SMTPUTF8 Email Addresses

Status of This Memo

This Internet-Draft is submitted in full conformance with the provisions of BCP 78 and BCP 79.¶

Internet-Drafts are working documents of the Internet Engineering Task Force (IETF). Note that other groups may also distribute working documents as Internet-Drafts. The list of current Internet-Drafts is at https://datatracker.ietf.org/drafts/current/.¶

Internet-Drafts are draft documents valid for a maximum of six months and may be updated, replaced, or obsoleted by other documents at any time. It is inappropriate to use Internet-Drafts as reference material or to cite them other than as "work in progress."¶

This Internet-Draft will expire on 16 April 2027.¶

Copyright Notice

Copyright (c) 2026 IETF Trust and the persons identified as the document authors. All rights reserved.¶

This document is subject to BCP 78 and the IETF Trust's Legal Provisions Relating to IETF Documents (https://trustee.ietf.org/license-info) in effect on the date of publication of this document. Please review these documents carefully, as they describe your rights and restrictions with respect to this document. Code Components extracted from this document must include Revised BSD License text as described in Section 4.e of the Trust Legal Provisions and are provided without warranty as described in the Revised BSD License.¶

Table of Contents

1. Introduction

[RFC6530]-[RFC6533] and [RFC6854]-[RFC6858] extend various aspects of the email system to support non-ASCII both in localparts and domain parts. In addition, some email software supports unicode in domain parts by using encoded domain parts in the SMTP transaction ("RCPT TO:info@xn--dmi-0na.fo") and presenting the unicode version (dømi.fo in this case) in the user interface.¶

The email address syntax extension is in [RFC6532], and allows almost all UTF8 strings as localparts. While this certainly allows everything users want to use, it is also flexible enough to allow many things that users and implementers find surprising and sometimes worrying. For example, the domain extracted by code may not match that shown to users on-screen.¶

2. Requirements Language

The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and "OPTIONAL" in this document are to be interpreted as described in BCP 14 [RFC2119] [RFC8174] when, and only when, they appear in all capitals, as shown here.¶

3. Terminology

Script, in this document, refers to the unicode script property (see [UAX24]). Each code point is assigned to one script ("a" is Latin), except that some are assigned to "Common" or a few other special values. Fraktur and /etc/rc.local aren't scripts in this document, but Latin is.¶

Latin refers to those code points that have the script property "Latin" in Unicode. Orléans in France and Münster in Germany both have Latin names in this document. It also refers to combinations of those code points and combining characters, and to strings that contain no code points from other scripts.¶

Han, Cyrillic etc. refer to those code points that have the respective script property in Unicode, as well as to strings that contain no code points from other scripts.¶

ASCII refers to the first 128 code points within unicode, which includes the letters A-Z but not É or Ü. It also refers to strings that contain only ASCII code points.¶

Non-ASCII refers to unicode code points except the first 128, and also to strings that contain at least one such code point.¶

By way of example, the address info@dømi.fo is latin and non-ASCII, its localpart is latin and ASCII, and its domain part is latin and non-ASCII. 中国 is a Han string in this document, but 阿Q正传 is neither a Latin string nor a Han string, because it contains a Latin Q and three Han code points.¶

The term a-label is defined in [RFC5890] section 2.3.2.1.¶

PRECIS is described in [RFC8264], and the IdentifierClass in section 4.2 of that document.¶

4. Rules

The following four rules apply to the [I-D.ietf-emailcore-rfc5322bis] mailbox production, as indirectly extended by [RFC6532].¶

1. An atom in an address MUST NOT be an a-label (e.g. xn--dmi-0na).¶

2. An address MUST contain only code points that the PRECIS IdentifierClass treats as valid, and " ". Where the IdentifierClass requires a contextual rule, the address MUST satisfy that rule ([RFC5892] Appendix A).¶

