source: draft-ietf-httpbis/latest/p1-messaging.xml @ 1514

Last change on this file since 1514 was 1514, checked in by julian.reschke@…, 8 years ago

spelling & grammar

  • Property svn:eol-style set to native
  • Property svn:mime-type set to text/xml
File size: 242.0 KB
[29]1<?xml version="1.0" encoding="utf-8"?>
[101]2<?xml-stylesheet type='text/xsl' href='../myxml2rfc.xslt'?>
[8]3<!DOCTYPE rfc [
4  <!ENTITY MAY "<bcp14 xmlns=''>MAY</bcp14>">
5  <!ENTITY MUST "<bcp14 xmlns=''>MUST</bcp14>">
6  <!ENTITY MUST-NOT "<bcp14 xmlns=''>MUST NOT</bcp14>">
7  <!ENTITY OPTIONAL "<bcp14 xmlns=''>OPTIONAL</bcp14>">
8  <!ENTITY RECOMMENDED "<bcp14 xmlns=''>RECOMMENDED</bcp14>">
9  <!ENTITY REQUIRED "<bcp14 xmlns=''>REQUIRED</bcp14>">
10  <!ENTITY SHALL "<bcp14 xmlns=''>SHALL</bcp14>">
11  <!ENTITY SHALL-NOT "<bcp14 xmlns=''>SHALL NOT</bcp14>">
12  <!ENTITY SHOULD "<bcp14 xmlns=''>SHOULD</bcp14>">
13  <!ENTITY SHOULD-NOT "<bcp14 xmlns=''>SHOULD NOT</bcp14>">
[29]14  <!ENTITY ID-VERSION "latest">
[1497]15  <!ENTITY ID-MONTH "January">
16  <!ENTITY ID-YEAR "2012">
[1101]17  <!ENTITY mdash "&#8212;">
[640]18  <!ENTITY caching-overview       "<xref target='Part6' x:rel='#caching.overview' xmlns:x=''/>">
[1374]19  <!ENTITY cache-incomplete       "<xref target='Part6' x:rel='#response.cacheability' xmlns:x=''/>">
[31]20  <!ENTITY payload                "<xref target='Part3' xmlns:x=''/>">
[115]21  <!ENTITY media-types            "<xref target='Part3' x:rel='#media.types' xmlns:x=''/>">
22  <!ENTITY content-codings        "<xref target='Part3' x:rel='#content.codings' xmlns:x=''/>">
[31]23  <!ENTITY CONNECT                "<xref target='Part2' x:rel='#CONNECT' xmlns:x=''/>">
24  <!ENTITY content.negotiation    "<xref target='Part3' x:rel='#content.negotiation' xmlns:x=''/>">
[874]25  <!ENTITY diff-mime              "<xref target='Part3' x:rel='#differences.between.http.and.mime' xmlns:x=''/>">
26  <!ENTITY representation         "<xref target='Part3' x:rel='#representation' xmlns:x=''/>">
[31]27  <!ENTITY header-cache-control   "<xref target='Part6' x:rel='#header.cache-control' xmlns:x=''/>">
[1454]28  <!ENTITY header-date            "<xref target='Part2' x:rel='' xmlns:x=''/>">
[31]29  <!ENTITY header-expect          "<xref target='Part2' x:rel='#header.expect' xmlns:x=''/>">
[937]30  <!ENTITY header-mime-version    "<xref target='Part3' x:rel='#mime-version' xmlns:x=''/>">
[31]31  <!ENTITY header-pragma          "<xref target='Part6' x:rel='#header.pragma' xmlns:x=''/>">
32  <!ENTITY header-warning         "<xref target='Part6' x:rel='#header.warning' xmlns:x=''/>">
33  <!ENTITY idempotent-methods     "<xref target='Part2' x:rel='#idempotent.methods' xmlns:x=''/>">
[1420]34  <!ENTITY method                 "<xref target='Part2' x:rel='#method' xmlns:x=''/>">
35  <!ENTITY status-code-reasonphr  "<xref target='Part2' x:rel='#status.code.and.reason.phrase' xmlns:x=''/>">
[31]36  <!ENTITY status-codes           "<xref target='Part2' x:rel='' xmlns:x=''/>">
37  <!ENTITY status-100             "<xref target='Part2' x:rel='#status.100' xmlns:x=''/>">
38  <!ENTITY status-1xx             "<xref target='Part2' x:rel='#status.1xx' xmlns:x=''/>">
[1309]39  <!ENTITY status-203             "<xref target='Part2' x:rel='#status.203' xmlns:x=''/>">
[1069]40  <!ENTITY status-3xx             "<xref target='Part2' x:rel='#status.3xx' xmlns:x=''/>">
[1323]41  <!ENTITY status-4xx             "<xref target='Part2' x:rel='#status.4xx' xmlns:x=''/>">
[31]42  <!ENTITY status-414             "<xref target='Part2' x:rel='#status.414' xmlns:x=''/>">
[1420]43  <!ENTITY cons-new-header-fields "<xref target='Part2' x:rel='#considerations.for.creating.header.fields' xmlns:x=''/>">
45<?rfc toc="yes" ?>
[29]46<?rfc symrefs="yes" ?>
47<?rfc sortrefs="yes" ?>
[8]48<?rfc compact="yes"?>
49<?rfc subcompact="no" ?>
50<?rfc linkmailto="no" ?>
51<?rfc editing="no" ?>
[203]52<?rfc comments="yes"?>
53<?rfc inline="yes"?>
[799]54<?rfc rfcedstyle="yes"?>
[8]55<?rfc-ext allow-markup-in-artwork="yes" ?>
56<?rfc-ext include-references-in-index="yes" ?>
[1477]57<rfc obsoletes="2145,2616" updates="2817" category="std" x:maturity-level="proposed"
[446]58     ipr="pre5378Trust200902" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
[153]59     xmlns:x=''>
[1472]60<x:link rel="next" basename="p2-semantics"/>
[120]63  <title abbrev="HTTP/1.1, Part 1">HTTP/1.1, part 1: URIs, Connections, and Message Parsing</title>
[29]65  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
[1106]66    <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
[8]67    <address>
68      <postal>
[1106]69        <street>345 Park Ave</street>
70        <city>San Jose</city>
[8]71        <region>CA</region>
[1106]72        <code>95110</code>
[29]73        <country>USA</country>
[8]74      </postal>
[29]75      <email></email>
76      <uri></uri>
[8]77    </address>
78  </author>
[29]80  <author initials="J." surname="Gettys" fullname="Jim Gettys">
[844]81    <organization abbrev="Alcatel-Lucent">Alcatel-Lucent Bell Labs</organization>
[8]82    <address>
83      <postal>
[29]84        <street>21 Oak Knoll Road</street>
85        <city>Carlisle</city>
[8]86        <region>MA</region>
[29]87        <code>01741</code>
88        <country>USA</country>
[8]89      </postal>
[844]90      <email></email>
91      <uri></uri>
[8]92    </address>
93  </author>
95  <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
[29]96    <organization abbrev="HP">Hewlett-Packard Company</organization>
[8]97    <address>
98      <postal>
[29]99        <street>HP Labs, Large Scale Systems Group</street>
100        <street>1501 Page Mill Road, MS 1177</street>
[8]101        <city>Palo Alto</city>
102        <region>CA</region>
[29]103        <code>94304</code>
104        <country>USA</country>
[8]105      </postal>
[29]106      <email></email>
[8]107    </address>
108  </author>
110  <author initials="H." surname="Frystyk" fullname="Henrik Frystyk Nielsen">
[29]111    <organization abbrev="Microsoft">Microsoft Corporation</organization>
[8]112    <address>
113      <postal>
[29]114        <street>1 Microsoft Way</street>
115        <city>Redmond</city>
116        <region>WA</region>
117        <code>98052</code>
118        <country>USA</country>
[8]119      </postal>
[29]120      <email></email>
[8]121    </address>
122  </author>
124  <author initials="L." surname="Masinter" fullname="Larry Masinter">
[1106]125    <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
[8]126    <address>
127      <postal>
[29]128        <street>345 Park Ave</street>
129        <city>San Jose</city>
[8]130        <region>CA</region>
[29]131        <code>95110</code>
132        <country>USA</country>
[8]133      </postal>
[29]134      <email></email>
135      <uri></uri>
[8]136    </address>
137  </author>
139  <author initials="P." surname="Leach" fullname="Paul J. Leach">
140    <organization abbrev="Microsoft">Microsoft Corporation</organization>
141    <address>
142      <postal>
143        <street>1 Microsoft Way</street>
144        <city>Redmond</city>
145        <region>WA</region>
146        <code>98052</code>
147      </postal>
148      <email></email>
149    </address>
150  </author>
152  <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
153    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
154    <address>
155      <postal>
[34]156        <street>MIT Computer Science and Artificial Intelligence Laboratory</street>
157        <street>The Stata Center, Building 32</street>
158        <street>32 Vassar Street</street>
[8]159        <city>Cambridge</city>
160        <region>MA</region>
161        <code>02139</code>
[29]162        <country>USA</country>
[8]163      </postal>
164      <email></email>
[34]165      <uri></uri>
[8]166    </address>
167  </author>
[95]169  <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
[94]170    <organization abbrev="W3C">World Wide Web Consortium</organization>
171    <address>
172      <postal>
173        <street>W3C / ERCIM</street>
174        <street>2004, rte des Lucioles</street>
175        <city>Sophia-Antipolis</city>
176        <region>AM</region>
177        <code>06902</code>
178        <country>France</country>
179      </postal>
180      <email></email>
181      <uri></uri>
182    </address>
183  </author>
[95]185  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
186    <organization abbrev="greenbytes">greenbytes GmbH</organization>
187    <address>
188      <postal>
189        <street>Hafenweg 16</street>
190        <city>Muenster</city><region>NW</region><code>48155</code>
191        <country>Germany</country>
192      </postal>
[609]193      <phone>+49 251 2807760</phone>
194      <facsimile>+49 251 2807761</facsimile>
195      <email></email>
196      <uri></uri>
[95]197    </address>
198  </author>
[31]200  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
[440]201  <workgroup>HTTPbis Working Group</workgroup>
[1373]205   The Hypertext Transfer Protocol (HTTP) is an application-level protocol for
206   distributed, collaborative, hypertext information systems. HTTP has been in
207   use by the World Wide Web global information initiative since 1990. This
208   document is Part 1 of the seven-part specification that defines the protocol
[1377]209   referred to as "HTTP/1.1" and, taken together, obsoletes
210   <xref target="RFC2616" x:fmt="none">RFC 2616</xref> and moves it to historic
211   status, along with its predecessor <xref target="RFC2068" x:fmt="none">RFC
212   2068</xref>.
215   Part 1 provides an overview of HTTP and its associated terminology, defines
216   the "http" and "https" Uniform Resource Identifier (URI) schemes, defines
217   the generic message syntax and parsing requirements for HTTP message frames,
218   and describes general security concerns for implementations.
