RFC2244 - ACAP -- Application Configuration Access Protocol

Network Working Group C. Newman
Request for Comments: 2244 Innosoft
Category: Standards Track J. G. Myers
Netscape
November 1997
ACAP -- Application Configuration Access Protocol
Status of this Memo
This document specifies an Internet standards track protocol for the
Internet community, and requests discussion and suggestions for
improvements. Please refer to the current edition of the "Internet
Official Protocol Standards" (STD 1) for the standardization state
and status of this protocol. Distribution of this memo is unlimited.
Copyright Notice
Copyright (C) The Internet Society 1997. All Rights Reserved.
Abstract
The Application Configuration Access Protocol (ACAP) is designed to
support remote storage and access of program option, configuration
and preference information. The data store model is designed to
allow a client relatively simple access to interesting data, to allow
new information to be easily added without server re-configuration,
and to promote the use of both standardized data and custom or
proprietary data. Key features include "inheritance" which can be
used to manage default values for configuration settings and access
control lists which allow interesting personal information to be
shared and group information to be restricted.
Table of Contents
Status of this Memo ............................................... i
Copyright Notice .................................................. i
Abstract .......................................................... i
ACAP Protocol Specification ....................................... 1
1. IntrodUCtion ............................................. 1
1.1. Conventions Used in this Document ........................ 1
1.2. ACAP Data Model .......................................... 1
1.3. ACAP Design Goals ........................................ 1
1.4. Validation ............................................... 2
1.5. Definitions .............................................. 2
1.6. ACAP Command Overview .................................... 4
2. Protocol Framework ....................................... 4
2.1. Link Level ............................................... 4
2.2. Commands and Responses ................................... 4
2.2.1. Client Protocol Sender and Server Protocol Receiver ...... 4
2.2.2. Server Protocol Sender and Client Protocol Receiver ...... 5
2.3. Server States ............................................ 6
2.3.1. Non-Authenticated State .................................. 6
2.3.2. Authenticated State ...................................... 6
2.3.3. Logout State ............................................. 6
2.4. Operational Considerations ............................... 7
2.4.1. Untagged Status Updates .................................. 7
2.4.2. Response when No Command in Progress ..................... 7
2.4.3. Auto-logout Timer ........................................ 7
2.4.4. Multiple Commands in Progress ............................ 8
2.5. Server Command Continuation Request ...................... 8
2.6. Data Formats ............................................. 8
2.6.1. Atom ..................................................... 9
2.6.2. Number ................................................... 9
2.6.3. String ................................................... 9
2.6.3.1. 8-bit and Binary Strings ................................. 10
2.6.4. Parenthesized List ....................................... 10
2.6.5. NIL ...................................................... 10
3. Protocol Elements ........................................ 10
3.1. Entries and Attributes ................................... 10
3.1.1. Predefined Attributes .................................... 11
3.1.2. Attribute Metadata ....................................... 12
3.2. ACAP URL scheme .......................................... 13
3.2.1. ACAP URL User Name and Authentication Mechanism .......... 13
3.2.2. Relative ACAP URLs ....................................... 14
3.3. Contexts ................................................. 14
3.4. Comparators .............................................. 15
3.5. Access Control Lists (ACLs) .............................. 17
3.6. Server Response Codes .................................... 18
4. Namespace Conventions .................................... 21
4.1. Dataset Namespace ........................................ 21
4.2. Attribute Namespace ...................................... 21
4.3. Formal Syntax for Dataset and Attribute Namespace ........ 22
5. Dataset Management ....................................... 23
5.1. Dataset Inheritance ...................................... 23
5.2. Dataset Attributes ....................................... 24
5.3. Dataset Creation ......................................... 25
5.4. Dataset Class Capabilities ............................... 25
5.5. Dataset Quotas ........................................... 26
6. Command and Response Specifications ...................... 26
6.1. Initial Connection ....................................... 26
6.1.1. ACAP Untagged Response ................................... 26
6.2. Any State ................................................ 27
6.2.1. NOOP Command ............................................. 27
6.2.2. LANG Command ............................................. 28
6.2.3. LANG Intermediate Response ............................... 28
6.2.4. LOGOUT Command ........................................... 29
6.2.5. OK Response .............................................. 29
6.2.6. NO Response .............................................. 29
6.2.7. BAD Response ............................................. 30
6.2.8. BYE Untagged Response .................................... 30
6.2.9. ALERT Untagged Response .................................. 31
6.3. Non-Authenticated State .................................. 31
6.3.1. AUTHENTICATE Command ..................................... 31
6.4. Searching ................................................ 33
6.4.1. SEARCH Command ........................................... 33
6.4.2. ENTRY Intermediate Response .............................. 37
6.4.3. MODTIME Intermediate Response ............................ 38
6.4.4. REFER Intermediate Response .............................. 38
6.4.5. Search Examples .......................................... 38
6.5. Contexts ................................................. 39
6.5.1. FREECONTEXT Command ...................................... 39
6.5.2. UPDATECONTEXT Command .................................... 40
6.5.3. ADDTO Untagged Response .................................. 40
6.5.4. REMOVEFROM Untagged Response ............................. 41
6.5.5. CHANGE Untagged Response ................................. 41
6.5.6. MODTIME Untagged Response ................................ 42
6.6. Dataset modification ..................................... 42
6.6.1. STORE Command ............................................ 42
6.6.2. DELETEDSINCE Command ..................................... 45
6.6.3. DELETED Intermediate Response ............................ 45
6.7. Access Control List Commands ............................. 45
6.7.1. SETACL Command ........................................... 46
6.7.2. DELETEACL Command ........................................ 46
6.7.3. MYRIGHTS Command ......................................... 47
6.7.4. MYRIGHTS Intermediate Response ........................... 47
6.7.5. LISTRIGHTS Command ....................................... 47
6.7.6. LISTRIGHTS Intermediate Response ......................... 48
6.8. Quotas ................................................... 48
6.8.1. GETQUOTA Command ......................................... 48
6.8.3. QUOTA Untagged Response .................................. 49
6.9. Extensions ............................................... 49
7. Registration Procedures .................................. 49
7.1. ACAP Capabilities ........................................ 50
7.2. ACAP Response Codes ...................................... 50
7.3. Dataset Classes .......................................... 51
7.4. Vendor SuBTree ........................................... 51
8. Formal Syntax ............................................ 52
9. Multi-lingual Considerations ............................. 61
10. Security Considerations .................................. 62
11. Acknowledgments .......................................... 63
12. Authors" Addresses ....................................... 63
Appendices ........................................................ 64
A. References ............................................... 64
B. ACAP KeyWord Index ....................................... 66
C. Full Copyright Statement
ACAP Protocol Specification
1. Introduction
1.1. Conventions Used in this Document
In examples, "C:" and "S:" indicate lines sent by the client and
server respectively. If such lines are wrapped without a new "C:" or
"S:" label, then the wrapping is for editorial clarity and is not
part of the command.
