objections to approving the minutes.

Eric Johnson1: Topic: action items

Eric Johnson1: No progress.

Eric Johnson1: Topic: New issues

Eric Johnson1: Subtopic: BINDINGS-138

Eric Johnson1: Simon pointing out that the normative statement is unclear as to *which* connectionFactory element it applies to.

Eric Johnson1: Bryan moves to open, Dave 2nds.

Eric Johnson1: No objections, motion passes, issue is opened.

Eric Johnson1: Simon review proposal.

Eric Johnson1: Simon: Does not appear that any new normative statements are required, just some additional wording required to clarify.

Eric Johnson1: Bryan: You don't believe this changes the meaning of the normative statement?

Eric Johnson1: Simon: No

Mike Edwards: +1

Eric Johnson1: Bryan moves to resolve BINDINGS-138 with proposal in JIRA, Dave 2nds.

Eric Johnson1: No objections, motion passes, BINDINGS-138 is resolved.

Eric Johnson1: Subtopic: BINDINGS-139

Eric Johnson1: Simon: As with previous issue, we have a normative statement that deals with a response, just need to narrow the meaning of BJM30019.

Eric Johnson1 morphed into Eric Johnson

Eric Johnson: Bryan moves to open, Mike 2nds.

Eric Johnson: No objections, motion passes and the issue is opened.

Eric Johnson: Simon reviews proposal

Eric Johnson: Bryan moves to resolve BINDINGS-139 with the proposal in JIRA, Dave 2nds.

Eric Johnson: No objections, motion passes, issue is resolved.

Eric Johnson: Topic: JMS Binding test assertions and test cases

Eric Johnson: Simon: Anyone looked at them?

Eric Johnson: (pregnant pause)

Eric Johnson: Simon: Apparently not - did a side by comparison of test assertions and test cases.

Eric Johnson: ... haven't had the time to look at the implementation of the test cases.

Eric Johnson: ... One concern to raise - syntax checking of URI.  Is it OK to just have a negative test, or do we need to validate different parts of the URI.

Eric Johnson: ... anything similar done in other TCs?

Eric Johnson: Bryan: We're not defining the format, this isn't the place to do exhaustive testing.

Eric Johnson: Simon: Currently we test if there's an invalid URI, we reject it.

Eric Johnson: ... Some assertions that deal with out headers get set.  We have a structure - if A, do this, or if B, do another, or if C do a third.

Eric Johnson: ... trouble is that the test cases don't test combinations - what if A & B are present.  For example, if the URI is present, and the destination is present, we don't have a test to confirm that the URI is used instead of the destination?

Eric Johnson: ... Is it reasonable to have the tests that we currently have?

Eric Johnson: Mike: Unreasonable to expect the test cases to cover all the possibilities.

Eric Johnson: ... Cover the assertions, but it is unrealistic to cover every feature.

Eric Johnson: Simon: We're missing the point slightly with the test cases that are there, because we don't verify the priority of a particular property.

Eric Johnson: Ant: That could be the case for some of them - there are certainly some missing.

Eric Johnson: Simon: I've not had a chance to look at all the test cases.  Specifically looked at the headers, and the order of that priority.

Eric Johnson: ... In principle, we don't think we need to test every case, but just verify that we're applying properties correctly.

Eric Johnson: ... There are a number of test cases where what was tested, and what was asserted are different.

Eric Johnson: ... Example: BJM_30010_TestCase - if you're using request response, only deliver one response.

Eric Johnson: ... Example: BJM_30010_TestCase - if you're using request response, only deliver one response.

Eric Johnson: ... possible to end up with multiple messages delivered.

Eric Johnson: ... Cannot test, because you cannot wait for eternity....

Eric Johnson: ... My feeling is that the assertion is untestable so we should remove the test case.

Eric Johnson: ... 60009 - "should" assertion.

Eric Johnson: Bryan: We've taken the view with other test cases that if they're marked "should", we won't test.

Eric Johnson: Simon: Really a comment on test assertions document.

Eric Johnson: Byran: Just should mark it as "not tested."

Eric Johnson: ... We had some discussions in the assembly TC, and I think we should follow that pattern.

Eric Johnson: Action: Simon - check on the pattern for not-having test cases, and verify that we've applied it.

Eric Johnson: ... make sure that assertions with "shoulds" don't have a mandatory test assertion, but "preferred".

Eric Johnson: Simon: Some more comments... need some more time to look at the test implementation.  Ant should look at Simon's list.

Eric Johnson: Action: Ant to look at Simon's email to see what should be done.

Eric Johnson: Simon: I think we should leave the vote until next week.  Everyone should look at the documents and provide comments.

Eric Johnson: ... any other comments.

Eric Johnson: Dave: Before we go on to 1.2, something else to comment on.

Eric Johnson: ... Status question: Where are we in terms of 1.1?  Assembly is ready, Policy is read ...

Eric Johnson: Simon: Web service binding - everything addressed, public review of test assertions & test cases is still ongoing?

Eric Johnson: Dave: Does Tuscany pass the suite?

Eric Johnson: Simon: Don't know, that would be a question for Mike.

Eric Johnson: Ant: Yes - web services binding passes under Tuscany.

Eric Johnson: Dave: So we're really just waiting for public review to finish

Eric Johnson: Bryan: After review, we can vote on final CDs of all documents.

Eric Johnson: Bryan: After review, we can vote on final CDs of all documents.

Eric Johnson: ... Sinking suspicion that we'll have to have another PR of CAA.

Eric Johnson: Simon: Action item that we have to resync the schema.  As far as the spec is concerned, everything is complete.  Need to do JMS test cases & test assertions PR.

Eric Johnson: Dave: Right.

Eric Johnson: Simon: JCA is still open.

Eric Johnson: Dave: Not sure we're going to hold up 1.1 for either JMS or JCA.

Eric Johnson: Topic: Impact from 1.2 work.

Eric Johnson: Simon: Have seen notes going by related to bindings, and where bindings are implied.

Eric Johnson: ... Are there going to be changes to bindings?

Eric Johnson: ... Is anyone able to speak to the known impact on the bindings?  Do we want to have a specific named person keeping the bindings TC up-to-date?

Eric Johnson: Bryan: To some degree, as some things become clearer in assembly, the information will come here, or more formal cross communication.

Eric Johnson: Dave: Or through common members as issues.

Eric Johnson: Simon: One approach is to not do anything as a TC.  What changes are coming, and of what magnitude, and when?

Eric Johnson: Dave: Maybe Eric will speak up - don't see a lot of change for existing bindings, but potential for new bindings.

Eric Johnson: Eric: Assembly TC might kick a number of things this way.  Complexity will lie with items like policy, and what it means to provide producer/consumer functionality.

Eric Johnson: Simon: Don't know how much we need to wait on what happens in assembly.

Eric Johnson: Bryan: For what it is worth, we're taking a wait-and-see attitude, but trying to keep up.  Officially taking a hands-off approach for consensus to appear.

Eric Johnson: Simon: Once we get the test cases done, do we work on HTTP binding, or are we waiting for other TCs to drive our work?

Eric Johnson: Bryan: Working on HTTP sounds like the right approach.

Eric Johnson: (oops, that was Dave)

Eric Johnson: Bryan: Issues have been raised in assembly about type system requirements - potentially has enabling impacts with binding.http.

Eric Johnson: Simon: any other discussion?

Eric Johnson: ... Next meeting will be focused on JMS test documents, and HTTP binding.  Please go look at test documents.

Eric Johnson: Topic: AOB?

Eric Johnson: Meeting adjourned.

sca-bindings message