Since we now have our own class for this, can't we make the test vm independent?

        <attribute name="namingFactoryInitial">org.apache.xbean.naming.global.GlobalContextManager</attribute>

from rmi-naming.

thanks
david jencks

On Nov 1, 2006, at 9:49 PM, Vamsavardhana Reddy wrote:

There is a private static final String NAMING_FACTORY_INITIAL = "com.sun.jndi.rmi.registry.RegistryContextFactory" in the class.  should it be

Should it be Class.forName(NAMING_FACTORY_INITIAL) ?

--vamsi

On 11/2/06, Jacek Laskowski <jacek@laskowski.net.pl> wrote:
On 11/1/06, rickmcguire@apache.org <rickmcguire@apache.org> wrote:
> Author: rickmcguire
> Date: Wed Nov  1 09:48:44 2006
> New Revision: 470013
>
> URL: http://svn.apache.org/viewvc?view=rev&rev=470013
> Log:
> GERONIMO-1840 NamingPropertiesTest is not compatible with non-Sun VMs.
...
> +        try {
> +            // the above assumes we're running on a Sun JVM.  If we can't load the class first,
> +            // we'll skip the attempt at creating the InitialContext.  We've already verified that the
> +            // system properties have been set to the correct values, so this last bit is largely a formality.
> +            Class.forName("NAME_FACTORY_INITIAL");
> +            new InitialContext();
> +        } catch (ClassNotFoundException e) {
> +        }

It's not the first time I ask for some help understanding a change and
this one is not an exception.

Why do we look up the "NAME_FACTORY_INITIAL" class since it's deemed
to fail every time?

Jacek

--
Jacek Laskowski
http://www.jaceklaskowski.pl