The key words "REQUIRED", "MUST", "MUST NOT", "SHOULD", "SHOULD NOT",
and "MAY" in this document are to be interpreted as described in "Key
words for use in RFCs to Indicate Requirement Levels" [KEYWORDS].
1.2. ACAP Data Model
An ACAP server eXPorts a hierarchical tree of entries. Each level of
the tree is called a dataset, and each dataset is made up of a list
of entries. Each entry has a unique name and may contain any number
of named attributes. Each attribute within an entry may be single
valued or multi-valued and may have associated metadata to assist
access and interpretation of the value.
The rules with which a client interprets the data within a portion of
ACAP"s tree of entries are called a dataset class.
1.3. ACAP Design Goals
ACAP"s primary purpose is to allow users access to their
configuration data from multiple network-connected computers. Users
can then sit down in front of any network-connected computer, run any
ACAP-enabled application and have access to their own configuration
data. Because it is hoped that many applications will become ACAP-
enabled, client simplicity was preferred to server or protocol
simplicity whenever reasonable.
ACAP is designed to be easily manageable. For this reason, it
includes "inheritance" which allows one dataset to inherit default
attributes from another dataset. In addition, access control lists
are included to permit delegation of management and quotas are
included to control storage. Finally, an ACAP server which is
conformant to this base specification should be able to support most
dataset classes defined in the future without requiring a server
reconfiguration or upgrade.
ACAP is designed to operate well with a client that only has
intermittent access to an ACAP server. For this reason, each entry
has a server maintained modification time so that the client may
detect changes. In addition, the client may ask the server for a
list of entries which have been removed since it last accessed the
server.
ACAP presumes that a dataset may be potentially large and/or the
client"s network connection may be slow, and thus offers server
sorting, selective fetching and change notification for entries
within a dataset.
As required for most Internet protocols, security, scalability and
internationalization were important design goals.
Given these design goals, an attempt was made to keep ACAP as simple
as possible. It is a traditional Internet text based protocol which
massively simplifies protocol debugging. It was designed based on
the successful IMAP [IMAP4] protocol framework, with a few
refinements.
1.4. Validation
By default, any value may be stored in any attribute for which the
user has appropriate permission and quota. This rule is necessary to
allow the addition of new simple dataset classes without
reconfiguring or upgrading the server.
In some cases, such as when the value has special meaning to the
server, it is useful to have the server enforce validation by
returning the INVALID response code to a STORE command. These cases
MUST be explicitly identified in the dataset class specification
which SHOULD include specific fixed rules for validation. Since a
given ACAP server may be unaware of any particular dataset class
specification, clients MUST NOT depend on the presence of enforced
validation on the server.
1.5. Definitions
access control list (ACL)
A set of identifier, rights pairs associated with an object. An
ACL is used to determine which operations a user is permitted to
perform on that object. See section 3.5.
attribute
A named value within an entry. See section 3.1.
comparator
A named function which can be used to perform one or more of
three comparison operations: ordering, equality and substring
matching. See section 3.4.
context
An ordered subset of entries in a dataset, created by a SEARCH
command with a MAKECONTEXT modifier. See section 3.3.
dataset
One level of hierarchy in ACAP"s tree of entries.
dataset class specification
The rules which allow a client to interpret the data within a
portion of ACAP"s tree of entries.
entry
A set of attributes with a unique entry name. See section 3.1.
metadata
Information describing an attribute, its value and any access
controls associated with that attribute. See section 3.1.2.
NIL This represents the non-existence of a particular data item.
NUL A control character encoded as 0 in US-ASCII [US-ASCII].
octet
An 8-bit value. On most modern computer systems, an octet is
one byte.
SASL Simple Authentication and Security Layer [SASL].
UTC Universal Coordinated Time as maintained by the Bureau
International des Poids et Mesures (BIPM).
UTF-8
An 8-bit transformation format of the Universal Character Set
[UTF8]. Note that an incompatible change was made to the coded
character set referenced by [UTF8], so for the purpose of this
document, UTF-8 refers to the UTF-8 encoding as defined by
version 2.0 of Unicode [UNICODE-2], or ISO 10646 [ISO-10646]
including amendments one through seven.
1.6. ACAP Command Overview
The AUTHENTICATE, NOOP, LANG and LOGOUT commands provide basic
protocol services. The SEARCH command is used to select, sort, fetch
and monitor changes to attribute values and metadata. The
UPDATECONTEXT and FREECONTEXT commands are also used to assist in
monitoring changes in attribute values and metadata. The STORE
command is used to add, modify and delete entries and attributes.
The DELETEDSINCE command is used to assist a client in
re-synchronizing a cache with the server. The GETQUOTA, SETACL,
DELETEACL, LISTRIGHTS and MYRIGHTS commands are used to examine
storage quotas and examine or modify access permissions.
2. Protocol Framework
2.1. Link Level
The ACAP protocol assumes a reliable data stream such as provided by
TCP. When TCP is used, an ACAP server listens on port 674.
2.2. Commands and Responses
An ACAP session consists of the establishment of a client/server
connection, an initial greeting from the server, and client/server
interactions. These client/server interactions consist of a client
command, server data, and a server completion result.
