[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: [xmlblaster] Socket protocol and plugin info
Hi Oliver,
thanks for the suggestion.
I have changed the code to:
// String tmp =
pluginInfo.getParameters().getProperty("useUdpForOneway",
""+this.useUdpForOneway);
// this.useUdpForOneway = Boolean.valueOf(tmp).booleanValue();
this.useUdpForOneway = this.glob.get("useUdpForOneWay",
this.useUdpForOneway, null, pluginInfo);
which should work. I would not commit the change however, I will test it
on the weekend and commit it only after successful tests.
Regards
Michele
Fels, Oliver wrote:
> Hi folks.
>
> While around with the socket protocol, I have come across the
> SocketConnection.init(Global, PluginInfo) method which is laid out as
> follows:
>
> public void init(org.xmlBlaster.util.Global glob, PluginInfo
> pluginInfo) throws XmlBlasterException
> {
> this.glob = (glob == null) ? Global.instance() : glob;
>
> this.pluginInfo = pluginInfo;
> String tmp =
> pluginInfo.getParameters().getProperty("useUdpForOneway",
> ""+this.useUdpForOneway);
> this.useUdpForOneway = Boolean.valueOf(tmp).booleanValue();
> if (log.isLoggable(Level.FINER)) log.finer("Entering
> init(useUdpForOneway="+this.useUdpForOneway+")");
> // Put this instance in the NameService, will be looked up by
> SocketCallbackImpl
>
> this.glob.addObjectEntry("org.xmlBlaster.client.protocol.socket.SocketCo
> nnection", this);
> }
>
> The difficult part is if not relying on a XmlBlasterAccess object but
> directly working with the SocketConnection class.
> In this case there is no PluginInfo object which leads to a
> NullPointerException when fetching the useUdpForOneWay parameter.
> I believe someone has already thought about this scenario because
> getVersion handles a non-existent PluginInfo:
>
> public String getVersion()
> {
> return (this.pluginInfo == null) ? "1.0" :
> this.pluginInfo.getVersion();
> }
>
> So that looks somehow like a bug. One solution could be to move the
> useUdpForOneway property into Global, the other to catch this faulty
> condition by testing the PluginInfo object for null reference.
>
> Any comments ?
>
> Best regards,
>
> Oliver
>