[an error occurred while processing this directive]
INTERNET-DRAFT                                           Ken A L Coar
draft-coar-cgi-v11-03.{html,txt}                      IBM Corporation
                                                      D.R.T. Robinson
                                                      E*TRADE UK Ltd.
                                                        25 June 1999

The WWW Common Gateway Interface
Version 1.1

[an error occurred while processing this directive]

Abstract

The Common Gateway Interface (CGI) is a simple interface for running external programs, software or gateways under an information server in a platform-independent manner. Currently, the supported information servers are HTTP servers.

The interface has been in use by the World-Wide Web since 1993. This specification defines the "current practice" parameters of the 'CGI/1.1' interface developed and documented at the U.S. National Centre for Supercomputing Applications [NCSA-CGI]. This document also defines the use of the CGI/1.1 interface on the Unix and AmigaDOS(tm) systems.

Discussion of this draft occurs on the CGI-WG mailing list; see the project Web page at <URL:http://CGI-Spec.Golux.Com/> for details on the mailing list and the status of the project.

[an error occurred while processing this directive]

Revision History

The revision history of this draft is being maintained using Web-based GUI notation, such as struck-through characters and colour-coded sections. The following legend describes how to determine the origin of a particular revision according to the colour of the text:

Black
Revision 00, released 28 May 1998
Green
Revision 01, released 28 December 1998
Major structure change: Section 4, "Request Metadata (Meta-Variables)" was moved entirely under Section 7, "Data Input to the CGI Script." Due to the size of this change, it is noted here and the text in its former location does not appear as struckthrough. This has caused major sections 5 and following to decrement by one. Other large text movements are likewise not marked up. References to RFC 1738 were changed to 2396 (1738's replacement).
Red
Revision 02, released 2 April, 1999
Added text to section 8.3 defining correct handling of HTTP/1.1 requests using "chunked" Transfer-Encoding. Labelled metavariable names in section 8 with the appropriate detail section numbers. Clarified allowed usage of Status and Location response header fields. Included new Internet-Draft language.
Fuchsia
Revision 03, released 25 June 1999
Changed references from "HTTP" to "Protocol-Specific" for the listing of things like HTTP_ACCEPT. Changed 'entity-body' and 'content-body' to 'message-body.' Added a note that response headers must comply with requirements of the protocol level in use. Added a lot of stuff about security (section 11). Clarified a bunch of productions. Pointed out that zero-length and omitted values are indistinguishable in this specification. Clarified production describing order of fields in script response header. Clarified issues surrounding encoding of data. Acknowledged additional contributors, and changed one of the authors' addresses.

Table of Contents

  1 Introduction..............................................TBD
   1.1 Purpose................................................TBD
   1.2 Requirements...........................................TBD
   1.3 Specifications.........................................TBD
   1.4 Terminology............................................TBD
  2 Notational Conventions and Generic Grammar................TBD
   2.1 Augmented BNF..........................................TBD
   2.2 Basic Rules............................................TBD
  3 Protocol Parameters.......................................TBD
   3.1 URL Encoding...........................................TBD
   3.2 The Script-URI.........................................TBD
  4 Invoking the Script.......................................TBD
  5 The CGI Script Command Line...............................TBD
  6 Data Input to the CGI Script..............................TBD
   6.1 Request Metadata (Metavariables).......................TBD
    6.1.1 AUTH_TYPE...........................................TBD
    6.1.2 CONTENT_LENGTH......................................TBD
    6.1.3 CONTENT_TYPE........................................TBD
    6.1.4 GATEWAY_INTERFACE...................................TBD
    6.1.5 Protocol-Specific Metavariables.....................TBD
    6.1.6 PATH_INFO...........................................TBD
    6.1.7 PATH_TRANSLATED.....................................TBD
    6.1.8 QUERY_STRING........................................TBD
    6.1.9 REMOTE_ADDR.........................................TBD
    6.1.10 REMOTE_HOST........................................TBD
    6.1.11 REMOTE_IDENT.......................................TBD
    6.1.12 REMOTE_USER........................................TBD
    6.1.13 REQUEST_METHOD.....................................TBD
    6.1.14 SCRIPT_NAME........................................TBD
    6.1.15 SERVER_NAME........................................TBD
    6.1.16 SERVER_PORT........................................TBD
    6.1.17 SERVER_PROTOCOL....................................TBD
    6.1.18 SERVER_SOFTWARE....................................TBD
    6.2 Request Message-Bodies................................TBD
  7 Data Output from the CGI Script...........................TBD
   7.1 Non-Parsed Header Output...............................TBD
   7.2 Parsed Header Output...................................TBD
    7.2.1 CGI header fields...................................TBD
     7.2.1.1 Content-Type.....................................TBD
     7.2.1.2 Location.........................................TBD
     7.2.1.3 Status...........................................TBD
     7.2.1.4 Extension header fields..........................TBD
    7.2.2 HTTP header fields..................................TBD
  8 Server Implementation.....................................TBD
   8.1 Requirements for Servers...............................TBD
    8.1.1 Script-URI..........................................TBD
    8.1.2 Request Message-body Handling.......................TBD
    8.1.3 Required Metavariables..............................TBD
    8.1.4 Response Compliance.................................TBD
   8.2 Recommendations for Servers............................TBD
   8.3 Summary of Metavariables...............................TBD
  9 Script Implementation.....................................TBD
   9.1 Requirements for Scripts...............................TBD
   9.2 Recommendations for Scripts............................TBD
  10 System Specifications....................................TBD
   10.1 AmigaDOS..............................................TBD
   10.2 Unix..................................................TBD
  11 Security Considerations..................................TBD
   11.1 Safe Methods..........................................TBD
   11.2 HTTP Header Fields Containing Sensitive Information...TBD
   11.3 Script Interference with the Server...................TBD
   11.4 Data Length and Buffering Considerations..............TBD
   11.5 Stateless Processing..................................TBD
  12 Acknowledgments..........................................TBD
  13 References...............................................TBD
  14 Authors' Addresses.......................................TBD
     

1. Introduction

1.1. Purpose

Together the HTTP [3,8] server and the CGI script are responsible for servicing a client request by sending back responses. The client request comprises a Universal Resource Identifier (URI) [1], a request method, and various ancillary information about the request provided by the transport mechanism.

The CGI defines the abstract parameters, known as metavariables, which describe the client's request. Together with a concrete programmer interface this specifies a platform-independent interface between the script and the HTTP server.