ACAP is a text-based line-oriented protocol. In general,
interactions transmitted by clients and servers are in the form of
lines; that is, sequences of characters that end with a CRLF. The
protocol receiver of an ACAP client or server is either reading a
line, or is reading a sequence of octets with a known count (a
literal) followed by a line. Both clients and servers must be
capable of handling lines of arbitrary length.
2.2.1. Client Protocol Sender and Server Protocol Receiver
The client command begins an operation. Each client command is
prefixed with a identifier (an alphanumeric string of no more than 32
characters, e.g., A0001, A0002, etc.) called a "tag". A different
tag SHOULD be generated by the client for each command.
There are two cases in which a line from the client does not
represent a complete command. In one case, a command argument is
quoted with an octet count (see the description of literal in section
2.6.3); in the other case, the command arguments require server
feedback (see the AUTHENTICATE command). In some of these cases, the
server sends a command continuation request if it is ready for the
next part of the command. This response is prefixed with the token
"+".
Note: If, instead, the server detected an error in a
command, it sends a BAD completion response with tag
matching the command (as described below) to reject the
command and prevent the client from sending any more of the
command.
It is also possible for the server to send a completion or
intermediate response for some other command (if multiple
commands are in progress), or untagged data. In either
case, the command continuation request is still pending;
the client takes the appropriate action for the response,
and reads another response from the server.
The ACAP server reads a command line from the client, parses the
command and its arguments, and transmits server data and a server
command completion result.
2.2.2. Server Protocol Sender and Client Protocol Receiver
Data transmitted by the server to the client come in four forms:
command continuation requests, command completion results,
intermediate responses, and untagged responses.
A command continuation request is prefixed with the token "+".
A command completion result indicates the success or failure of the
operation. It is tagged with the same tag as the client command
which began the operation. Thus, if more than one command is in
progress, the tag in a server completion response identifies the
command to which the response applies. There are three possible
server completion responses: OK (indicating success), NO (indicating
failure), or BAD (indicating protocol error such as unrecognized
command or command syntax error).
An intermediate response returns data which can only be interpreted
within the context of a command in progress. It is tagged with the
same tag as the client command which began the operation. Thus, if
more than one command is in progress, the tag in an intermediate
response identifies the command to which the response applies. A
tagged response other than "OK", "NO", or "BAD" is an intermediate
response.
An untagged response returns data or status messages which may be
interpreted outside the context of a command in progress. It is
prefixed with the token "*". Untagged data may be sent as a result
of a client command, or may be sent unilaterally by the server.
There is no syntactic difference between untagged data that resulted
from a specific command and untagged data that were sent
unilaterally.
The protocol receiver of an ACAP client reads a response line from
the server. It then takes action on the response based upon the
first token of the response, which may be a tag, a "*", or a "+" as
described above.
A client MUST be prepared to accept any server response at all times.
This includes untagged data that it may not have requested.
This topic is discussed in greater detail in the Server Responses
section.
2.3. Server States
An ACAP server is in one of three states. Most commands are valid in
only certain states. It is a protocol error for the client to
attempt a command while the server is in an inappropriate state for
that command. In this case, a server will respond with a BAD command
completion result.
2.3.1. Non-Authenticated State
In non-authenticated state, the user must supply authentication
credentials before most commands will be permitted. This state is
entered when a connection starts.
2.3.2. Authenticated State
In authenticated state, the user is authenticated and most commands
will be permitted. This state is entered when acceptable
authentication credentials have been provided.
2.3.3. Logout State
In logout state, the session is being terminated, and the server will
close the connection. This state can be entered as a result of a
client request or by unilateral server decision.
+--------------------------------------+
initial connection and server greeting
+--------------------------------------+
(1) (2)
VV
+-----------------+
non-authenticated
+-----------------+
(4) (3)
VV
+----------------+
authenticated
+----------------+
(4)
VV VV VV
+--------------------------------------+
logout and close connection
+--------------------------------------+
(1) connection (ACAP greeting)
(2) rejected connection (BYE greeting)
(3) successful AUTHENTICATE command
(4) LOGOUT command, server shutdown, or connection closed
2.4. Operational Considerations
2.4.1. Untagged Status Updates
At any time, a server can send data that the client did not request.
2.4.2. Response when No Command in Progress
Server implementations are permitted to send an untagged response
while there is no command in progress. Server implementations that
send such responses MUST deal with flow control considerations.
Specifically, they must either (1) verify that the size of the data
does not exceed the underlying transport"s available window size, or
(2) use non-blocking writes.
2.4.3. Auto-logout Timer
If a server has an inactivity auto-logout timer, that timer MUST be
of at least 30 minutes duration. The receipt of ANY command from the
client during that interval MUST suffice to reset the auto-logout
timer.
2.4.4. Multiple Commands in Progress
The client is not required to wait for the completion result of a
command before sending another command, subject to flow control
constraints on the underlying data stream. Similarly, a server is
not required to process a command to completion before beginning
processing of the next command, unless an ambiguity would result
because of a command that would affect the results of other commands.
If there is such an ambiguity, the server executes commands to
completion in the order given by the client.
2.5. Server Command Continuation Request
The command continuation request is indicated by a "+" token instead
of a tag. This indicates that the server is ready to accept the
continuation of a command from the client.
This response is used in the AUTHENTICATE command to transmit server
data to the client, and request additional client data. This
response is also used if an argument to any command is a
synchronizing literal (see section 2.6.3).
The client is not permitted to send the octets of a synchronizing
literal unless the server indicates that it expects it. This permits
the server to process commands and reject errors on a line-by-line
basis, assuming it checks for non-synchronizing literals at the end
of each line. The remainder of the command, including the CRLF that
terminates a command, follows the octets of the literal. If there
are any additional command arguments the literal octets are followed
by a space and those arguments.
Example: C: A099 FREECONTEXT {10}
S: + "Ready for additional command text"
C: FRED
C: FOOB
S: A099 OK "FREECONTEXT completed"
C: A044 BLURDYBLOOP {102856}
S: A044 BAD "No such command as "BLURDYBLOOP""
2.6. Data Formats
ACAP uses textual commands and responses. Data in ACAP can be in one
of five forms: atom, number, string, parenthesized list or NIL.
