On 2024/05/10 11:40, Matthieu Herrb wrote:
> On Fri, May 10, 2024 at 10:19:22AM +0100, Kirill A. Korinsky wrote:
> > On Fri, 10 May 2024 06:57:20 +0100,
> > Matthieu Herrb <matth...@openbsd.org> wrote:
> > > 
> > > https://www.rfc-editor.org/rfc/rfc6376#section-3.3 says that
> > > rsa-sha256 SHOULD be used. Unfortunatly Mail::DKIM::Signer uses
> > > rsa-sha1 by default when no algorithm is specifed.
> > > 
> > > Update the dkimproxy.out sample config...
> > > 
> > > Make aboutmy.email (and other checkers) happier, and hopefully less
> > > rejects by hotmail/google and co...
> > > 
> > > comments? ok?
> > >
> > 
> > I'd like to point that using anything else whan RSA with SHA256 leads to
> > issues. The cause is OpenDKIM which is widley used. It had well known issue
> > with ed25519 [1] which probably will be fixed in the next release.
> > 
> > Anyway, the last release had happened in 2015 and this project seems to be
> > not that alive, so, no hope that it will be released and distributed soon.
> > 
> > My point: let add reference to this issue and suggest to use only 
> > RSA/SHA256.
> > 
> > Footnotes:
> > [1]  https://github.com/trusteddomainproject/OpenDKIM/issues/6
> > 
> Hi,
> 
> Afaict dkimpproxy is not using opendkim but p5-Mail-DKIM. dkimproxy
> itself also hasn't seen a update since many years, but the underlying
> perl lib has been last updated last january (and could use an update
> in the port).
> 
> So unless you imply that because many people use opendkim, ed25519
> based signatures shouldn't be used at all I'm not sure I understand
> what you're saying.

ed25519 can be used, but at the moment if you do use it, you probably
want to be double-signing with both that + rsa-sha256.

Reply via email to