Potential issue with COSString and UTF-16-encoded Strings.
----------------------------------------------------------
Key: PDFBOX-904
URL: https://issues.apache.org/jira/browse/PDFBOX-904
Project: PDFBox
Issue Type: Bug
Components: PDModel
Affects Versions: 1.3.1
Reporter: Neil McErlean
I've been looking into PDFBOX-903 and I came across a potential issue with the
COSString class.
The issue occurs when you construct an instance of COSString and pass a
UTF-16-encoded String.
The current code (trunk) checks the passed String parameter in the constructor
to see if it is UTF-16. It does this by looking for char values above 255.
Whilst a String that contains char values greater than 255 is likely to be
UTF-16, it is possible to have UTF-16-encoded Strings whose characters do not
exceed this limit.
These Strings would be incorrectly marked as being not unicode16. An example
(from the upcoming patch)
/**してく */
String textHighBits = "\u3057\u3066\u304f";
Furthermore, if you construct a COSString using the COSString(byte[])
constructor, then the COSString class cannot know what the encoding is.
I will attach a patch in a moment which includes a test case to reproduce the
issue and a fix for the product code.
--
This message is automatically generated by JIRA.
-
You can reply to this email to add a comment to the issue online.
