[PDFBOX-583] TextPosition#getIndividualWidths returns negative values

• Type: Bug
• Status: Closed
• Priority: Critical
• Resolution: Fixed
• Affects Version/s: 1.0.0
• Fix Version/s: 1.0.0
• Component/s: Text extraction, Utilities
• Assignee: Unassigned
• Reporter: Villu Ruusmann
• Created: 12/Dec/09 16:47
• Updated: 22/Feb/10 18:28
• Resolved: 18/Dec/09 18:53

Description

I was notified by Peter Costello (peterwcostello at gmail dot com) that there has been a regression in PDFStreamSource, which results in TextPosition returning negative individual character widths. This could happen, because I only tested with TextPosition#getWidth(), but not with TextPosition#getIndividualWidths(). Below is the original message:

Bug introduced to /pdfbox/trunk/src/main/java/org/apache/pdfbox/util/PDFStreamEngine.java on revision 885904

Line 422 changed to:
float widthText = glyphMatrixStDisp.getXPosition() - glyphMatrixEndDisp.getXPosition();

Polarity of width now negative. Should be
float widthText = glyphMatrixEndDisp.getXPosition() - glyphMatrixStDisp.getXPosition();

Issue Links

• is related to: PDFBOX-571 Dubious handling of word spacing (Tw)