[RFC]: UTF-8 (SMTPUTF8) support

[boesing]

RFC

Q	A
Proposed Version(s)	3.0.0
BC Break?	Yes

Goal

Providing proper UTF-8 support for laminas-mail.
Would fix #71
Would fix #67
Would fix #57

Background

I've found myself working on a fork from Zend_Mail which was basically created to allow UTF-8 characters when parsing E-Mails from string. Main purpose was to allow UTF-8 subjects and UTF-8 E-Mail address.

Considerations

I've played a little bit around and since almost everything is finally verified by HeaderValue#isValid, I think we have to tackle that one.

I was thinking about adding another argument to HeaderValue#isValid to add "extensions".
So maybe something like:

final class HeaderValue
{
    public const EXTENSION_NONE = 0;
    public const EXTENSION_SMTPUTF8 = 1;

    /**
     * @psalm-param int-mask-of<HeaderValue::EXTENSION_*> $withExtensions
     */
    public static function isValid(string $value, int $withExtensions = self::EXTENSION_NONE): bool
    {
    }
}

Yes - another bitmask (but TBH - I don't see how to add something like this without actually having a bitmask).
This would actually allow us allowing the SMTPUTF8 extension from outside.

I'd say, when reading an E-Mail from Imap or Pop3 (is this still a thing?) - we should allow SMTPUTF8 since if the mail does contain UTF-8 headers, it is due to the fact that a server actually delivered that mail - right?

So now we have to somehow pass that extension information through all headers - and also to the Message (in case we are parsing an existing mail).

Proposal(s)

Implement a generic way to allow SMTP extensions from these methods:

• Headers#fromString
• Message#fromString

Appendix

None so far. Happy to hear your thoughts about this as I definitely would love to get rid of that Zend_Mail fork.