I've committed a fix that checks for null before doing the insert. This means that the unit tests now pass.
I've re-opened TUSCANY-2069 so the change can be validated. If the code really must insert null then we need to do something different - e.g. revert collection or add a "marker value" to represent null. Mark > -----Original Message----- > From: Ramkumar R [mailto:[EMAIL PROTECTED] > Sent: 17 April 2008 09:52 > To: [email protected] > Subject: Re: svn commit: r648763 - > /incubator/tuscany/java/sca/modules/databinding/src/main/java/org/apache/t > uscany/sca/databinding/impl/DirectedGraph.java > > Hi Simon, > I am also facing this issue, here Raymond have used ConcurrentHashMap to > resolve the issue as mentioned in TUSCANY-2069, but we can't be using > ConcurrentHashMap in this case as this collection (ConcurrentHashMap) does > not allow null value to be inserted in the Map. Since the > DirectedGraph.java > class explicitly tries to insert null value in its getShortestPath() > method > we see this exception. For now we can revert to the old code of having > HashMap to resolve this issue. > > -- > Thanks & Regards, > Ramkumar Ramalingam --------------------------------------------------------------------- To unsubscribe, e-mail: [EMAIL PROTECTED] For additional commands, e-mail: [EMAIL PROTECTED]
