QueryIterator should use setImmediate for callbacks to avoid max call stack

QueryIterator’s _toArrayImplementation is a recursive/looping async implementation that calls its callback synchronously. This can lead to RangeError: Maximum call stack size exceeded for large result sets.

See more: http://stackoverflow.com/questions/20936486/node-js-maximum-call-stack-size-exceeded

The fix is to wrap the callback() in setImmediate or process.nextTick: http://stackoverflow.com/questions/15349733/setimmediate-vs-nexttick

[ctstone]

@mehdijazayeri3000, @moderakh, please see repro at https://github.com/ctstone/repro-documentdb-176

The key, I think, is the large maxItemCount, (5000) combined with large-ish documents. I added some logging to defaultQueryExecutionContext.fetchMore to see how many documents are coming down over the wire, and it fails when the number reaches around 4500 in one response. My average document size is around 2200 bytes.

In general, async methods should never be called recursively without using one of setImmediate, process.nextTick, or setTimeout. See the StackOverflow link in the OP for more details.

Thanks!

[mehdijazayeri3000]

@ctstone Thank you so much. We will fix this as soon as possible.