fix: validate truncate widths in utility entrypoints

[fallintoplace]

Summary

• Reject non-positive widths in the TruncateLiteral and TruncateLiteralMax utility entrypoints.
• Reject zero width in TruncateUTF8Max, which returns a Result and can report invalid input directly.
• Add regression coverage for zero and negative widths across numeric, decimal, string, and binary literal truncation.

Why

The truncate transform validates its width before creating a transform function, but the utility functions can also be called directly. Without a guard there, invalid widths can reach modulo operations or iterator arithmetic in the type-specific implementations.

Testing

• cmake --build build --target util_test
• ctest --test-dir build -R util_test --output-on-failure

[wgtmac]

Thank you, @fallintoplace!