Add chunked processing to diff.js so patcher supports multi-GB files

[Jiyath5516F]

Replace whole-file buffering with chunked read/write and incremental MD5 so the patch generator/applier can handle files larger than 1GB. Adds CHUNK_SIZE-based processing, incremental checksum, progress hooks, and file-size warnings.

Key changes

• diff.js: use File.slice + chunked FileReader instead of loading entire files
• Incremental MD5 (SparkMD5.ArrayBuffer) and chunked XOR patch generation/application
• Progress updates via window.updateProgress and user warnings for very large files