[PDFBOX-592] please accommodate '-' where a number is expected

• Type: Improvement
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 0.8.0-incubator
• Fix Version/s: 1.3.1
• Component/s: Parsing
• Assignee: Jukka Zitting
• Reporter: Godmar Back
• Created: 07/Jan/10 05:54
• Updated: 26/Oct/10 09:33
• Resolved: 24/Aug/10 12:31

Description

PDF's created by ScanSoft's PDF driver have '-' where a number is needed, as in 1 0 0 1 - 783 Tm. See http://markmail.org/message/r63jfd5wybejzbkr for details.

Proposal: interpret - as 0. Patch:

— /tmp/x/pdfbox-0.8.0-incubating/src/main/java/org/apache/pdfbox/pdfparser/PDFStreamParser.java 2009-09-14 19:39:44.000000000 -0400
+++ src/main/java/org/apache/pdfbox/pdfparser/PDFStreamParser.java 2010-01-07 00:14:45.000000000 -0500
@@ -252,7 +252,12 @@
dotNotRead = false;
}
}

• retval = COSNumber.get( buf.toString() );
  + String number = buf.toString();
  + /* accommodate PDF files (such as ScanSoft-created ones that output '-'
  + * where a number is expected. Substitute a 0 */
  + if ("-".equals(number))
  + number = "0";
  + retval = COSNumber.get( number );
  break;
  }
  case 'B':