[PDFBOX-4762] Inconsistent handling of incorrect data

• Type: Improvement
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Affects Version/s: 2.0.18
• Fix Version/s: 2.0.19, 3.0.0 PDFBox
• Component/s: Rendering
• Labels: patch
• Flags: Patch
• Assignee: Tilman Hausherr
• Reporter: Daniel Persson
• Created: 04/Feb/20 09:09
• Updated: 30/Jun/20 06:13
• Resolved: 05/Feb/20 17:07

Description

We had a PDF that had a strange page with 200Mb+ of text to extract and the deflate function did not work correctly. 

This created a fatal in PDFBox and I did some debugging and noticed that we handle  SetNonStrokingColorSpace and SetStrokingColorSpace in different ways. One of them had a check if the in data was incorrect and returned and the other one did not have this check.

I made this small patch that I will include in this issue to rectify this inconsistency.

 

Added the crashing pdf on my google drive if you want it to test with

https://drive.google.com/open?id=1bcT27NoqNM-pphYiFCy13bq81potqUc6

 

Best regards

Daniel

Issue Links

• relates to: PDFBOX-3691 ClassCastException in SetStrokingColorSpace.process