Function preprocess 

pub fn preprocess(
    input: &[u8],
    limits: &Limits,
) -> HedlResult<PreprocessedInput>

Preprocess raw input bytes into lines.

This handles:

• UTF-8 validation
• BOM skipping
• CRLF normalization
• Bare CR rejection
• Control character validation (SIMD-optimized)
• Size and line length limits
• Line boundary detection (SIMD-accelerated with memchr)

Performance

Uses SIMD-accelerated newline scanning via memchr for 4-20x faster preprocessing on large files (> 1 MB).