221   This part also obsoletes RFCs <xref target="RFC2145" x:fmt="none">2145</xref>
222   (on HTTP version numbers) and <xref target="RFC2817" x:fmt="none">2817</xref>
223   (on using CONNECT for TLS upgrades) and moves them to historic status.
227<note title="Editorial Note (To be removed by RFC Editor)">
228  <t>
229    Discussion of this draft should take place on the HTTPBIS working group
[1268]230    mailing list (, which is archived at
231    <eref target=""/>.
232  </t>
233  <t>
234    The current issues list is at
235    <eref target=""/> and related
236    documents (including fancy diffs) can be found at
[324]237    <eref target=""/>.
[36]238  </t>
[153]239  <t>
[1499]240    The changes in this draft are summarized in <xref target="changes.since.18"/>.
[153]241  </t>
245<section title="Introduction" anchor="introduction">
[8]247   The Hypertext Transfer Protocol (HTTP) is an application-level
[374]248   request/response protocol that uses extensible semantics and MIME-like
[391]249   message payloads for flexible interaction with network-based hypertext
[374]250   information systems. HTTP relies upon the Uniform Resource Identifier (URI)
[1171]251   standard <xref target="RFC3986"/> to indicate the target resource and
[391]252   relationships between resources.
[374]253   Messages are passed in a format similar to that used by Internet mail
254   <xref target="RFC5322"/> and the Multipurpose Internet Mail Extensions
[852]255   (MIME) <xref target="RFC2045"/> (see &diff-mime; for the differences
[374]256   between HTTP and MIME messages).
[544]259   HTTP is a generic interface protocol for information systems. It is
[391]260   designed to hide the details of how a service is implemented by presenting
261   a uniform interface to clients that is independent of the types of
262   resources provided. Likewise, servers do not need to be aware of each
263   client's purpose: an HTTP request can be considered in isolation rather
264   than being associated with a specific type of client or a predetermined
265   sequence of application steps. The result is a protocol that can be used
266   effectively in many different contexts and for which implementations can
267   evolve independently over time.
[849]270   HTTP is also designed for use as an intermediation protocol for translating
271   communication to and from non-HTTP information systems.
272   HTTP proxies and gateways can provide access to alternative information
[451]273   services by translating their diverse protocols into a hypertext
[374]274   format that can be viewed and manipulated by clients in the same way
275   as HTTP services.
[544]278   One consequence of HTTP flexibility is that the protocol cannot be
279   defined in terms of what occurs behind the interface. Instead, we
280   are limited to defining the syntax of communication, the intent
281   of received communication, and the expected behavior of recipients.
282   If the communication is considered in isolation, then successful
[969]283   actions ought to be reflected in corresponding changes to the
[544]284   observable interface provided by servers. However, since multiple
[901]285   clients might act in parallel and perhaps at cross-purposes, we
[544]286   cannot require that such changes be observable beyond the scope
287   of a single response.
[374]290   This document is Part 1 of the seven-part specification of HTTP,
[1148]291   defining the protocol referred to as "HTTP/1.1", obsoleting
292   <xref target="RFC2616"/> and <xref target="RFC2145"/>.
[544]293   Part 1 describes the architectural elements that are used or
[621]294   referred to in HTTP, defines the "http" and "https" URI schemes,
295   describes overall network operation and connection management,
296   and defines HTTP message framing and forwarding requirements.
[374]297   Our goal is to define all of the mechanisms necessary for HTTP message
298   handling that are independent of message semantics, thereby defining the
[544]299   complete set of requirements for message parsers and
[391]300   message-forwarding intermediaries.
[1452]303<section title="Conformance and Error Handling" anchor="intro.conformance.and.error.handling">
305   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
306   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
[96]307   document are to be interpreted as described in <xref target="RFC2119"/>.
[1452]310   This document defines conformance criteria for several roles in HTTP
311   communication, including Senders, Recipients, Clients, Servers, User-Agents,
312   Origin Servers, Intermediaries, Proxies and Gateways. See <xref target="architecture"/>
313   for definitions of these terms.
316   An implementation is considered conformant if it complies with all of the
317   requirements associated with its role(s). Note that SHOULD-level requirements
318   are relevant here, unless one of the documented exceptions is applicable.
321   This document also uses ABNF to define valid protocol elements
322   (<xref target="notation"/>). In addition to the prose requirements placed
323   upon them, Senders &MUST-NOT; generate protocol elements that are invalid.
326   Unless noted otherwise, Recipients &MAY; take steps to recover a usable
327   protocol element from an invalid construct. However, HTTP does not define
328   specific error handling mechanisms, except in cases where it has direct
329   impact on security. This is because different uses of the protocol require
330   different error handling strategies; for example, a Web browser may wish to
331   transparently recover from a response where the Location header field
332   doesn't parse according to the ABNF, whereby in a systems control protocol
333   using HTTP, this type of error recovery could lead to dangerous consequences.
[390]337<section title="Syntax Notation" anchor="notation">
338<iref primary="true" item="Grammar" subitem="ALPHA"/>
339<iref primary="true" item="Grammar" subitem="CR"/>
340<iref primary="true" item="Grammar" subitem="CRLF"/>
341<iref primary="true" item="Grammar" subitem="CTL"/>
342<iref primary="true" item="Grammar" subitem="DIGIT"/>
343<iref primary="true" item="Grammar" subitem="DQUOTE"/>
344<iref primary="true" item="Grammar" subitem="HEXDIG"/>
[1425]345<iref primary="true" item="Grammar" subitem="HTAB"/>
[390]346<iref primary="true" item="Grammar" subitem="LF"/>
347<iref primary="true" item="Grammar" subitem="OCTET"/>
348<iref primary="true" item="Grammar" subitem="SP"/>
[395]349<iref primary="true" item="Grammar" subitem="VCHAR"/>
351   This specification uses the Augmented Backus-Naur Form (ABNF) notation
352   of <xref target="RFC5234"/>.
[390]354<t anchor="core.rules">
355  <x:anchor-alias value="ALPHA"/>
356  <x:anchor-alias value="CTL"/>
357  <x:anchor-alias value="CR"/>
358  <x:anchor-alias value="CRLF"/>
359  <x:anchor-alias value="DIGIT"/>
360  <x:anchor-alias value="DQUOTE"/>
361  <x:anchor-alias value="HEXDIG"/>
[1425]362  <x:anchor-alias value="HTAB"/>
[390]363  <x:anchor-alias value="LF"/>
364  <x:anchor-alias value="OCTET"/>
365  <x:anchor-alias value="SP"/>
[395]366  <x:anchor-alias value="VCHAR"/>
[543]367   The following core rules are included by
[390]368   reference, as defined in <xref target="RFC5234" x:fmt="," x:sec="B.1"/>:
[395]369   ALPHA (letters), CR (carriage return), CRLF (CR LF), CTL (controls),
[390]370   DIGIT (decimal 0-9), DQUOTE (double quote),
[1425]371   HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line feed),
372   OCTET (any 8-bit sequence of data), SP (space), and
373   VCHAR (any visible <xref target="USASCII"/> character).
[849]376   As a syntactic convention, ABNF rule names prefixed with "obs-" denote
[738]377   "obsolete" grammar rules that appear for historical reasons.
[368]380<section title="ABNF Extension: #rule" anchor="notation.abnf">
382  The #rule extension to the ABNF rules of <xref target="RFC5234"/> is used to
383  improve readability.
386  A construct "#" is defined, similar to "*", for defining comma-delimited
387  lists of elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating
388  at least &lt;n&gt; and at most &lt;m&gt; elements, each separated by a single
[1396]389  comma (",") and optional whitespace (OWS, <xref target="basic.rules"/>).   
392  Thus,
[400]393</preamble><artwork type="example">
394  1#element =&gt; element *( OWS "," OWS element )
397  and:
[400]398</preamble><artwork type="example">
399  #element =&gt; [ 1#element ]
402  and for n &gt;= 1 and m &gt; 1:
[400]403</preamble><artwork type="example">
404  &lt;n&gt;#&lt;m&gt;element =&gt; element &lt;n-1&gt;*&lt;m-1&gt;( OWS "," OWS element )
407  For compatibility with legacy list rules, recipients &SHOULD; accept empty
408  list elements. In other words, consumers would follow the list productions:
[400]410<figure><artwork type="example">
[458]411  #element =&gt; [ ( "," / element ) *( OWS "," [ OWS element ] ) ]
413  1#element =&gt; *( "," OWS ) element *( OWS "," [ OWS element ] )
[738]416  Note that empty elements do not contribute to the count of elements present,
417  though.
420  For example, given these ABNF productions:
422<figure><artwork type="example">
423  example-list      = 1#example-list-elmt
[1396]424  example-list-elmt = token ; see <xref target="field.rules"/> 
427  Then these are valid values for example-list (not including the double
428  quotes, which are present for delimitation only):
430<figure><artwork type="example">
431  "foo,bar"
[1462]432  "foo ,bar,"
433  "foo , ,bar,charlie   "
436  But these values would be invalid, as at least one non-empty element is
437  required:
439<figure><artwork type="example">
440  ""
441  ","
442  ",   ,"
[421]445  <xref target="collected.abnf"/> shows the collected ABNF, with the list rules
446  expanded as explained above.
[8]450<section title="Basic Rules" anchor="basic.rules">
[229]451<t anchor="rule.LWS">
[395]452   This specification uses three rules to denote the use of linear
453   whitespace: OWS (optional whitespace), RWS (required whitespace), and
454   BWS ("bad" whitespace).
[1384]456<t anchor="rule.OWS">
[1176]457   The OWS rule is used where zero or more linear whitespace octets might
458   appear. OWS &SHOULD; either not be produced or be produced as a single
[1392]459   SP. Multiple OWS octets that occur within field-content &SHOULD; either
[1425]460   be replaced with a single SP or transformed to all SP octets (each
[1392]461   octet other than SP replaced with SP) before interpreting the field value
462   or forwarding the message downstream.
[1384]464<t anchor="rule.RWS">
[1176]465   RWS is used when at least one linear whitespace octet is required to
466   separate field tokens. RWS &SHOULD; be produced as a single SP.
[1404]467   Multiple RWS octets that occur within field-content &SHOULD; either
468   be replaced with a single SP or transformed to all SP octets before
469   interpreting the field value or forwarding the message downstream.
[1384]471<t anchor="rule.BWS">
[395]472   BWS is used where the grammar allows optional whitespace for historical
473   reasons but senders &SHOULD-NOT; produce it in messages. HTTP/1.1
474   recipients &MUST; accept such bad optional whitespace and remove it before
475   interpreting the field value or forwarding the message downstream.
