Adding a case for END_ARRAY in JsonExt.

[malcolmsparks]

I'm using clj-json to parse very large amounts of JSON (>1Gb) so it's very useful to have access to Jackson via clj-json, and laziness is critical, so parsed-seq* is very useful (and more so if it were public!). However, at the end of a large lazy sequence within a large JSON structure, the END_ARRAY token causes a "Cannot parse" exception to be thrown from JsonExt. This amendment is unlikely to have compatibility issues unless users were handling this exception in their logic. I think the overall benefit of this change to those like me who need to parse very large datasets makes it worthwhile - so please consider merging this.