1.2. Requirements

This specification uses the same words as RFC 1123 [5] to define the significance of each particular requirement. These are:

[an error occurred while processing this directive]

MUST

This word or the adjective 'required' means that the item is an absolute requirement of the specification.

SHOULD

This word or the adjective 'recommended' means that there may exist valid reasons in particular circumstances to ignore this item, but the full implications should be understood and the case carefully weighed before choosing a different course.

MAY

This word or the adjective 'optional' means that this item is truly optional. One vendor may choose to include the item because a particular marketplace requires it or because it enhances the product, for example; another vendor may omit the same item.

An implementation is not compliant if it fails to satisfy one or more of the 'must' requirements for the protocols it implements. An implementation that satisfies all of the 'must' and all of the 'should' requirements for its features is said to be 'unconditionally compliant'; one that satisfies all of the 'must' requirements but not all of the 'should' requirements for its features is said to be 'conditionally compliant.'

1.3. Specifications

Not all of the functions and features of the CGI are defined in the main part of this specification. The following phrases are used to describe the features which are not specified:

system defined

The feature may differ between systems, but must be the same for different implementations using the same system. A system will usually identify a class of operating-systems. Some systems are defined in section 10 of this document. New systems may be defined by new specifications without revision of this document.

implementation defined

The behaviour of the feature may vary from implementation to implementation, but a particular implementation must document its behaviour.

1.4. Terminology

This specification uses many terms defined in the HTTP/1.1 specification [8]; however, the following terms are used here in a sense which may not accord with their definitions in that document, or with their common meaning.

metavariable

A named parameter that carries information from the server to the script. It is not necessarily a variable in the operating-system's environment, although that is the most common implementation.

script

The software which is invoked by the server via this interface. It need not be a standalone program, but could be a dynamically-loaded or shared library, or even a subroutine in the server. It may be a set of statements interpreted at run-time, as the term 'script' is frequently understood, but that is not a requirement and within the context of this specification the term has the broader definition stated.

server

The application program which invokes the script in order to service requests.

2. Notational Conventions and Generic Grammar

2.1. Augmented BNF

All of the mechanisms specified in this document are described in both prose and an augmented Backus-Naur Form (BNF) similar to that used by RFC 822 [6]. This augmented BNF contains the following constructs:

name = definition

The definition by the equal character ("="). Whitespace is only significant in that continuation lines of a definition are indented.

"literal"

Quotation marks (") surround literal text, except for a literal quotation mark, which is surrounded by angle-brackets ("<" and ">"). Unless stated otherwise, the text is case-sensitive.

rule1 | rule2

Alternative rules are separated by a vertical bar ("|").

(rule1 rule2 rule3)

Elements enclosed in parentheses are treated as a single element.

*rule

A rule preceded by an asterisk ("*") may have zero or more occurrences. A rule preceded by an integer followed by an asterisk must occur at least the specified number of times.

[rule]

An element enclosed in square brackets ("[" and "]") is optional.

2.2. Basic Rules

The following rules are used throughout this specification to describe basic parsing constructs.

[an error occurred while processing this directive]

    alpha         = lowalpha | hialpha
    alphanum      = alpha | digit
    lowalpha      = "a" | "b" | "c" | "d" | "e" | "f" | "g" | "h"
                    | "i" | "j" | "k" | "l" | "m" | "n" | "o" | "p"
                    | "q" | "r" | "s" | "t" | "u" | "v" | "w" | "x"
                    | "y" | "z"
    hialpha       = "A" | "B" | "C" | "D" | "E" | "F" | "G" | "H"
                    | "I" | "J" | "K" | "L" | "M" | "N" | "O" | "P"
                    | "Q" | "R" | "S" | "T" | "U" | "V" | "W" | "X"
                    | "Y" | "Z"
    digit         = "0" | "1" | "2" | "3" | "4" | "5" | "6" | "7"
                    | "8" | "9"
    hex           = digit | "A" | "B" | "C" | "D" | "E" | "F" | "a"
                    | "b" | "c" | "d" | "e" | "f"
    escaped       = "%" hex hex
    OCTET         = <any 8-bit sequence of data>
    CHAR          = <any US-ASCII character (octets 0 - 127)>
    CTL           = <any US-ASCII control character
                    (octets 0 - 31) and DEL (127)>
    CR            = <US-ASCII CR, carriage return (13)>
    LF            = <US-ASCII LF, linefeed (10)>
    SP            = <US-ASCII SP, space (32)>
    HT            = <US-ASCII HT, horizontal tab (9)>
    NL            = CR | LF
    LWSP          = SP | HT | NL
    tspecial      = "(" | ")" | "@" | "," | ";" | ":" | "\" | <">
                    | "/" | "[" | "]" | "?" | "<" | ">" | "{" | "}"
                    | SP | HT | NL
    token         = 1*<any CHAR except CTLs or tspecials>
    quoted-string = ( <"> *qdtext <"> ) | ( "<" *qatext ">")
    qdtext        = <any CHAR except <"> and CTLs but including LWSP>
    qatext        = <any CHAR except "<", ">" and CTLs but
                    including LWSP>
    mark          = "-" | "_" | "." | "!" | "~" | "*" | "'" | "(" | ")"
    unreserved    = alphanum | mark
    reserved      = ";" | "/" | "?" | ":" | "@" | "&" | "=" |
                    "$" | ","
    uric          = reserved | unreserved | escaped
  

Note that newline (NL) need not be a single character, but can be a character sequence.

3. Protocol Parameters

3.1. URL Encoding

Some variables and constructs used here are described as being 'URL-encoded'. This encoding is described in section 2 of RFC 2396 [4].

An alternate "shortcut" encoding for representing the space character exists and is in common use. Scripts MUST be prepared to recognise both '+' and '%20' as an encoded space in a URL-encoded value.

Note that some unsafe characters may have different semantics if they are encoded. The definition of which characters are unsafe depends on the context. For example, the following two URLs do not necessarily refer to the same resource:

[an error occurred while processing this directive]

    http://somehost.com/somedir%2Fvalue
    http://somehost.com/somedir/value
  

See section 2 of RFC 2396 [4] for authoritative treatment of this issue.

3.2. The Script-URI

The 'Script-URI' is defined as the URI of the resource identified by the metavariables. Often, this URI will be the same as the URI requested by the client (the 'Client-URI'); however, it need not be. Instead, it could be a URI invented by the server, and so it can only be used in the context of the server and its CGI interface.

