Re: CXF client startup time

[Kevin Priebe <ke...@realtyserver.com>]

I just signed up for the users list to add my comment to Benson's email.

 

My CXF client startup experience is similar, but seems to be worse when
testing CXF 2.3 compared to 2.2.  We use JAX-WS and Aegis and the startup
time for our largest service used to be 2200ms, which is already quite long.
However, when testing 2.3, the same service takes 3000ms to initialize!
This is a noticeable difference and also affects the startup time of our
application.  The release notes for 2.3 mentioned some work on speeding up
CXF, so I'm not sure why we are seeing this slow down (was hoping to see the
numbers go the other way).  Nothing was changed except the CXF version
update for this test.  For now we are staying on 2.2.11 just to avoid the
speed hit.  Any idea why this might take so long?  The test includes
creating the client and calling 1 test method (that does nothing on the
server to ensure it's not our fault).

 

Kevin

 

Re: WSDL defines I/O mep but how is all types of responses handled?

[Daniel Kulp <dk...@apache.org>]

On Wednesday 03 November 2010 3:36:04 pm am am wrote:
> Hi,
> 
> If a web service defines in the WSDL an operation with input/output message
> exchange pattern (all elements optional for response object type), if the
> web service returns a 202 Accepted, is this handled by the framework
> silently, or an exception will be raised? I would expect a null pointer to
> be returned to the calling code. Is this the case?
> 
> Thanks!

Well, I wouldn't expect an NPE.   I WOULD expect some sort of 
WebServiceException or other exception related to not being able to parse the 
non-existent XML.

-- 
Daniel Kulp
dkulp@apache.org
http://dankulp.com/blog

WSDL defines I/O mep but how is all types of responses handled?

[am am <ak...@yahoo.com>]

Hi,

If a web service defines in the WSDL an operation with input/output message 
exchange pattern (all elements optional for response object type), if the web 
service returns a 202 Accepted, is this handled by the framework silently, or an 
exception will be raised? I would expect a null pointer to be returned to the 
calling code. Is this the case?

Thanks!



      

Re: CXF client startup time

[Benson Margulies <bi...@gmail.com>]

I did some preliminary measurement, but nothing jumped out. Dan and I
are discussing a relatively radical attack.

On Mon, Nov 1, 2010 at 4:11 PM, Kevin Priebe <ke...@realtyserver.com> wrote:
> I just signed up for the users list to add my comment to Benson's email.
>
>
>
> My CXF client startup experience is similar, but seems to be worse when
> testing CXF 2.3 compared to 2.2.  We use JAX-WS and Aegis and the startup
> time for our largest service used to be 2200ms, which is already quite long.
> However, when testing 2.3, the same service takes 3000ms to initialize!
> This is a noticeable difference and also affects the startup time of our
> application.  The release notes for 2.3 mentioned some work on speeding up
> CXF, so I'm not sure why we are seeing this slow down (was hoping to see the
> numbers go the other way).  Nothing was changed except the CXF version
> update for this test.  For now we are staying on 2.2.11 just to avoid the
> speed hit.  Any idea why this might take so long?  The test includes
> creating the client and calling 1 test method (that does nothing on the
> server to ensure it's not our fault).
>
>
>
> Kevin
>
>
>
>