You make a good point, there are other ways of doing this, but
performance in communication is the most important reason I started
using a Socket.  I was originally using an HTTPService to deal with
two-way communication but the performance was not nearly as good as a
direct socket.  Don't get me wrong, it wasn't really noticeable running
the application straight, but compared to a Socket connection in a
side-by-side comparison the performance was SO much better on a Socket.

I've got a very nice system developed if I can just get over this hurdle
of Flash denying me access back to the socket I originally got the SWF
from.  Like I said, it works beautifully if I run on port 8080, but when
I fall below 1024 it throws the security error.

Granted what I'm trying to do may be impossible (it shouldn't be
though), but I would say this is the "best practice" if I can accomplish
it.  I just have a hard time believing that it won't let me do something
as simple and basic as this.

Thanks again

--- In [email protected], "Johannes Nel" <[EMAIL PROTECTED]>
wrote:
>
> as i said its a guess. red5 implemented a mina server to listen to the
rtmp
> (and rtmpt) protocol, trying that with a custom socket instead of a
> netconnection might fail, but then again you could send data via a
> netconnection as well (remote shared objects do this). actually a good
hard
> look at red5 would come in handy as they have serializers for amf and
amf3
> as well, which is a very efficient way to send data from and too
flash.
> a suggestion i have heard (this was mike nimer's idea) was to write a
flv
> which streamed only metadata, then using function calls back to red5
to do
> what you need to do the other way (or even a remote shared object).
> so many ways to do this apart from a plain socket to be honest
>
> On 18 Apr 2007 06:43:56 -0700, Matt [EMAIL PROTECTED] wrote:
> >
> >   I've never used RTMP, so I'm looking into that now. Essentially my
> > goal here is to have a two-way communication between the flash
client
> > and my web server via port 80. I don't particularly care how I
> > accomplish that so long as it works. ;)
> >
> > Thanks for the tip and if you have any examples using RTMP I'd love
to
> > see them.
> >
> > --- In [email protected] <flexcoders%40yahoogroups.com>,
> > "Johannes Nel" johannes.nel@
> > wrote:
> > >
> > > try specifying a random protocol name and get your servlet to
interpret
> > > that. its just a guess but it mighyt work as rtmpt and rtmp can be
> > used over
> > > port 80
> > >
> > > On 18 Apr 2007 02:29:02 -0700, Matt matt@ wrote:
> > > >
> > > > I am trying to use ActionScript (in Flex) to communicate back to
the
> > > > HTTP server that the SWF was loaded from on port 80 using a
Socket. If
> > > > my server is running from port 8080 everything works fine, but
when
> > > > it's on port 80 I get a sandbox violation.
> > > >
> > > > I'm calling
> > > > Security.loadPolicy('
> > > >
> >
> >
http://localhost/jseamless-test/?application=basic&resource=crossdomain.\
xml
> > > > ')
> > > > (this hits a Servlet, but references the file...also, I've tried
> > > > placing the crossdomain.xml file in the root of the server as
well and
> > > > I get the exact same result) and I can verify it is actually
loading
> > > > the XML file from the server, but I still get the sandbox
violation.
> > > >
> > > > My cross-domain file looks like this:
> > > >
> > > > <?xml version="1.0"?>
> > > > <cross-domain-policy>
> > > > <allow-access-from domain="*" to-ports="*"/>
> > > > </cross-domain-policy>
> > > >
> > > > What am I missing here? I must be able to connect to port 80 for
this
> > > > application and can't seem to get around the sandbox no matter
what I
> > > > do. Help would be extremely appreciated.
> > > >
> > > > I posted on the Flex forums at Adobe, but I never got any
response and
> > > > heard you guys were much more knowledgeable, so I hoped you
could
> > help. :)
> > > >
> > > > Thanks
> > > >
> > > >
> > > >
> > >
> > >
> > >
> > > --
> > > j:pn
> > > http://www.lennel.org
> > >
> >
> >
> >
>
>
>
> --
> j:pn
> http://www.lennel.org
>


Reply via email to