Desktop app. coms. via sockets - test
Andrew Thompson
launching JWS* based apps. from web pages.
This example tests for the problems that would be
encountered if using sockets (from the JWS app.
back to the applet) for that reporting.
<http://www.physci.org/test/kk/applet.html>
* The technique should also work for non-JWS apps.
launched from a web page, but JWS apps. have a
security sandbox - they are both a more interesting
challenge, and more relevant to my needs.
More test results would be appreciated - especially
for machines with a firewall (I tried to enable the
Firewall on my box - and got a Win. message
saying 'no can do' - I'll try again later..)
The sort of detail I'd like, is expressed in the results table
<http://www.physci.org/test/kk/applet.html#result>
All results/comments welcome.
--
Andrew Thompson
http://www.athompson.info/andrew/
Message posted via JavaKB.com
http://www.javakb.com/Uwe/Forums.aspx/java-general/200705/1
Andrew Thompson
Andrew Thompson
Note: this repost* is X-posted to
comp.lang.java.programmer and comp.lang.java.help
but *follow-ups* (replies) have been directed
to c.l.j.programmer only. If you reply, please
follow it on c.l.j.p.
> I want a better (more information feedback) way of
> launching JWS* based apps. from web pages.
>
> This example tests for the problems that would be
> encountered if using sockets(from the JWSapp.
> back to the applet) for that reporting.
> <http://www.physci.org/test/kk/applet.html>
>
> * The technique should also work for non-JWS apps.
> launched from a web page, but JWS apps. have a
> security sandbox - they are both a more interesting
> challenge, and more relevant to my needs.
>
> More test results would be appreciated - especially
> for machines with a firewall (I tried to enable the
> Firewall on my box - and got a Win. message
> saying 'no can do' - I'll try again later..)
>
> The sort of detail I'd like, is expressed in the results table
> <http://www.physci.org/test/kk/applet.html#result>
>
> All results/comments welcome.
* ..and yes, I feel entirely justified in
reposting the original message completely
untrimmed. I have waited around 2 weeks to
get any replies, and the original may have
dropped off some people's news clients.
I am willing to supply the code for this.
It is not secret, I just thought that being
sandboxed and based closely on Sun's KnockKnock
server/client example, it was not especially
interesting.
So how about it? Detailed, fragmentary, or
..*any* other test results would be appreciated.
Andrew Thompson
Tom Hawtin
> On May 27, 9:11 am, "Andrew Thompson" <u32984@uwe> wrote:
>> I want a better (more information feedback) way of
>> launching JWS* based apps. from web pages.
> I am willing to supply the code for this.
> It is not secret, I just thought that being
> sandboxed and based closely on Sun's KnockKnock
> server/client example, it was not especially
> interesting.
>
> So how about it? Detailed, fragmentary, or
> ..*any* other test results would be appreciated.
On my laptop running ye olde 32-bit software (Windows Vista Home
Premium, Opera and JDK1.6), it does the same as FireFox up as far as the
security dialog. Obviously I deny the random security request from an
unknown party.
A token in the URL would work better, but wouldn't give a constant URL
for the application. I don't think cookies get shared between applets
and WebStart applications unfortunately.
But show us your source anyway.
Tom Hawtin
Andrew Thompson
...
> On my laptop running ye olde 32-bit software (Windows Vista Home
> Premium, Opera and JDK1.6), it does the same as FireFox up as far as the
> security dialog.
Do you have a firewall enabled on that machine?
...
> A token in the URL would work better, but wouldn't give a constant URL
> for the application. I don't think cookies get shared between applets
> andWebStartapplications unfortunately.
I considered using URL parameters in various forms.
The problem is that I cannot guarantee that the original
page with the applet, and any page opened by the
BasicService.showDocument() (in which I would put the
URL parameters to indicate 'success'), would open in
the same browser, let alone the same VM (and thereby
be in the 'same AppletContext as'/'be easily accessible
to' the first applet).
So it then comes back to how two untrusted applets,
potentially in different VM's, might interact.
AFAIU, there is no way.
> But show us your source anyway.
Umm.. OK, sure..
<http://www.physci.org/test/kk/source.zip>
(About 8Kb, made with the Jar tool. Includes
the java source, JNLP launch files and the
applet page)
But note.. I am *not* asking for a code critique
on this *code*. It is horrid - hacked together,
rather than designed. It is the security environment
that controls it, that interests me.
Andrew (the 'unknown party';) T.