Classification: UNCLASSIFIED
Caveats: NONE
Simon,
Sorry about the delay - worked from home yesterday and could not respond
(at least did not know how) to respond to existing threads without email
history - I normally use my work email to post and home email did not
get all the mailing copies.
The message is getting too long and hard to locate current discussion
point ("cursor"). My response is embedded with "GY:" prefix following
your most recent text.
Thanks,
Gang
-----Original Message-----
From: Simon Nash [mailto:n...@apache.org]
Sent: Wednesday, February 02, 2011 2:40 AM
To: dev@tuscany.apache.org
Subject: Re: An issue with PolicyHandler extension framework - a bug or
design flaw? (UNCLASSIFIED)
Simon Nash wrote:
> Yang, Gang CTR US USA wrote:
>> Classification: UNCLASSIFIED
>> Caveats: NONE
>>
>> Simon,
>>
>> Don't worry about if I'm offended - I'm not. I'm only focused on one
>> aspect and I'm sure there are other aspects that need to be
considered.
>> Discussion only make the product better.
>>
>> The "in-out: beforeInvoke(inMC) / afterInvoke(inMC) /
>> beforeInvoke(outMC)" pattern seems confusing though - just my
opinion. I
>> think "in-out: beforeInvoke(inMC) / afterInvoke(inMC)" is more
>> understandable. afterInvoke() can be called with both inMC and outMC
>> (contexts is an array) so post invocation logging can be done in
>> afterInvoke(). Actually, if Axis2 is used, you can get the inbound
>> context from the outbound context. I'm using it to pass states from
>> inbound to outbound.
>>
> OK, then we could do the following:
> in-only: beforeInvoke(inMC) / afterInvoke(inMC)
> in-out: beforeInvoke(inMC) / afterInvoke(outMC)
> out-only: beforeInvoke(outMC)
>
Thinking about this some more, I believe it's also possible for
the policy handler to get outMC from inMC for the in-out case. This
means that we could still call afterInvoke(inMC) for the in-out case
as we do at present, but make sure that outMC is fully populated with
the response data before we make this call. The policy handler could
get outMC from inMC and do whatever processing it wants on the
contents of outMC.
GY: I guess it's concept and design issue. I think there are two things
that we have been putting together in our discussion. The first is the
message pattern: in-only, in-out and out-only. What goes with this is
the direction of the message. For in-only, there's only one direction,
inbound. For in-out, there's inbound and outbound. For out-only, there
is only outbound. This is where I'm interested for WS-security. From a
service viewpoint looking at the in-out pattern, I would like to think
processes in terms of inbound and outbound, which would correspond to
two handler methods in the design, let's say inInvoke() and outInvoke()
for the inbound and outbound respectively. Naturally I would like to
access inbound context, inMC, and outbound context, outMC in these two
methods respectively rather than always inbound context, inMC. That's
where I came from.
GY: Then there's the second thing in our discussion, which the timing of
invocation of the handler's methods relative to the invocation of the
endpoint via wire. I guess that's why the current methods names are
called before/afterInvoke(). I had mapped "before" to "in" and "after"
to "out". This probably explains why I see it the way I did.
GY: Back to what should it be moving forward, my opinion is that if we
keep the methods names as is, then
In-only: beforeInvoke(inMC) / afterInvoke(inMC)
In-out: beforeInvoke(inMC) / afterInvoke(outMC)
Out-only: beforeInvoke(outMC) / afterInvoke(outMC)
GY: Here we combined the two semantics, inbound/outbound and
before/after into one. One other note on afterInvoke() calling position,
it needs to be moved from Axis2ServiceProvider.invokeTarget() to a later
time because outMC, if it's directly passed into afterInvoke() or
obtained from inMC, has not be setup correctly.
Simon
> When afterInvoke() is called with outMC, it can do postprocessing of
> the request (by getting inMC from outMC) and preprocessing of the
> response.
