Added :preserve_key_order option to maintain original hash key order

[robkiessling]

This is in response to issue #98, which requests an option to preserve the original key order of compared hashes.

The existing implementation always outputs changes in a fixed order:

1. Deletions (-)
2. Updates (~)
3. Additions (+)

And within each group, keys are sorted alphabetically.

Unfortunately, this means that simply disabling sort_by here isn’t enough:

hashdiff/lib/hashdiff/compare_hashes.rb

Lines 23 to 25 in b7734e3

	added_keys = (obj2_keys - obj1_keys).sort_by(&:to_s)
	common_keys = (obj1_keys & obj2_keys).sort_by(&:to_s)
	deleted_keys = (obj1_keys - obj2_keys).sort_by(&:to_s)

because grouping by operation type still forces a reordering that doesn’t reflect the original hash structure.

My goal for this option is to have all deletions and updates listed in the order they appear in the first hash. Then, similar to how Ruby’s Hash#merge appends new keys, any keys found only in the second hash are appended in their original order.

To support this behavior, the logic for handling deletions, updates, and additions was unified into a handle_key lambda. This allows each key to be evaluated in sequence, rather than processing all deletions, then all updates, then all additions in separate passes. I chose a lambda (instead of a separate method) to avoid having to pass in the many options and lookups needed within handle_key.

[liufengyun]

Thank you @robkiessling , LGTM 🎉