[NUTCH-643] ClassCastException in PdfParser on encrypted PDF with empty password

• Type: Bug
• Status: Closed
• Priority: Major
• Resolution: Fixed
• Affects Version/s: 1.0.0
• Fix Version/s: 1.0.0
• Component/s: fetcher

Environment

This problem affects the current trunk too.

• Patch Info: Patch Available
• Assignee: Andrzej Bialecki
• Reporter: Guillaume Smet
• Created: 08/Aug/08 13:48
• Updated: 10/Apr/09 12:29
• Resolved: 06/Feb/09 13:12

Description

Hi,

If a PDF document is encrypted with an empty password, the PdfParser should decrypt it using the empty password.

This behaviour is implemented with the following code:
if (pdf.isEncrypted())

{ DocumentEncryption decryptor = new DocumentEncryption(pdf); //Just try using the default password and move on decryptor.decryptDocument(""); }

It uses a deprecated API and moreover it seems there is a bug in PDFBox in this deprecated API (we have a ClassCastException in PDFBox) as we have the following error:

2008-08-07 19:15:56,860 WARN parse.pdf - General exception in PDF parser: org.pdfbox.pdmodel.encryption.PDEncryptionDictionary cannot be cast to org.pdfbox.pdmodel.encryption.PDStandardEncryption
2008-08-07 19:15:56,862 WARN parse.pdf - java.lang.ClassCastException: org.pdfbox.pdmodel.encryption.PDEncryptionDictionary cannot be cast to org.pdfbox.pdmodel.encryption.PDStandardEncryption
2008-08-07 19:15:56,862 WARN parse.pdf - at org.pdfbox.encryption.DocumentEncryption.decryptDocument(DocumentEncryption.java:197)
2008-08-07 19:15:56,862 WARN parse.pdf - at org.apache.nutch.parse.pdf.PdfParser.getParse(PdfParser.java:98)
2008-08-07 19:15:56,862 WARN parse.pdf - at org.apache.nutch.parse.ParseUtil.parse(ParseUtil.java:82)
2008-08-07 19:15:56,862 WARN parse.pdf - at org.apache.nutch.fetcher.Fetcher$FetcherThread.output(Fetcher.java:336)
2008-08-07 19:15:56,862 WARN parse.pdf - at org.apache.nutch.fetcher.Fetcher$FetcherThread.run(Fetcher.java:178)
2008-08-07 19:15:56,874 WARN fetcher.Fetcher - Error parsing: http://www2.culture.gouv.fr/deps/fr/stateurope071.pdf: failed(2,0): Can't be handled as pdf document. java.lang.ClassCastException: org.pdfbox.pdmodel.encryption.PDEncryptionDictionary cannot be cast to org.pdfbox.pdmodel.encryption.PDStandardEncryption

Using the new security API, we don't have any error parsing this document and we can get its content:
if (pdf.isEncrypted())

{ // Just try using the default password and move on pdf.openProtection(new StandardDecryptionMaterial("")); }

I attached the patch fixing this problem: it works perfectly with the above document and get rids of the deprecated API.

Regards,

–
Guillaume