The Script-URI has the syntax of generic-RL as defined in section 2.1 of RFC 1808 [7], with the exception that object parameters and fragment identifiers are not permitted:

[an error occurred while processing this directive]

    <scheme>://<host><port>/<path>?<query>
  

The various components of the Script-URI are defined by some of the metavariables (see section 4 below);

[an error occurred while processing this directive]

    script-uri = protocol "://" SERVER_NAME ":" SERVER_PORT enc-script
                 enc-path-info "?" QUERY_STRING
  

where 'protocol' is obtained from SERVER_PROTOCOL, 'enc-script' is a URL-encoded version of SCRIPT_NAME and 'enc-path-info' is a URL-encoded version of PATH_INFO. See section 4.6 for more information about the PATH_INFO metavariable.

Note that the scheme and the protocol are not identical; for instance, a resource accessed via an SSL mechanism may have a Client-URI with a scheme of "https" rather than "http". CGI/1.1 provides no means for the script to reconstruct this, and therefore the Script-URI includes the base protocol used.

4. Invoking the Script

The script is invoked in a system defined manner. Unless specified otherwise, the file containing the script will be invoked as an executable program.

5. The CGI Script Command Line

Some systems support a method for supplying an array of strings to the CGI script. This is only used in the case of an 'indexed' query. This is identified by a "GET" or "HEAD" HTTP request with a URL query string not containing any unencoded "=" characters. For such a request, servers SHOULD parse the search string into words, using the following rules:

[an error occurred while processing this directive]

    search-string = search-word *( "+" search-word )
    search-word   = 1*schar
    schar         = xunreserved | escaped | xreserved
    xunreserved   = alpha | digit | xsafe | extra
    xsafe         = "$" | "-" | "_" | "."
    xreserved     = ";" | "/" | "?" | ":" | "@" | "&"
  

After parsing, each word is URL-decoded, optionally encoded in a system defined manner, and then the argument list is set to the list of words.

If the server cannot create any part of the argument list, then the server SHOULD NOT generate any command line information. For example, the number of arguments may be greater than operating system or server limitations permit, or one of the words may not be representable as an argument.

Scripts SHOULD check to see if the QUERY_STRING value contains an unencoded "=" character, and SHOULD NOT use the command line arguments if it does.

6. Data Input to the CGI Script

Information about a request comes from two different sources: the request header, and any associated message-body. Servers MUST make portions of this information available to scripts.

6.1. Request Metadata (Metavariables)

Each CGI server implementation MUST define a mechanism to pass data about the request from the server to the script. The metavariables containing these data are accessed by the script in a system defined manner. The representation of the characters in the metavariables is system defined.

This specification does not distinguish between the representation of null values and missing ones. Whether null or missing values (such as a query component of "?" or "", respectively) are represented by undefined metavariables or by metavariables with values of "" is implementation-defined.

Case is not significant in the metavariable names, in that there cannot be two different variables whose names differ in case only. Here they are shown using a canonical representation of capitals plus underscore ("_"). The actual representation of the names is system defined; for a particular system the representation MAY be defined differently than this.

Metavariable values MUST be considered case-sensitive except as noted otherwise.

The canonical metavariables defined by this specification are:

[an error occurred while processing this directive]

    AUTH_TYPE
    CONTENT_LENGTH
    CONTENT_TYPE
    GATEWAY_INTERFACE
    PATH_INFO
    PATH_TRANSLATED
    QUERY_STRING
    REMOTE_ADDR
    REMOTE_HOST
    REMOTE_IDENT
    REMOTE_USER
    REQUEST_METHOD
    SCRIPT_NAME
    SERVER_NAME
    SERVER_PORT
    SERVER_PROTOCOL
    SERVER_SOFTWARE
  

Metavariables with names beginning with the protocol name (e.g., "HTTP_ACCEPT") are also canonical in their description of request header fields. The number and meaning of these fields may change independently of this specification. (See also section 6.1.5.)

6.1.1. AUTH_TYPE

This variable is specific to requests made via the "http" scheme.

If the Script-URI required access authentication for external access, then the server MUST set the value of this variable from the 'auth-scheme' token in the request's "Authorization" header field. Otherwise it is set to NULL.

[an error occurred while processing this directive]

    AUTH_TYPE   = "" | auth-scheme
    auth-scheme = "Basic" | "Digest" | token
  

HTTP access authentication schemes are described in section 11 of the HTTP/1.1 specification [8]. The auth-scheme is not case-sensitive.

Servers MUST provide this metavariable to scripts if the request header included an "Authorization" field that was authenticated.

6.1.2. CONTENT_LENGTH

This metavariable is set to the size of the message-body entity attached to the request, if any, in decimal number of octets. If no data are attached, then this metavariable is either NULL or not defined. The syntax is the same as for the HTTP "Content-Length" header field (section 14.14, HTTP/1.1 specification [8]).

[an error occurred while processing this directive]

    CONTENT_LENGTH = "" | 1*digit
  

Servers MUST provide this metavariable to scripts if the request was accompanied by a message-body entity.

6.1.3. CONTENT_TYPE

If the request includes a message-body, CONTENT_TYPE is set to the Internet Media Type [9] of the attached entity if the type was provided via a "Content-type" field in the request header, or if the server can determine it in the absence of a supplied "Content-type" field. The syntax is the same as for the HTTP "Content-Type" header field.

[an error occurred while processing this directive]

    CONTENT_TYPE = "" | media-type
    media-type   = type "/" subtype *( ";" parameter)
    type         = token
    subtype      = token
    parameter    = attribute "=" value
    attribute    = token
    value        = token | quoted-string
  

The type, subtype, and parameter attribute names are not case-sensitive. Parameter values MAY be case sensitive. Media types and their use in HTTP are described in section 3.7 of the HTTP/1.1 specification [8].

Example:

[an error occurred while processing this directive]

    application/x-www-form-urlencoded
  

There is no default value for this variable. If and only if it is unset, then the script MAY attempt to determine the media type from the data received. If the type remains unknown, then the script MAY choose to either assume a content-type of application/octet-stream or reject the request with a 415 ("Unsupported Media Type") error. See section 7.2.1.3 for more information about returning error status values.

