Unicode characters(Korean) corrupted in EXIF metadata

[gerosyab]

When writing EXIF metadata with Korean characters, they get corrupted.

Example:

• Input: "안녕하세요" (Korean)
• Output: Something like "HUX8�" or other garbled text

Environment

• @uswriting/exiftool v1.0.4
• Browser: Chrome/Safari on Windows/Mac
• File type: JPEG

Test Results

❌ 안녕하세요: Got "HUX8�"

Reproduction

import { parseMetadata, writeMetadata } from '@uswriting/exiftool';

const result = await writeMetadata(jpegFile, {
  Artist: "안녕하세요"
}, {
  charset: 'utf8',
  args: ['-charset', 'utf8', '-L']
});

const modified = new File([result.data], 'test.jpg');
const readBack = await parseMetadata(modified);
// Korean text is corrupted here

Seems like the WASM ExifTool isn't handling UTF-8 properly. English works fine, but CJK characters get mangled.