2.6.1. Atom
An atom consists of one to 1024 non-special characters. It must
begin with a letter. Atoms are used for protocol keywords.
2.6.2. Number
A number consists of one or more digit characters, and represents a
numeric value. Numbers are restricted to the range of an unsigned
32-bit integer: 0 < number < 4,294,967,296.
2.6.3. String
A string is in one of two forms: literal and quoted string. The
literal form is the general form of string. The quoted string form
is an alternative that avoids the overhead of processing a literal at
the cost of restrictions of what may be in a quoted string.
A literal is a sequence of zero or more octets (including CR and LF),
prefix-quoted with an octet count in the form of an open brace ("{"),
the number of octets, close brace ("}"), and CRLF. In the case of
literals transmitted from server to client, the CRLF is immediately
followed by the octet data.
There are two forms of literals transmitted from client to server.
The form where the open brace ("{") and number of octets is
immediately followed by a close brace ("}") and CRLF is called a
synchronizing literal. When sending a synchronizing literal, the
client must wait to receive a command continuation request before
sending the octet data (and the remainder of the command). The other
form of literal, the non-synchronizing literal, is used to transmit a
string from client to server without waiting for a command
continuation request. The non-synchronizing literal differs from the
synchronizing literal by having a plus ("+") between the number of
octets and the close brace ("}") and by having the octet data
immediately following the CRLF.
A quoted string is a sequence of zero to 1024 octets excluding NUL,
CR and LF, with double quote (<">) characters at each end.
The empty string is represented as "" (a quoted string with zero
characters between double quotes), as {0} followed by CRLF (a
synchronizing literal with an octet count of 0), or as {0+} followed
by a CRLF (a non-synchronizing literal with an octet count of 0).
Note: Even if the octet count is 0, a client transmitting a
synchronizing literal must wait to receive a command
continuation request.
2.6.3.1. 8-bit and Binary Strings
Most strings in ACAP are restricted to UTF-8 characters and may not
contain NUL octets. Attribute values MAY contain any octets
including NUL.
2.6.4. Parenthesized List
Data structures are represented as a "parenthesized list"; a sequence
of data items, delimited by space, and bounded at each end by
parentheses. A parenthesized list can contain other parenthesized
lists, using multiple levels of parentheses to indicate nesting.
The empty list is represented as () -- a parenthesized list with no
members.
2.6.5. NIL
The special atom "NIL" represents the non-existence of a particular
data item that is represented as a string or parenthesized list, as
distinct from the empty string "" or the empty parenthesized list ().
3. Protocol Elements
This section defines data formats and other protocol elements used
throughout the ACAP protocol.
3.1. Entries and Attributes
Within a dataset, each entry name is made up of zero or more UTF-8
characters other than slash ("/"). A slash separated list of
entries, one at each level of the hierarchy, forms the full path to
an entry.
Each entry is made up of a set of attributes. Each attribute has a
hierarchical name in UTF-8, with each component of the name separated
by a period (".").
The value of an attribute is either single or multi-valued. A single
value is NIL (has no value), or a string of zero or more octets. A
multi-value is a list of zero or more strings, each of zero or more
octets.
Attribute names are not permitted to contain asterisk ("*") or
percent ("%") and MUST be valid UTF-8 strings which do not contain
NUL. Invalid attribute names result in a BAD response. Entry names
are not permitted to begin with "." or contain slash ("/") and MUST
be valid UTF-8 strings which do not contain NUL. Invalid entry names
in the entry field of a command result in a BAD response.
Use of non-visible UTF-8 characters in attribute and entry names is
discouraged.
3.1.1. Predefined Attributes
Attribute names which do not contain a dot (".") are reserved for
standardized attributes which have meaning in any dataset. The
following attributes are defined by the ACAP protocol.
entry
Contains the name of the entry. MUST be single valued.
Attempts to use illegal or multi-valued values for the entry
attribute are protocol errors and MUST result in a BAD
completion response. This is a special case.
modtime
Contains the date and time any read-write metadata in the entry
was last modified. This value MUST be in UTC, MUST be
automatically updated by the server.
The value consists of 14 or more US-ASCII digits. The first
four indicate the year, the next two indicate the month, the
next two indicate the day of month, the next two indicate the
hour (0 - 23), the next two indicate the minute, and the next
two indicate the second. Any further digits indicate fractions
of a second.
The time, particularly fractions of a second, need not be
accurate. It is REQUIRED, however, that any two entries in a
dataset changed by successive modifications have strictly
ascending modtime values. In addition, each STORE command
within a dataset (including simultaneous stores from different
connections) MUST use different modtime values.
This attribute has enforced validation, so any attempt to STORE
a value in this attribute MAY result in a NO response with an
INVALID response code.
subdataset
If this attribute is set, it indicates the existence of a sub-
dataset of this entry.
The value consists of a list of relative ACAP URLs (see section
3.2) which may be used to locate the sub-dataset. The base URL
is the full path to the entry followed by a slash ("/"). The
value "." indicates a subdataset is located directly under this
one. Multiple values indicate replicated copies of the
subdataset.
For example, if the dataset "/folder/site/" has an entry
"public-folder" with a subdataset attribute of ".", then there
exists a dataset "/folder/site/public-folder/". If the value of
the subdataset attribute was instead
"//other.acap.domain//folder/site/public-folder/", that would
indicate the dataset is actually located on a different ACAP
server.
A dataset can be created by storing a "subdataset" attribute
including ".", and a sub-hierarchy of datasets is deleted by
storing a NIL value to the "subdataset" attribute on the entry
in the parent dataset.
This attribute has enforced syntax validation. Specifically, if
an attempt is made to STORE a non-list value (other than NIL),
an empty list, or one of the values does not follow the URL
syntax rules [BASIC-URL, REL-URL], then this will result in a NO
response with an INVALID response code.
3.1.2. Attribute Metadata
Each attribute is made up of metadata items which describe that
attribute, its value and any associated access controls. Metadata
items may be either read-only, in which case the client is never
permitted to modify the item, or read-write, in which case the client
may modify the item if the access control list (ACL) permits.
