That looks fine. Thanks,
Ed.

On 23/05/06, Pete Robbins <[EMAIL PROTECTED]> wrote:

Ed, please take a look at the code I checked in for Jira TUSCANY-401.

There is a TuscanyRuntime class which allows you to specify the default
module and also the path to the system root. The constructor takes these
in
the <order root, default_module> but I think it is probably better to have
these the other way round as the most common thing will be to specify the
module and not the root. This would allow you to code:

TuscanyRuntime runtime("myModule");

instead of

TuscanyRuntime runtime("", "myModule"); // root set to "" will use env var

What do you think?


On 18/05/06, Pete Robbins <[EMAIL PROTECTED]> wrote:
>
>  Java has a TuscanyRuntime to specify the default module so I will do
the
> same. Something like:
> TuscanyRuntime::start(<default_moduleComponent_name>);
>
> You will still need to set the TUSCANY_SCACPP_SYSTEM_ROOT to the root
> deployment folder for the system.
>
> Cheers,
>
>
>
> On 17/05/06, Pete Robbins <[EMAIL PROTECTED]> wrote:
> >
> > That's a good point. A call to initialize the runtime passing the
system
> > root and default module is a possibility. I'll look into it.
> >
> >
> > On 17/05/06, Edward Slattery <[EMAIL PROTECTED] > wrote:
> > >
> > > I am using the tuscany c++ SCA, and am finding it quite inconvenient
> > > that
> > > the root and default module are set by environment variables. As I
use
> > > a
> > > development environment which reads the env at startup, I have to
> > > alter
> > > these variables , close my studio, re-open every time I want to
change
> > > tests.
> > >
> > > I would really appreciate it if I could have a call to the runtime
> > > which
> > > overrides the environment settings, and maybe even another call
which
> > > re-instates the use of env vars. Then I could hop about between
tests,
> > > and
> > > even write a generic test client which takes the names of the root
and
> > >
> > > module as parameters.
> > >
> > > Any chance?
> > >
> > >
> >
> >
> > --
> >
> > Pete
> >
>
>
>
> --
>
> Pete
>



--
Pete


Reply via email to