Yes, but why would you want to? You shouldn't use RSA keys less than 1024
bits long anyway because of insufficient security.

On Wed, Apr 07, 2004 at 09:38:43PM -0700, David Brownell wrote:
> Actually, I spent more time working with the library, and found that keys <
> 361 bits are not able to sign messages.  I found this value by generating a
> key of N bits, and then attempting to sign a 1 byte message using that key.
> All key sizes failed < 361.  Is there a way I could have computed this value
> programmatically rather than through trial and error?
> 
> Thanks!
> 
> -----Original Message-----
> From: Wei Dai [mailto:[EMAIL PROTECTED] 
> Sent: Wednesday, April 07, 2004 8:09 PM
> To: [EMAIL PROTECTED]
> Subject: Re: Max Message Length When Signing?
> 
> When you use RSASS, you should be able to sign arbitrarily long messages, as
> long as the key size is sufficient to handle the hash size (which is fixed
> for all message lengths). If, using the same key, you're able to sign short
> messages but not long ones, then something is wrong. Please post sample code
> if that's the case.
> 
> On Wed, Apr 07, 2004 at 08:51:54AM -0700, David Brownell wrote:
> > I am attempting to sign a message using the test application that 
> > ships with the Crypto++ library (which uses RSASS to sign a message).  
> > If I use a very small key to sign a very large file, I get a 
> > KeyTooSmall exception.  I know that in the real world hashes of 
> > messages are signed rather than the messages themselves to alleviate 
> > this problem, but for the sake of argument, is there a way to tell at 
> > run time that maximum message size that can be signed for a given private
> key?
> > 
> > Thanks in advance for your help!
> > David Brownell

Reply via email to