The following metadata items are defined in this specification:
acl The access control list for the attribute, if one exists. If
the attribute does not have an ACL, NIL is returned.
Read-write. See section 3.5 for the contents of an ACL.
attribute
The attribute name. Read-only.
myrights
The set of rights that the client has to the attribute.
Read-only. See section 3.5 for the possible rights.
size This is the length of the value. In the case of a
multi-value, this is a list of lengths for each of the values.
Read-only.
value The value. For a multi-value, this is a list of single
values. Read-write.
Additional items of metadata may be defined in extensions to this
protocol. Servers MUST respond to unrecognized metadata by returning
a BAD command completion result.
3.2. ACAP URL scheme
ACAP URLs are used within the ACAP protocol for the "subdataset"
attribute, referrals and inheritance. They provide a convenient
syntax for referring to other ACAP datasets. The ACAP URL follows
the common Internet scheme syntax as defined in [BASIC-URL] except
that plaintext passwords are not permitted. If :<port> is omitted,
the port defaults to 674.
An ACAP URL has the following general form:
url-acap = "acap://" url-server "/" url-enc-entry [url-filter]
[url-extension]
The <url-server> element includes the hostname, and optional user
name, authentication mechanism and port number. The <url-enc-entry>
element contains the name of an entry path encoded according to the
rules in [BASIC-URL].
The <url-filter> element is an optional list of interesting attribute
names. If omitted, the URL refers to all attributes of the named
entry. The <url-extension> element is reserved for extensions to
this URL scheme.
Note that unsafe or reserved characters such as " " or "?" MUST be
hex encoded as described in the URL specification [BASIC-URL]. Hex
encoded octets are interpreted according to UTF-8 [UTF8].
3.2.1. ACAP URL User Name and Authentication Mechanism
A user name and/or authentication mechanism may be supplied. They
are used in the "AUTHENTICATE" command after making the connection to
the ACAP server. If no user name or authentication mechanism is
supplied, then the SASL ANONYMOUS [SASL-ANON] mechanism is used by
default. If an authentication mechanism is supplied without a user
name, then one SHOULD be obtained from the specified mechanism or
requested from the user as appropriate. If a user name is supplied
without an authentication mechanism then ";AUTH=*" is assumed.
The ";AUTH=" authentication parameter is interpreted as described in
the IMAP URL Scheme [IMAP-URL].
Note that if unsafe or reserved characters such as " " or ";" are
present in the user name or authentication mechanism, they MUST be
encoded as described in the URL specification [BASIC-URL].
3.2.2. Relative ACAP URLs
Because ACAP uses "/" as the hierarchy separator for dataset paths,
it works well with the relative URL rules defined in the relative URL
specification [REL-URL].
The <aauth> grammar element is considered part of the user name for
purposes of resolving relative ACAP URLs.
The base URL for a relative URL stored in an attribute"s value is
formed by taking the path to the dataset containing that attribute,
appending a "/" followed by the entry name of the entry containing
that attribute followed by "/".
3.3. Contexts
A context is subset of entries in a dataset or datasets, created by a
SEARCH command with a MAKECONTEXT modifier. Context names are
client-generated strings and must not start with the slash ("/")
character.
When a client creates a context, it may request automatic
notification of changes. A client may also request enumeration of
entries within a context. Enumeration simplifies the implementation
of a "virtual scrollbar" by the client.
A context exists only within the ACAP session in which it was
created. When the connection is closed, all contexts associated with
that connection are automatically discarded. A server is required to
support at least 100 active contexts within a session. If the server
supports a larger limit it must advertise it in a CONTEXTLIMIT
capability.
3.4. Comparators
A comparator is a named function which takes two input values and can
be used to perform one or more of four comparison operations:
ordering, equality, prefix and substring matching.
The ordering operation is used both for the SORT search modifier and
the COMPARE and COMPARESTRICT search keys. Ordering comparators can
determine the ordinal precedence of any two values. When used for
ordering, a comparator"s name can be prefixed with "+" or "-" to
indicate that the ordering should be normal order or reversed order
respectively. If no prefix is included, "+" is assumed.
For the purpose of ordering, a comparator may designate certain
values as having an undefined ordinal precedence. Such values always
collate with equal value after all other values regardless of whether
normal or reversed ordering is used. Unless the comparator
definition specifies otherwise, multi-values and NIL values have an
undefined ordinal precedence.
The equality operation is used for the EQUAL search modifier, and
simply determines if the two values are considered equal under the
comparator function. When comparing a single value to a multi-value,
the two are considered equal if any one of the multiple values is
equal to the single value.
The prefix match operation is used for the PREFIX search modifier,
and simply determines if the search value is a prefix of the item
being searched. In the case of prefix search on a multi-value, the
match is successful if the value is a prefix of any one of the
multiple values.
The substring match operation is used for the SUBSTRING search
modifier, and simply determines if search value is a substring of the
item being searched. In the case of substring search on a multi-
value, the match is successful if the value is a substring of any one
of the multiple values.
Rules for naming and registering comparators will be defined in a
future specification. Servers MUST respond to unknown or improperly
used comparators with a BAD command completion result.
The following comparators are defined by this standard and MUST be
implemented:
i;octet
Operations: Ordering, Equality, Prefix match, Substring match
For collation, the i;octet comparator interprets the value of
an attribute as a series of unsigned octets with ordinal
values from 0 to 255. When ordering two strings, each octet
pair is compared in sequence until the octets are unequal or
the end of the string is reached. When collating two strings
where the shorter is a prefix of the longer, the shorter
string is interpreted as having a smaller ordinal value. The
"i;octet" or "+i;octet" forms collate smaller ordinal values
earlier, and the "-i;octet" form collates larger ordinal
values earlier.
For the equality function, two strings are equal if they are
the same length and contain the same octets in the same
order. NIL is equal only to itself.