Servers MUST provide this metavariable to scripts if a "Content-Type" field was present in the original request header. If the server receives a request with an attached entity but no "Content-Type" header field, it MAY attempt to determine the correct datatype, or it MAY omit this metavariable when communicating the request information to the script.

6.1.4. GATEWAY_INTERFACE

This metavariable is set to the dialect of CGI being used by the server to communicate with the script. Syntax:

[an error occurred while processing this directive]

    GATEWAY_INTERFACE = "CGI" "/" major "." minor
    major             = 1*digit
    minor             = 1*digit
  

Note that the major and minor numbers are treated as separate integers and hence each may be more than a single digit. Thus CGI/2.4 is a lower version than CGI/2.13 which in turn is lower than CGI/12.3. Leading zeros in either the major or the minor number MUST be ignored by scripts and SHOULD NOT be generated by servers.

This document defines the 1.1 version of the CGI interface ("CGI/1.1").

Servers MUST provide this metavariable to scripts.

6.1.5. Protocol-Specific Metavariables

These metavariables are specific to the protocol via which the request is made. Interpretation of these variables depends on the value of the SERVER_PROTOCOL metavariable (see section 6.1.17).

Metavariables with names beginning with "HTTP_" contain values from the request header, if the scheme used was HTTP. Each HTTP header field name is converted to upper case, has all occurrences of "-" replaced with "_", and has "HTTP_" prepended to form the metavariable name. Similar transformations are applied for other protocols. The header data MAY be presented as sent by the client, or MAY be rewritten in ways which do not change its semantics. If multiple header fields with the same field-name are received then the server MUST rewrite them as though they had been received as a single header field having the same semantics before being represented in a metavariable. Similarly, a header field that is received on more than one line MUST be merged into a single line. The server MUST, if necessary, change the representation of the data (for example, the character set) to be appropriate for a CGI metavariable.

Servers are not required to create metavariables for all the request header fields that they receive. In particular, they MAY decline to make available any header fields carrying authentication information, such as "Authorization", or which are available to the script via other metavariables, such as "Content-Length" and "Content-Type".

6.1.6. PATH_INFO

The PATH_INFO metavariable specifies a path to be interpreted by the CGI script. It identifies the resource or sub-resource to be returned by the CGI script, and it is derived from the portion of the URI path following the script name but preceding any query data. The syntax and semantics are similar to a decoded HTTP URL 'path' token (defined in RFC 2396 [4]), with the exception that a PATH_INFO of "/" represents a single void path segment.

[an error occurred while processing this directive]

    PATH_INFO = "" | ( "/" path )
    path      = segment *( "/" segment )
    segment   = *pchar
    pchar     = <any CHAR except "/">
  

The PATH_INFO string is the trailing part of the <path> component of the Script-URI (see section 3.2) that follows the SCRIPT_NAME portion of the path.

Servers MAY impose their own restrictions and limitations on what values they will accept for PATH_INFO, and MAY reject or edit any values they consider objectionable before passing them to the script.

Servers MUST make this URI component available to CGI scripts. The PATH_INFO value is case-sensitive, and the server MUST preserve the case of the PATH_INFO element of the URI when making it available to scripts.

6.1.7. PATH_TRANSLATED

PATH_TRANSLATED is derived by taking any path-info component of the request URI (see section 6.1.6), decoding it (see section 3.1), parsing it as a URI in its own right, and performing any virtual-to-physical translation appropriate to map it onto the server's document repository structure. If the request URI includes no path-info component, the PATH_TRANSLATED metavariable SHOULD NOT be defined.

[an error occurred while processing this directive]

    PATH_TRANSLATED = *CHAR
  

For a request such as the following:

[an error occurred while processing this directive]

    http://somehost.com/cgi-bin/somescript/this%2eis%2epath%2einfo
  

the PATH_INFO component would be decoded, and the result parsed as though it were a request for the following:

[an error occurred while processing this directive]

    http://somehost.com/this.is.the.path.info
  

This would then be translated to a location in the server's document repository, perhaps a filesystem path something like this:

[an error occurred while processing this directive]

    /usr/local/www/htdocs/this.is.the.path.info
  

The result of the translation is the value of PATH_TRANSLATED.

The value of PATH_TRANSLATED may or may not map to a valid repository location. Servers MUST preserve the case of the path-info segment if and only if the underlying repository supports case-sensitive names. If the repository is only case-aware, case-preserving, or case-blind with regard to document names, servers are not required to preserve the case of the original segment through the translation.

The translation algorithm the server uses to derive PATH_TRANSLATED is implementation defined; CGI scripts which use this variable may suffer limited portability.

Servers SHOULD provide this metavariable to scripts if and only if the request URI includes a path-info component.

6.1.8. QUERY_STRING

A URL-encoded string; the <query> part of the Script-URI. (See section 3.2.)

[an error occurred while processing this directive]

    QUERY_STRING = query-string
    query-string = *uric
  

The URL syntax for a query string is described in section 3 of RFC 2396 [4].

Servers MUST supply this value to scripts. The QUERY_STRING value is case-sensitive. If the Script-URI does not include a query component, the QUERY_STRING metavariable MUST be defined as an empty string ("").

6.1.9. REMOTE_ADDR

The IP address of the client sending the request to the server. This is not necessarily that of the user agent (such as if the request came through a proxy).

[an error occurred while processing this directive]

    REMOTE_ADDR  = hostnumber
    hostnumber   = ipv4-address | ipv6-address
  

The definitions of ipv4-address and ipv6-address are provided in Appendix B of RFC 2373 [13].

Servers MUST supply this value to scripts.

6.1.10. REMOTE_HOST

The fully qualified domain name of the client sending the request to the server, if available, otherwise NULL. (See section 6.1.9.) Fully qualified domain names take the form as described in section 3.5 of RFC 1034 [10] and section 2.1 of RFC 1123 [5]. Domain names are not case sensitive.

Servers SHOULD provide this information to scripts.

6.1.11. REMOTE_IDENT

The identity information reported about the connection by a RFC 1413 [11] request to the remote agent, if available. Servers MAY choose not to support this feature, or not to request the data for efficiency reasons.

[an error occurred while processing this directive]

    REMOTE_IDENT = *CHAR
  

The data returned may be used for authentication purposes, but the level of trust reposed in them should be minimal.

Servers MAY supply this information to scripts if the RFC1413 [11] lookup is performed.

