It looks like all digits and capital letters. And some kind of spaces or tabs. It's not a bomb. These file come in routinely. All the ones I have looked at (ftp'd size vs the unencrypted file sitting in archive) are right about 2-to-1.
To get a still encrypted file, I would have to file a request to modify a script and at least 3 groups would have to approve the request. And I would have to wait at least 1 week before I actually made the change. I remember freedom. . . . . Thanks. On Fri, Jan 9, 2009 at 7:50 PM, David Shaw <[email protected]> wrote: > On Jan 9, 2009, at 7:07 PM, Robert J. Hansen wrote: > > Scott Lambdin wrote: >> >>> Someone sends us a big ~700MB pgp encrypted file and when we decrypt it >>> the resulting file is about half that size. Anyone have an idea what >>> they might be doing to swell it up like that? >>> >> >> Option 1: they're not using compression and they're ASCII-armoring the >> file. You can expect to see a large size swell. >> > > Not double. By definition ASCII armor is around 1/3 larger (actually 137%) > than the original document (not counting headers and such, but they only > amount to a few hundred bytes, not megs). > > Option 2: they're sending a file that's carefully crafted to blow up. >> I've seen a ridiculously tiny zip archive (a couple of K) that expands >> into hundreds of terabytes. There are sixteen zip archives in that zip >> archive, each zip archive expands into another sixteen zip archives, >> each of those zip archives expands into several gigs of zeros, etc., etc. >> > > Other way around - the original file was ~700MB. The decrypted file was > ~350MB. > > Incidentally, GPG has code to deal with the potential denial of service > from a "bzip bomb" like you mention. See the --max-output option. > > Scott, do you know what OpenPGP program created the file that was sent to > you? Can you tell us what sort of data it it? (text? binary? image file? > (if so, jpeg? mpeg? other?) Also please try decrypting the file again and > add "-v -v" to the command line. Please send us anything you can that isn't > sensitive (specifically the compressed packet algo number, and the raw data > size and mode from the literal data packet). > > David > -- There's a box?
_______________________________________________ Gnupg-users mailing list [email protected] http://lists.gnupg.org/mailman/listinfo/gnupg-users