For non-binary, non-nil single values, i;octet ordering is
equivalent to the ANSI C [ISO-C] strcmp() function applied to
C string representations of the values. For non-binary,
non-nil single values, i;octet substring match is equivalent
to the ANSI C strstr() function applied to the C string
representations of the values.
i;ascii-casemap
Operations: Ordering, Equality, Prefix match, Substring match
The i;ascii-casemap comparator first applies a mapping to the
attribute values which translates all US-ASCII letters to
uppercase (octet values 0x61 to 0x7A are translated to octet
values 0x41 to 0x5A respectively), then applies the i;octet
comparator as described above. With this function the values
"hello" and "HELLO" have the same ordinal value and are
considered equal.
i;ascii-numeric
Operations: Ordering, Equality
The i;ascii-numeric comparator interprets strings as decimal
positive integers represented as US-ASCII digits. All values
which do not begin with a US-ASCII digit are considered equal
with an ordinal value higher than all non-NIL single-valued
attributes. Otherwise, all US-ASCII digits (octet values
0x30 to 0x39) are interpreted starting from the beginning of
the string to the first non-digit or the end of the string.
3.5. Access Control Lists (ACLs)
An access control list is a set of identifier, rights pairs used to
restrict access to a given dataset, attribute or attribute within an
entry. An ACL is represented by a multi-value with each value
containing an identifier followed by a tab character followed by the
rights. The syntax is defined by the "acl" rule in the formal syntax
in section 8.
Identifier is a UTF-8 string. The identifier "anyone" is reserved to
refer to the universal identity (all authentications, including
anonymous). All user name strings accepted by the AUTHENTICATE
command to authenticate to the ACAP server are reserved as
identifiers for the corresponding user. Identifiers starting with a
slash ("/") character are reserved for authorization groups which
will be defined in a future specification. Identifiers MAY be
prefixed with a dash ("-") to indicate a revocation of rights. All
other identifiers have implementation-defined meanings.
Rights is a string listing a (possibly empty) set of alphanumeric
characters, each character listing a set of operations which is being
controlled. Letters are reserved for "standard" rights, listed
below. The set of standard rights may only be extended by a
standards-track or IESG approved experimental RFC. Digits are
reserved for implementation or site defined rights. The currently
defined standard rights are:
x - search (use EQUAL search key with i;octet comparator)
r - read (access with SEARCH command)
w - write (modify with STORE command)
i - insert (perform STORE on a previously NIL value)
a - administer (perform SETACL or STORE on ACL attribute/metadata)
An implementation may force rights to always or never be granted. In
particular, implementations are expected to grant implicit read and
administer rights to a user"s personal dataset storage in order to
avoid denial of service problems. Rights are never tied, unlike the
IMAP ACL extension [IMAP-ACL].
It is possible for multiple identifiers in an access control list to
apply to a given user (or other authentication identity). For
example, an ACL may include rights to be granted to the identifier
matching the user, one or more implementation-defined identifiers
matching groups which include the user, and/or the identifier
"anyone". These rights are combined by taking the union of all
positive rights which apply to a given user and subtracting the union
of all negative rights which apply to that user. A client MAY avoid
this calculation by using the MYRIGHTS command and metadata items.
Each attribute of each entry of a dataset may potentially have an
ACL. If an attribute in an entry does not have an ACL, then access
is controlled by a default ACL for that attribute in the dataset, if
it exists. If there is no default ACL for that attribute in the
dataset, access is controlled by a default ACL for that dataset. The
default ACL for a dataset must exist.
In order to perform any access or manipulation on an entry in a
dataset, the client must have "r" rights on the "entry" attribute of
the entry. Implementations should take care not to reveal via error
messages the existence of an entry for which the client does not have
"r" rights. A client does not need access to the "subdataset"
attribute of the parent dataset in order to access the contents of a
dataset.
Many of the ACL commands and responses include an "acl object"
parameter, for specifying what the ACL applies to. This is a
parenthesized list. The list contains just the dataset name when
referring to the default ACL for a dataset. The list contains a
dataset name and an attribute name when referring to the default ACL
for an attribute in a dataset. The list contains a dataset name, an
attribute name, and an entry name when referring to the ACL for an
attribute of an entry of a dataset.
3.6. Server Response Codes
An OK, NO, BAD, ALERT or BYE response from the server MAY contain a
response code to describe the event in a more detailed machine
parsable fashion. A response code consists of data inside
parentheses in the form of an atom, possibly followed by a space and
arguments. Response codes are defined when there is a specific
action that a client can take based upon the additional information.
In order to support future extension, the response code is
represented as a slash-separated hierarchy with each level of
hierarchy representing increasing detail about the error. Clients
MUST tolerate additional hierarchical response code detail which they
don"t understand.
The currently defined response codes are:
AUTH-TOO-WEAK
This response code is returned on a tagged NO result from an
AUTHENTICATE command. It indicates that site security policy
forbids the use of the requested mechanism for the specified
authentication identity.
ENCRYPT-NEEDED
This response code is returned on a tagged NO result from an
AUTHENTICATE command. It indicates that site security policy
requires the use of a strong encryption mechanism for the
specified authentication identity and mechanism.
INVALID
This response code indicates that a STORE command included
data which the server implementation does not permit. It
MUST NOT be used unless the dataset class specification for
the attribute in question explicitly permits enforced server
validation. The argument is the attribute which was invalid.
MODIFIED
This response code indicates that a conditional store failed
because the modtime on the entry is later than the modtime
specified with the STORE command UNCHANGEDSINCE modifier.
The argument is the entry which had been modified.
NOEXIST
This response code indicates that a search or NOCREATE store
failed because a specified dataset did not exist. The
argument is the dataset which does not exist.
PERMISSION
A command failed due to insufficient permission based on the
access control list or implicit rights. The argument is the
acl-object which caused the permission failure.
QUOTA
A STORE or SETACL command which would have increased the size
of the dataset failed due to insufficient quota.
REFER
This response code may be returned in a tagged NO response to
any command that takes a dataset name as a parameter. It has
one or more arguments with the syntax of relative URLs. It
is a referral, indicating that the command should be retried
using one of the relative URLs.
