feat(bytesbuf): reduce unnecessary reference count jiggling

[sandersaares]

Summary

Eliminate unnecessary atomic reference count operations in BytesView by avoiding clones in PartialEq and BytesViewSlices.

Changes

• PartialEq<BytesView>: Iterate over spans_reversed using indices instead of cloning both views (which incremented atomic ref counts for every span).
• PartialEq<&[u8]>: Same index-based iteration pattern.
• BytesViewSlices: Store a reference and index instead of a cloned view, removing unsafe transmute calls in the process.
• Benchmarks: Add early-exit inequality and slices iteration benchmarks.

Benchmark results

• slices_iterate: ~32% faster
• ne_view_early_exit: ~7.5% faster (early-exit benefits most since clone overhead is avoided entirely)

(Low confidence, noisy - mostly just to sanity check)

Closes #279

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 100.0%. Comparing base (56d48cf) to head (e387866).
⚠️ Report is 1 commits behind head on main.

Additional details and impacted files

@@           Coverage Diff           @@
##             main     #283   +/-   ##
=======================================
  Coverage   100.0%   100.0%           
=======================================
  Files         141      141           
  Lines        8640     8664   +24     
=======================================
+ Hits         8640     8664   +24     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.