Fix substring handling for non-BMP Unicode characters - #462

Fixes Unicode substring handling for non-BMP characters in $substring, $substringAfter, and $substringBefore.

Closes #461.

Changes

• Updated JSONataUtils.substr() to consistently use Unicode code-point indexing.
• Correctly convert code-point positions into UTF-16 offsets.
• Clamp substring lengths to the remaining code points.
• Removed duplicated substring helper implementations.
• Added regression tests for supplementary Unicode characters and boundary conditions.

Testing

• Focused substring tests: 134 passed.
• Full Maven test suite: 1,223 passed, 0 failures, 0 errors, 93 skipped.
• git diff --check: clean.