SASL This response code can occur in the tagged OK response to a
successful AUTHENTICATE command and includes the optional
final server response data from the server as specified by
SASL [SASL].
TOOMANY
This response code may be returned in a tagged OK response to
a SEARCH command which includes the LIMIT modifier. The
argument returns the total number of matching entries.
TOOOLD
The modtime specified in the DELETEDSINCE command is too old,
so deletedsince information is no longer available.
TRANSITION-NEEDED
This response code occurs on a NO response to an AUTHENTICATE
command. It indicates that the user name is valid, but the
entry in the authentication database needs to be updated in
order to permit authentication with the specified mechanism.
This can happen if a user has an entry in a system
authentication database such as Unix /etc/passwd, but does
not have credentials suitable for use by the specified
mechanism.
TRYLATER
A command failed due to a temporary server failure. The
client MAY continue using local information and try the
command later.
TRYFREECONTEXT
This response code may be returned in a tagged NO response to
a SEARCH command which includes the MAKECONTEXT modifier. It
indicates that a new context may not be created due to the
server"s limit on the number of existing contexts.
WAYTOOMANY
This response code may be returned in a tagged NO response to
a SEARCH command which includes a HARDLIMIT search modifier.
It indicates that the SEARCH would have returned more entries
than the HARDLIMIT permitted.
Additional response codes MUST be registered with IANA according
to the proceedures in section 7.2. Client implementations MUST
tolerate response codes that they do not recognize.
4. Namespace Conventions
4.1. Dataset Namespace
The dataset namespace is a slash-separated hierarchy. The first
component of the dataset namespace is a dataset class. Dataset
classes MUST have a vendor prefix (vendor.<vendor/product>) or be
specified in a standards track or IESG approved experimental RFC.
See section 7.3 for the registration template.
The second component of the dataset name is "site", "group", "host",
or "user" referring to server-wide data, administrative group data,
per-host data and per-user data respectively.
For "group", "host", and "user" areas, the third component of the
path is the group name, the fully qualified host domain name, or the
user name. A path of the form "/<dataset-class>/~/" is a convenient
abbreviation for "/<dataset-class>/user/<current-user>/".
Dataset names which begin with "/byowner/" are reserved as an
alternate view of the namespace. This provides a way to see all the
dataset classes which a particular owner uses. For example,
"/byowner/~/<dataset-class>/" is an alternate name for
"/<dataset-class>/~/". Byowner provides a way to view a list of
dataset classes owned by a given user; this is done using the dataset
"/byowner/user/<current-user>/" with the NOINHERIT SEARCH modifier.
The dataset "/" may be used to find all dataset classes visible to
the current user. A dataset of the form "/<dataset-class>/user/" may
be used to find all users which have made a dataset or entry of that
class visible to the current user.
The formal syntax for a dataset name is defined by the "dataset-name"
rule in section 4.3.
4.2. Attribute Namespace
Attribute names which do not contain a dot (".") are reserved for
standardized attributes which have meaning in any dataset. In order
to simplify client implementations, the attribute namespace is
intended to be unique across all datasets. To achieve this,
attribute names are prefixed with the dataset class name followed by
a dot ("."). Attributes which affect management of the dataset are
prefixed with "dataset.". In addition, a subtree of the "vendor."
attribute namespace may be registered with IANA according to the
rules in section 7.4. ACAP implementors are encouraged to help
define interoperable dataset classes specifications rather than using
the private attribute namespace.
Some users or sites may wish to add their own private attributes to
certain dataset classes. In order to enable this, the "user.<user-
name>." and "site." subtrees of the attribute namespace are reserved
for user-specific and site-specific attributes respectively and will
not be standardized. Such attributes are not interoperable so are
discouraged in favor of defining standard attributes. A future
extension is expected to permit discovery of syntax for user or
site-specific attributes. Clients wishing to support display of user
or site-specific attributes should display the value of any non-NIL
single-valued "user.<user-name>." or "site." attribute which has
valid UTF-8 syntax.
The formal syntax for an attribute name is defined by the
"attribute-name" rule in the next section.
4.3. Formal Syntax for Dataset and Attribute Namespace
The naming conventions for datasets and attributes are defined by the
following ABNF. Note that this grammar is not part of the ACAP
protocol syntax in section 8, as dataset names and attribute names
are encoded as strings within the ACAP protocol.
attribute-dacl = "dataset.acl" *("." name-component)
attribute-dset = dataset-std 1*("." name-component)
MUST be defined in a dataset class specification
attribute-name = attribute-std / attr-site / attr-user / vendor-name
attribute-std = "entry" / "subdataset" / "modtime" /
"dataset.inherit" / attribute-dacl / attribute-dset
attr-site = "site" 1*("." name-component)
attr-user = "user." name-component 1*("." name-component)
byowner = "/byowner/" owner "/"
[dataset-class "/" dataset-sub]
dataset-class = dataset-std / vendor-name
dataset-normal = "/" [dataset-class "/"
(owner-prefix / dataset-tail)]
dataset-name = byowner / dataset-normal
dataset-std = name-component
MUST be registered with IANA and the spec MUST
be published as a standards track or
IESG-approved experimental RFC
dataset-sub = *(dname-component "/")
The rules for this portion of the namespace may
be further restricted by the dataset class
specification.
dataset-tail = owner "/" dataset-sub
dname-component = 1*UTF8-CHAR
MUST NOT begin with "." or contain "/"
name-component = 1*UTF8-CHAR
MUST NOT contain ".", "/", "%", or "*"
owner = "site" / owner-host / owner-group /
owner-user / "~"
owner-group = "group/" dname-component
owner-host = "host/" dname-component
owner-prefix = "group/" / "host/" / "user/"
owner-user = "user/" dname-component
vendor-name = vendor-token *("." name-component)
vendor-token = "vendor." name-component
MUST be registered with IANA
5. Dataset Management
The entry with an empty name ("") in the dataset is used to hold
management information for the dataset as a whole.
5.1. Dataset Inheritance
It is possible for one dataset to inherit data from another. The
dataset from which the data is inherited is called the base dataset.
Data in the base dataset appears in the inheriting dataset, except
when overridden by a STORE to the inheriting dataset.
