Broken KIO connections

[Dave Feustel]

I also have for quite a while suspected the presence of an exploitable hack in
kio code, although I have absolutely no proof, just the gut feeling.

-----Original Message-----
>From: Harri Porten <porten at kde.org>
>Sent: Apr 14, 2007 10:24 AM
>To: kde-core-devel at kde.org
>Subject: Broken KIO connections
>
>Hi!
>
>Am I the only one suffering from KIO networking problems that go along 
>with zillions
>
>   QAbstractSocket::waitForBytesWritten() is not allowed in UnconnectedState
>
>warnings? As at least the HTTP slave is affected I can hardly use KHTML 
>right now.
>
>I attached gdb to the HTTP slave and found that the following code in 
>HTTPProtocol::httpOpen()
>
>   if ( !isConnectionValid() )
>   {
>     if (!httpOpenConnection())
>     {
>        kDebug(7113) << "Couldn't connect, oopsie!" << endl;
>        return false;
>     }
>   }
>
>does not lead to a valid connection. Either httpOpenConnect() or 
>isConnectionValid() must be broken since the move to QTcpSocket as 
>isConnection() still returns false although httpOpenConnection() returned 
>true.
>
>This silenced the warning for this one
>
>   if (!isConnectionValid()) {
>        kDebug(7113) << "Thought we are connected be we are not!" << endl;
>        return false;
>   }
>
>but other places/slaves seem to suffer from the same problem.
>
>Harri.


http://RepublicBroadcasting.org - Because You CAN Handle The Truth!
http://iceagenow.com - Because Global Warming Is Yet Another NWO Scam!