On 11/28/06, Andrew Borley <[EMAIL PROTECTED]> wrote:

On 11/28/06, Simon Laws <[EMAIL PROTECTED]> wrote:
> On 11/27/06, Jean-Sebastien Delfino <[EMAIL PROTECTED]> wrote:
> >
> > Simon Laws wrote:
> > > On 11/27/06, Simon Laws <[EMAIL PROTECTED]> wrote:
> > >>
> > >>
> > >>
> > >> On 11/27/06, Pete Robbins <[EMAIL PROTECTED]> wrote:
> > >> >
> > >> > On 27/11/06, Pete Robbins <[EMAIL PROTECTED]> wrote:
> > >> > >
> > >> > >
> > >> > >
> > >> > >  On 27/11/06, Simon Laws <[EMAIL PROTECTED] > wrote:
> > >> > > >
> > >> > > > I've just checked out SCA and SDO from SVN and I get and
error
> > >> in VC
> > >> > > > Express
> > >> > > > when building SCA...
> > >> > > >
> > >> > > > ------ Build started: Project: tuscany_sca_ws_dispatcher,
> > >> > Configuration:
> > >> > > > Debug Win32 ------
> > >> > > > Compiling...
> > >> > > > Axis2Dispatcher.cpp
> > >> > > > Axis2DispatcherModule.cpp
> > >> > > >
> > >> >
> > >>
> >
c:\simon\projects\eclipse3.2\cpp\sca\runtime\extensions\ws\service\axis2c\src\tuscany\sca\ws\axis2dispatchermodule.cpp(66)
> > >>
> > >> >
> > >> > > >
> > >> > > > : error C2440: '=' : cannot convert from 'axis2_status_t
> > >> (__stdcall
> > >> > > > *)(axis2_module_t *,const axis2_env_t *,axis2_conf_ctx_t
> > >> > > > *,axis2_module_desc_t *)' to 'axis2_status_t (__stdcall
> > >> > *)(axis2_module
> > >> > > > *,const axis2_env_t *)'
> > >> > > >        This conversion requires a reinterpret_cast, a C-style
> > cast
> > >> > or
> > >> > > > function-style cast
> > >> > > > Generating Code...
> > >> > > >
> > >> > > > Does this mean I have an out of date Axis install on my
machine?
> > >> > > >
> > >> > > > Simon
> > >> > > >
> > >> > > >
> > >> > > Possibly. It builds fine for me against Axis 0.94 and 0.95
> > >> > >
> > >> > > --
> > >> > > Pete
> > >> >
> > >> >
> > >> > actually it's 0.94 I'm using
> > >> > --
> > >> > Pete
> > >>
> > >>
> > >> Thanks Pete,
> > >>
> > >> I just upgraded to 0.95 and it built. Yet to test it.
> > >>
> > >> S
> > >>
> > >> So, I tried running the calculator sample with the web serivces
> > >> client and
> > > it crashes. The local client works fine. So it is likely an issue
with
> > my
> > > environment but has anyone else tried with Axis2 C 0.95?
> > >
> > > For me it crashes in the axis2c DLL called from the line:
> > >
> > >   stub = (axis2_stub_t *)
> > >        axis2_stub_create_with_endpoint_uri_and_client_home(env,
> > > endpoint_uri,
> > >            client_home);
> > >
> > > Which is in the axis2_Calculator_stub.cpp file.
> > >
> > > I'm fairly sure that my environment is set properly to use Axis2C
0.95
> > > (i.e.
> > > I've updated my path to point to the right version). I'll keep
looking
> > > but
> > > It would be useful to know if some has used it successfully.
> > >
> > > Simon
> > >
> > > Simon
> > >
> >
> > Hi Simon,
> >
> > I'm using Axis2C 0.94. Most of us are probably still using Axis2C 0.94
> > as it's what we've used for our M2 release.
> >
> > The error you're getting is not in the SCA runtime. This Web Services
> > client is not an SCA client and is using the Axis2C API directly. We
may
> > have to update this client to work with Axis2C 0.95... In the
meantime,
> > could you try to see if a real SCA client using an SCA reference with
a
> > WS binding works? There's such a client in the RubyCalculator sample
> > under sample.calculator.wsclient.
> >
> > Switching to Axis2C 0.95 soon is probably a good idea anyway. Axis2C
> > 0.94 is not on the Apache mirrors anymore and you have to dig to find
it
> > in the Apache backup archive at http://www.apache.org/dist/ws/axis2/c/
.
> >
> > --
> > Jean-Sebastien
> >
> >
> > ---------------------------------------------------------------------
> > To unsubscribe, e-mail: [EMAIL PROTECTED]
> > For additional commands, e-mail: [EMAIL PROTECTED]
> >
> > Hi Sebastien
>
> Thanks for the suggestion. I've been on a little journey getting the
Ruby
> extension to work but I can confirm that Axis C 0.95 does work when used
> between the ruby calculator ws client and the cpp calculator server. So
it
> sounds like you are correct and that the native C++ axis client is
suspect.
>
> This is what I found using the Ruby extension:
>
> 1/ Don't try and use Ruby provided by Cygwin. Ruby runs fine but because
I'm
> using a SVN checkout of Tuscany C++ I need to compile up the Ruby
extension
> and of course the Cygwin Ruby is all set to use the GCC tool chain. I
did
> chase this for quite a while but gave up eventually and installed
windows
> (MSVC) version of Ruby. Maybe we can just update the documentation to
> discuss this.

