Fix iTXt chunk parsing: dedicated parser with zlib decompression support

[Copilot]

The previous implementation incorrectly routed iTXt chunks through the tEXt parser, which decodes bytes as a string and trims null terminators—corrupting the binary header fields in iTXt payloads (compression flag, method, language tag, translated keyword).

Changes

• Dedicated ReadInternationalTextualData parser — parses iTXt payload at byte level:

  • Extracts null-terminated keyword
  • Reads compression flag and method bytes
  • Skips language tag and translated keyword (both null-terminated)
  • Decodes remaining bytes as UTF-8 text, with optional ZLibStream decompression when compression_flag == 1
  • Throws InvalidDataException for unknown compression methods (only zlib/method 0 is defined by spec)

• Switch case split — iTXt now routes to ReadInternationalTextualData; tEXt continues using ReadTextualData unchanged

• Test fixtures — two minimal 1×1 PNGs added: itxt-uncompressed.png and itxt-compressed.png, with corresponding test methods covering both code paths

// iTXt layout parsed at byte level:
// [keyword]\0 [comp_flag:1] [comp_method:1] [language_tag]\0 [translated_keyword]\0 [text...]
var keyword = Encoding.UTF8.GetString(buffer, 0, keywordEnd);
// ...
if (compressionFlag == 1)
{
    using var zlib = new ZLibStream(new MemoryStream(textBytes), CompressionMode.Decompress);
    // decompress into UTF-8 text
}

---

✨ Let Copilot coding agent set things up for you — coding agent works faster and does higher quality work when set up for your repo.