[ https://issues.apache.org/jira/browse/PDFBOX-2342?page=com.atlassian.jira.plugin.system.issuetabpanels:comment-tabpanel&focusedCommentId=14130386#comment-14130386 ]
Tilman Hausherr commented on PDFBOX-2342: ----------------------------------------- The non sequential parser doesn't decrypt recursively like the sequential one does (why?). In a COSDictionary, only COSStrings are decrypted, the rest is left untouched. There's already a suspicious TODO there. For now, I've just added the decryption of COSArray which solves [~ssteiner1]s problem. But I wonder what else is incorrectly (not) decrypted, e.g. a COSDictionary within a COSDictionary. Why aren't we using this nice "does everything" method? {code} private void decrypt(COSBase obj, long objNum, long genNum) throws IOException {code} > WriteDecodedDoc cant decrypt pdf form correctly > ----------------------------------------------- > > Key: PDFBOX-2342 > URL: https://issues.apache.org/jira/browse/PDFBOX-2342 > Project: PDFBox > Issue Type: Bug > Components: Parsing > Affects Versions: 2.0.0 > Reporter: simon steiner > Attachments: test.pdf > > > java -jar ~/pdf-box-svn/app/target/pdfbox-app-2.0.0-SNAPSHOT.jar > WriteDecodedDoc -nonSeq test.pdf > country selection is wrong -- This message was sent by Atlassian JIRA (v6.3.4#6332)