Sieve Email Filtering: delivery by mailboxid

Sieve rules are sometimes created using graphical interfaces which allow users to select the mailbox to be used as a target for a rule. If that mailbox is renamed, the client may also update its internal representation of the rule and update the sieve script to match, however this is a multi-step process and subject to partial failures. Also, if the folder is renamed by a different mechanism (e.g. another IMAP client) the rules will get out of sync. By extending fileinto to reference an immutable mailboxid, sieve rules can continue to target the same mailbox, regardless of how it gets renamed.

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 when, and only when, they appear in all capitals, as shown here.

The server must advertise the capability "mailboxid", and scripts which use the following extensions must explicitly request the capability "mailboxid". Example:

require "mailboxid";

Normally, the positional mailbox argument to a fileinto action delivers the message in a mailbox with the matching name. This document adds a new tagged argument :mailboxid. If the :mailboxid argument is provided, the positional mailbox argument refers to an mailboxid rather than an name. If there is no mailbox with the specified :mailboxid then the action is the same as if there was no named mailbox, and implementations are free to make the same choices given in Section 4.1 when the specified mailbox doesn't exist. The tagged argument :mailboxid to fileinto consumes zero additional tokens. Example:

require "fileinto"; require "mailboxid"; if header :contains ["from"] "coyote" { fileinto :mailboxid "F6352ae03-b7f5-463c-896f-d8b48ee3"; }

This test behaves identically to the mailboxexists test defined in but operates on mailboxids rather than mailbox names. Usage: mailboxidexists <mailboxids: string-list> The "mailboxidexists" test is true if all mailboxes listed in the "mailboxids" argument exist in the mailstore, and each allows the user in whose context the Sieve script runs to "deliver" messages into it. When the mailstore is an IMAP server, "delivery" of messages is possible if: a) the READ-WRITE response code is present for the mailbox (see Section 7.1 of ), if IMAP Access Control List (ACL) is not supported by the server, or b) the user has 'p' or 'i' rights for the mailbox (see Section 5.2 of ). Note that a successful "mailboxidexists" test for a mailbox doesn't necessarily mean that a "fileinto :mailboxid" action on this mailbox would succeed. For example, the "fileinto" action might put user over quota. The "mailboxidexists" only verifies existence of the mailbox and whether the user in whose context the Sieve script runs has permissions to execute "fileinto" on it. Example:

require "fileinto"; require "mailboxid"; if header :contains ["from"] "coyote" { if mailboxidexists "F6352ae03-b7f5-463c-896f-d8b48ee3" { fileinto :mailboxid "F6352ae03-b7f5-463c-896f-d8b48ee3"; } else { fileinto "INBOX.harassment"; } }

test /= ":mailboxidexists" string-list tag /= ":mailboxid"

This document does not add any security considerations beyond those in , and . Implementers are referred to the security considerations sections of those documents.

IANA are requested to add a capability to the sieve-extensions registry:

To: iana@iana.org Subject: Registration of new Sieve extension Capability name: mailboxid Description: adds test for checking for mailbox existence by objectid and a new optional argument to fileinto to select the destination mailbox using objectid. RFC number: this RFC Contact address: The EXTRA discussion list <extra@ietf.org>

This document borrows heavily from for the matching mailboxexists test, and from for an example of modifying the fileinto command.

Is there a more explicit way to update the grammar? It seems less fully specified than IMAP.