>
> Simon
>
>> Gang
>>
>> -----Original Message-----
>> From: Simon Nash [mailto:n...@apache.org] Sent: Tuesday, February 01,
>> 2011 12:59 PM
>> To: dev@tuscany.apache.org
>> Subject: Re: An issue with PolicyHandler extension framework - a bug
or
>> design flaw? (UNCLASSIFIED)
>>
>> Yang, Gang CTR US USA wrote:
>>> Classification: UNCLASSIFIED
>>> Caveats: NONE
>>>
>>> Simon,
>>>
>>> I certainly do not have broad-enough knowledge about Tuscany
internals
>>> to determine if the changes I made was consistent. It only serves as
a
>>> code snippet to demonstrate my issue, but not necessarily final code
>> to
>>> fix the problem. Simon L. said he was going to play with it to see
if
>>> there's any problem. The key issue is that for the outbound
(response)
>>> processing, the outbound context should be provided.
>>>
>> Hi Gang,
>> My comment wasn't intended as any criticism of the code you have
>> written, which fixes an important problem in the case of adding
>> WS-Security to response messages. I was trying to progress the
>> discussion to reach a conclusion among the Tuscany community on
>> what final code change needs to be applied.
>>
>>> Regarding to the Axis2ServiceInOutSyncMessageReceiver vs
>>> Axis2ServiceInMessageReceiver, why would afterInvoke() be even
called
>>> for inbound-only message pattern?
>> >
>> Possibly because someone might want to write log entries both
>> before and after the service method has been invoked, to record the
>> fact that the service request has completed and returned.
>>
>>> Is there a
>>> Axis2ServiceOutMessageReceiver that process the outbound-only
message
>>> pattern? I would think that the outbound-only processing needs some
>>> consideration.
>>>
>> That's a good point. Tuscany currently doesn't support out-only.
>> If this were to be added, it would probably be necessary to call
>> beforeInvoke() with outMC to give the policy handler the opportunity
>> to add the necessary security information to the outbound message.
>> So perhaps the sequence should be:
>> in-only: beforeInvoke(inMC) / afterInvoke(inMC)
>> in-out: beforeInvoke(inMC) / afterInvoke(inMC) /
beforeInvoke(outMC)
>> out-only: beforeInvoke(outMC)
>>
>> Simon
>>
>>> Thanks,
>>> Gang
>>>
>>> -----Original Message-----
>>> From: Simon Nash [mailto:n...@apache.org] Sent: Tuesday, February
01,
>>> 2011 10:58 AM
>>> To: dev@tuscany.apache.org
>>> Subject: Re: An issue with PolicyHandler extension framework - a bug
>> or
>>> design flaw? (UNCLASSIFIED)
>>>
>>> Yang, Gang CTR US USA wrote:
>>>> Classification: UNCLASSIFIED
>>>> Caveats: NONE
>>>>
>>>> Hi, Simon,
>>>>
>>>> I have created the JIRA, TUSCANY-3822
>>> (https://issues.apache.org/jira/browse/TUSCANY-3822), and included
the
>>> two classes that I changed in order to access the outbound
>>> MessageContext in afterInvoke().
>>>> I've also successfully integrated our WS-Security solution with
>>> Tuscany 1.6.1 (with the fix) using the PolicyHandler approach in a
>>> prototype and have indicated to our project management that there's
>>> solution based on Tuscany 1.6.1 (with the fix). We got a "go" to
>> proceed
>>> down this path and our delivery date is the end of March. Our
>>> WS-Security project is to support other projects that are already
>> using
>>> Tuscany 1.6 and needed the government-specific WS-Security solution.
>>>> I'm not sure if there's another more "formal" channel to push for
the
>>> fix, but I would appreciate it if you or someone from Tuscany could
>>> provide the next release information (Tuscany 1.6.2? and date) and
>> would
>>> like to make sure that the fix be included. If there's anything else
I
>>> could do to further help on this, please let me know.
>>> I've been thinking about the consequences of moving the position
>>> of the afterInvoke() call. I can see that it works for a security
>>> policy handler, but it could cause problems for some other kind of
>>> policy handler such as logging.
>>>
>>> At present there's a pair of beforeInvoke() and afterInvoke() calls
>>> for every inbound method invocation, whether or not the method
>>> produces a response. These calls are made by the invokeTarget()
>>> method of Axis2ServiceProvider, which can be called from either
>>> Axis2ServiceInOutSyncMessageReceiver or
Axis2ServiceInMessageReceiver.