The base dataset is usually a system-wide or group-wide set of
defaults. A system-wide dataset usually has one inheriting dataset
per user, allowing each user to add to or modify the defaults as
appropriate.
An entry which exists in both the inheriting and base dataset
inherits a modtime equal to the greater of the two modtimes. An
attribute in such an entry is inherited from the base dataset if it
was never modified by a STORE command in the inheriting dataset or if
DEFAULT was stored to that attribute. This permits default entries
to be amended rather than replaced in the inheriting dataset.
The "subdataset" attribute is not directly inherited. If the base
dataset includes a "subdataset" attribute and the inheriting dataset
does not, then the "subdataset" attribute will inherit a virtual
value of a list containing a ".". The subdataset at that node is
said to be a "virtual" dataset as it is simply a virtual copy of the
appropriate base dataset with all "subdataset" attributes changed to
a list containing a ".". A virtual dataset is not visible if
NOINHERIT is specified on the SEARCH command.
Servers MUST support at least two levels of inheritance. This
permits a user"s dataset such as "/options/user/fred/common" to
inherit from a group dataset such as "/options/group/dinosaur
operators/common" which in turn inherits from a server-wide dataset
such as "/options/site/common".
5.2. Dataset Attributes
The following attributes apply to management of the dataset when
stored in the "" entry of a dataset. These attributes are not
inherited.
dataset.acl
This holds the default access control list for the dataset.
This attribute is validated, so an invalid access control list
in a STORE command will result in a NO response with an INVALID
response code.
dataset.acl.<attribute>
This holds the default access control list for an attribute
within the dataset. This attribute is validated, so an invalid
access control list in a STORE command will result in a NO
response with an INVALID response code.
dataset.inherit
This holds the name of a dataset from which to inherit according
to the rules in the previous section. This attribute MAY refer
to a non-existent dataset, in which case nothing is inherited.
This attribute is validated, so illegal dataset syntax or an
attempt to store a multi-value will result in a NO response with
an INVALID response code.
5.3. Dataset Creation
When a dataset is first created (by storing a "." in the subdataset
attribute or storing an entry in a previously non-existent dataset),
the dataset attributes are initialized with the values from the
parent dataset in the "/byowner/" hierarchy. In the case of the
"dataset.inherit" attribute, the appropriate hierarchy component is
added. For example, given the following entry (note that t refers
to the US-ASCII horizontal tab character):
entry path "/byowner/user/joe/"
dataset.acl ("joetxrwia" "fredtxr")
dataset.inherit "/byowner/site"
If a new dataset class "/byowner/user/joe/new" is created, it will
have the following dataset attributes:
entry path "/byowner/user/joe/new/"
dataset.acl ("joetxrwia" "fredtxr")
dataset.inherit "/byowner/site/new"
Note that the dataset "/byowner/user/joe/new/" is equivalent to
"/new/user/joe/".
5.4. Dataset Class Capabilities
Certain dataset classes or dataset class features may only be useful
if there is an active updating client or integrated server support
for the feature. The dataset class "capability" is reserved to allow
clients or servers to advertise such features. The "entry" attribute
within this dataset class is the name of the dataset class whose
features are being described. The attributes are prefixed with
"capability.<dataset-class>." and are defined by the appropriate
dataset class specification.
Since it is possible for an unprivileged user to run an active client
for himself, a per-user capability dataset is useful. The dataset
"/capability/~/" holds information about all features available to
the user (via inheritance), and the dataset "/capability/site/" holds
information about all features supported by the site.
5.5. Dataset Quotas
Management and scope of quotas is implementation dependent. Clients
can check the applicable quota limit and usage (in bytes) with the
GETQUOTA command. Servers can notify the client of a low quota
situation with the QUOTA untagged response.
6. Command and Response Specifications
ACAP commands and responses are described in this section. Commands
are organized first by the state in which the command is permitted,
then by a general category of command type.
Command arguments, identified by "Arguments:" in the command
descriptions below, are described by function, not by syntax. The
precise syntax of command arguments is described in the Formal Syntax
section.
Some commands cause specific server data to be returned; these are
identified by "Data:" in the command descriptions below. See the
response descriptions in the Responses section for information on
these responses, and the Formal Syntax section for the precise syntax
of these responses. It is possible for server data to be transmitted
as a result of any command; thus, commands that do not specifically
require server data specify "no specific data for this command"
instead of "none".
The "Result:" in the command description refers to the possible
tagged status responses to a command, and any special interpretation
of these status responses.
6.1. Initial Connection
Upon session startup, the server sends one of two untagged responses:
ACAP or BYE. The untagged BYE response is described in section
6.2.8.
6.1.1. ACAP Untagged Response
Data: capability list
The untagged ACAP response indicates the session is ready to
accept commands and contains a space-separated listing of
capabilities that the server supports. Each capability is
represented by a list containing the capability name optionally
followed by capability specific string arguments.
ACAP capability names MUST be registered with IANA according to
the rules in section 7.1.
Client implementations SHOULD NOT require any capability name
beyond those defined in this specification, and MUST tolerate any
unknown capability names. A client implementation MAY be
configurable to require SASL mechanisms other than CRAM-MD5
[CRAM-MD5] for site security policy reasons.
The following initial capabilities are defined:
CONTEXTLIMIT
The CONTEXTLIMIT capability has one argument which is a
number describing the maximum number of contexts the server
supports per connection. The number 0 indicates the server
has no limit, otherwise this number MUST be greater than
100.
IMPLEMENTATION
The IMPLEMENTATION capability has one argument which is a
string describing the server implementation. ACAP clients
MUST NOT alter their behavior based on this value. It is
intended primarily for debugging purposes.
SASL The SASL capability includes a list of the authentication
mechanisms supported by the server. See section 6.3.1.
Example: S: * ACAP (IMPLEMENTATION "ACME v3.5")
(SASL "CRAM-MD5") (CONTEXTLIMIT "200")
6.2. Any State
The following commands and responses are valid in any state.
6.2.1. NOOP Command
Arguments: none
Data: no sp