6.1.12. REMOTE_USER

If the request required authentication using the "Basic" mechanism (i.e., the AUTH_TYPE metavariable is set to "Basic"), then the value of the REMOTE_USER metavariable is set to the user-ID supplied. In all other cases the value of this metavariable is undefined.

[an error occurred while processing this directive]

    REMOTE_USER = *OCTET
  

This variable is specific to requests made via the HTTP protocol.

Servers SHOULD provide this metavariable to scripts.

6.1.13. REQUEST_METHOD

The REQUEST_METHOD metavariable is set to the method with which the request was made, as described in section 5.1.1 of the HTTP/1.0 specification [3] and section 5.1.1 of the HTTP/1.1 specification [8].

[an error occurred while processing this directive]

    REQUEST_METHOD   = http-method
    http-method      = "GET" | "HEAD" | "POST" | "PUT" | "DELETE"
                       | "OPTIONS" | "TRACE" | extension-method
    extension-method = token
  

The method is case sensitive. CGI/1.1 servers MAY choose to process some methods directly rather than passing them to scripts.

This variable is specific to requests made with HTTP.

Servers MUST provide this metavariable to scripts.

6.1.14. SCRIPT_NAME

The SCRIPT_NAME metavariable is set to a URL path that could identify the CGI script (rather than the script's output). The syntax and semantics are identical to a decoded HTTP URL 'path' token (see RFC 2396 [4]).

[an error occurred while processing this directive]

    SCRIPT_NAME = "" | ( "/" [ path ] )
  

The SCRIPT_NAME string is some leading part of the <path> component of the Script-URI derived in some implementation defined manner. No PATH_INFO or QUERY_STRING segments (see sections 6.1.6 and 6.1.8) are included in the SCRIPT_NAME value.

Servers MUST provide this metavariable to scripts.

6.1.15. SERVER_NAME

The SERVER_NAME metavariable is set to the name of the server, as derived from the <host> part of the Script-URI (see section 3.2).

[an error occurred while processing this directive]

    SERVER_NAME = hostname | hostnumber
  

Servers MUST provide this metavariable to scripts.

6.1.16. SERVER_PORT

The SERVER_PORT metavariable is set to the port on which the request was received, as used in the <port> part of the Script-URI.

[an error occurred while processing this directive]

    SERVER_PORT = 1*digit
  

If the <port> portion of the script-URI is blank, the actual port number upon which the request was received MUST be supplied.

Servers MUST provide this metavariable to scripts.

6.1.17. SERVER_PROTOCOL

The SERVER_PROTOCOL metavariable is set to the name and revision of the information protocol with which the request arrived. This is not necessarily the same as the protocol version used by the server in its response to the client.

[an error occurred while processing this directive]

    SERVER_PROTOCOL   = HTTP-Version | extension-version
                        | extension-token
    HTTP-Version      = "HTTP" "/" 1*digit "." 1*digit
    extension-version = protocol "/" 1*digit "." 1*digit
    protocol          = 1*( alpha | digit | "+" | "-" | "." )
    extension-token   = token
  

'protocol' is a version of the <scheme> part of the Script-URI, but is not identical to it. For example, the scheme of a request may be "https" while the protocol remains "http". The protocol is not case sensitive, but by convention, 'protocol' is in upper case.

A well-known extension token value is "INCLUDED", which signals that the current document is being included as part of a composite document, rather than being the direct target of the client request.

Servers MUST provide this metavariable to scripts.

6.1.18. SERVER_SOFTWARE

The SERVER_SOFTWARE metavariable is set to the name and version of the information server software answering the request (and running the gateway).

[an error occurred while processing this directive]

    SERVER_SOFTWARE = 1*product
    product         = token [ "/" product-version ]
    product-version = token
  

Servers MUST provide this metavariable to scripts.

6.2. Request Message-Bodies

As there may be a data entity attached to the request, there MUST be a system defined method for the script to read these data. Unless defined otherwise, this will be via the 'standard input' file descriptor.

If the CONTENT_LENGTH value (see section 6.1.2) is non-NULL, the server MUST supply at least that many bytes to scripts on the standard input stream. Scripts are not obliged to read the data. Servers MAY signal an EOF condition after CONTENT_LENGTH bytes have been read, but are not obligated to do so. Therefore, scripts MUST NOT attempt to read more than CONTENT_LENGTH bytes, even if more data are available.

For non-parsed header (NPH) scripts (see section 7.1 below), servers SHOULD attempt to ensure that the data supplied to the script are precisely as supplied by the client and unaltered by the server.

Section 8.1.2 describes the requirements of servers with regard to requests that include message-bodies.

7. Data Output from the CGI Script

There MUST be a system defined method for the script to send data back to the server or client; a script MUST always return some data. Unless defined otherwise, this will be via the 'standard output' file descriptor.

There are two forms of output that scripts can supply to servers: non-parsed header (NPH) output, and parsed header output. Servers MUST support parsed header output and MAY support NPH output. The method of distinguishing between the two types of output (or scripts) is implementation defined.

Servers MAY implement a timeout period within which data must be received from scripts. If a server implementation defines such a timeout and receives no data from a script within the timeout period, the server MAY terminate the script process and SHOULD abort the client request with either a '504 Gateway Timed Out' or a '500 Internal Server Error' response.

7.1. Non-Parsed Header Output

Scripts using the NPH output form MUST return a complete HTTP response message, as described in Section 6 of the HTTP specifications [3,8]. NPH scripts MUST use the SERVER_PROTOCOL variable to determine the appropriate format for a response.

Servers SHOULD attempt to ensure that the script output is sent directly to the client, with minimal internal and no transport-visible buffering.

7.2. Parsed Header Output

Scripts using the parsed header output form MUST supply a CGI response message to the server as follows:

[an error occurred while processing this directive]

    CGI-Response   = *optional-field CGI-Field *optional-field NL [ Message-Body ]
    optional-field = ( CGI-Field | HTTP-Field )
    CGI-Field      = Content-type
                   | Location
                   | Status
                   | extension-header
  

The response comprises a header and a body, separated by a blank line. The body may be NULL. The header fields are either CGI header fields to be interpreted by the server, or HTTP header fields to be included in the response returned to the client if the request method is HTTP. At least one CGI-Field MUST be supplied, but no CGI field name may be used more than once in a response. If a body is supplied, then a "Content-type" header field MUST be supplied by the script, otherwise the script MUST send a "Location" or "Status" header field. If a Location CGI-Field is returned, then the script MUST NOT supply any HTTP-Fields.

