[fix][ml] Fix issues in estimateEntryCountBySize

[lhotari]

Motivation

The changes introduced in PR #23931 caused a regression in the entry count estimation logic for branch-3.0. While PR #24055 attempted to address this regression, it did not fully resolve the issues, leaving Pulsar's branch-3.0 in a broken state. A NullPointerException could occur when using ReadOnlyManagedLedgerImpl instances, as the currentLedger is null in these cases. ReadOnlyManagedLedgerImpl is used in branch-3.0 by PulsarSQL. In master branch, it's currently unused, but since the code exists, it makes sense to fix this bug in the master branch. This PR also contains other improvements.

A problem was also detected in using ManagedLedgerImpl.getNextValidPosition method to find the next valid position. This method is relatively costly and takes the current LAC into account which doesn't make much sense just for estimation purposes.

Modifications

1. NPE protection: Added null checks for currentLedger to handle ReadOnlyManagedLedgerImpl instances properly
2. Code reorganization: Extracted the estimation logic into a separate EntryCountEstimator class to improve testability and code clarity
3. Parameter enhancements: Modified method signature to accept maxEntries and stop calculations when this limit is reached
4. Return type change: Changed return type from long to int to match the actual usage in the codebase
5. Documentation: Added clear comments about implementation details
6. Tests: Added comprehensive unit tests in EntryCountEstimatorTest that validate various scenarios.

Documentation

• doc
• doc-required
• doc-not-needed
• doc-complete

[codecov-commenter]

Codecov Report

Attention: Patch coverage is 96.15385% with 2 lines in your changes missing coverage. Please review.

Project coverage is 74.16%. Comparing base (bbc6224) to head (0b0c6a1).
Report is 1213 commits behind head on master.

Files with missing lines	Patch %	Lines
...e/bookkeeper/mledger/impl/EntryCountEstimator.java	96.00%	1 Missing and 1 partial ⚠️

Additional details and impacted files

@@             Coverage Diff              @@
##             master   #24089      +/-   ##
============================================
+ Coverage     73.57%   74.16%   +0.59%     
+ Complexity    32624    32070     -554     
============================================
  Files          1877     1864      -13     
  Lines        139502   144334    +4832     
  Branches      15299    16463    +1164     
============================================
+ Hits         102638   107052    +4414     
+ Misses        28908    28824      -84     
- Partials       7956     8458     +502     

Flag	Coverage Δ
inttests	26.63% <63.46%> (+2.04%)	⬆️
systests	23.12% <63.46%> (-1.21%)	⬇️
unittests	73.67% <96.15%> (+0.83%)	⬆️

Flags with carried forward coverage won't be shown. Click here to find out more.

Files with missing lines	Coverage Δ
...che/bookkeeper/mledger/impl/ManagedCursorImpl.java	78.94% <100.00%> (-0.36%)	⬇️
...keeper/mledger/impl/cache/RangeEntryCacheImpl.java	63.59% <ø> (+4.84%)	⬆️
...e/bookkeeper/mledger/impl/EntryCountEstimator.java	96.00% <96.00%> (ø)

... and 1064 files with indirect coverage changes

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.
• 📦 JS Bundle Analysis: Save yourself from yourself by tracking and limiting bundle sizes in JS merges.

[lhotari]

There seem to be other gaps in the estimation logic. I'll revisit this in this PR in further commits.

[lhotari]

follow up in #24125, please review