Internet Engineering Task Force (IETF)                          C. Daboo
Request for Comments: 6609                                   Apple, Inc.
Category: Standards Track                                       A. Stone
ISSN: 2070-1721                                              Serendipity
                                                                May 2012


                Sieve Email Filtering: Include Extension

Abstract

   The Sieve Email Filtering "include" extension permits users to
   include one Sieve script inside another.  This can make managing
   large scripts or multiple sets of scripts much easier, and allows a
   site and its users to build up libraries of scripts.  Users are able
   to include their own personal scripts or site-wide scripts.

Status of This Memo

   This is an Internet Standards Track document.

   This document is a product of the Internet Engineering Task Force
   (IETF).  It represents the consensus of the IETF community.  It has
   received public review and has been approved for publication by the
   Internet Engineering Steering Group (IESG).  Further information on
   Internet Standards is available in Section 2 of RFC 5741.

   Information about the current status of this document, any errata,
   and how to provide feedback on it may be obtained at
   http://www.rfc-editor.org/info/rfc6609.

Copyright Notice

   Copyright (c) 2012 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
   (http://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 Simplified BSD License text as described in Section 4.e of
   the Trust Legal Provisions and are provided without warranty as
   described in the Simplified BSD License.

Table of Contents

   1. Introduction and Overview .......................................2
   2. Conventions Used in This Document ...............................2
   3. Include Extension ...............................................3
      3.1. General Considerations .....................................3
      3.2. Control Structure "include" ................................4
      3.3. Control Structure "return" .................................7
      3.4. Interaction with the "variables" Extension .................8
           3.4.1. Control Structure "global" ..........................8
           3.4.2. Variables Namespace global .........................10
      3.5. Interaction with Other Extensions .........................11
   4. Security Considerations ........................................12
   5. IANA Considerations ............................................12
   6. References .....................................................13
      6.1. Normative References ......................................13
      6.2. Informative References ....................................13
   Appendix A. Acknowledgments .......................................14

1.  Introduction and Overview

   It's convenient to be able to break Sieve [RFC5228] scripts down into
   smaller components that can be reused in a variety of different
   circumstances.  For example, users may want to have a default script
   and a special 'vacation' script, the latter being activated when the
   user goes on vacation.  In that case, the default actions should
   continue to be run, but a vacation command should be executed first.
   One option is to edit the default script to add or remove the
   vacation command as needed.  Another is to have a vacation script
   that simply has a vacation command and then includes the default
   script.

   This document defines the Sieve Email Filtering "include" extension,
   which permits users to include one Sieve script inside another.

2.  Conventions Used in This Document

   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
   document are to be interpreted as described in [RFC2119].

   Conventions for notations are as in Sieve [RFC5228], Section 1.1.

   The following key phrases are used to describe scripts and script
   execution:

   script
      a valid Sieve script.

   script execution
      an instance of a Sieve interpreter invoked for a given message
      delivery, starting with the user's active script and continuing
      through any included scripts until the final disposition of the
      message (e.g., delivered, forwarded, discarded, rejected, etc.).

   immediate script
      the individual Sieve script file being executed.

   including script
      the individual Sieve script file that had an include statement
      that included the immediate script.

3.  Include Extension

3.1.  General Considerations

   Sieve implementations that implement the "include", "return", and
   "global" commands described below have an identifier of "include" for
   use with the capability mechanism.  If any of the "include",
   "return", or "global" commands are used in a script, the "include"
   capability MUST be listed in the "require" statement in that script.

   Sieve implementations need to track the use of actions in included
   scripts so that implicit "keep" behavior can be properly determined
   based on whether any actions have executed in any script.

   Sieve implementations are allowed to limit the total number of nested
   included scripts, but MUST provide for a total of at least three
   levels of nested scripts including the top-level script.  An error
   MUST be generated either when the script is uploaded to the Sieve
   repository, or when the script is executed, if any nesting limit is
   exceeded.  If such an error is detected whilst processing a Sieve
   script, an implicit "keep" action MUST be executed to prevent loss of
   any messages.

   Sieve implementations MUST NOT allow recursive script inclusion.
   Both direct recursion, where script A includes script A (itself), and
   indirect recursion, where script A includes script B which includes
   script A once again, are prohibited.

   Sieve implementations MUST generate an error at execution time if an
   included script is a recursive inclusion.  Implementations MUST NOT
   generate errors for recursive includes at upload time, as this would
   force an upload ordering requirement upon script authors and
   generators.

   Sieve implementations MUST generate an error at execution time if an
   included script does not exist, except when the ":optional" parameter
   is specified.  Implementations MUST NOT generate errors for scripts
   missing at upload time, as this would force an upload ordering
   requirement upon script authors and generators.

