[PDFBOX-3881] Handling of Byte Order Mark with Metadata-Fields

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.0.7
• Fix Version/s: 2.0.8, 3.0.0 PDFBox
• Component/s: Parsing
• Labels: BOM
• Environment: Windows
• Assignee: Tilman Hausherr
• Reporter: Nico Prenzel
• Created: 27/Jul/17 10:59
• Updated: 02/Nov/17 21:00
• Resolved: 28/Jul/17 17:47

Description

PDDocumentInformation e.g. getAuthor() honors the byte order of the extracted string and removes the byte order mark signs.

But if the extracted string does only contain the byte order mark signs the corresponding string "þÿ" is returned.

Is this the intended solution?
I'd appreciate to remove the byte order mark signs also, if the extracted string does only contain these signs.

Problematic code:

public String getString()
  {
  if (this.bytes.length > 2)
    {
      if (((this.bytes[0] & 0xFF) == 254) && ((this.bytes[1] & 0xFF) == 255))
      {

        return new String(this.bytes, 2, this.bytes.length - 2, Charsets.UTF_16BE);
      }
      if (((this.bytes[0] & 0xFF) == 255) && ((this.bytes[1] & 0xFF) == 254))
      {

        return new String(this.bytes, 2, this.bytes.length - 2, Charsets.UTF_16LE);
      }
    }
    

    return PDFDocEncoding.toString(this.bytes);
  }

Attachment has an example pdf