After an upgrade from version 2.0-beta4 to beta7 the NDCPatternConverter writes an object-ID instead of the content of the NDC-stack.
We are using an pattern with "[%0.50x]". In beta4 the resulting output looks like "[cbi@CE03178]" which means username and machine. Now in beta7 it looks like "[logging.log4j.spi.MutableThreadContextStack@875ef7]".
I analysed the issue in NDCPatternConverter.format(...) method, where event.getContextStack() is called and the result is passed to StringBuilder.append(...), which means, that the toString()-method will be invoked.
In beta4 getContextStack() returns an instance of ImmutableStack. This class inherits its toString() method from AbstractList, where the elements of the collection will be formatted human-readable.
Now in beta7 there comes an instance of MutableThreadContextStack which isn't derived from AbstractList but implements the Collection-Interface. The toString() method comes from Object and returns the name of the class and an object-ID instead of the context of the unterlying stack/collection.
In my opinion you just need to copy or derive the toString() method from AbstractList to solve this issue. Thank you in advance!