QFJ-1283 Deduplicate checksum calculation

[kangdaeun1022]

Summary

• centralize the string-equivalent checksum loop in MessageUtils
• reuse the allocation-free CharSequence path from Message#setChecksum and FieldMap#calculateChecksum
• preserve byte-based checksum calculation for multibyte charsets
• add regression coverage for the StringBuilder fast path, UTF-8 encoded bytes, and repeating-group checksum consistency

Optimization rationale

The shared helper accepts CharSequence, so Message#setChecksum operates directly on its existing StringBuilder without creating a String or byte[]. String-equivalent charsets continue to use direct character traversal, while non-string-equivalent charsets continue to calculate the checksum from encoded bytes.

Testing

• ./mvnw -pl quickfixj-base -Dtest=FieldMapTest,MessageUtilsTest test -Dmaven.javadoc.skip=true -PskipBundlePlugin (16 tests)
• ./mvnw -pl quickfixj-base test -Dmaven.javadoc.skip=true -PskipBundlePlugin (170 tests)

Fixes #1283

[codecov]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 70.58%. Comparing base (d63552e) to head (29042a4).

Additional details and impacted files

@@             Coverage Diff              @@
##             master    #1309      +/-   ##
============================================
+ Coverage     70.38%   70.58%   +0.20%     
- Complexity     2247     2254       +7     
============================================
  Files           159      159              
  Lines          9065     9061       -4     
  Branches       1192     1189       -3     
============================================
+ Hits           6380     6396      +16     
+ Misses         2220     2197      -23     
- Partials        465      468       +3     

☔ View full report in Codecov by Harness.
📢 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.