On Feb 2, 2012, at 8:38 AM, Tankred Hase wrote: > > I think this has little to do with the FileAPI and more to do with what is > > happening inside of the > > PGP encryption code. > > Yes I have made a qunit test for just the pgp logic which reproduces this > problem. >
Good idea. > > With up-crypt I was able to upload and encrypt files up to 10 MB easily, > > but the encryption was > > being done in AES. > > Doesn't pgp also use aes for the file encryption and rsa only for > signing/encryption of the session key? If openpgp.js uses the same aes > implementation as up-crypt, could it be that the problem is somewhere in the > compression or signature logic? How is the usual process with defects in this > project... should i file a bug in github? Yes, that is correct, though up-crypt uses a different AES implementation (https://code.google.com/p/crypto-js/#AES) than that of openpgpjs (http://www.hanewin.net/encrypt/aes/aes.htm). I guess an experiment could be of swapping out the AES code inside of openpgpjs with that of crypto-js. - Art. _______________________________________________ http://openpgpjs.org