[351]477<t anchor="rule.whitespace">
478  <x:anchor-alias value="BWS"/>
479  <x:anchor-alias value="OWS"/>
480  <x:anchor-alias value="RWS"/>
481  <x:anchor-alias value="obs-fold"/>
[351]483<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="OWS"/><iref primary="true" item="Grammar" subitem="RWS"/><iref primary="true" item="Grammar" subitem="BWS"/>
[1425]484  <x:ref>OWS</x:ref>            = *( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> / obs-fold )
[401]485                 ; "optional" whitespace
[1425]486  <x:ref>RWS</x:ref>            = 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> / obs-fold )
[401]487                 ; "required" whitespace
[351]488  <x:ref>BWS</x:ref>            = <x:ref>OWS</x:ref>
[401]489                 ; "bad" whitespace
[1425]490  <x:ref>obs-fold</x:ref>       = <x:ref>CRLF</x:ref> ( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
491                 ; obsolete line folding
492                 ; see <xref target="field.parsing"/>
[1431]498<section title="Architecture" anchor="architecture">
[621]500   HTTP was created for the World Wide Web architecture
[391]501   and has evolved over time to support the scalability needs of a worldwide
502   hypertext system. Much of that architecture is reflected in the terminology
503   and syntax productions used to define HTTP.
[899]506<section title="Client/Server Messaging" anchor="operation">
[965]507<iref primary="true" item="client"/>
508<iref primary="true" item="server"/>
509<iref primary="true" item="connection"/>
[899]511   HTTP is a stateless request/response protocol that operates by exchanging
[1435]512   <x:dfn>messages</x:dfn> (<xref target="http.message"/>) across a reliable
513   transport or session-layer
[1227]514   "<x:dfn>connection</x:dfn>". An HTTP "<x:dfn>client</x:dfn>" is a
515   program that establishes a connection to a server for the purpose of
516   sending one or more HTTP requests.  An HTTP "<x:dfn>server</x:dfn>" is a
[899]517   program that accepts connections in order to service HTTP requests by
518   sending HTTP responses.
[965]520<iref primary="true" item="user agent"/>
521<iref primary="true" item="origin server"/>
522<iref primary="true" item="browser"/>
523<iref primary="true" item="spider"/>
[1227]524<iref primary="true" item="sender"/>
525<iref primary="true" item="recipient"/>
[899]527   Note that the terms client and server refer only to the roles that
[630]528   these programs perform for a particular connection.  The same program
[901]529   might act as a client on some connections and a server on others.  We use
[1227]530   the term "<x:dfn>user agent</x:dfn>" to refer to the program that initiates a request,
[630]531   such as a WWW browser, editor, or spider (web-traversing robot), and
[1227]532   the term "<x:dfn>origin server</x:dfn>" to refer to the program that can originate
[965]533   authoritative responses to a request.  For general requirements, we use
[1227]534   the term "<x:dfn>sender</x:dfn>" to refer to whichever component sent a given message
535   and the term "<x:dfn>recipient</x:dfn>" to refer to any component that receives the
[965]536   message.
539  <t>
540    <x:h>Note:</x:h> The term 'user agent' covers both those situations where
541    there is a user (human) interacting with the software agent (and for which
542    user interface or interactive suggestions might be made, e.g., warning the
543    user or given the user an option in the case of security or privacy
544    options) and also those where the software agent may act autonomously.
545  </t>
548   Most HTTP communication consists of a retrieval request (GET) for
549   a representation of some resource identified by a URI.  In the
[901]550   simplest case, this might be accomplished via a single bidirectional
[899]551   connection (===) between the user agent (UA) and the origin server (O).
553<figure><artwork type="drawing">
[899]554         request   &gt;
555    UA ======================================= O
556                                &lt;   response
[965]558<iref primary="true" item="message"/>
559<iref primary="true" item="request"/>
560<iref primary="true" item="response"/>
[1227]562   A client sends an HTTP request to the server in the form of a <x:dfn>request</x:dfn>
[1435]563   message, beginning with a request-line that includes a method, URI, and
564   protocol version (<xref target="request.line"/>),
565   followed by MIME-like header fields containing
566   request modifiers, client information, and payload metadata
567   (<xref target="header.fields"/>),
568   an empty line to indicate the end of the header section, and finally
569   a message body containing the payload body (if any,
570   <xref target="message.body"/>).
[1227]573   A server responds to the client's request by sending an HTTP <x:dfn>response</x:dfn>
[1435]574   message, beginning with a status line that
[677]575   includes the protocol version, a success or error code, and textual
[1435]576   reason phrase (<xref target="status.line"/>),
577   followed by MIME-like header fields containing server
578   information, resource metadata, and payload metadata
579   (<xref target="header.fields"/>),
580   an empty line to indicate the end of the header section, and finally
581   a message body containing the payload body (if any,
582   <xref target="message.body"/>).
[1471]585   Note that 1xx responses (&status-1xx;) are not final; therefore, a server
586   can send zero or more 1xx responses, followed by exactly one final response
587   (with any other status code).
[630]590   The following example illustrates a typical message exchange for a
591   GET request on the URI "":
[630]594client request:
[803]595</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
[633]596GET /hello.txt HTTP/1.1
597User-Agent: curl/7.16.3 libcurl/7.16.3 OpenSSL/0.9.7l zlib/1.2.3
599Accept: */*
[630]603server response:
[633]604</preamble><artwork type="message/http; msgtype=&#34;response&#34;" x:indent-with="  ">
605HTTP/1.1 200 OK
606Date: Mon, 27 Jul 2009 12:28:53 GMT
607Server: Apache
608Last-Modified: Wed, 22 Jul 2009 19:15:56 GMT
609ETag: "34aa387-d-1568eb00"
610Accept-Ranges: bytes
611Content-Length: <x:length-of target="exbody"/>
612Vary: Accept-Encoding
613Content-Type: text/plain
[633]615<x:span anchor="exbody">Hello World!
[1316]619<section title="Message Orientation and Buffering" anchor="message-orientation-and-buffering">
621   Fundamentally, HTTP is a message-based protocol. Although message bodies can
622   be chunked (<xref target="chunked.encoding"/>) and implementations often
623   make parts of a message available progressively, this is not required, and
624   some widely-used implementations only make a message available when it is
625   complete. Furthermore, while most proxies will progressively stream messages,
626   some amount of buffering will take place, and some proxies might buffer
627   messages to perform transformations, check content or provide other services.
630   Therefore, extensions to and uses of HTTP cannot rely on the availability of
631   a partial message, or assume that messages will not be buffered. There are
632   strategies that can be used to test for buffering in a given connection, but
633   it should be understood that behaviors can differ across connections, and
634   between requests and responses.
637   Recipients &MUST; consider every message in a connection in isolation;
638   because HTTP is a stateless protocol, it cannot be assumed that two requests
639   on the same connection are from the same client or share any other common
[1324]640   attributes. In particular, intermediaries might mix requests from different
641   clients into a single server connection. Note that some existing HTTP
642   extensions (e.g., <xref target="RFC4559"/>) violate this requirement, thereby
643   potentially causing interoperability and security problems.
[1148]647<section title="Connections and Transport Independence" anchor="transport-independence">
649   HTTP messaging is independent of the underlying transport or
650   session-layer connection protocol(s).  HTTP only presumes a reliable
651   transport with in-order delivery of requests and the corresponding
652   in-order delivery of responses.  The mapping of HTTP request and
653   response structures onto the data units of the underlying transport
654   protocol is outside the scope of this specification.
657   The specific connection protocols to be used for an interaction
[1171]658   are determined by client configuration and the target resource's URI.
659   For example, the "http" URI scheme
[1148]660   (<xref target="http.uri"/>) indicates a default connection of TCP
661   over IP, with a default TCP port of 80, but the client might be
662   configured to use a proxy via some other connection port or protocol
663   instead of using the defaults.
666   A connection might be used for multiple HTTP request/response exchanges,
667   as defined in <xref target="persistent.connections"/>.
[630]671<section title="Intermediaries" anchor="intermediaries">
[965]672<iref primary="true" item="intermediary"/>
[1148]674   HTTP enables the use of intermediaries to satisfy requests through
675   a chain of connections.  There are three common forms of HTTP
[1227]676   <x:dfn>intermediary</x:dfn>: proxy, gateway, and tunnel.  In some cases,
[901]677   a single intermediary might act as an origin server, proxy, gateway,
[630]678   or tunnel, switching behavior based on the nature of each request.
680<figure><artwork type="drawing">
[899]681         &gt;             &gt;             &gt;             &gt;
[1306]682    <x:highlight>UA</x:highlight> =========== <x:highlight>A</x:highlight> =========== <x:highlight>B</x:highlight> =========== <x:highlight>C</x:highlight> =========== <x:highlight>O</x:highlight>
[899]683               &lt;             &lt;             &lt;             &lt;
686   The figure above shows three intermediaries (A, B, and C) between the
687   user agent and origin server. A request or response message that
688   travels the whole chain will pass through four separate connections.
[630]689   Some HTTP communication options
[901]690   might apply only to the connection with the nearest, non-tunnel
[624]691   neighbor, only to the end-points of the chain, or to all connections
[901]692   along the chain. Although the diagram is linear, each participant might
[624]693   be engaged in multiple, simultaneous communications. For example, B
[901]694   might be receiving requests from many clients other than A, and/or
[624]695   forwarding requests to servers other than C, at the same time that it
696   is handling A's request.
[965]699<iref primary="true" item="upstream"/><iref primary="true" item="downstream"/>
700<iref primary="true" item="inbound"/><iref primary="true" item="outbound"/>
[1227]701   We use the terms "<x:dfn>upstream</x:dfn>" and "<x:dfn>downstream</x:dfn>"
702   to describe various requirements in relation to the directional flow of a
703   message: all messages flow from upstream to downstream.
704   Likewise, we use the terms inbound and outbound to refer to
705   directions in relation to the request path:
706   "<x:dfn>inbound</x:dfn>" means toward the origin server and
707   "<x:dfn>outbound</x:dfn>" means toward the user agent.
[965]709<t><iref primary="true" item="proxy"/>
[1227]710   A "<x:dfn>proxy</x:dfn>" is a message forwarding agent that is selected by the
[630]711   client, usually via local configuration rules, to receive requests
712   for some type(s) of absolute URI and attempt to satisfy those
713   requests via translation through the HTTP interface.  Some translations
714   are minimal, such as for proxy requests for "http" URIs, whereas
[901]715   other requests might require translation to and from entirely different
[630]716   application-layer protocols. Proxies are often used to group an
717   organization's HTTP requests through a common intermediary for the
718   sake of security, annotation services, or shared caching.