>>>
>>> With the proposed change, the afterInvoke() call would be part of
>>> Axis2ServiceInOutSyncMessageReceiver and would therefore be made
only
>>> for inbound methods that produce responses, not for inbound methods
>>> that don't produce responses. For a logging policy handler, this
>>> doesn't seem correct.
>>>
>>> Perhaps it would be better to arrange the logic so that
afterInvoke()
>>> is still called by Axis2ServiceProvider for both the in-only case
>>> and the in-out case, and change the signature of the afterInvoke()
>>> call to include outMC if the method is in-out. This would require
>>> passing outMC to Axis2ServiceProvider.invokeTarget() and moving the
>>> logic that sets up outMC from Axis2ServiceInOutSyncMessageReceiver
>>> into Axis2ServiceProvider.
>>>
>>> Simon N.
>>>
>>>> Thanks again.
>>>> Gang
>>>>
>>>> -----Original Message-----
>>>> From: Simon Laws [mailto:simonsl...@googlemail.com] Sent: Monday,
>>>> January 24, 2011 5:01 PM
>>>> To: dev@tuscany.apache.org
>>>> Subject: Re: An issue with PolicyHandler extension framework - a
bug
>>> or design flaw? (UNCLASSIFIED)
>>>> On Mon, Jan 24, 2011 at 9:51 PM, Yang, Gang CTR US USA
>>>> <gang.y...@us.army.mil> wrote:
>>>>> Classification: UNCLASSIFIED
>>>>> Caveats: NONE
>>>>>
>>>>> Hi, Simon and others,
>>>>>
>>>>>
>>>>>
>>>>> I'm reposting this because I found my earlier post was grouped
under
>>> an old
>>>>> topic and it may be why I did not receive any response, which I
>> badly
>>> hope
>>>>> to hear. To provide a background, I'm trying to write a Policy
>>> extension to
>>>>> add a new WS-security implementation to WS binding in Tuscany. In
my
>>> earlier
>>>>> post (see
>>> http://www.mail-archive.com/dev@tuscany.apache.org/msg15333.html),
>>>>> I described the problems with Tuscany 1.6.x Policy extension
>>> framework using
>>>>> PolicyInterceptor approach and got some pointers from Simon, one
is
>>> to try
>>>>> Tuscany 2.0 and the other is to use the PolicyHandler extension in
>>> Tuscany
>>>>> 1.6.x.
>>>>>
>>>>>
>>>>>
>>>>> I did not get any response for my request for help on the 2.0
Policy
>>>>>
>>>>> issues (see
>>>>> http://www.mail-archive.com/dev@tuscany.apache.org/msg15353.html),
>> so
>>> I went
>>>>> ahead down to
>>>>>
>>>>> the second path Simon suggested - to work with PolicyHandler
>>> extension
>>>>> point in 1.6.x.
>>>>>
>>>>>
>>>>>
>>>>> Everything went well on the reference side until I was testing on
>> the
>>>>> service side. I think the current PolicyHandler extension
framework
>>> has
>>>>> a defect when processing the PolicyHandler.afterInvoke() call for
>> the
>>>>> service side and the model does not fit the need to add
WS-security
>>> to the
>>>>> binding.ws.axis2 binding implementation. Today
>>> PolicyHandler.afterInvoke()
>>>>> is call with the OMElement response and MessageContext inMC, which
>> is
>>> still
>>>>> the message context for the inbound request. In order to add
>>> WS-security to
>>>>> the response, I need to access the message context for the
outbound
>>>>> response.
>>>>>
>>>>>
>>>>>
>>>>> So I modified Axis2ServiceInOutSyncMessageReceiver and
>>>>>
>>>>> Axis2ServiceProvider classes to delay the calling of
>>>>> PolicyHandler.afterInvoke()
>>>>>
>>>>> from inside Axis2ServiceProvider.invokeTarget() before outMC is
>>> constructed
>>>>> to after outMC has been constructed in
>>>>> Axis2ServiceInOutSyncMessageReceiver.invokeBusinessLogic(). After
>> the
>>>>> change, my prototype worked and it also passed all unit tests.
>>>>>
>>>>>
>>>>>
>>>>> Now I would like to hear Tuscany developers' opinions and/or get
>> your
>>>>> blessing. If this change is fine, how can it be pulled back into
the
>>> code
>>>>> base? If not, what can be done to support what I would like to do,
>>> which is
>>>>> pretty natural and generic.
