Re: [mdr-users] NullPointerException with IBM JRE

Martin wrote:
> Thanks for the patch. I have to admit I do not understand (even after 
> your explanation) how the IBM JRE works (when it makes the entries 
> invalid and why your patch fixes that). Could you please point me to 
> some web page containing the description of the problem?

The only place I can think of is Sun's Javadoc for Map.Entry
itself, which says:

    public static interface Map.Entry

    A map entry (key-value pair).  The Map.entrySet method returns a
    collection-view of the map, whose elements are of this class.  The
    only way to obtain a reference to a map entry is from the iterator
    of this collection-view.  These Map.Entry objects are valid only
    for the duration of the iteration; more formally, the behavior of
    a map entry is undefined if the backing map has been modified after
    the entry was returned by the iterator, except through the iterator's
    own remove operation, or through the setValue operation on a map
    entry returned by the iterator.

In other words, the Map.Entry objects in an entry set are not
necessarily copies of the entry data, they're actually backed
by the entries in the map (note that Map.Entry.setValue()
actually updates its backing entry).

When one says:

    ArrayList result = new ArrayList(dirty.entrySet());

'result' is filled with Map.Entry objects but those objects
are still backed by the entries in the 'dirty' map.  If we
then do:

    dirty.clear();

we've just invalidated all those Entry objects according to
the rule above.  By changing it to:

    ArrayList result = new ArrayList(new HashMap(dirty).entrySet());

we copy the entries to another map which is *not* subsequently
modified, so the Entry objects will remain valid indefinitely.

BTW, I have another patch I'd like to submit re. XMI writing
and the XML encoding attribute.  Is this a good place for us
to correspond on patches or is somewhere else preferred?

Mike