   If the Sieve "variables" extension [RFC5229] is present, an issue
   arises with the "scope" of variables defined in scripts that may
   include each other.  For example, if a script defines the variable
   "${status}" with one particular meaning or usage, and another defines
   "${status}" with a different meaning, then if one script includes the
   other there is an issue as to which "${status}" is being referenced.
   To solve this problem, Sieve implementations MUST follow the scoping
   rules defined in Section 3.4 and support the "global" command defined
   there.

3.2.  Control Structure "include"

      Usage:  include [LOCATION] [":once"] [":optional"] <value: string>

              LOCATION = ":personal" / ":global"

   The "include" command takes an optional "location" parameter, an
   optional ":once" parameter, an optional ":optional" parameter, and a
   single string argument representing the name of the script to include
   for processing at that point.  Implementations MUST restrict script
   names according to ManageSieve [RFC5804], Section 1.6.  The script
   name argument MUST be a constant string as defined in [RFC5229],
   Section 3; implementations MUST NOT expand variables in the script
   name argument.

   The "location" parameter MUST default to ":personal" if not
   specified.  The "location" parameter MUST NOT be specified more than
   once.  The "location" has the following meanings:

   :personal
      Indicates that the named script is stored in the user's own
      personal (private) Sieve repository.

   :global
      Indicates that the named script is stored in a site-wide Sieve
      repository, accessible to all users of the Sieve system.

   The ":once" parameter tells the interpreter only to include the named
   script if it has not already been included at any other point during
   script execution.  If the script has already been included,
   processing continues immediately following the "include" command.
   Implementations MUST NOT generate an error if an "include :once"
   command names a script whose inclusion would be recursive; in this
   case, the script MUST be considered previously included, and
   therefore "include :once" will not include it again.

   Note: It is RECOMMENDED that script authors and generators use the
   ":once" parameter only when including a script that performs general
   duties such as declaring global variables and making sanity checks of
   the environment.

   The ":optional" parameter indicates that the script may be missing.
   Ordinarily, an implementation MUST generate an error during execution
   if an "include" command specifies a script that does not exist.  When
   ":optional" is specified, implementations MUST NOT generate an error
   for a missing script, and MUST continue as if the "include" command
   had not been present.

   The included script MUST be a valid Sieve script.  Implementations
   MUST validate that each script has its own "require" statements for
   all optional capabilities used by that script.  The scope of a
   "require" statement is the script in which it immediately appears,
   and neither inherits nor passes on capabilities to other scripts
   during the course of execution.

   A "stop" command in an included script MUST stop all script
   processing, including the processing of the scripts that include the
   immediate one.  The "return" command (described below) stops
   processing of the immediate script only, and allows the scripts that
   include it to continue.

   The "include" command MAY appear anywhere in a script where a control
   structure is legal, and MAY be used within another control structure,
   e.g., an "if" block.

   Examples:

   The user has four scripts stored in their personal repository:

   "default"

      This is the default active script that includes several others.

      require ["include"];

      include :personal "always_allow";
      include :global "spam_tests";
      include :personal "spam_tests";
      include :personal "mailing_lists";

   Personal script "always_allow"

      This script special-cases some correspondent email addresses and
      makes sure any message containing those addresses is always kept.

      if address :is "from" "boss@example.com"
      {
          keep;
      }
      elsif address :is "from" "ceo@example.com"
      {
          keep;
      }

   Personal script "spam_tests" (uses "reject" [RFC5429])

      This script does some user-specific spam tests to catch spam
      messages not caught by the site-wide spam tests.

      require ["reject"];

      if header :contains "Subject" "XXXX"
      {
          reject "Subject XXXX is unacceptable.";
      }
      elsif address :is "from" "money@example.com"
      {
          reject "Mail from this sender is unwelcome.";
      }

   Personal script "mailing_lists"

      This script looks for messages from different mailing lists and
      files each into a mailbox specific to the mailing list.

      require ["fileinto"];

      if header :is "List-ID" "sieve.ietf.org"
      {
          fileinto "lists.sieve";
      }
      elsif header :is "List-ID" "ietf-imapext.imc.org"
      {
          fileinto "lists.imapext";
      }

   There is one script stored in the global repository:

   Site script "spam_tests" (uses "reject" [RFC5429])

      This script does some site-wide spam tests that any user at the
      site can include in their own scripts at a suitable point.  The
      script content is kept up to date by the site administrator.

      require ["reject"];

      if anyof (header :contains "Subject" "$$",
                header :contains "Subject" "Make money")
      {
          reject "No thank you.";
      }

3.3.  Control Structure "return"

      Usage:  return

   The "return" command stops processing of the immediately included
   script only and returns processing control to the script that
   includes it.  If used in the main script (i.e., not in an included
   script), it has the same effect as the "stop" command, including the
   appropriate "keep" action if no other actions have been executed up
   to that point.