3. An address MUST NOT contain more than one script, when ASCII is disregarded. Han, Hiragana and Katakana count as one script here, since Japanese uses all three together; so do Han and Hangul, and Han and Bopomofo. (For example: In the word Orléans, Orl and ans are ASCII and é is non-ASCII. The single é is in a single script, so the word contains only one script. まり子 is Hiragana and Han, which count as one.)¶

4. An address MUST be in the unicode NFC form.¶

This does not affect rules in other documents. For example, [RFC952], [RFC5891] and others specify rules on domain names, which implicitly affect which email addresses are permitted on the public internet. Domain registries often have policies regarding which names may be registered.¶

5. Examples

example@example.com is permitted, because 1) it does not contain any a-label, 2) it consists entirely of permissible code points and 3) it contains no non-ASCII code points at all.¶

The address dømi@dømi.fo is permitted, because 1) it does not contain any a-label, 2) the IdentifierClass permits every code point in it and 3) it consists entirely of 'Latin' and 'Common' code points (and ./@).¶

The address U+200E '@' U+200F '.' U+200E is not permitted, because 2) the IdentifierClass disallows U+200E and U+200F as ignorable code points.¶

The address 'admin' U+3164 '@example.com' is not permitted for the same reason. U+3164 HANGUL FILLER renders as nothing at all, so the address looks exactly like admin@example.com. U+0640 ARABIC TATWEEL is disallowed too, although it is a letter by category: it only elongates the preceding letter, and would let one address be spelled in any number of ways.¶

The address まり子@まり子.日本 is permitted, because 1) it contains no a-label, 2) the IdentifierClass permits every code point in it and 3) Hiragana and Han count as one script.¶

阿Q正传@阿Q正传.example is permitted because it contains ASCII and Han, dømi@dømi.fo is legal because it contains ASCII and Latin, but 阿Q正传@dømi.fo is not, because it contains both Han (阿) and Latin non-ASCII (ø).¶

example@-example-.example.com is permitted by the rules above. However, since [RFC952] and its successors prohibit both of the hyphens, that email address is effectively not permitted on the internet.¶

6. IANA Considerations

This document does not require any actions from the IANA.¶

7. Security Considerations

When a program renders a non-ASCII email address string on-screen or stores it as a string (e.g. in a system log file), the address shown to the eventual human may not be as intended. One possible way to diminish the risk is to replace impermissible code points with U+FFFD, REPLACEMENT CHARACTER. There are also other options.¶

The rules in this document permit a number of code points that can make an address difficult to cut and paste.¶

This document does not change the longstanding rule that addresses may be either case-sensitive or case-insensitive. However, UTF8 adds an extra wrinkle: case conversion depends on the locale. When processing addresses in remote domains, one cannot be sure which conversion rules apply. Implementers may want to consider unusual cases such as the Turkish dotless ı.¶

8. References

Appendix A. Acknowledgments

The authors wish to thank John C. Klensin, Daniel Eggert, Andrew Gallagher, Jeffrey Stedfast, Pete Resnick, (your name here, please)¶

(Wow, the ack section is already outdated)¶

Dømi.fo and 例子.中国 are reserved by nic.fo and CNNIC for use in examples and documentation.¶

阿Q正传 is a famous Chinese novella, 阿Q is the main character.¶

Appendix B. Instructions to the RFC editor

Please remove all mentions of the Protocol Police before publication (including this sentence).¶

Please remove the Open Issues section.¶

Appendix C. Testing

This is a set of test addresses in JSON format.¶

Below is a verbatim copy of https://github.com/arnt/mailmaint-smtputf8/tests.json as it was on (date here). It contains a number of strange and unusual code points, so cutting and pasting this may not work. Rather, it is recommended to either use the rfcstrip tool or download the tests using a command such as curl https://github.com/arnt/mailmaint-smtputf8/tests.json > tests.json.¶

Note to IETF reviewers: The tests will be included here shortly before publication (and after IETF Last Call).¶

Appendix D. Open issues

1. More words on the relationship between this and the companion. There are several parallel differences, maybe this warrants a section of its own.¶

2. SHOULD or MUST not contain a-labels?¶

Authors' Addresses