fix: Taking slicing into account when writing BooleanBuffers as fast-encoding format

[Kontinuation]

Which issue does this PR close?

Closes #1520.

Rationale for this change

This is a problem I found when working on #1511, the null bits were not correctly written and caused test failures. This patch is an attempt to fix it.

This patch is only aiming for fixing correctness problems. As #1190 (comment) pointed out, the fast BatchWriter may write full data buffer for sliced Utf8 arrays, so there's still some performance implications when working with sliced arrays.

What changes are included in this PR?

Correctly take slicing indices and length into account when writing BooleanBuffers. This applies to null bits of all arrays, and the values of boolean arrays.

How are these changes tested?

Added a new round-trip test for sliced record batches.

[codecov-commenter]

Codecov Report

All modified and coverable lines are covered by tests ✅

Project coverage is 58.97%. Comparing base (f09f8af) to head (05a4b04).
Report is 74 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #1522      +/-   ##
============================================
+ Coverage     56.12%   58.97%   +2.84%     
- Complexity      976     1028      +52     
============================================
  Files           119      122       +3     
  Lines         11743    12268     +525     
  Branches       2251     2309      +58     
============================================
+ Hits           6591     7235     +644     
+ Misses         4012     3875     -137     
- Partials       1140     1158      +18     

☔ 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.