Yes, this bug is already fixed in CVS.
On Tue, Apr 27, 2004 at 05:06:47PM +0200, Renzo Tomaselli wrote:
> Hi,
> oh well, I reply to myself :-)
> I found a solution to the question below by using NewSignatureAccumulator to
> get the hash, then I can use method Sign instead of SignMessage.
> Pretty good, but then the compiler complains about method
> NewSignatureAccumulator in TF_SignerImpl (file pubkey.h) being declared as
> private. The same holds true for DL_SignerImpl as well.
> If I declare both methods as public, compilation succeeds.
> Is that a Crypto++ bug ?
> Thanks,
>
> Renzo Tomaselli
>
> ----- Original Message -----
> From: "Renzo Tomaselli" <[EMAIL PROTECTED]>
> To: "crypto++ MailList" <[EMAIL PROTECTED]>
> Sent: Tuesday, April 27, 2004 11:42 AM
> Subject: signing long messages
>
>
> > Hi,
> > assume to sign long messages, where contents must be digested in
> > advance. Then the final digest result should be encrypted.
> > Until Crypto++ v. 5.0 we had a chance to do it through method
> > AccessDigestSignatureScheme().SignDigest, called on DSA::Signer or
> > RSA::Signer.
> > Now (5.1) only SignMessage seems surviving. It requires original message
> > contents only, no digest.
> > Any trick to do it in two steps ?
> > Thanks,
> >
> > Renzo Tomaselli
> >
> >
