Next in thread →
Next in month →
RE: [ebxml-iic] about 3.5 in spec
Title: about 3.5 in spec
Mike:
agree
with your comments.
I see
one remaining snag with $MessageId the way we use it:
test
cases 182 and 187 (multi-hop) actually need $MessageId in step #3, to come
from
the
message send in step #1, not from message sent in step #2, I believe (we want to
correlate the
If
thats teh case, two ways to handle this:
(1)
could use either an assignment notation, i.e. a second
expression at the end of a message expression, add: $myMessageId =
<XPath>
(2)
reintroduce "$RefToMessageId" as coming from RefToMessageId of the message in
previous step (either sent or received)
Opinion?
Jacques
-----Original Message-----
From: Michael Kass
[mailto:]
Sent: Monday, September 29, 2003
10:26 AM
To: Jacques Durand;
Subject: Re: [ebxml-iic] about 3.5 in
spec
Jacques,
Attached is the latest draft
MS Conformance Spec, version 0.96
Please see comments below.
Thanks,
Mike
----- Original Message -----
From:
Jacques Durand
To: 'Michael Kass' ;
Sent: Friday, September 26, 2003 1:19
PM
Subject: RE: [ebxml-iic] about 3.5 in
spec
Mike:
Front page should show right date, and also
version
(and it is called now a "Committee Draft" V0.95)
Comments mostly on the "parameters", which was the main remaining
issue:
---------------------------------
3.3.2: Message
content parameters -
Preceded with a '$' symbol, these parameters represent dynamic message
content used to both construct an ebXML message as well as evaluate returned
message conformance. These parameters and their description are listed in
section 3.5.2.
>should
be : "...are listed in section 3.5.1"
[MIKE]
- Fixed
---------------------------------
3.3.2:
we need to say a word on the notion of "test step". We can reuse def
from Test Framework 3.3.1, e.g.:
"A
Test Case is a sequence of Test Steps. A Test Step is an aggregate of one or
more operations on the implementation under test. A Test Step usually
involves a single message sending or receiving operation, plus some data
verification, like checking or enforcing a condition on message header. A Test Step may also include
conditional actions that are a basis for the execution of the assertion
within the Test Step itself."
---------------------------------
[MIKE]
- Added similar wording for Test Case, Test Step..etc
descriptions
About
parameters: we should get rid of $CPAId:
testcase_27.4:
"eb:CPAId=$CPAId
" , why not refer to one of our predefined CPAId ike in other test cases?
Same for 27.5 and other places. I guess we can do without
$CPAId.
---------------------------------
Table
of parameters in 3.5.1 below need a few words about *when* these parameters
are set, as the expressions where they are used assume they are
set,
and
we should get rid of those that are
unused.
CPAId (suggest to get rid
of)
[MIKE] -
Removed
Reference to one of 13 MSH configuration
instances described in section
3.5.3
ConversationId
As defined in [ebMS]. Represents the
ConversationId associated with a particular test case (will remain the
same for all test steps of the test
case).
[MIKE] -
Aded
Service (not used here)
[MIKE] -
Removed
For this abstract test suite:
urn:ebxml:iic:test
Action (not used here)
[MIKE] -
Removed
As defined in
[ebMS]
SenderParty
(instead of using a parameter, e.g. in test case
#90, can't we just directly refer to the From/PartyId element in the
received message? Same for multi-hop test
cases)
[MIKE] - I
believe that using such a syntax would "break" our XPath meaning, and
confuse the test writer, since we really have no logical way
to describe this relationship in an XPath expression. I
believe that a simple $SenderParty parameter is much easier to
understand.
The From/PartyId message
value
ReceiverParty
(not used
here)
[MIKE] -
Removed
The To/PartyId message
value
MessageId
(we still have expressions with some of these
left: I thought we decided to get rid of
them)
[I removed
those that were not necessary however $MessageId is still necessary
for RefToMessageId, StatusResponse and Acknowledgment test case
verification ( verification must be done that they are referring to
the correct previous message), so I have left those in for tests 41,
43, 101, 134, 182, 183, 184, 187, 188,
As defined in
[ebMS]
RefToMessageId
[MIKE] - I changed $RefToMessageId to
$MessageId, since they are semantically equivalent. Therefore
there is no need for both of these
parameters.
As defined in [ebMS]. Represents the message
ID value of the message that was manipulated in the previous test
step, e.g. of the latest message sent or
received.
---------------------------------------
The parameterized example in Sec 3.5.2 should not show parameters
that are not used in the test suite.
[MIKE] - Fixed that.
Regards,
Jacques
-----Original Message-----
From: Michael Kass
[mailto:]
Sent: Wednesday, September 24,
2003 8:56 AM
To: Jacques Durand;
Subject: Re: [ebxml-iic] about
3.5 in spec
Jacques and all,
Please see my comments
below. Also, I've attached V0.95 based on comments
below.
Regards,
Mike
----- Original Message -----
Next in thread →
Next in month →