Use object ids instead of hashes for Jekyll documents

[qwtel]

Proposed solution for #16.

This change drastically improves performance when passing posts, documents or pages, e.g.

{% for post of site.posts %}
  {% included_cached post.html post=post %}
{% endfor %}

The problem with doing the above is the long time it takes to calculate the .hash of a large Jekyll document, such as a post.

However, it should be safe to assume that documents don't change during a page build, so using the object_id instead of the hash for comparison results in a large performance boost.

The implementation only applies to known Jekyll types, so that other complex data still get the old treatment, e.g.

---
data:
  a: 3
  b: 4
---
{% include_cached partial.html data=page.data %}

still works, even if data is passed in multiple places with different object ids.

[welcome]

Welcome! Congrats on your first pull request to Jekyll Include Cache. If you haven't already, please be sure to check out the contributing guidelines.

[benbalter]

Nice! Thanks for this.

[welcome]

Congrats on getting your first pull request to Jekyll Include Cache merged! Without amazing humans like you submitting pull requests, we couldn’t run this project. You rock! 🎉

If you're interested in tackling another bug or feature, take a look at the open issues, especially those labeled help wanted.