Re: pgcrypto: PGP signatures

On 2014-09-03 10:33 PM, Jeff Janes wrote:
> On Wed, Sep 3, 2014 at 12:43 PM, Marko Tiikkaja <marko@joh.to> wrote:
>> Right.  This patch only adds support for signing data when encrypting it
>> at the same time.  There's no support for detached signatures, nor is there
>> support for anything other than signatures of encrypted data.  I should
>> have been more clear on that in my initial email. :-(
>>
>>
> OK, thanks.  How hard do you think it would to allow NULL (or empty
> string?) passwords to gpg_sym_signatures and gpg_sym_decrypt_verify to
> accommodate this?

To sign without encrypting?  I think those should really be a different 
set of functions altogether.  But this patch is already humongous (on my 
standards, at least), so I think that should be done separately.

> I think docs section F.25.3 needs to be re-titled and expanded to reflect
> signatures as well as encryption, and an explanation added about signatures
> only being processed on encrypted data if that restriction can't be removed.

I don't have an objection to that.  I fully acknowledge that the 
documentation doesn't state the limitations of signing should this patch 
be applied.

> I've switched to using a signed plus symmetrically encrypted message for
> testing.
>
> One surprising thing so far is that the 3rd argument to
> gpg_sym_decrypt_verify must be dearmored. I thought it would detect and
> dearmor automatically.

I can't see that as an improvement to be honest.

> Once I wrap it in dearmor, I get the ERROR:  No signature matching the key
> id present in the message
>
> The public key block I am giving it is for the keyid that is reported
> by pgp_sym_signatures, so I don't know what the problem might be.

Have you tried with the debug=1 option?  (It's undocumented, but it was 
like that before this patch and I didn't touch it).

> When I get more time, I'll look at your examples from the regression tests
> to see if I can figure it out.

Thanks!  I'm happy to help if you run into any trouble!


.marko