721<iref primary="true" item="transforming proxy"/>
722<iref primary="true" item="non-transforming proxy"/>
[1227]723   An HTTP-to-HTTP proxy is called a "<x:dfn>transforming proxy</x:dfn>" if it is designed
[1107]724   or configured to modify request or response messages in a semantically
725   meaningful way (i.e., modifications, beyond those required by normal
726   HTTP processing, that change the message in a way that would be
727   significant to the original sender or potentially significant to
728   downstream recipients).  For example, a transforming proxy might be
729   acting as a shared annotation server (modifying responses to include
730   references to a local annotation database), a malware filter, a
731   format transcoder, or an intranet-to-Internet privacy filter.  Such
732   transformations are presumed to be desired by the client (or client
733   organization) that selected the proxy and are beyond the scope of
734   this specification.  However, when a proxy is not intended to transform
[1227]735   a given message, we use the term "<x:dfn>non-transforming proxy</x:dfn>" to target
[1309]736   requirements that preserve HTTP message semantics. See &status-203; and
737   &header-warning; for status and warning codes related to transformations.
[965]739<t><iref primary="true" item="gateway"/><iref primary="true" item="reverse proxy"/>
[1170]740<iref primary="true" item="accelerator"/>
[1227]741   A "<x:dfn>gateway</x:dfn>" (a.k.a., "<x:dfn>reverse proxy</x:dfn>")
742   is a receiving agent that acts
[630]743   as a layer above some other server(s) and translates the received
744   requests to the underlying server's protocol.  Gateways are often
[1175]745   used to encapsulate legacy or untrusted information services, to
[1227]746   improve server performance through "<x:dfn>accelerator</x:dfn>" caching, and to
[1175]747   enable partitioning or load-balancing of HTTP services across
748   multiple machines.
751   A gateway behaves as an origin server on its outbound connection and
752   as a user agent on its inbound connection.
753   All HTTP requirements applicable to an origin server
754   also apply to the outbound communication of a gateway.
755   A gateway communicates with inbound servers using any protocol that
756   it desires, including private extensions to HTTP that are outside
757   the scope of this specification.  However, an HTTP-to-HTTP gateway
758   that wishes to interoperate with third-party HTTP servers &MUST;
759   comply with HTTP user agent requirements on the gateway's inbound
760   connection and &MUST; implement the Connection
761   (<xref target="header.connection"/>) and Via (<xref target="header.via"/>)
762   header fields for both connections.
[965]764<t><iref primary="true" item="tunnel"/>
[1227]765   A "<x:dfn>tunnel</x:dfn>" acts as a blind relay between two connections
[630]766   without changing the messages. Once active, a tunnel is not
[901]767   considered a party to the HTTP communication, though the tunnel might
[630]768   have been initiated by an HTTP request. A tunnel ceases to exist when
769   both ends of the relayed connection are closed. Tunnels are used to
770   extend a virtual connection through an intermediary, such as when
771   transport-layer security is used to establish private communication
772   through a shared firewall proxy.
[1108]774<t><iref primary="true" item="interception proxy"/><iref primary="true" item="transparent proxy"/>
[1170]775<iref primary="true" item="captive portal"/>
[1107]776   In addition, there may exist network intermediaries that are not
777   considered part of the HTTP communication but nevertheless act as
778   filters or redirecting agents (usually violating HTTP semantics,
779   causing security problems, and otherwise making a mess of things).
[1227]780   Such a network intermediary, often referred to as an "<x:dfn>interception proxy</x:dfn>"
781   <xref target="RFC3040"/>, "<x:dfn>transparent proxy</x:dfn>" <xref target="RFC1919"/>,
782   or "<x:dfn>captive portal</x:dfn>",
[1108]783   differs from an HTTP proxy because it has not been selected by the client.
784   Instead, the network intermediary redirects outgoing TCP port 80 packets
785   (and occasionally other common port traffic) to an internal HTTP server.
[1170]786   Interception proxies are commonly found on public network access points,
[1108]787   as a means of enforcing account subscription prior to allowing use of
[1170]788   non-local Internet services, and within corporate firewalls to enforce
789   network usage policies.
790   They are indistinguishable from a man-in-the-middle attack.
794<section title="Caches" anchor="caches">
[965]795<iref primary="true" item="cache"/>
[1227]797   A "<x:dfn>cache</x:dfn>" is a local store of previous response messages and the
[630]798   subsystem that controls its message storage, retrieval, and deletion.
799   A cache stores cacheable responses in order to reduce the response
800   time and network bandwidth consumption on future, equivalent
[901]801   requests. Any client or server &MAY; employ a cache, though a cache
[630]802   cannot be used by a server while it is acting as a tunnel.
805   The effect of a cache is that the request/response chain is shortened
806   if one of the participants along the chain has a cached response
807   applicable to that request. The following illustrates the resulting
808   chain if B has a cached copy of an earlier response from O (via C)
809   for a request which has not been cached by UA or A.
[624]811<figure><artwork type="drawing">
[899]812            &gt;             &gt;
813       UA =========== A =========== B - - - - - - C - - - - - - O
814                  &lt;             &lt;
[965]816<t><iref primary="true" item="cacheable"/>
[1227]817   A response is "<x:dfn>cacheable</x:dfn>" if a cache is allowed to store a copy of
[630]818   the response message for use in answering subsequent requests.
[901]819   Even when a response is cacheable, there might be additional
[630]820   constraints placed by the client or by the origin server on when
821   that cached response can be used for a particular request. HTTP
822   requirements for cache behavior and cacheable responses are
[640]823   defined in &caching-overview;
[630]826   There are a wide variety of architectures and configurations
827   of caches and proxies deployed across the World Wide Web and
828   inside large organizations. These systems include national hierarchies
[624]829   of proxy caches to save transoceanic bandwidth, systems that
830   broadcast or multicast cache entries, organizations that distribute
[639]831   subsets of cached data via optical media, and so on.
[1148]835<section title="Protocol Versioning" anchor="http.version">
[625]836  <x:anchor-alias value="HTTP-Version"/>
837  <x:anchor-alias value="HTTP-Prot-Name"/>
[1148]839   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate
840   versions of the protocol. This specification defines version "1.1".
841   The protocol version as a whole indicates the sender's compliance
842   with the set of requirements laid out in that version's corresponding
843   specification of HTTP.
846   The version of an HTTP message is indicated by an HTTP-Version field
847   in the first line of the message. HTTP-Version is case-sensitive.
849<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-Version"/><iref primary="true" item="Grammar" subitem="HTTP-Prot-Name"/>
[1313]850  <x:ref>HTTP-Version</x:ref>   = <x:ref>HTTP-Prot-Name</x:ref> "/" <x:ref>DIGIT</x:ref> "." <x:ref>DIGIT</x:ref>
[625]851  <x:ref>HTTP-Prot-Name</x:ref> = <x:abnf-char-sequence>"HTTP"</x:abnf-char-sequence> ; "HTTP", case-sensitive
[1313]854   The HTTP version number consists of two decimal digits separated by a "."
855   (period or decimal point).  The first digit ("major version") indicates the
856   HTTP messaging syntax, whereas the second digit ("minor version") indicates
857   the highest minor version to which the sender is at least conditionally
858   compliant and able to understand for future communication.  The minor
859   version advertises the sender's communication capabilities even when the
[1148]860   sender is only using a backwards-compatible subset of the protocol,
861   thereby letting the recipient know that more advanced features can
862   be used in response (by servers) or in future requests (by clients).
[1148]865   When an HTTP/1.1 message is sent to an HTTP/1.0 recipient
[1170]866   <xref target="RFC1945"/> or a recipient whose version is unknown,
867   the HTTP/1.1 message is constructed such that it can be interpreted
868   as a valid HTTP/1.0 message if all of the newer features are ignored.
869   This specification places recipient-version requirements on some
870   new features so that a compliant sender will only use compatible
871   features until it has determined, through configuration or the
872   receipt of a message, that the recipient supports HTTP/1.1.
[1170]875   The interpretation of an HTTP header field does not change
[1148]876   between minor versions of the same major version, though the default
877   behavior of a recipient in the absence of such a field can change.
878   Unless specified otherwise, header fields defined in HTTP/1.1 are
[1170]879   defined for all versions of HTTP/1.x.  In particular, the Host and
880   Connection header fields ought to be implemented by all HTTP/1.x
881   implementations whether or not they advertise compliance with HTTP/1.1.
[1170]884   New header fields can be defined such that, when they are
[1148]885   understood by a recipient, they might override or enhance the
886   interpretation of previously defined header fields.  When an
887   implementation receives an unrecognized header field, the recipient
888   &MUST; ignore that header field for local processing regardless of
889   the message's HTTP version.  An unrecognized header field received
890   by a proxy &MUST; be forwarded downstream unless the header field's
891   field-name is listed in the message's Connection header-field
892   (see <xref target="header.connection"/>).
893   These requirements allow HTTP's functionality to be enhanced without
894   requiring prior update of all compliant intermediaries.
897   Intermediaries that process HTTP messages (i.e., all intermediaries
[1514]898   other than those acting as tunnels) &MUST; send their own HTTP-Version
[1148]899   in forwarded messages.  In other words, they &MUST-NOT; blindly
900   forward the first line of an HTTP message without ensuring that the
901   protocol version matches what the intermediary understands, and
902   is at least conditionally compliant to, for both the receiving and
903   sending of messages.  Forwarding an HTTP message without rewriting
904   the HTTP-Version might result in communication errors when downstream
905   recipients use the message sender's version to determine what features
906   are safe to use for later communication with that sender.
909   An HTTP client &SHOULD; send a request version equal to the highest
910   version for which the client is at least conditionally compliant and
911   whose major version is no higher than the highest version supported
912   by the server, if this is known.  An HTTP client &MUST-NOT; send a
913   version for which it is not at least conditionally compliant.
916   An HTTP client &MAY; send a lower request version if it is known that
917   the server incorrectly implements the HTTP specification, but only
918   after the client has attempted at least one normal request and determined
919   from the response status or header fields (e.g., Server) that the
920   server improperly handles higher request versions.
923   An HTTP server &SHOULD; send a response version equal to the highest
924   version for which the server is at least conditionally compliant and
925   whose major version is less than or equal to the one received in the
926   request.  An HTTP server &MUST-NOT; send a version for which it is not
927   at least conditionally compliant.  A server &MAY; send a 505 (HTTP
928   Version Not Supported) response if it cannot send a response using the
929   major version used in the client's request.
932   An HTTP server &MAY; send an HTTP/1.0 response to an HTTP/1.0 request
933   if it is known or suspected that the client incorrectly implements the
934   HTTP specification and is incapable of correctly processing later
935   version responses, such as when a client fails to parse the version
936   number correctly or when an intermediary is known to blindly forward
937   the HTTP-Version even when it doesn't comply with the given minor
938   version of the protocol. Such protocol downgrades &SHOULD-NOT; be
939   performed unless triggered by specific client attributes, such as when
940   one or more of the request header fields (e.g., User-Agent) uniquely
941   match the values sent by a client known to be in error.