Yep, we should document this.

> 2/ The ruby extension install doc describes the install steps for
windows (
> cpp/sca/doc/RubyExtension.html#winbin). This needs a little updating as
>
>    1. Use the following command sequence to enable the extension:
>       - cd <tuscany_sca_install_dir>\extensions\ruby
>       - deploy.bat
>
> doesn't match the deploy.bat script which for windows expects two
> parameters.
>
> :usage
> echo Usage: deploy <sca-root> <build-output>
> :end
>
> - it's not clear what sca-root and build-output are intended to be
> - this echo statement needs to be updated to something like
>
> :usage
> echo "Usage: deploy <sca-root> <build-output>"
> :end
>
> otherwise the < > characters confuse the DOS shell.

I think you're looking at the wrong deploy.bat! We unfortunately have
2 for both the ruby and python extensions - one that is used by the
Ruby extension Windows build to put the compiled libraries, xsd, etc
in sca/deploy/extensions/ruby and is not included in a binary
distribution (found at [1]) and one that is to be used by users to
enable the ruby extension by simply renaming the library from
tuscany_sca_ruby.dll.disabled to tuscany_sca_ruby.dll (found at [2]).
It looks like you tried to use the first one when you should have been
using the second! Even more confusingly, deploy.bat[2] is copied to
the sca/deploy/extensions/ruby directory, so it's very easy to pick
the wrong path and get the wrong deploy.bat!

I think we need to rename one or both of these - perhaps change the
second one to EnableExtension.bat or similar?

> 3/ As I was building from source I ran with the SCA VCExpress project
which
> did the install for me (using the deploy.bat file with suitable
parameters I
> believe). However it did leave me with tuscany_sca_ruby.dll.disabled in
> cpp\sca\deploy\extensions\ruby\bin. As a ruby novice it took me a while
to
> work out that the "require" statement in the rubly script was referring
> directly to a DLL and the fact that it was marked as ".disable" was
stopping
> it working.

Yep, this is a symptom of the issue above! Running deploy.bat should
have taken that .disabled suffix away..

> Generally as a newcomer to ruby (but not to Tuscany C++ in my case) it
would
> be useful to have some text or preferably a picture describing the file
> layout that you expect to be in place for
> the Ruby extension to work properly. Currently it just suggests the
> environment variables that need setting but doesn't tell you what the
> results of the various steps should be. Maybe we can come up with some
> scheme for the documentation of the various extensions that tells you
that
> you should end up with.  Perhaps something like:
>
> deploy/
>   bin/
>   extensions/
>      ruby/
>         bin/
>            tuscany_sca_ruby.dll
> etc.

Agreed. What would be even nicer would be to have a decent
installer/configurer that just lets you choose what extensions you
want to have.

> Anyhow. If I have some time I'll help to address my moans but I need to
get
> on with some other things at the moment.
>
Cheers
Andy

[1]
https://svn.apache.org/repos/asf/incubator/tuscany/cpp/sca/runtime/extensions/ruby/deploy.bat
[2]
https://svn.apache.org/repos/asf/incubator/tuscany/cpp/sca/runtime/extensions/ruby/src/deploy.bat

---------------------------------------------------------------------
To unsubscribe, e-mail: [EMAIL PROTECTED]
For additional commands, e-mail: [EMAIL PROTECTED]

Thanks Guys. Sorry for being dim. First time through with the new
extension layout:-).

+1 for renaming the second deploy.bat to enable.bat.

S

Reply via email to