Each header field in a CGI-Response MUST be specified on a single line; CGI/1.1 does not support continuation lines.

7.2.1. CGI header fields

The CGI header fields have the generic syntax:

[an error occurred while processing this directive]

    generic-field  = field-name ":" [ field-value ] NL
    field-name     = token
    field-value    = *( field-content | LWSP )
    field-content  = *( token | tspecial | quoted-string )
  

The field-name is not case sensitive; a NULL field value is equivalent to the header field not being sent.

7.2.1.1. Content-Type

The Internet Media Type [9] of the entity body, which is to be sent unmodified to the client.

[an error occurred while processing this directive]

    Content-Type = "Content-Type" ":" media-type NL
  

This is actually an HTTP-Field rather than a CGI-Field, but it is listed here because of its importance in the CGI dialogue as a member of the "one of these is required" set of header fields.

7.2.1.2. Location

This is used to specify to the server that the script is returning a reference to a document rather than an actual document.

[an error occurred while processing this directive]

    Location         = "Location" ":"
                       ( fragment-URI | rel-URL-abs-path ) NL
    fragment-URI     = URI [ # fragmentid ]
    URI              = scheme ":" *qchar
    fragmentid       = *qchar
    rel-URL-abs-path = "/" [ hpath ] [ "?" query-string ]
    hpath            = fpsegment *( "/" psegment )
    fpsegment        = 1*hchar
    psegment         = *hchar
    hchar            = alpha | digit | safe | extra
                       | ":" | "@" | "& | "="
  

The Location value is either an absolute URI with optional fragment, as defined in RFC 1630 [1], or an absolute path within the server's URI space (i.e., omitting the scheme and network-related fields) and optional query-string. If an absolute URI is returned by the script, then the server MUST generate a '302 redirect' HTTP response message unless the script has supplied an explicit Status response header field. Scripts returning an absolute URI MAY choose to provide a message-body. Servers MUST make any appropriate modifications to the script's output to ensure the response to the user-agent complies with the response protocol version. If the Location value is a path, then the server MUST generate the response that it would have produced in response to a request containing the URL

[an error occurred while processing this directive]

    scheme "://" SERVER_NAME ":" SERVER_PORT rel-URL-abs-path
  

Note: If the request was accompanied by a message-body (such as for a POST request), and the script redirects the request with a Location field, the message-body may not be available to the resource that is the target of the redirect.

7.2.1.3. Status

The "Status" header field is used to indicate to the server what status code the server MUST use in the response message.

[an error occurred while processing this directive]

    Status        = "Status" ":" digit digit digit SP reason-phrase NL
    reason-phrase = *<CHAR, excluding CTLs, NL>
  

The valid status codes are listed in section 6.1.1 of the HTTP/1.0 specifications [3]. If the SERVER_PROTOCOL is "HTTP/1.1", then the status codes defined in the HTTP/1.1 specification [8] may be used. If the script does not return a "Status" header field, then "200 OK" SHOULD be assumed by the server.

If a script is being used to handle a particular error or condition encountered by the server, such as a '404 Not Found' error, the script SHOULD use the "Status" CGI header field to propagate the error condition back to the client. E.g., in the example mentioned it SHOULD include a "Status: 404 Not Found" in the header data returned to the server.

7.2.1.4. Extension header fields

Scripts MAY include in their CGI response header additional fields not defined in this or the HTTP specification. These are called "extension" fields, and have the syntax of a generic-field as defined in section 7.2.1. The name of an extension field MUST NOT conflict with a field name defined in this or any other specification; extension field names SHOULD begin with "X-CGI-" to ensure uniqueness.

7.2.2. HTTP header fields

The script MAY return any other header fields defined by the specification for the SERVER_PROTOCOL (HTTP/1.0 [3] or HTTP/1.1 [8]). Servers MUST resolve conflicts beteen CGI header and HTTP header formats or names (see section 8).

8. Server Implementation

This section defines the requirements that must be met by HTTP servers in order to provide a coherent and correct CGI/1.1 environment in which scripts may function. It is intended primarily for server implementors, but it is useful for script authors to be familiar with the information as well.

8.1. Requirements for Servers

In order to be considered CGI/1.1-compliant, a server must meet certain basic criteria and provide certain minimal functionality. The details of these requirements are described in the following sections.

8.1.1. Script-URI

Servers MUST support the standard mechanism (described below) which allows script authors to determine what URL to use in documents which reference the script; specifically, what URL to use in order to achieve particular settings of the metavariables. This mechanism is as follows:

The server MUST translate the header data from the CGI header field syntax to the HTTP header field syntax if these differ. For example, the character sequence for newline (such as Unix's ASCII NL) used by CGI scripts may not be the same as that used by HTTP (ASCII CR followed by LF). The server MUST also resolve any conflicts between header fields returned by the script and header fields that it would otherwise send itself.

8.1.2. Request Message-body Handling

These are the requirements for server handling of message-bodies directed to CGI/1.1 resources:

  1. The message-body the server provides to the CGI script MUST have any transfer encodings removed.
  2. The server MUST derive and provide a value for the CONTENT_LENGTH metavariable that reflects the length of the message-body after any transfer decoding.
  3. The server MUST leave intact any content-encodings of the message-body.

8.1.3. Required Metavariables

Servers MUST provide scripts with certain information and metavariables as described in section 8.3.

8.1.4. Response Compliance

Servers MUST ensure that responses sent to the user-agent meet all requirements of the protocol level in effect. This may involve modifying, deleting, or augmenting any header fields and/or message-body supplied by the script.

8.2. Recommendations for Servers

Servers SHOULD provide the "query" component of the script-URI as command-line arguments to scripts if it does not contain any unencoded '=' characters and the command-line arguments can be generated in an unambiguous manner. (See section 5.)

Servers SHOULD set the AUTH_TYPE metavariable to the value of the 'auth-scheme' token of the "Authorization" field if it was supplied as part of the request header. (See section 6.1.1.)

Where applicable, servers SHOULD set the current working directory to the directory in which the script is located before invoking it.

Servers MAY reject with error '404 Not Found' any requests that would result in an encoded "/" being decoded into PATH_INFO or SCRIPT_NAME, as this might represent a loss of information to the script.

