When executing an aggregate pipeline, the Java driver gives me a NullPointerException without further explanation.

While debugging I found that it's happening while trying to unbox AggregateIterableImpl.batchSize of type java.lang.Integer to be written to field FindIterableImpl.findOptions.batchSize of type int.

When I specify a batchSize() at the Aggregate other than null, e.g. explicitly setting it to "0" (the documented default value), everything is fine.

So, instead of just using

List<Document> result = collection.aggregate(pipeline).into(new LinkedList<Document>);

I need to write

List<Document> result = collection.aggregate(pipeline).batchSize(0).into(new LinkedList<Document>);

which is rather unusual and smells a bit bad .

I think it would help to change the type of the batchSize attribute of AggregateIterableImpl from java.lang.Integer to int, but I cannot tell whether it has any side-effects...