Re: Signature [was: BIOS time fine, Linux/Debian's isn't! ...]

[Reco <recoverym4n@enotuniq.net>]

On Fri, Jul 31, 2020 at 08:47:50AM -0400, The Wanderer wrote:
> On 2020-07-31 at 08:37, Reco wrote:
> >>> Quirks are fine, when things continue to work.  Your sig
> >>> separator fails completely.
> >> 
> >> "Fails" in what way?  He doesn't actually have a "signature file". 
> >> He doesn't have a signature separator ("-- \n") followed by a
> >> signature message.  He simply uses "-- t\n\n" at the end of each
> >> email.
> >> 
> >> You're not missing out on any content.
> > 
> > It's a kmail thing. mutt, being a superior MUA, is not affected.
> 
> Can you clarify both A: in what way this is kmail-specific,

kmail apparently mistakes '-- t' for the MIME-separator.

The "offending" e-mail actually look like this:

== cut ==

-- t

--gKMricLos+KVdGMg
Content-Type: application/pgp-signature; name="signature.asc"
Content-Description: Digital signature

-----BEGIN PGP SIGNATURE-----

== cut ==

Where "-- t" is a signature and is contained inside of a MIME-part, and
"--gKMricLos+KVdGMg" is an actual RFC2045-compilant separator.


> and B: what behaviors you're seeing in respect to mutt that are
> relevant here?

The lack of aforementioned confusion. The signature shows as the author
intended, gpg validation happens as expected.


> I see failure on my end, with Thunderbird, which is similar to what I
> understand the problem being complained about to be. It is my
> understanding that the problem is rooted in (lack of) compliance with a
> particular RFC, such that the resulting behavior will be basically
> universal across all compliant mail clients.

RFC2045 is a standard, but we all know that certain proprietary MUA
which deliberately violates almost all standards if it comes to e-mail.
Apparently both KDE and Mozilla consider more important to be compatible
with this certain MUA than to follow actual standards closely.

Reco