3.4.  Interaction with the "variables" Extension

   In order to avoid problems of variables in an included script
   "overwriting" those from the script that includes it, this
   specification requires that all variables defined in a script MUST be
   kept "private" to the immediate script by default -- that is, they
   are not "visible" to other scripts.  This ensures that two script
   authors cannot inadvertently cause problems by choosing the same name
   for a variable.

   However, sometimes there is a need to make a variable defined in one
   script available to others.  This specification defines the new
   command "global" to declare that a variable is shared among scripts.
   Effectively, two namespaces are defined: one local to the immediate
   script, and another shared among all scripts.  Implementations MUST
   allow a non-global variable to have the same name as a global
   variable but have no interaction between them.

3.4.1.  Control Structure "global"

      Usage:  global <value: string-list>

   The "global" command accepts a string list argument that defines one
   or more names of variables to be stored in the global variable space.
   Each name MUST be a constant string and conform to the syntax of
   variable-name as defined in the "variables" extension document
   [RFC5229], Section 3.  Match variables cannot be specified, and
   namespace prefixes are not allowed.  An invalid name MUST be detected
   as a syntax error.

   The "global" command is only available when the script has both
   "include" and "variables" in its require line.  If the "global"
   command appears when only "include" or only "variables" has been
   required, an error MUST be generated when the script is uploaded.

   If a "global" command is given the name of a variable that has
   previously been defined in the immediate script with "set", an error
   MUST be generated either when the script is uploaded or at execution
   time.

   If a "global" command lists a variable that has not been defined in
   the "global" namespace, the name of the variable is now marked as
   global, and any subsequent "set" command will set the value of the
   variable in global scope.

   A variable has global scope in all scripts that have declared it with
   the "global" command.  If a script uses that variable name without
   declaring it global, the name specifies a separate, non-global
   variable within that script.

   Interpretation of a string containing a variable marked as global,
   but without any value set, SHALL behave as any other access to an
   unknown variable, as specified in the "variables" extension document
   [RFC5229], Section 3 (i.e., evaluates to an empty string).

   Example:

   The active script

      The included script may contain repetitive code that is
      effectively a subroutine that can be factored out.  In this
      script, the test that matches last will leave its value in the
      test_mailbox variable, and the top-level script will file the
      message into that mailbox.  If no tests matched, the message will
      be implicitly kept in the INBOX.

      require ["fileinto", "include", "variables", "relational"];
      global "test";
      global "test_mailbox";

      set "test" "$$";
      include "subject_tests";

      set "test" "Make money";
      include "subject_tests";

      if string :count "eq" "${test_mailbox}" "1"
      {
          fileinto "${test_mailbox}";
          stop;
      }

   Personal script "subject_tests"

      This script performs a number of tests against the message, sets
      the global test_mailbox variable with a folder to file the message
      into, and then falls back to the top-level script.


      require ["include", "variables"];
      global ["test", "test_mailbox"];

      if header :contains "Subject" "${test}"
      {
          set "test_mailbox" "spam-${test}";
      }

3.4.2.  Variables Namespace global

   In addition to the "global" command, this document defines the
   variables namespace "global", in accordance with the "variables"
   extension document [RFC5229], Section 3.  The "global" namespace has
   no sub-namespaces (e.g., 'set "global.data.from" "me@example.com";'
   is not allowed).  The variable-name part MUST be a valid identifier
   (e.g., 'set "global.12" "value";' is not valid because "12" is not a
   valid identifier).

   Note that the "variables" extension document [RFC5229], Section 3
   suggests that extensions should define a namespace that is the same
   as its capability string (in this case, "include" rather than
   "global").  Nevertheless, references to the "global" namespace
   without a prior require statement for the "include" extension MUST
   cause an error.

   Example:

      require ["variables", "include"];

      set "global.i_am_on_vacation" "1";

   Variables declared global and variables accessed via the "global"
   namespace MUST each be one and the same.  In the following example
   script, we see the variable "i_am_on_vacation" used in a "global"
   command, and again with the "global" namespace.  Consider these as
   two syntaxes with identical meaning.

   Example:

      require ["variables", "include", "vacation"];
      global "i_am_on_vacation";

      set "global.i_am_on_vacation" "1";

      if string :is "${i_am_on_vacation}" "1"
      {
          vacation "It's true, I am on vacation.";
      }

3.5.  Interaction with Other Extensions

   When "include" is used with the "editheader" extension [RFC5293], any
   changes made to headers in a script MUST be propagated both to and
   from included scripts.  By way of example, if a script deletes one
   header and adds another, then includes a second script, the included
   script MUST NOT see the removed header, and MUST see the added
   header.  Likewise, if the included script adds or removes a header,
   upon returning to the including script, subsequent actions MUST see
   the added headers and MUST NOT see the removed headers.

