Apparently we all saw this at the same time.. weird
Paul
On 8/24/07, Mondain <[EMAIL PROTECTED]> wrote:
>
> Adam,
> I changed your fix just a little and added it to trunk; my quick
> smoke-test seemed to work fine. Thanks for the patch!
>
> public void ping() {
> long newPingTime = System.currentTimeMillis();
> if (lastPingSent == 0) {
> lastPongReceived = newPingTime;
> }
> Ping pingRequest = new Ping();
> pingRequest.setValue1((short) Ping.PING_CLIENT);
> lastPingSent = newPingTime;
> int now = (int) (lastPingSent & 0xffffffff);
> pingRequest.setValue2(now);
> pingRequest.setValue3(Ping.UNDEFINED);
> ping(pingRequest);
> }
>
> Paul
>
> On 8/24/07, Adam <[EMAIL PROTECTED]> wrote:
> >
> > So I found my app was taking a long time to index these
> > files, which meant that it was timing out. No problem,
> > thinks I, I'll just increase the timeout value in the
> > red5.properties file.
> >
> > But it was to no avail! It coninued to timout, even when
> > the rtmp.max_inactivity was set well high.
> >
> > Further investgation shows that RTMPConnection.java
> > compares (lastPingSent - lastPongReceived > maxInactivity)
> > to see if there's a timeout. Which is fine if it's
> > ever recieved a ping reply, but if not even the first
> > ping is replied to then lastPongReceived is still zero,
> > so the comparison drops through to the code to close
> > the inactive connection whatever the value of max_inactivity.
> >
> > This can be fixed with one line of code inserted into the
> > ping() function to set lastPongReceived as well as lastPingSent
> > if it's setting the value for the first time.
> >
> > One line does the trick:
> > if(lastPingSent==0){lastPongReceived = System.currentTimeMillis();}
> >
> >
> >
> > IE the pong function was:
> > --------------
> > public void ping() {
> > Ping pingRequest = new Ping();
> > pingRequest.setValue1((short) Ping.PING_CLIENT);
> > lastPingSent = System.currentTimeMillis();
> > int now = (int) (lastPingSent & 0xffffffff);
> > pingRequest.setValue2(now);
> > pingRequest.setValue3(Ping.UNDEFINED);
> > ping(pingRequest);
> > }
> > --------------
> >
> > But a working version would read:
> >
> > --------------
> > public void ping() {
> > if(lastPingSent==0){ lastPongReceived =
> > System.currentTimeMillis(); }
> > Ping pingRequest = new Ping();
> > pingRequest.setValue1((short) Ping.PING_CLIENT);
> > lastPingSent = System.currentTimeMillis();
> > int now = (int) (lastPingSent & 0xffffffff);
> > pingRequest.setValue2(now);
> > pingRequest.setValue3(Ping.UNDEFINED);
> > ping(pingRequest);
> > }
> > --------------
> >
> >
> > Can someone drop that into the main trunk? Or let me know
> > where I should post the fix if this maillist isn't good enough?
> >
> > Thanks.
> >
> > Adam..........
> >
> > _______________________________________________
> > Red5 mailing list
> > [email protected]
> > http://osflash.org/mailman/listinfo/red5_osflash.org
> >
>
>
>
> --
> It is difficult to free fools from the chains they revere. - Voltaire
--
It is difficult to free fools from the chains they revere. - Voltaire
_______________________________________________
Red5 mailing list
[email protected]
http://osflash.org/mailman/listinfo/red5_osflash.org