>>>>>
>>>>>
>>>>>
>>>>> I would really appreciate responses so our project can make a
>>> decision.
>>>>>
>>>>> Thanks in advance,
>>>>>
>>>>> Gang
>>>>>
>>>>>
>>>>>
>>>>> PS:
>>>>>
>>>>> The following is the code segments to high light the change I made
>>> (with
>>>>> "Gang: ..." comments):
>>>>>
>>>>>
>>>>>
>>>>> Axis2ServiceInOutSyncMessageReceiver.invokeBusinessLogic(...) {
>>>>>
>>>>> ....
>>>>>
>>>>> OMElement responseOM =
>>>>> (OMElement)provider.invokeTarget(operation, args, inMC);
>>>>>
>>>>>
>>>>>
>>>>> /*
>>>>>
>>>>> for ( PolicyHandler policyHandler : policyHandlerList
)
>> {
>>>>> policyHandler.afterInvoke(operation, args, inMC,
>>>>> responseOM);
>>>>>
>>>>> }
>>>>>
>>>>> */
>>>>>
>>>>>
>>>>>
>>>>> SOAPEnvelope soapEnvelope =
>>>>> getSOAPFactory(inMC).getDefaultEnvelope();
>>>>>
>>>>> if (null != responseOM ) {
>>>>>
>>>>> soapEnvelope.getBody().addChild(responseOM);
>>>>>
>>>>> }
>>>>>
>>>>> outMC.setEnvelope(soapEnvelope);
>>>>>
>>>>>
>>>>>
>>> outMC.getOperationContext().setProperty(Constants.RESPONSE_WRITTEN,
>>>>> Constants.VALUE_TRUE);
>>>>>
>>>>>
>>>>>
>>>>> // Gang: Added
>>>>>
>>>>> for ( PolicyHandler policyHandler : policyHandlerList
)
>> {
>>>>> policyHandler.afterInvoke(responseOM, outMC);
>>>>>
>>>>> }
>>>>>
>>>>> ...
>>>>>
>>>>> }
>>>>>
>>>>>
>>>>>
>>>>> Axis2ServiceProvider.invokeTarget(...) {
>>>>>
>>>>> ...
>>>>>
>>>>> // find the runtime wire and invoke it with the message
>>>>>
>>>>> RuntimeWire wire =
>>>>> ((RuntimeComponentService)contract).getRuntimeWire(getBinding());
>>>>>
>>>>> Object response = wire.invoke(op, msg);
>>>>>
>>>>>
>>>>>
>>>>> // Gang: Commented out
>>>>>
>>>>> // for ( PolicyHandler policyHandler : policyHandlerList )
{
>>>>>
>>>>> // policyHandler.afterInvoke(response, inMC);
>>>>>
>>>>> //}
>>>>>
>>>>>
>>>>>
>>>>> return response;
>>>>>
>>>>> }
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>>
>>>>> Classification: UNCLASSIFIED
>>>>> Caveats: NONE
>>>>>
>>>>>
>>>> Hi Gang
>>>>
>>>> It seems reasonable to me at first glance.
>>>>
>>>> I'd like to see it running. A way for me to do this and to get it
>> into
>>>> Tuscany is for you to create a JIRA and attach your patch to it. I
>> can
>>>> then apply it locally and take a look at it.
>>>>
>>>> Sorry I didn't notice your 2.x question. It is in my reader but I
>> just
>>>> missed it amongst other stuff. I'd like to make sure that we can do
>>>> the same thing in 2.x as that is where most development is
happening.
>>>> What I could possibly do when I get 5 minutes is create some policy
>>>> interceptors which shows you how to get hold of the MC at the
before
>>>> and after locations on reference and server. If in the future you
>> want
>>>> to move up to 2.x then we'd be more comfortable that the code does
>>>> what you need it to.
>>>>
>>>> Regards
>>>>
>>>> Simon
>>>>
>>>
>>> Classification: UNCLASSIFIED
>>> Caveats: NONE
>>>
>>>
>>>
>>>
>>
>>
>> Classification: UNCLASSIFIED
>> Caveats: NONE
>>
>>
>>
>>
>
>
>
Classification: UNCLASSIFIED
Caveats: NONE
