multiple signers

David Shaw dshaw at jabberwocky.com
Wed Dec 11 15:54:01 CET 2002


On Wed, Dec 11, 2002 at 03:07:18PM +0100, Werner Koch wrote:
> On Wed, 11 Dec 2002 10:40:01 +0100, Timo Schulz said:
> 
> > Yes, I thought the same. When the file is very large, it takes a long
> > time to verify two or more signatures. What about parallel hashing.
> 
> Even worse, you can't do this when the data comes from a pipe.  
> We already use several hash contexts to work around PGP bugs.  I am
> not in favor of setting up even more.  If the message is correctly
> encoded, we can hash the data using more than one algorithms, it is
> just a matter of enabling the algorithms for the hash context and
> doing a md_copy right before hashing the signature specific part. 

The fix I did was to enable each algorithm needed and then hash with
only one context (except for the PGP bug case, when there are two
contexts).  This doesn't work (I added an error return) if one of the
sigs is 0x00 and one is 0x01 since the textmode filter is in place
before the data is even read to be hashed, but it's somewhat crazy to
verify the same data with and without textmode at the same time so I
don't think this matters much.

David

-- 
   David Shaw  |  dshaw at jabberwocky.com  |  WWW http://www.jabberwocky.com/
+---------------------------------------------------------------------------+
   "There are two major products that come out of Berkeley: LSD and UNIX.
      We don't believe this to be a coincidence." - Jeremy S. Anderson




More information about the Gnupg-devel mailing list