Feat/add api guidelines fixed history

[TKorr]

Description

Aligns FixedHistory, SlotArena, and FrequencyBuckets with the
Rust API Guidelines —
specifically C-ITER-TY (iterator types are named structs, not impl Iterator)
and C-COMMON-TRAITS (standard trait impls).

FixedHistory<K> (src/ds/fixed_history.rs)

• Derives Copy — the backing [u64; K] array is always Copy
• Adds hand-rolled PartialEq, Eq, and Hash that compare logical MRU content
  rather than raw backing-array state (stale slots are invisible to equality/hashing)
• Adds iter() -> Iter<'_, K> for borrowed iteration in MRU order
• Implements IntoIterator for both FixedHistory<K> (owned → IntoIter<K>) and
  &FixedHistory<K> (borrowed → Iter<'_, K>)
• Both iterator types implement ExactSizeIterator

SlotArena<T> (src/ds/slot_arena.rs)

• iter() now returns the named Iter<'_, T> struct instead of impl Iterator,
  making the type nameable by downstream code

FrequencyBuckets<K> / FrequencyBucketsHandle<H> (src/ds/frequency_buckets.rs)

• iter_entries() now returns the named EntryIter<'_, K> struct
• Derives Debug on FrequencyBucketIdIter and FrequencyBucketEntryIter
• Adds FrequencyBucketsHandle::into_inner() to unwrap the handle back to the
  inner FrequencyBuckets<H>
• Updates module-level doc tables to use intra-doc links throughout

Related Issue

Fixes #

Type of Change

• 🐛 Bug fix (non-breaking change that fixes an issue)
• ✨ New feature (non-breaking change that adds functionality)
• 💥 Breaking change (fix or feature that would cause existing functionality to change)
• 📝 Documentation update
• 🔧 Refactoring (no functional changes)
• ⚡ Performance improvement
• 🧪 Test addition or modification

Breaking: SlotArena::iter(), FrequencyBuckets::iter_entries(), and
FrequencyBucketsHandle::iter_entries() now return named concrete types instead
of impl Iterator. Code that stored the iterator in a typed binding or threaded
the type through a generic will need to update to the new names (Iter,
EntryIter). Callers that simply iterate in a for loop or call .collect()
are unaffected.

How Has This Been Tested?

• Unit tests
• Integration tests
• Manual testing

New tests added in fixed_history.rs:

Test	What it covers
iter_yields_mru_order	Basic MRU ordering
iter_on_empty / iter_on_zero_capacity	Edge cases
iter_after_wrap	Correct ordering after ring wraps
iter_partially_filled	Sub-capacity buffer
iter_count_matches_len	Iterator length invariant
iter_exact_size	ExactSizeIterator contract
iter_matches_to_vec_mru	Parity with existing to_vec_mru
ref_into_iter_for_loop	&FixedHistory in for loop without consuming
owned_into_iter_for_loop	Owned IntoIterator
partial_eq_*	Logical equality across cursor positions
hash_equal_histories_same_hash	Same logical content → same hash
hash_usable_in_hashmap	Works as HashMap key
copy_produces_independent_value	Copy doesn't alias
copy_can_be_passed_by_value	Can pass by value multiple times

Test environment:

• OS: macOS 15
• Rust version: stable

Checklist

• My code follows the project's coding standards
• I have run cargo fmt and cargo clippy
• I have added tests for my changes
• All new and existing tests pass (cargo test)
• I have updated the documentation as needed
• I have added an entry to CHANGELOG.md (if applicable)

Additional Notes

The PartialEq/Hash implementations on FixedHistory deliberately iterate
through kth_most_recent rather than comparing the raw data array — two
histories with identical logical content but different cursor positions (e.g. one
that has wrapped) correctly compare equal and produce the same hash. This is
verified by hash_equal_histories_same_hash.