axis-c-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From cara <>
Subject Re: How best to transform a XML file into a message.
Date Thu, 02 Apr 2009 15:37:01 GMT

IMHO, it isn't worth doing web services unless you use the autogenerated
code. I learned from experience to stop hand jamming xml. If your WSDL
changes, you want to re-autogenerate the plumbing code on both your client
and server side ... and with any luck, any code changes you have are
isolated in the code you had to write in order to use the objects generated
on your client side.

But, I did find your question hard to answer. I've been successful hand
jamming xml (and doing xslt) and wrapping it in a SOAP env for old fashioned
rpc wsdls. But with doc-literal, you should try to use the objects generated
by axis.

did this help?

On Thu, Apr 2, 2009 at 4:44 AM, Sam Carleton <>wrote:

> Folks,
> I know this is a free forum thus folks are free to ignore questions.  On
> the question below, the big question I have comes from my ignorance of the
> plumbing of Web Services.  Any feedback, even feedback such as, "Sam go
> learn more about the SOAP envelope to find out your own confort level, fore
> this is something only you can answer." would be much appreciated.
> Sam
> P.S.  I am starting to talk to WSO2 about some paid support, I am hoping to
> get some feedback on this sooner rather then later, though ;)
> On Wed, Apr 1, 2009 at 7:34 AM, Sam Carleton <>wrote:
>> My first operation I need to implement with Axis2/C is sort of
>> interesting...  I have an XML file on the server that needs to be
>> transformed and returned to the client.  The question is how best to do this
>> transformation?
>> A: Do I use the code generated by WSDL2C and in the call to the operation
>> populate the response using all the generated code and allowing the
>> generated code to serialize it into the SOAP message?
>> B: Do I write some XSLT (which I do know) to do the transformation in the
>> services Invoke method?
>> I like option B better, but I have a feeling there is some namespace and
>> endpoint data that is magically done for me in option A.  Myself being new
>> to this whole thing, which would be faster for me to implement?  I am
>> thinking A, but would the learning experience of implementing B end up
>> making my life a lot easier down the road?  As I just posed, I have a lot
>> more to learn about the whole WS-* stuff and to figure out where and/or if
>> that will fit into my solution, would option B prevent me from using the
>> other WS-* or do those things kick in before/after the services invoke?
>> Sam

View raw message