944   The intention of HTTP's versioning design is that the major number
945   will only be incremented if an incompatible message syntax is
946   introduced, and that the minor number will only be incremented when
947   changes made to the protocol have the effect of adding to the message
948   semantics or implying additional capabilities of the sender.  However,
[1170]949   the minor version was not incremented for the changes introduced between
950   <xref target="RFC2068"/> and <xref target="RFC2616"/>, and this revision
951   is specifically avoiding any such changes to the protocol.
[391]955<section title="Uniform Resource Identifiers" anchor="uri">
[621]956<iref primary="true" item="resource"/>
958   Uniform Resource Identifiers (URIs) <xref target="RFC3986"/> are used
959   throughout HTTP as the means for identifying resources. URI references
[621]960   are used to target requests, indicate redirects, and define relationships.
[901]961   HTTP does not limit what a resource might be; it merely defines an interface
[391]962   that can be used to interact with a resource via HTTP. More information on
963   the scope of URIs and resources can be found in <xref target="RFC3986"/>.
965  <x:anchor-alias value="URI-reference"/>
966  <x:anchor-alias value="absolute-URI"/>
967  <x:anchor-alias value="relative-part"/>
968  <x:anchor-alias value="authority"/>
969  <x:anchor-alias value="path-abempty"/>
970  <x:anchor-alias value="path-absolute"/>
971  <x:anchor-alias value="port"/>
972  <x:anchor-alias value="query"/>
973  <x:anchor-alias value="uri-host"/>
974  <x:anchor-alias value="partial-URI"/>
976   This specification adopts the definitions of "URI-reference",
[649]977   "absolute-URI", "relative-part", "port", "host",
[1171]978   "path-abempty", "path-absolute", "query", and "authority" from the
979   URI generic syntax <xref target="RFC3986"/>.
980   In addition, we define a partial-URI rule for protocol elements
981   that allow a relative URI but not a fragment.
983<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="URI-reference"/><iref primary="true" item="Grammar" subitem="absolute-URI"/><iref primary="true" item="Grammar" subitem="authority"/><iref primary="true" item="Grammar" subitem="path-absolute"/><iref primary="true" item="Grammar" subitem="port"/><iref primary="true" item="Grammar" subitem="query"/><iref primary="true" item="Grammar" subitem="uri-host"/>
[395]984  <x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in <xref target="RFC3986" x:fmt="," x:sec="4.1"/>&gt;
985  <x:ref>absolute-URI</x:ref>  = &lt;absolute-URI, defined in <xref target="RFC3986" x:fmt="," x:sec="4.3"/>&gt;
986  <x:ref>relative-part</x:ref> = &lt;relative-part, defined in <xref target="RFC3986" x:fmt="," x:sec="4.2"/>&gt;
987  <x:ref>authority</x:ref>     = &lt;authority, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2"/>&gt;
988  <x:ref>path-abempty</x:ref>  = &lt;path-abempty, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
989  <x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
990  <x:ref>port</x:ref>          = &lt;port, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.3"/>&gt;
991  <x:ref>query</x:ref>         = &lt;query, defined in <xref target="RFC3986" x:fmt="," x:sec="3.4"/>&gt;
992  <x:ref>uri-host</x:ref>      = &lt;host, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>&gt;
994  <x:ref>partial-URI</x:ref>   = relative-part [ "?" query ]
[1171]997   Each protocol element in HTTP that allows a URI reference will indicate
998   in its ABNF production whether the element allows any form of reference
999   (URI-reference), only a URI in absolute form (absolute-URI), only the
1000   path and optional query components, or some combination of the above.
1001   Unless otherwise indicated, URI references are parsed relative to the
1002   effective request URI, which defines the default base URI for references
1003   in both the request and its corresponding response.
1006<section title="http URI scheme" anchor="http.uri">
1007  <x:anchor-alias value="http-URI"/>
1008  <iref item="http URI scheme" primary="true"/>
1009  <iref item="URI scheme" subitem="http" primary="true"/>
[621]1011   The "http" URI scheme is hereby defined for the purpose of minting
1012   identifiers according to their association with the hierarchical
1013   namespace governed by a potential HTTP origin server listening for
1014   TCP connections on a given port.
1016<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http-URI"/>
1017  <x:ref>http-URI</x:ref> = "http:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
[1171]1020   The HTTP origin server is identified by the generic syntax's
1021   <x:ref>authority</x:ref> component, which includes a host identifier
1022   and optional TCP port (<xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>).
1023   The remainder of the URI, consisting of both the hierarchical path
1024   component and optional query component, serves as an identifier for
1025   a potential resource within that origin server's name space.
1028   If the host identifier is provided as an IP literal or IPv4 address,
1029   then the origin server is any listener on the indicated TCP port at
1030   that IP address. If host is a registered name, then that name is
1031   considered an indirect identifier and the recipient might use a name
1032   resolution service, such as DNS, to find the address of a listener
1033   for that host.
[621]1034   The host &MUST-NOT; be empty; if an "http" URI is received with an
1035   empty host, then it &MUST; be rejected as invalid.
1036   If the port subcomponent is empty or not given, then TCP port 80 is
1037   assumed (the default reserved port for WWW services).
1040   Regardless of the form of host identifier, access to that host is not
[901]1041   implied by the mere presence of its name or address. The host might or might
1042   not exist and, even when it does exist, might or might not be running an
[621]1043   HTTP server or listening to the indicated port. The "http" URI scheme
1044   makes use of the delegated nature of Internet names and addresses to
1045   establish a naming authority (whatever entity has the ability to place
1046   an HTTP server at that Internet name or address) and allows that
1047   authority to determine which names are valid and how they might be used.
1050   When an "http" URI is used within a context that calls for access to the
1051   indicated resource, a client &MAY; attempt access by resolving
1052   the host to an IP address, establishing a TCP connection to that address
[1435]1053   on the indicated port, and sending an HTTP request message
1054   (<xref target="http.message"/>) containing the URI's identifying data
1055   (<xref target="message.routing"/>) to the server.
[621]1056   If the server responds to that request with a non-interim HTTP response
[1435]1057   message, as described in &status-code-reasonphr;, then that response
[621]1058   is considered an authoritative answer to the client's request.
1061   Although HTTP is independent of the transport protocol, the "http"
1062   scheme is specific to TCP-based services because the name delegation
1063   process depends on TCP for establishing authority.
1064   An HTTP service based on some other underlying connection protocol
1065   would presumably be identified using a different URI scheme, just as
1066   the "https" scheme (below) is used for servers that require an SSL/TLS
[901]1067   transport layer on a connection. Other protocols might also be used to
[1101]1068   provide access to "http" identified resources &mdash; it is only the
[621]1069   authoritative interface used for mapping the namespace that is
1070   specific to TCP.
1073   The URI generic syntax for authority also includes a deprecated
1074   userinfo subcomponent (<xref target="RFC3986" x:fmt="," x:sec="3.2.1"/>)
[1157]1075   for including user authentication information in the URI.  Some
1076   implementations make use of the userinfo component for internal
1077   configuration of authentication information, such as within command
1078   invocation options, configuration files, or bookmark lists, even
1079   though such usage might expose a user identifier or password.
1080   Senders &MUST-NOT; include a userinfo subcomponent (and its "@"
1081   delimiter) when transmitting an "http" URI in a message.  Recipients
1082   of HTTP messages that contain a URI reference &SHOULD; parse for the
1083   existence of userinfo and treat its presence as an error, likely
1084   indicating that the deprecated subcomponent is being used to obscure
1085   the authority for the sake of phishing attacks.
1089<section title="https URI scheme" anchor="https.uri">
[622]1090   <x:anchor-alias value="https-URI"/>
[452]1091   <iref item="https URI scheme"/>
1092   <iref item="URI scheme" subitem="https"/>
[621]1094   The "https" URI scheme is hereby defined for the purpose of minting
1095   identifiers according to their association with the hierarchical
1096   namespace governed by a potential HTTP origin server listening for
1097   SSL/TLS-secured connections on a given TCP port.
1100   All of the requirements listed above for the "http" scheme are also
1101   requirements for the "https" scheme, except that a default TCP port
1102   of 443 is assumed if the port subcomponent is empty or not given,
1103   and the TCP connection &MUST; be secured for privacy through the
1104   use of strong encryption prior to sending the first HTTP request.
[621]1106<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="https-URI"/>
1107  <x:ref>https-URI</x:ref> = "https:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
[877]1110   Unlike the "http" scheme, responses to "https" identified requests
[1171]1111   are never "public" and thus &MUST-NOT; be reused for shared caching.
1112   They can, however, be reused in a private cache if the message is
1113   cacheable by default in HTTP or specifically indicated as such by
1114   the Cache-Control header field (&header-cache-control;).
[877]1117   Resources made available via the "https" scheme have no shared
1118   identity with the "http" scheme even if their resource identifiers
[1171]1119   indicate the same authority (the same host listening to the same
1120   TCP port).  They are distinct name spaces and are considered to be
1121   distinct origin servers.  However, an extension to HTTP that is
1122   defined to apply to entire host domains, such as the Cookie protocol
[1275]1123   <xref target="RFC6265"/>, can allow information
[1171]1124   set by one service to impact communication with other services
1125   within a matching group of host domains.
[621]1128   The process for authoritative access to an "https" identified
1129   resource is defined in <xref target="RFC2818"/>.
[621]1133<section title="http and https URI Normalization and Comparison" anchor="uri.comparison">
[621]1135   Since the "http" and "https" schemes conform to the URI generic syntax,
1136   such URIs are normalized and compared according to the algorithm defined
1137   in <xref target="RFC3986" x:fmt="," x:sec="6"/>, using the defaults
1138   described above for each scheme.
[621]1141   If the port is equal to the default port for a scheme, the normal
1142   form is to elide the port subcomponent. Likewise, an empty path
1143   component is equivalent to an absolute path of "/", so the normal
1144   form is to provide a path of "/" instead. The scheme and host
1145   are case-insensitive and normally provided in lowercase; all
1146   other components are compared in a case-sensitive manner.
1147   Characters other than those in the "reserved" set are equivalent
1148   to their percent-encoded octets (see <xref target="RFC3986"
1149   x:fmt="," x:sec="2.1"/>): the normal form is to not encode them.
[391]1152   For example, the following three URIs are equivalent:
1154<figure><artwork type="example">
[1176]1163<section title="Message Format" anchor="http.message">
[647]1164<x:anchor-alias value="generic-message"/>
1165<x:anchor-alias value="message.types"/>
1166<x:anchor-alias value="HTTP-message"/>
1167<x:anchor-alias value="start-line"/>
1168<iref item="header section"/>
1169<iref item="headers"/>
1170<iref item="header field"/>
[647]1172   All HTTP/1.1 messages consist of a start-line followed by a sequence of
[1176]1173   octets in a format similar to the Internet Message Format
[647]1174   <xref target="RFC5322"/>: zero or more header fields (collectively
1175   referred to as the "headers" or the "header section"), an empty line
1176   indicating the end of the header section, and an optional message-body.
