[PDFBOX-4778] Avoid illegal matrix values

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 2.0.18, 3.0.0 PDFBox
• Fix Version/s: 2.0.19, 3.0.0 PDFBox
• Component/s: Parsing
• External issue URL: https://bugs.ghostscript.com/show_bug.cgi?id=695040
• Assignee: Andreas Lehmkühler
• Reporter: Andreas Lehmkühler
• Created: 16/Feb/20 14:57
• Updated: 28/Apr/20 05:26
• Resolved: 16/Feb/20 15:26

Description

Tim Allison found a bunch of problematic pdfs and one of them runs very slow/forever.

The attached pdf has a broken content stream so that a lot of matrix operations are skipped due to missing values which lead to illegal values when multiplying matrices. Some of the calculated text positions are broken and in the end text extraction with sorting is very slow or runs infinite.

Issue Links

• is related to: PDFBOX-4783 empty annotation appearance BBox brings "Multiplying two matrices produces illegal values"