Optimize chunkOnModifiedUtf8ByteSize from O(n²) to O(n)

[Nepomuk5665]

Summary

• Fixes O(n²) performance issue in chunkOnModifiedUtf8ByteSize that caused significant build time overhead
• Implements incremental byte size calculation instead of creating substrings for each character
• Adds a helper function to calculate Modified UTF-8 byte size per character

Background

This addresses the performance issue reported in #350. The original implementation created a new substring and converted it to a byte array for every character in the input string:

// Before: O(n²) - substring + toByteArray for each iteration
val charModifiedUtf8ByteArraySize =
    this
        .substring(nextChunkStart, i + 1)
        .let { chunk -> chunk.toByteArray().size + chunk.count { char -> char == '\u0000' } }

This is now replaced with an O(n) approach that calculates byte sizes incrementally:

// After: O(n) - direct calculation per character
val charByteSize = this[i].modifiedUtf8ByteSize()

Technical Details

The new modifiedUtf8ByteSize() helper function handles Modified UTF-8 encoding rules:

• U+0000 (null): 2 bytes (C0 80 encoding - special for Modified UTF-8)
• U+0001-U+007F: 1 byte
• U+0080-U+07FF: 2 bytes
• U+0800-U+FFFF: 3 bytes

Testing

All existing tests in CharSequenceUtf8Test pass, confirming behavioral correctness.

[rossbacher]

This is great! thank you very much!

Can you fix the Ci issues? Just run ./gradlew check javaDoc and fix the lint issues by running formatKotlin on the broken module,.

[rossbacher]

Thank you! Great fix!