[openpgp] text signatures

["Neal H. Walfield" <[e-mail]>]

4880 says:

> 5.2.4.  Computing Signatures
>
>    For text document signatures (type 0x01), the
>    document is canonicalized by converting line endings to <CR><LF>,
>    and the resulting data is hashed.

https://www.rfc-editor.org/rfc/rfc4880.html#section-5.2.4

The crypto refresh changes this to say:

> 5.2.4.  Computing Signatures
>
>    For text document signatures (type 0x01), the
>    implementation MUST first canonicalize the document by converting
>    line endings to <CR><LF> and encoding it in UTF-8 (see [RFC3629]).

https://datatracker.ietf.org/doc/html/draft-ietf-openpgp-crypto-refresh-07#section-5.2.4

Specifically, the OpenPGP implementation must now also convert the
text to UTF-8.

I wonder if this will break anything?  I still communicate with users
of Thunderbird 68 (because they rely on the web of trust), which uses
windows-1252 and text signatures:

   $ sq packet dump --hex --session-key ... /tmp/encrypted.asc
   └── Signature Packet, old CTB, 3 header bytes + 563 bytes
           Version: 4
           Type: Text

My understanding is that windows-1252 is not a subset of UTF-8.

I suspect that if deployed as-is, these signatures will no longer
validate.  That seems unfortunate to me.

Thoughts?

Neal