Before the Flash transport, it wasn't even theoretically possible to support
gadget-gadget without a relay, and nobody requested the functionality
either, so no time was invested. It's also worth noting that gadget-gadget
by definition demands an addressing scheme, which is under the purview of
the containing context (ie IFRAME ID and name), so enforcing the security
requirement underlying gadget-gadget (container inability to snoop/spoof) is
tricky.

Re: the Flash transport, it's public:
http://svn.apache.org/repos/asf/shindig/trunk/features/src/main/javascript/features/rpc//flash.transport.js
...and:
http://svn.apache.org/repos/asf/shindig/trunk/content/
...built from:
http://svn.apache.org/repos/asf/shindig/trunk/features/src/main/flex/Main.as

Not complete yet.

--j

On Fri, Jan 14, 2011 at 2:19 PM, Maxwell <[email protected]> wrote:

> why do not support gadget-gadget transparently in rpc, just like pubsub
> does?
>
> is that flash public, are you putting it on shindig or other place?
>
> Thanks
>
> On Mon, Jan 10, 2011 at 10:40 PM, John Hjelmstad <[email protected]> wrote:
>
> > Very late response, back from vacation. Gadget-to-gadget is typically
> only
> > supported by proxying messages through the container. The pubsub and
> > pubsub-2 features do this but other ad hoc solutions have been built as
> > well.
> >
> > Early work is being done to allow direct sibling addressing ie.
> > to="/mySiblingsId" though this only works on window.postMessage at the
> > moment.
> >
> > I'm implementing a transport that utilizes Flash, with which it *may* be
> > possible to provide similar functionality, but it's too early to say.
> >
> > -j
> >
> > On Wed, Dec 22, 2010 at 11:16 AM, Maxwell <[email protected]> wrote:
> >
> > > I see, and what do you usually use for gadget to gadget communication,
> > when
> > > they are not in the same container?Implements own service or use
> another
> > > feature from opensocial/shindig?
> > >
> > > Thanks
> > > Maxwell
> > >
> > > Sent from my iPhone
> > >
> > > On 22/12/2010, at 16:51, Henry Saputra <[email protected]>
> wrote:
> > >
> > > > inner communication meaning between gadget and container? yes
> > > >
> > > > - Henry
> > > >
> > > > On Wed, Dec 22, 2010 at 5:11 AM, Maxwell <[email protected]>
> wrote:
> > > >> I was looking "gadgets.rpc" spec, and looks like this feature also
> > > >> just supports "inner communication" is that right?
> > > >>
> > > >> On Tuesday, December 21, 2010, Maxwell <[email protected]> wrote:
> > > >>> So, what do you recommend me to make a "communication" in different
> > > containers? Could I use the MessageService?
> > > >>>
> > > >>> On Tue, Dec 21, 2010 at 5:15 PM, Maxwell <[email protected]>
> > wrote:
> > > >>> Nothing specific yet, I'm just writing some tests, and looking if
> we
> > > will be able to use shindig, one of the requirements, it's gadgets
> > talking
> > > in different containers, (or browsers).
> > > >>>
> > > >>> My test will be a simple chat (2 gadgets sending messages), it's
> easy
> > > with pub/sub, i already wrote one, but now we would like to try with
> > > different containers.
> > > >>>
> > > >>>
> > > >>> Maxwell
> > > >>>
> > > >>> On Tue, Dec 21, 2010 at 4:22 PM, John Hjelmstad <[email protected]>
> > > wrote:
> > > >>>
> > > >>> It doesn't -- the "pubsub" feature ferries messages between gadgets
> > by
> > > way
> > > >>> of the container itself, using in-browser communication. That's
> true
> > as
> > > well
> > > >>> of the "pubsub-2" feature as well. Communication btw different
> > gadgets
> > > >>> requires a server component, which has typically been built ad hoc.
> > > What's
> > > >>> your use case?
> > > >>>
> > > >>> --John
> > > >>>
> > > >>> On Tue, Dec 21, 2010 at 9:02 AM, Maxwell <[email protected]>
> > wrote:
> > > >>>
> > > >>>> Does pub/sub provides support to communication between gadgets
> that
> > > are not
> > > >>>> in the same container? Maybe another browser window.
> > > >>>>
> > > >>>> Thanks.
> > > >>>
> > > >>>
> > > >>>
> > > >>
> > > >
> > > >
> > > >
> > > > --
> > > > Thanks,
> > > > Henry
> > >
> >
>

Reply via email to