axis-c-user mailing list archives

Site index · List index
Message view « Date » · « Thread »
Top « Date » · « Thread »
From Olivier Mengué <omen...@oxymel.com>
Subject RE : Need suggestions on Axis/c client stub
Date Wed, 10 Feb 2010 10:41:51 GMT
Hi

I'm a recent Axis2/C user. I don't know about Axis-C/C++.

Ashok P wrote:

>3) As per the  documentation, axis-c supports "ADB" binding for client code
>generations and "none" option for raw xml in/out. Is ADB is efficient to
>consume non-axis web services?  

I don't know what is "ADB binding".
With Axis2/C I had to do manual XML tree building. The Axis2 codegen tool was not able to
generate the XML tree building for my WSDL.

>5) Does it mandatory to use axis2.xml? I like all these settings through
>code by settings the options - like settings the protocol "http" or https
>(ssl), port , ssl certificate, client timeout and more. Does such provision
>is available in both C/C++ client generated code?

I'm interested by an answer to this question too.

>6)  From the other posting on forum, it's not clear that "AXIS2C_HOME"
>deplorer folder is mandatory or not for client side. Can anybody confirm
>this?

It seems to be mandatory to use AXIS2C_HOME as it is the way to find the axis2.xml.

axis2.xml is AFAIK mantadory for client side.

> I agree few Dlls are required for client distribution. I like my
>client application (along with client stub) should have small footprint for
>distribution, so can I build stub statically for depended libraries like
>axiom.dll, axis2_http_sender.dll etc?

It looks like the transport layers (http_sender, http_receiver...) are loaded dynamically
by Axis2/C.
And there is no option in the build system to make a build with statically linked transports.

>7) Axis-C++ is not recommended for production use. Does I need to test
>client generated (stub) for stability considering the recommendation? Any
>suggestions on axis-c++ clients to be used on production? Otherwise I need
>go with axis-c only.

The recommandation is to use Axis2/C (which is not axis-c).
I understood it as "Axis-C++ is buggy and is broken by design, we choose to rewrite everything
as Axis2/C". Is it really the case? I don't know. That is just the feeling I got by reading
the site.

Olivier Mengué

Mime
View raw message