[647]1178<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"/>
1179  <x:ref>HTTP-message</x:ref>    = <x:ref>start-line</x:ref>
1180                    *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
[229]1181                    <x:ref>CRLF</x:ref>
1182                    [ <x:ref>message-body</x:ref> ]
[1393]1185   The normal procedure for parsing an HTTP message is to read the
1186   start-line into a structure, read each header field into a hash
1187   table by field name until the empty line, and then use the parsed
1188   data to determine if a message-body is expected.  If a message-body
1189   has been indicated, then it is read as a stream until an amount
1190   of octets equal to the message-body length is read or the connection
1191   is closed.
[1424]1194   Recipients &MUST; parse an HTTP message as a sequence of octets in an
1195   encoding that is a superset of US-ASCII <xref target="USASCII"/>.
1196   Parsing an HTTP message as a stream of Unicode characters, without regard
1197   for the specific encoding, creates security vulnerabilities due to the
1198   varying ways that string processing libraries handle invalid multibyte
1199   character sequences that contain the octet LF (%x0A).  String-based
1200   parsers can only be safely used within protocol elements after the element
1201   has been extracted from the message, such as within a header field-value
1202   after message parsing has delineated the individual fields.
[1435]1205<section title="Start Line" anchor="start.line">
1206  <x:anchor-alias value="Start-Line"/>
[1431]1208   An HTTP message can either be a request from client to server or a
1209   response from server to client.  Syntactically, the two types of message
1210   differ only in the start-line, which is either a Request-Line (for requests)
1211   or a Status-Line (for responses), and in the algorithm for determining
1212   the length of the message-body (<xref target="message.body"/>).
1213   In theory, a client could receive requests and a server could receive
1214   responses, distinguishing them by their different start-line formats,
1215   but in practice servers are implemented to only expect a request
1216   (a response is interpreted as an unknown or invalid request method)
1217   and clients are implemented to only expect a response.
[1431]1219<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="start-line"/>
1220  <x:ref>start-line</x:ref>      = <x:ref>Request-Line</x:ref> / <x:ref>Status-Line</x:ref>
[1431]1225   Implementations &MUST-NOT; send whitespace between the start-line and
1226   the first header field. The presence of such whitespace in a request
1227   might be an attempt to trick a server into ignoring that field or
1228   processing the line after it as a new request, either of which might
1229   result in a security vulnerability if other implementations within
1230   the request chain interpret the same message differently.
1231   Likewise, the presence of such whitespace in a response might be
1232   ignored by some clients or cause others to cease parsing.
[1435]1235<section title="Request-Line" anchor="request.line">
1236  <x:anchor-alias value="Request"/>
[1432]1237  <x:anchor-alias value="Request-Line"/>
1239   The Request-Line begins with a method token, followed by a single
1240   space (SP), the request-target, another single space (SP), the
1241   protocol version, and ending with CRLF.
1243<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Request-Line"/>
1244  <x:ref>Request-Line</x:ref>   = <x:ref>Method</x:ref> <x:ref>SP</x:ref> <x:ref>request-target</x:ref> <x:ref>SP</x:ref> <x:ref>HTTP-Version</x:ref> <x:ref>CRLF</x:ref>
1247<section title="Method" anchor="method">
1248  <x:anchor-alias value="Method"/>
1250   The Method token indicates the request method to be performed on the
1251   target resource. The request method is case-sensitive.
1253<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Method"/>
1254  <x:ref>Method</x:ref>         = <x:ref>token</x:ref>
1257   See &method; for further information, such as the list of methods defined
1258   by this specification, the IANA registry, and considerations for new methods.
[1432]1262<section title="request-target" anchor="request-target">
1263  <x:anchor-alias value="request-target"/>
1265   The request-target identifies the target resource upon which to apply
1266   the request.  The four options for request-target are described in
1267   <xref target="request-target-types"/>.
1269<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-target"/>
1270  <x:ref>request-target</x:ref> = "*"
1271                 / <x:ref>absolute-URI</x:ref>
1272                 / ( <x:ref>path-absolute</x:ref> [ "?" <x:ref>query</x:ref> ] )
1273                 / <x:ref>authority</x:ref>
1276   HTTP does not place a pre-defined limit on the length of a request-target.
1277   A server &MUST; be prepared to receive URIs of unbounded length and
1278   respond with the 414 (URI Too Long) status code if the received
1279   request-target would be longer than the server wishes to handle
1280   (see &status-414;).
1283   Various ad-hoc limitations on request-target length are found in practice.
1284   It is &RECOMMENDED; that all HTTP senders and recipients support
1285   request-target lengths of 8000 or more octets.
1288  <t>
1289    <x:h>Note:</x:h> Fragments (<xref target="RFC3986" x:fmt="," x:sec="3.5"/>)
1290    are not part of the request-target and thus will not be transmitted
1291    in an HTTP request.
1292  </t>
[1435]1297<section title="Response Status-Line" anchor="status.line">
1298  <x:anchor-alias value="Response"/>
[1432]1299  <x:anchor-alias value="Status-Line"/>
1301   The first line of a Response message is the Status-Line, consisting
1302   of the protocol version, a space (SP), the status code, another space,
1303   a possibly-empty textual phrase describing the status code, and
1304   ending with CRLF.
1306<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Line"/>
1307  <x:ref>Status-Line</x:ref> = <x:ref>HTTP-Version</x:ref> <x:ref>SP</x:ref> <x:ref>Status-Code</x:ref> <x:ref>SP</x:ref> <x:ref>Reason-Phrase</x:ref> <x:ref>CRLF</x:ref>
1310<section title="Status Code" anchor="status.code">
1311  <x:anchor-alias value="Status-Code"/>
1313   The Status-Code element is a 3-digit integer result code of the attempt to
1314   understand and satisfy the request. See &status-code-reasonphr; for
1315   further information, such as the list of status codes defined by this
1316   specification, the IANA registry, and considerations for new status codes.
1318<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Status-Code"/>
1319  <x:ref>Status-Code</x:ref>    = 3<x:ref>DIGIT</x:ref>
1323<section title="Reason Phrase" anchor="reason.phrase">
1324  <x:anchor-alias value="Reason-Phrase"/>
1326   The Reason Phrase exists for the sole purpose of providing a textual
1327   description associated with the numeric status code, out of deference to
1328   earlier Internet application protocols that were more frequently used with
1329   interactive text clients. A client &SHOULD; ignore the content of the Reason
1330   Phrase.
1332<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Reason-Phrase"/>
1333  <x:ref>Reason-Phrase</x:ref>  = *( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
[647]1340<section title="Header Fields" anchor="header.fields">
1341  <x:anchor-alias value="header-field"/>
[229]1342  <x:anchor-alias value="field-content"/>
1343  <x:anchor-alias value="field-name"/>
1344  <x:anchor-alias value="field-value"/>
[647]1345  <x:anchor-alias value="OWS"/>
[647]1347   Each HTTP header field consists of a case-insensitive field name
1348   followed by a colon (":"), optional whitespace, and the field value.
[647]1350<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="header-field"/><iref primary="true" item="Grammar" subitem="field-name"/><iref primary="true" item="Grammar" subitem="field-value"/><iref primary="true" item="Grammar" subitem="field-content"/>
[1425]1351  <x:ref>header-field</x:ref>   = <x:ref>field-name</x:ref> ":" <x:ref>OWS</x:ref> <x:ref>field-value</x:ref> <x:ref>BWS</x:ref>
[229]1352  <x:ref>field-name</x:ref>     = <x:ref>token</x:ref>
[1425]1353  <x:ref>field-value</x:ref>    = *( <x:ref>field-content</x:ref> / <x:ref>obs-fold</x:ref> )
1354  <x:ref>field-content</x:ref>  = *( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
[1393]1357   The field-name token labels the corresponding field-value as having the
1358   semantics defined by that header field.  For example, the Date header field
[1436]1359   is defined in &header-date; as containing the origination
[1393]1360   timestamp for the message in which it appears.
[1393]1363   HTTP header fields are fully extensible: there is no limit on the
1364   introduction of new field names, each presumably defining new semantics,
1365   or on the number of header fields used in a given message.  Existing
1366   fields are defined in each part of this specification and in many other
1367   specifications outside the standards process.
1368   New header fields can be introduced without changing the protocol version
1369   if their defined semantics allow them to be safely ignored by recipients
1370   that do not recognize them.
[1393]1373   New HTTP header fields &SHOULD; be registered with IANA according
[1420]1374   to the procedures in &cons-new-header-fields;.
[1393]1375   Unrecognized header fields &MUST; be forwarded by a proxy unless the
1376   field-name is listed in the Connection header field
1377   (<xref target="header.connection"/>) or the proxy is specifically
1378   configured to block or otherwise transform such fields.
1379   Unrecognized header fields &SHOULD; be ignored by other recipients.
[647]1382   The order in which header fields with differing field names are
1383   received is not significant. However, it is "good practice" to send
1384   header fields that contain control data first, such as Host on
1385   requests and Date on responses, so that implementations can decide
1386   when not to handle a message as early as possible.  A server &MUST;
1387   wait until the entire header section is received before interpreting
1388   a request message, since later header fields might include conditionals,
1389   authentication credentials, or deliberately misleading duplicate
1390   header fields that would impact request processing.
[651]1393   Multiple header fields with the same field name &MUST-NOT; be
1394   sent in a message unless the entire field value for that
[647]1395   header field is defined as a comma-separated list [i.e., #(values)].
1396   Multiple header fields with the same field name can be combined into
1397   one "field-name: field-value" pair, without changing the semantics of the
1398   message, by appending each subsequent field value to the combined
1399   field value in order, separated by a comma. The order in which
1400   header fields with the same field name are received is therefore
1401   significant to the interpretation of the combined field value;
1402   a proxy &MUST-NOT; change the order of these field values when
1403   forwarding a message.
1406  <t>
[994]1407   <x:h>Note:</x:h> The "Set-Cookie" header field as implemented in
[1151]1408   practice can occur multiple times, but does not use the list syntax, and
[1275]1409   thus cannot be combined into a single line (<xref target="RFC6265"/>). (See Appendix A.2.3 of <xref target="Kri2001"/>
[994]1410   for details.) Also note that the Set-Cookie2 header field specified in
[647]1411   <xref target="RFC2965"/> does not share this problem.
1412  </t>
1415<section title="Field Parsing" anchor="field.parsing">
[1393]1417   No whitespace is allowed between the header field-name and colon.
1418   In the past, differences in the handling of such whitespace have led to
1419   security vulnerabilities in request routing and response handling.
1420   Any received request message that contains whitespace between a header
1421   field-name and colon &MUST; be rejected with a response code of 400
1422   (Bad Request).  A proxy &MUST; remove any such whitespace from a response
1423   message before forwarding the message downstream.