   When "include" is used with the MIME extension [RFC5703]
   "foreverypart" control structure, the included script MUST be
   presented with the current MIME part as though it were the entire
   message.  A script SHALL NOT have any special control over the
   control structure it was included from.  The "break" command in an
   included script is not valid on its own and may not terminate a
   "foreverypart" iteration in another script.  The included script can
   use "return" to transfer control back to the including script.  A
   global variable can be used to convey results to the including
   script.  A "stop" in an included script, even within a "foreverypart"
   loop, still halts all script execution, per Section 3.2.

   When "include" is used with the "reject" extension [RFC5429], calling
   "reject" or "ereject" at any time sets the reject action on the
   message, and continues script execution.  Apropos of the MIME
   extension, if an included script sees only a portion of the message
   and calls a reject, it is the entire message and not the single MIME
   part that carries the rejection.

4.  Security Considerations

   Sieve implementations MUST ensure adequate security for the global
   script repository to prevent unauthorized changes to global scripts.
   For example, a site policy might enable only certain users with
   administrative privileges to modify the global scripts.  Sites are
   advised against allowing all users to have write access to the sites'
   global scripts.

   Sieve implementations MUST ensure that script names are checked for
   validity and proper permissions prior to inclusion, in order to
   prevent a malicious user from gaining access to files accessible to
   the mail server software that should not be accessible to the user.

   Sieve implementations MUST ensure that script names are safe for use
   with their storage system.  An error MUST be generated either when
   the script is uploaded or at execution time for a script including a
   name that could be used as a vector to attack the storage system.  By
   way of example, the following include commands should be considered
   hostile: 'include "./../..//etc/passwd"', 'include "foo$(`rm
   star`)"'.

   Beyond these, the "include" extension does not raise any security
   considerations that are not discussed in the base Sieve [RFC5228]
   document and the "variables" extension document [RFC5229].

5.  IANA Considerations

   The following template specifies the IANA registration of the Sieve
   extension specified in this document:

      To: iana@iana.org
      Subject: Registration of new Sieve extension

      Capability name: include
      Description:     adds the "include" command to execute other Sieve
                       scripts, the "return" action from an included
                       script, and the "global" command and "global"
                       variables namespace to access variables shared
                       among included scripts.
      RFC number:      this RFC
      Contact address: the Sieve discussion list <sieve@ietf.org>

   This information has been added to IANA's "Sieve Extensions" registry
   (http://www.iana.org).

6.  References

6.1.  Normative References

   [RFC2119]  Bradner, S., "Key words for use in RFCs to Indicate
              Requirement Levels", BCP 14, RFC 2119, March 1997.

   [RFC5228]  Guenther, P., Ed., and T. Showalter, Ed., "Sieve: An Email
              Filtering Language", RFC 5228, January 2008.

   [RFC5229]  Homme, K., "Sieve Email Filtering: Variables Extension",
              RFC 5229, January 2008.

   [RFC5804]  Melnikov, A., Ed., and T. Martin, "A Protocol for Remotely
              Managing Sieve Scripts", RFC 5804, July 2010.

6.2.  Informative References

   [RFC5293]  Degener, J. and P. Guenther, "Sieve Email Filtering:
              Editheader Extension", RFC 5293, August 2008.

   [RFC5429]  Stone, A., Ed., "Sieve Email Filtering: Reject and
              Extended Reject Extensions", RFC 5429, March 2009.

   [RFC5703]  Hansen, T. and C. Daboo, "Sieve Email Filtering: MIME Part
              Tests, Iteration, Extraction, Replacement, and Enclosure",
              RFC 5703, October 2009.

Appendix A.  Acknowledgments

   Thanks to Stephan Bosch, Ned Freed, Arnt Gulbrandsen, Tony Hansen,
   Kjetil Torgrim Homme, Jeffrey Hutzelman, Barry Leiba, Alexey
   Melnikov, Ken Murchison, Marc Mutz, and Rob Siemborski, for comments
   and corrections.

Authors' Addresses

   Cyrus Daboo
   Apple Inc.
   1 Infinite Loop
   Cupertino, CA  95014
   USA

   EMail: cyrus@daboo.name
   URI:   http://www.apple.com/


   Aaron Stone
   Serendipity
   1817 California St. #104
   San Francisco, CA  94109
   USA

   EMail: aaron@serendipity.cx