[PDFBOX-3864] UTF16 encoded string to PDFDocEncoding

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.6
• Fix Version/s: 2.0.7, 3.0.0 PDFBox
• Component/s: PDModel
• Assignee: Tilman Hausherr
• Reporter: Tilman Hausherr
• Created: 11/Jul/17 14:20
• Updated: 21/Jul/17 04:40
• Resolved: 11/Jul/17 14:56

Description

From torakiki in the mailing list:

Hi, we came across this case where we are basically cloning outline items
where the original outline title is a UTF16BE encoded text string
containing the value 00A0 (non break space). We later use the string to
assign the title in a new outline item and the A0 is recognised as a € sign.
Here is a simple test:

        COSString victim = COSString
                .parseHex("FEFF004300680061007000740065007200A0");
        PDOutlineItem node = new PDOutlineItem();
        node.setTitle(victim.getString());

If you look at the node dictionary you'll see that the title value is
Chapter€

The cause is that in the initialization of PDFDocEncoding it was forgotten that there are "holes" in the 0..255 sequence. I'll add that and a test.