1426   A field value &MAY; be preceded by optional whitespace (OWS); a single SP is
1427   preferred. The field value does not include any leading or trailing white
1428   space: OWS occurring before the first non-whitespace octet of the
1429   field value or after the last non-whitespace octet of the field value
1430   is ignored and &SHOULD; be removed before further processing (as this does
1431   not change the meaning of the header field).
[395]1434   Historically, HTTP header field values could be extended over multiple
1435   lines by preceding each extra line with at least one space or horizontal
[1425]1436   tab (obs-fold). This specification deprecates such line
[395]1437   folding except within the message/http media type
1438   (<xref target=""/>).
[1393]1439   HTTP senders &MUST-NOT; produce messages that include line folding
[395]1440   (i.e., that contain any field-content that matches the obs-fold rule) unless
1441   the message is intended for packaging within the message/http media type.
[1393]1442   HTTP recipients &SHOULD; accept line folding and replace any embedded
1443   obs-fold whitespace with either a single SP or a matching number of SP
1444   octets (to avoid buffer copying) prior to interpreting the field value or
1445   forwarding the message downstream.
1448   Historically, HTTP has allowed field content with text in the ISO-8859-1
1449   <xref target="ISO-8859-1"/> character encoding and supported other
1450   character sets only through use of <xref target="RFC2047"/> encoding.
1451   In practice, most HTTP header field values use only a subset of the
1452   US-ASCII character encoding <xref target="USASCII"/>. Newly defined
[1176]1453   header fields &SHOULD; limit their field values to US-ASCII octets.
[647]1454   Recipients &SHOULD; treat other (obs-text) octets in field content as
1455   opaque data.
1459<section title="Field Length" anchor="field.length">
1461   HTTP does not place a pre-defined limit on the length of header fields,
1462   either in isolation or as a set. A server &MUST; be prepared to receive
1463   request header fields of unbounded length and respond with a 4xx status
1464   code if the received header field(s) would be longer than the server wishes
1465   to handle.
1468   A client that receives response headers that are longer than it wishes to
1469   handle can only treat it as a server error.
1472   Various ad-hoc limitations on header length are found in practice. It is
1473   &RECOMMENDED; that all HTTP senders and recipients support messages whose
1474   combined header fields have 4000 or more octets.
1478<section title="Common Field ABNF Rules" anchor="field.rules">
1479<t anchor="rule.token.separators">
1480  <x:anchor-alias value="tchar"/>
1481  <x:anchor-alias value="token"/>
1482  <x:anchor-alias value="special"/>
1483  <x:anchor-alias value="word"/>
1484   Many HTTP/1.1 header field values consist of words (token or quoted-string)
1485   separated by whitespace or special characters. These special characters
1486   &MUST; be in a quoted string to be used within a parameter value (as defined
1487   in <xref target="transfer.codings"/>).
1489<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="word"/><iref primary="true" item="Grammar" subitem="token"/><iref primary="true" item="Grammar" subitem="tchar"/><iref primary="true" item="Grammar" subitem="special"/>
1490  <x:ref>word</x:ref>           = <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref>
1492  <x:ref>token</x:ref>          = 1*<x:ref>tchar</x:ref>
1494  IMPORTANT: when editing "tchar" make sure that "special" is updated accordingly!!!
1495 -->
1496  <x:ref>tchar</x:ref>          = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*"
1497                 / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~"
1498                 / <x:ref>DIGIT</x:ref> / <x:ref>ALPHA</x:ref>
1499                 ; any <x:ref>VCHAR</x:ref>, except <x:ref>special</x:ref>
1501  <x:ref>special</x:ref>        = "(" / ")" / "&lt;" / ">" / "@" / ","
1502                 / ";" / ":" / "\" / DQUOTE / "/" / "["
1503                 / "]" / "?" / "=" / "{" / "}"
1505<t anchor="rule.quoted-string">
1506  <x:anchor-alias value="quoted-string"/>
1507  <x:anchor-alias value="qdtext"/>
1508  <x:anchor-alias value="obs-text"/>
1509   A string of text is parsed as a single word if it is quoted using
1510   double-quote marks.
1512<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-string"/><iref primary="true" item="Grammar" subitem="qdtext"/><iref primary="true" item="Grammar" subitem="obs-text"/>
1513  <x:ref>quoted-string</x:ref>  = <x:ref>DQUOTE</x:ref> *( <x:ref>qdtext</x:ref> / <x:ref>quoted-pair</x:ref> ) <x:ref>DQUOTE</x:ref>
1514  <x:ref>qdtext</x:ref>         = <x:ref>OWS</x:ref> / %x21 / %x23-5B / %x5D-7E / <x:ref>obs-text</x:ref>
1515  <x:ref>obs-text</x:ref>       = %x80-FF
1517<t anchor="rule.quoted-pair">
1518  <x:anchor-alias value="quoted-pair"/>
1519   The backslash octet ("\") can be used as a single-octet
1520   quoting mechanism within quoted-string constructs:
1522<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
[1450]1523  <x:ref>quoted-pair</x:ref>    = "\" ( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
1526   Recipients that process the value of the quoted-string &MUST; handle a
1527   quoted-pair as if it were replaced by the octet following the backslash.
1530   Senders &SHOULD-NOT; escape octets in quoted-strings that do not require
1531   escaping (i.e., other than DQUOTE and the backslash octet).
[395]1533<t anchor="rule.comment">
1534  <x:anchor-alias value="comment"/>
1535  <x:anchor-alias value="ctext"/>
1536   Comments can be included in some HTTP header fields by surrounding
1537   the comment text with parentheses. Comments are only allowed in
1538   fields containing "comment" as part of their field value definition.
1540<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
[702]1541  <x:ref>comment</x:ref>        = "(" *( <x:ref>ctext</x:ref> / <x:ref>quoted-cpair</x:ref> / <x:ref>comment</x:ref> ) ")"
[687]1542  <x:ref>ctext</x:ref>          = <x:ref>OWS</x:ref> / %x21-27 / %x2A-5B / %x5D-7E / <x:ref>obs-text</x:ref>
[702]1544<t anchor="rule.quoted-cpair">
1545  <x:anchor-alias value="quoted-cpair"/>
[1176]1546   The backslash octet ("\") can be used as a single-octet
[703]1547   quoting mechanism within comment constructs:
1549<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-cpair"/>
[1450]1550  <x:ref>quoted-cpair</x:ref>    = "\" ( <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / <x:ref>VCHAR</x:ref> / <x:ref>obs-text</x:ref> )
[1393]1553   Senders &SHOULD-NOT; escape octets in comments that do not require escaping
1554   (i.e., other than the backslash octet "\" and the parentheses "(" and ")").
1559<section title="Message Body" anchor="message.body">
[229]1560  <x:anchor-alias value="message-body"/>
1562   The message-body (if any) of an HTTP message is used to carry the
[958]1563   payload body associated with the request or response.
1565<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
[852]1566  <x:ref>message-body</x:ref> = *OCTET
[958]1569   The message-body differs from the payload body only when a transfer-coding
[1156]1570   has been applied, as indicated by the Transfer-Encoding header field
1571   (<xref target="header.transfer-encoding"/>).  If more than one
1572   Transfer-Encoding header field is present in a message, the multiple
1573   field-values &MUST; be combined into one field-value, according to the
1574   algorithm defined in <xref target="header.fields"/>, before determining
1575   the message-body length.
[1156]1578   When one or more transfer-codings are applied to a payload in order to
1579   form the message-body, the Transfer-Encoding header field &MUST; contain
1580   the list of transfer-codings applied. Transfer-Encoding is a property of
1581   the message, not of the payload, and thus &MAY; be added or removed by
1582   any implementation along the request/response chain under the constraints
1583   found in <xref target="transfer.codings"/>.
1586   If a message is received that has multiple Content-Length header fields
1587   (<xref target="header.content-length"/>) with field-values consisting
1588   of the same decimal value, or a single Content-Length header field with
1589   a field value containing a list of identical decimal values (e.g.,
1590   "Content-Length: 42, 42"), indicating that duplicate Content-Length
1591   header fields have been generated or combined by an upstream message
[1214]1592   processor, then the recipient &MUST; either reject the message as invalid
1593   or replace the duplicated field-values with a single valid Content-Length
1594   field containing that decimal value prior to determining the message-body
1595   length.
[8]1598   The rules for when a message-body is allowed in a message differ for
1599   requests and responses.
1602   The presence of a message-body in a request is signaled by the
1603   inclusion of a Content-Length or Transfer-Encoding header field in
[852]1604   the request's header fields, even if the request method does not
1605   define any use for a message-body.  This allows the request
1606   message framing algorithm to be independent of method semantics.
1609   For response messages, whether or not a message-body is included with
1610   a message is dependent on both the request method and the response
[1432]1611   status code (<xref target="status.code"/>).
[852]1612   Responses to the HEAD request method never include a message-body
1613   because the associated response header fields (e.g., Transfer-Encoding,
1614   Content-Length, etc.) only indicate what their values would have been
[1161]1615   if the request method had been GET.  All 1xx (Informational), 204 (No Content),
[852]1616   and 304 (Not Modified) responses &MUST-NOT; include a message-body.
1617   All other responses do include a message-body, although the body
1618   &MAY; be of zero length.
[852]1621   The length of the message-body is determined by one of the following
[8]1622   (in order of precedence):
1625  <list style="numbers">
1626    <x:lt><t>
[852]1627     Any response to a HEAD request and any response with a status
1628     code of 100-199, 204, or 304 is always terminated by the first
1629     empty line after the header fields, regardless of the header
1630     fields present in the message, and thus cannot contain a message-body.
[8]1631    </t></x:lt>
1632    <x:lt><t>
[1156]1633     If a Transfer-Encoding header field is present
1634     and the "chunked" transfer-coding (<xref target="transfer.codings"/>)
[957]1635     is the final encoding, the message-body length is determined by reading
1636     and decoding the chunked data until the transfer-coding indicates the
1637     data is complete.
[852]1638    </t>
1639    <t>
[957]1640     If a Transfer-Encoding header field is present in a response and the
1641     "chunked" transfer-coding is not the final encoding, the message-body
1642     length is determined by reading the connection until it is closed by
1643     the server.
1644     If a Transfer-Encoding header field is present in a request and the
1645     "chunked" transfer-coding is not the final encoding, the message-body
1646     length cannot be determined reliably; the server &MUST; respond with
1647     the 400 (Bad Request) status code and then close the connection.
1648    </t>
1649    <t>
[1156]1650     If a message is received with both a Transfer-Encoding header field
1651     and a Content-Length header field, the Transfer-Encoding overrides
1652     the Content-Length.
