Hi guys, so I have finally found some time to get the tests and a proof of concept of pgpbox running, since my arrival in Sydney. You can try it out at: http://0-0-6-snapshot.pgpbox-org.appspot.com/pgpbox.html
Just drag a PDF onto the dotted area it's contents will be read via the FileReader, encrypted, decrypted and finally rendered with pdf.js. This works reasonably fast with PDFs of a couple hundred KB. One problem I'm facing is that the FileReader and pdf.js store the PDF in an ArrayBuffer, but openpgp.js only takes Strings. Also, it seems to have trouble with unicode Strings, so I also convert the string to base64 before encryption. Here's the current pipeline: FileReader -> *ArrayBuffer -> Unicode -> Base64 -> Encrypt -> Decrypt -> Base64 -> Unicode -> ArrayBuffer* -> pdf.js (the encrypted PDF will later be uploaded to a server after encryption of course ;) * * Apart from the fact that this is not very efficient, as you can imagine, it requires quite a large amount of memory (in bytes) to store the encrypted PDF: ArrayBuffer size: 678 Unicode size: 678 Base64 size: 2712 Cipher size: 4826 Base64Dec size: 2712 UnicodeDec size: 678 Decrypted ArrayBuffer size: 678 I put the qunit tests up under the following URL: http://0-0-6-snapshot.pgpbox-org.appspot.com/qunit.html The firsts 6 tests should run just fine in Chrome 16.0.912.77 (you may have to refresh twice, since the qunit tests are asychronous and sometimes have trouble the first time around???). The 7th test fails, since I dont encode the unicode string to base64 before encryption. First question: Is this a bug or a feature? Second question: Is it somehow possible or is it outside the scope of openpgp.js to take an ArrayBuffer as input/output? Aside from the fact, that this would be a killer feature for the encryption of large files, it would also make the library way more flexible ;) I hope I'm not sounding too demanding :/ Looking forward to your feedback. Thanks! Kind regards, Tankred
_______________________________________________ http://openpgpjs.org