Although the server and the CGI script need not be consistent in their handling of URL paths (client URLs and the PATH_INFO data, respectively), server authors may wish to impose consistency. So the server implementation SHOULD define its behaviour for the following cases:

  1. define any restrictions on allowed characters, in particular whether ASCII NUL is permitted;
  2. define any restrictions on allowed path segments, in particular whether non-terminal NULL segments are permitted;
  3. define the behaviour for "." or ".." path segments; i.e., whether they are prohibited, treated as ordinary path segments or interpreted in accordance with the relative URL specification [7];
  4. define any limits of the implementation, including limits on path or search string lengths, and limits on the volume of header data the server will parse.

Servers MAY generate the Script-URI in any way from the client URI, or from any other data (but the behaviour SHOULD be documented).

For non-parsed header (NPH) scripts (see section 7.1), servers SHOULD attempt to ensure that the script input comes directly from the client, with minimal buffering. For all scripts the data will be as supplied by the client.

8.3. Summary of MetaVariables

Servers MUST provide the following metavariables to scripts. See the individual descriptions for exceptions and semantics.

[an error occurred while processing this directive]

    CONTENT_LENGTH (section 6.1.2)
    CONTENT_TYPE (section 6.1.3)
    GATEWAY_INTERFACE (section 6.1.4)
    PATH_INFO (section 6.1.6)
    QUERY_STRING (section 6.1.8)
    REMOTE_ADDR (section 6.1.9)
    REQUEST_METHOD (section 6.1.13)
    SCRIPT_NAME (section 6.1.14)
    SERVER_NAME (section 6.1.15)
    SERVER_PORT (section 6.1.16)
    SERVER_PROTOCOL (section 6.1.17)
    SERVER_SOFTWARE (section 6.1.18)
  

Servers SHOULD define the following metavariables for scripts. See the individual descriptions for exceptions and semantics.

[an error occurred while processing this directive]

    AUTH_TYPE (section 6.1.1)
    REMOTE_HOST (section 6.1.10)
  

In addition, servers SHOULD provide metavariables for all fields present in the HTTP request header, with the exception of those involved with access control. Servers MAY at their discretion provide metavariables for access control fields.

Servers MAY define the following metavariables. See the individual descriptions for exceptions and semantics.

[an error occurred while processing this directive]

    PATH_TRANSLATED (section 6.1.7)
    REMOTE_IDENT (section 6.1.11)
    REMOTE_USER (section 6.1.12)
  

Servers MAY at their discretion define additional implementation-specific extension metavariables provided their names do not conflict with defined header field names. Implementation-specific metavariable names SHOULD be prefixed with "X_" (e.g., "X_DBA") to avoid the potential for such conflicts.

9. Script Implementation

This section defines the requirements and recommendations for scripts that are intended to function in a CGI/1.1 environment. It is intended primarily as a reference for script authors, but server implementors should be familiar with these issues as well.

9.1. Requirements for Scripts

Scripts using the parsed-header method to communicate with servers MUST supply a response header to the server. (See section 7.)

Scripts using the NPH method to communicate with servers MUST provide complete HTTP responses, and MUST use the value of the SERVER_PROTOCOL metavariable to determine the appropriate format. (See section 7.1.)

Scripts MUST check the value of the REQUEST_METHOD metavariable in order to provide an appropriate response. (See section 6.1.13.)

Scripts MUST be prepared to handled URL-encoded values in metavariables. In addition, they MUST recognise both "+" and "%20" in URL-encoded quantities as representing the space character. (See section 3.1.)

Scripts MUST ignore leading zeros in the major and minor version numbers in the GATEWAY_INTERFACE metavariable value. (See section 6.1.4.)

When processing requests that include a message-body, scripts MUST NOT read more than CONTENT_LENGTH bytes from the input stream. (See sections 6.1.2 and 6.2.)

9.2. Recommendations for Scripts

Servers may interrupt or terminate script execution at any time and without warning, so scripts SHOULD be prepared to deal with abnormal termination.

Scripts MUST reject with error '405 Method Not Allowed' requests made using methods that they do not support. If the script does not intend processing the PATH_INFO data, then it SHOULD reject the request with '404 Not Found' if PATH_INFO is not NULL.

If a script is processing the output of a form, it SHOULD verify that the CONTENT_TYPE is "application/x-www-form-urlencoded" [2] or whatever other media type is expected.

Scripts parsing PATH_INFO, PATH_TRANSLATED, or SCRIPT_NAME SHOULD be careful of void path segments ("//") and special path segments ("." and ".."). They SHOULD either be removed from the path before use in OS system calls, or the request SHOULD be rejected with '404 Not Found'.

As it is impossible for scripts to determine the client URI that initiated a request without knowledge of the specific server in use, the script SHOULD NOT return "text/html" documents containing relative URL links without including a "<BASE>" tag in the document.

When returning header fields, scripts SHOULD try to send the CGI header fields (see section 7.2) as soon as possible, and SHOULD send them before any HTTP header fields. This may help reduce the server's memory requirements.

10. System Specifications

10.1. AmigaDOS

The implementation of the CGI on an AmigaDOS operating system platform SHOULD use environment variables as the mechanism of providing request metadata to CGI scripts.

Environment variables

These are accessed by the DOS library routine GetVar. The flags argument SHOULD be 0. Case is ignored, but upper case is recommended for compatibility with case-sensitive systems.

The current working directory

The current working directory for the script is set to the directory containing the script.

Character set

The US-ASCII character set is used for the definition of environment variable names and header field names; the newline (NL) sequence is LF; servers SHOULD also accept CR LF as a newline.

10.2. Unix

The implementation of the CGI on a UNIX operating system platform SHOULD use environment variables as the mechanism of providing request metadata to CGI scripts.

For Unix compatible operating systems, the following are defined:

Environment variables

These are accessed by the C library routine getenv.

The command line

This is accessed using the argc and argv arguments to main(). The words have any characters that are 'active' in the Bourne shell escaped with a backslash. If the value of the QUERY_STRING metavariable contains an unencoded equals-sign '=', then the command line SHOULD NOT be used by the script.

The current working directory

The current working directory for the script SHOULD be set to the directory containing the script.

Character set

The US-ASCII character set is used for the definition of environment variable names and header field names; the newline (NL) sequence is LF; servers SHOULD also accept CR LF as a newline.

11. Security Considerations

11.1. Safe Methods

As discussed in the security considerations of the HTTP specifications [3,8], the convention has been established that the GET and HEAD methods should be 'safe'; they should cause no side-effects and only have the significance of resource retrieval.

