RMI callbacks over the internet a fallacy?

[lkeller]

From all my searching on the java.sun.com site, I have
been unable to find one example of anybody
who has successfully implemented a RMI callback over
the internet with a stock web browser.

From the search of this forum, I can see that I'm not
the only one having a problem implementing
RMI callbacks over the internet. The part that is scaring
me is that I haven't found anybody that has
said "Hay, it works!". What I hope to accomplish with
this thread is the fact that it either "works", "it
works - but..." or "Don't do RMI callbacks using a browser
clients". Hopefully, it will show up in a
search by other people with the same problems. It would
be nice if some of this makes it into the
FAQ.

Of the people that have posted previously, they all seem
to be having what appears to be the same
problem I'm having, a security exception when the client
tries to register the callback object. Some
people have suggested and tried to implement a security
manager on the client side. However, from
java/docs/api/java/rmi/RMISecurityManager.html , it states
that "RMISecurityManager does not apply
to applets, which run under the protection of their browser
security manager.", which makes sense.
Other people are always quick to point out that you need
a security policy file for the registry. This
file seems to benefit only locally executed appletviewers
and local RMI objects possessed by remote
clients. (someone please confirm this for me) With a
broad policy file, if your client is able to connect
without an exception, your policy file is in place and
working.

So, what could be causing this exception?

I have created a simple RMI callback test page and have
placed all the source code, compilation
scripts and command lines on it. I have striped anything
that could cause any ambiguity and
implemented a functionality that simply takes a number
and increments it. This is a classic case
described by other people where it works using a local
appletviewer but not over the web. I will try to
keep the server side running but for the moment, you
can assume it throws no exceptions when is is
run or whenever you make a connection.

http://24.1.12.134:5829/java/rmi/ or http://C515671-A.lakwod2.co.home.com:5829/java/rmi/
(the fully
qualified name)

Anybody?

Thank you in advance for your time.

[syed]

Hi,
Ur applet has to be signed by netscape signed tool. U should get a personal
certificate or create ur own certificate from the signtool to do that. visit


www.versigin.com
www.thewate.com
www.keywitness.com
For more about signing, visit

developer.netscape.com

Cheers
syed zulfiqar
query@edmail.com


"lkeller" <lkeller@hotmail.com> wrote:
>
>
>
> From all my searching on the java.sun.com site, I have
>been unable to find one example of anybody
> who has successfully implemented a RMI callback over
>the internet with a stock web browser.
>
> From the search of this forum, I can see that I'm not
>the only one having a problem implementing
> RMI callbacks over the internet. The part that is scaring
>me is that I haven't found anybody that has
> said "Hay, it works!". What I hope to accomplish with
>this thread is the fact that it either "works", "it
> works - but..." or "Don't do RMI callbacks using a browser
>clients". Hopefully, it will show up in a
> search by other people with the same problems. It would
>be nice if some of this makes it into the
> FAQ.
>
> Of the people that have posted previously, they all
seem
>to be having what appears to be the same
> problem I'm having, a security exception when the client
>tries to register the callback object. Some
> people have suggested and tried to implement a security
>manager on the client side. However, from
> java/docs/api/java/rmi/RMISecurityManager.html , it
states
>that "RMISecurityManager does not apply
> to applets, which run under the protection of their
browser
>security manager.", which makes sense.
> Other people are always quick to point out that you
need
>a security policy file for the registry. This
> file seems to benefit only locally executed appletviewers
>and local RMI objects possessed by remote
> clients. (someone please confirm this for me) With a
>broad policy file, if your client is able to connect
> without an exception, your policy file is in place and
>working.
>
> So, what could be causing this exception?
>
> I have created a simple RMI callback test page and have
>placed all the source code, compilation
> scripts and command lines on it. I have striped anything
>that could cause any ambiguity and
> implemented a functionality that simply takes a number
>and increments it. This is a classic case
> described by other people where it works using a local
>appletviewer but not over the web. I will try to
> keep the server side running but for the moment, you
>can assume it throws no exceptions when is is
> run or whenever you make a connection.
>
> http://24.1.12.134:5829/java/rmi/ or http://C515671-A.lakwod2.co.home.com:5829/java/rmi/
>(the fully
> qualified name)
>
> Anybody?
>
> Thank you in advance for your time.