Can't lazily parse large JSON objects

[conan]

Cheshire has support for lazily parsing JSON files containing large numbers of objects, but not for parsing large objects themselves.

For example:

{"stars" : [
    {"name" : "Betelgeuse", "magnitude" : 0.58},
    {"name" : "Rigel", "magnitude" : 0.12},
    ... 300 billion others ...
    {"name" : "Antares", "magnitude" : 0.92}]}

Trying to use Cheshire to lazily get the stars in the Milky Way one by one from this JSON doesn't work - if I try to parse the "stars" object using core/parsed-seq, it calls directly into parse/parse*, which bypasses the lazy array functionality in parse/parse, and attempts to evaluate the entire galaxy. Even using parse/parse directly doesn't seem to work (but I'm not sure why).

[carylee]

I'd love to be able to do this as well. https://github.com/dominictarr/JSONStream does this really nicely.

[jgalrito]

Any chance of this happening?