CGI scripts are responsible for enforcing any HTTP security considerations [3,8] with respect to the protocol version level of the request and any side effects generated by the scripts on behalf of the server. Primary among these are the considerations of safe and idempotent methods. Idempotent requests are those that may be repeated an arbitrary number of times and produce side effects identical to a single request.

11.2. HTTP Header Fields Containing Sensitive Information

Some HTTP header fields may carry sensitive information which the server SHOULD NOT pass on to the script unless explicitly configured to do so. For example, if the server protects the script using the "Basic" authentication scheme, then the client will send an "Authorization" header field containing a username and password. If the server, rather than the script, validates this information then the password SHOULD NOT be passed on to the script via the HTTP_AUTHORIZATION metavariable without careful consideration. This also applies to the Proxy-Authorization header field and the corresponding HTTP_PROXY_AUTHORIZATION metavariable.

11.3. Script Interference with the Server

The most common implementation of CGI invokes the script as a child process using the same user and group as the server process. It SHOULD therefore be ensured that the script cannot interfere with the server process, its configuration, or documents.

If the script is executed by calling a function linked in to the server software (either at compile-time or run-time) then precautions SHOULD be taken to protect the core memory of the server, or to ensure that untrusted code cannot be executed.

11.4. Data Length and Buffering Considerations

This specification places no limits on the length of message-bodies presented to the script. Scripts should not assume that statically allocated buffers of any size are sufficient to contain the entire submission at one time. Use of a fixed length buffer without careful overflow checking may result in an attacker exploiting 'stack-smashing' or 'stack-overflow' vulnerabilities of the operating system. Scripts may spool large submissions to disk or other buffering media, but a rapid succession of large submissions may result in denial of service conditions. If the CONTENT_LENGTH of a message-body is larger than resource considerations allow, scripts should respond with an error status appropriate for the protocol version; potentially applicable status codes include '503 Service Unavailable' (HTTP/1.0 and HTTP/1.1), '413 Request Entity Too Large' (HTTP/1.1), and '414 Request-URI Too Long' (HTTP/1.1).

11.5. Stateless Processing

The stateless nature of the Web makes each script execution and resource retrieval independent of all others even when multiple requests constitute a single conceptual Web transaction. Because of this, a script should not make any assumptions about the context of the user-agent submitting a request. In particular, scripts should examine data obtained from the client and verify that they are valid, both in form and content, before allowing them to be used for sensitive purposes such as input to other applications, commands, or operating system services. These uses include, but are not limited to: system call arguments, database writes, dynamically evaluated source code, and input to billing or other secure processes. It is important that applications be protected from invalid input regardless of whether the invalidity is the result of user error, logic error, or malicious action.

Authors of scripts involved in multi-request transactions should be particularly cautios about validating the state information; undesirable effects may result from the substitution of dangerous values for portions of the submission which might otherwise be presumed safe. Subversion of this type occurs when alterations are made to data from a prior stage of the transaction that were not meant to be controlled by the client (e.g., hidden HTML form elements, cookies, embedded URLs, etc.).

12. Acknowledgements

This work is based on a draft published in 1997 by David R. Robinson, which in turn was based on the original CGI interface that arose out of discussions on the www-talk mailing list. In particular, Rob McCool, John Franks, Ari Luotonen, George Phillips and Tony Sanders deserve special recognition for their efforts in defining and implementing the early versions of this interface.

This document has also greatly benefited from the comments and suggestions made by Chris Adie, Dave Kristol, Mike Meyer, David Morris, Jeremy Madea, Patrick McManus, Adam Donahue, Ross Patterson, and Harald Alvestrand.

13. References

[1]
Berners-Lee, T., 'Universal Resource Identifiers in WWW: A Unifying Syntax for the Expression of Names and Addresses of Objects on the Network as used in the World-Wide Web', RFC 1630, CERN, June 1994.

[2]
Berners-Lee, T. and Connolly, D., 'Hypertext Markup Language - 2.0', RFC 1866, MIT/W3C, November 1995.

[3]
Berners-Lee, T., Fielding, R. T. and Frystyk, H., 'Hypertext Transfer Protocol -- HTTP/1.0', RFC 1945, MIT/LCS, UC Irvine, May 1996.

[4]
Berners-Lee, T., Fielding, R., and Masinter, L., Editors, 'Uniform Resource Identifiers (URI): Generic Syntax', RFC 2396, MIT, U.C. Irvine, Xerox Corporation, August 1996.

[5]
Braden, R., Editor, 'Requirements for Internet Hosts -- Application and Support', STD 3, RFC 1123, IETF, October 1989.

[6]
Crocker, D.H., 'Standard for the Format of ARPA Internet Text Messages', STD 11, RFC 822, University of Delaware, August 1982.

[7]
Fielding, R., 'Relative Uniform Resource Locators', RFC 1808, UC Irvine, June 1995.

[8]
Fielding, R., Gettys, J., Mogul, J., Frystyk, H. and Berners-Lee, T., 'Hypertext Transfer Protocol -- HTTP/1.1', RFC 2068, UC Irvine, DEC, MIT/LCS, January 1997.

[9]
Freed, N. and Borenstein N., 'Multipurpose Internet Mail Extensions (MIME) Part Two: Media Types', RFC 2046, Innosoft, First Virtual, November 1996.

[10]
Mockapetris, P., 'Domain Names - Concepts and Facilities', STD 13, RFC 1034, ISI, November 1987.

[11]
St. Johns, M., 'Identification Protocol', RFC 1431, US Department of Defense, February 1993.

[12]
'Coded Character Set -- 7-bit American Standard Code for Information Interchange', ANSI X3.4-1986.

[13]
Hinden, R. and Deering, S., 'IP Version 6 Addressing Architecture', RFC 2373, Nokia, Cisco Systems, July 1998.

14. Authors' Addresses

Ken A L Coar
MeepZor Consulting
7824 Mayfaire Crest Lane, Suite 202
Raleigh, NC 27615-4875
U.S.A.

Tel: +1 (919) 254.4237
Fax: +1 (919) 254.5250
Email: Ken.Coar@Golux.Com

David Robinson
E*TRADE UK Ltd
Mount Pleasant House
2 Mount Pleasant
Huntingdon Road
Cambridge CB3 0RN
UK

Tel: +44 (1223) 566926
Fax: +44 (1223) 506288
Email: drtr@etrade.co.uk