[901]1653     Such a message might indicate an attempt to perform request or response
[852]1654     smuggling (bypass of security-related checks on message routing or content)
[969]1655     and thus ought to be handled as an error.  The provided Content-Length &MUST;
[852]1656     be removed, prior to forwarding the message downstream, or replaced with
1657     the real message-body length after the transfer-coding is decoded.
[8]1658    </t></x:lt>
1659    <x:lt><t>
[957]1660     If a message is received without Transfer-Encoding and with either
[1156]1661     multiple Content-Length header fields having differing field-values or
1662     a single Content-Length header field having an invalid value, then the
1663     message framing is invalid and &MUST; be treated as an error to
1664     prevent request or response smuggling.
[852]1665     If this is a request message, the server &MUST; respond with
[919]1666     a 400 (Bad Request) status code and then close the connection.
[1173]1667     If this is a response message received by a proxy, the proxy
1668     &MUST; discard the received response, send a 502 (Bad Gateway)
[919]1669     status code as its downstream response, and then close the connection.
[1156]1670     If this is a response message received by a user-agent, it &MUST; be
[1031]1671     treated as an error by discarding the message and closing the connection.
[8]1672    </t></x:lt>
1673    <x:lt><t>
[1156]1674     If a valid Content-Length header field
[957]1675     is present without Transfer-Encoding, its decimal value defines the
1676     message-body length in octets.  If the actual number of octets sent in
1677     the message is less than the indicated Content-Length, the recipient
1678     &MUST; consider the message to be incomplete and treat the connection
1679     as no longer usable.
1680     If the actual number of octets sent in the message is more than the indicated
1681     Content-Length, the recipient &MUST; only process the message-body up to the
1682     field value's number of octets; the remainder of the message &MUST; either
1683     be discarded or treated as the next message in a pipeline.  For the sake of
1684     robustness, a user-agent &MAY; attempt to detect and correct such an error
1685     in message framing if it is parsing the response to the last request on
[1366]1686     a connection and the connection has been closed by the server.
[957]1687    </t></x:lt>
1688    <x:lt><t>
[852]1689     If this is a request message and none of the above are true, then the
1690     message-body length is zero (no message-body is present).
[8]1691    </t></x:lt>
1692    <x:lt><t>
[852]1693     Otherwise, this is a response message without a declared message-body
1694     length, so the message-body length is determined by the number of octets
1695     received prior to the server closing the connection.
[8]1696    </t></x:lt>
1697  </list>
[852]1700   Since there is no way to distinguish a successfully completed,
1701   close-delimited message from a partially-received message interrupted
1702   by network failure, implementations &SHOULD; use encoding or
1703   length-delimited messages whenever possible.  The close-delimiting
1704   feature exists primarily for backwards compatibility with HTTP/1.0.
[852]1707   A server &MAY; reject a request that contains a message-body but
1708   not a Content-Length by responding with 411 (Length Required).
[872]1711   Unless a transfer-coding other than "chunked" has been applied,
1712   a client that sends a request containing a message-body &SHOULD;
1713   use a valid Content-Length header field if the message-body length
1714   is known in advance, rather than the "chunked" encoding, since some
1715   existing services respond to "chunked" with a 411 (Length Required)
1716   status code even though they understand the chunked encoding.  This
1717   is typically because such services are implemented via a gateway that
1718   requires a content-length in advance of being called and the server
1719   is unable or unwilling to buffer the entire request before processing.
[852]1722   A client that sends a request containing a message-body &MUST; include a
1723   valid Content-Length header field if it does not know the server will
1724   handle HTTP/1.1 (or later) requests; such knowledge can be in the form
1725   of specific user configuration or by remembering the version of a prior
1726   received response.
[1431]1730<section anchor="incomplete.messages" title="Handling Incomplete Messages">
1732   Request messages that are prematurely terminated, possibly due to a
1733   cancelled connection or a server-imposed time-out exception, &MUST;
1734   result in closure of the connection; sending an HTTP/1.1 error response
1735   prior to closing the connection is &OPTIONAL;.
[852]1738   Response messages that are prematurely terminated, usually by closure
[872]1739   of the connection prior to receiving the expected number of octets or by
[852]1740   failure to decode a transfer-encoded message-body, &MUST; be recorded
[1374]1741   as incomplete.  A response that terminates in the middle of the header
1742   block (before the empty line is received) cannot be assumed to convey the
[1390]1743   full semantics of the response and &MUST; be treated as an error.
[1374]1746   A message-body that uses the chunked transfer encoding is
1747   incomplete if the zero-sized chunk that terminates the encoding has not
1748   been received.  A message that uses a valid Content-Length is incomplete
1749   if the size of the message-body received (in octets) is less than the
1750   value given by Content-Length.  A response that has neither chunked
1751   transfer encoding nor Content-Length is terminated by closure of the
1752   connection, and thus is considered complete regardless of the number of
1753   message-body octets received, provided that the header block was received
1754   intact.
1757   A user agent &MUST-NOT; render an incomplete response message-body as if
1758   it were complete (i.e., some indication must be given to the user that an
1759   error occurred).  Cache requirements for incomplete responses are defined
1760   in &cache-incomplete;.
[957]1763   A server &MUST; read the entire request message-body or close
1764   the connection after sending its response, since otherwise the
1765   remaining data on a persistent connection would be misinterpreted
1766   as the next request.  Likewise,
1767   a client &MUST; read the entire response message-body if it intends
1768   to reuse the same connection for a subsequent request.  Pipelining
1769   multiple requests on a connection is described in <xref target="pipelining"/>.
[1431]1773<section title="Message Parsing Robustness" anchor="message.robustness">
[1431]1775   Older HTTP/1.0 client implementations might send an extra CRLF
1776   after a POST request as a lame workaround for some early server
1777   applications that failed to read message-body content that was
1778   not terminated by a line-ending. An HTTP/1.1 client &MUST-NOT;
1779   preface or follow a request with an extra CRLF.  If terminating
1780   the request message-body with a line-ending is desired, then the
1781   client &MUST; include the terminating CRLF octets as part of the
1782   message-body length.
1785   In the interest of robustness, servers &SHOULD; ignore at least one
1786   empty line received where a Request-Line is expected. In other words, if
1787   the server is reading the protocol stream at the beginning of a
1788   message and receives a CRLF first, it &SHOULD; ignore the CRLF.
1789   Likewise, although the line terminator for the start-line and header
1790   fields is the sequence CRLF, we recommend that recipients recognize a
1791   single LF as a line terminator and ignore any CR.
1794   When a server listening only for HTTP request messages, or processing
1795   what appears from the start-line to be an HTTP request message,
1796   receives a sequence of octets that does not match the HTTP-message
1797   grammar aside from the robustness exceptions listed above, the
1798   server &MUST; respond with an HTTP/1.1 400 (Bad Request) response. 
[1435]1803<section title="Message Routing" anchor="message.routing">
[1432]1805   In most cases, the user agent is provided a URI reference
[1171]1806   from which it determines an absolute URI for identifying the target
1807   resource.  When a request to the resource is initiated, all or part
1808   of that URI is used to construct the HTTP request-target.
1811<section title="Types of Request Target" anchor="request-target-types">
[391]1813   The four options for request-target are dependent on the nature of the
[809]1814   request.
[1480]1816<t anchor="asterix-form"><iref item="asterisk form (of request-target)"/>
[1171]1817   The asterisk "*" form of request-target, which &MUST-NOT; be used
1818   with any request method other than OPTIONS, means that the request
1819   applies to the server as a whole (the listening process) rather than
1820   to a specific named resource at that server.  For example,
[803]1822<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
1823OPTIONS * HTTP/1.1
[1480]1825<t anchor="absolute-URI-form"><iref item="absolute-URI form (of request-target)"/>
[1063]1826   The "absolute-URI" form is &REQUIRED; when the request is being made to a
[1171]1827   proxy. The proxy is requested to either forward the request or service it
1828   from a valid cache, and then return the response. Note that the proxy &MAY;
[8]1829   forward the request on to another proxy or directly to the server
[374]1830   specified by the absolute-URI. In order to avoid request loops, a
[1171]1831   proxy that forwards requests to other proxies &MUST; be able to
1832   recognize and exclude all of its own server names, including
[8]1833   any aliases, local variations, and the numeric IP address. An example
1834   Request-Line would be:
[803]1836<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
1837GET HTTP/1.1
[374]1840   To allow for transition to absolute-URIs in all requests in future
1841   versions of HTTP, all HTTP/1.1 servers &MUST; accept the absolute-URI
[8]1842   form in requests, even though HTTP/1.1 clients will only generate
1843   them in requests to proxies.
1846   If a proxy receives a host name that is not a fully qualified domain
1847   name, it &MAY; add its domain to the host name it received. If a proxy
1848   receives a fully qualified domain name, the proxy &MUST-NOT; change
1849   the host name.
[1480]1851<t anchor="authority-form"><iref item="authority form (of request-target)"/>
[1161]1852   The "authority form" is only used by the CONNECT request method (&CONNECT;).
[1480]1854<t anchor="origin-form"><iref item="origin form (of request-target)"/>
[1171]1855   The most common form of request-target is that used when making
[1173]1856   a request to an origin server ("origin form").
[1171]1857   In this case, the absolute path and query components of the URI
1858   &MUST; be transmitted as the request-target, and the authority component
1859   &MUST; be transmitted in a Host header field. For example, a client wishing
1860   to retrieve a representation of the resource, as identified above,
1861   directly from the origin server would open (or reuse) a TCP connection
1862   to port 80 of the host "" and send the lines:
[803]1864<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
1865GET /pub/WWW/TheProject.html HTTP/1.1
[1171]1869   followed by the remainder of the Request. Note that the origin form
1870   of request-target always starts with an absolute path; if the target
1871   resource's URI path is empty, then an absolute path of "/" &MUST; be
1872   provided in the request-target.
[1171]1875   If a proxy receives an OPTIONS request with an absolute-URI form of
1876   request-target in which the URI has an empty path and no query component,
1877   then the last proxy on the request chain &MUST; use a request-target
1878   of "*" when it forwards the request to the indicated origin server.
1881   For example, the request
[803]1882</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
[1161]1886  would be forwarded by the final proxy as
[803]1887</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
1888OPTIONS * HTTP/1.1
1892   after connecting to port 8001 of host "".
[391]1896   The request-target is transmitted in the format specified in
[452]1897   <xref target="http.uri"/>. If the request-target is percent-encoded
1898   (<xref target="RFC3986" x:fmt="," x:sec="2.1"/>), the origin server
[391]1899   &MUST; decode the request-target in order to
[8]1900   properly interpret the request. Servers &SHOULD; respond to invalid
[391]1901   request-targets with an appropriate status code.
[1479]1904   A non-transforming proxy &MUST-NOT; rewrite the "path-absolute" and "query"
1905   parts of the received request-target when forwarding it to the next inbound
1906   server, except as noted above to replace a null path-absolute with "/" or
1907   "*".
1910  <t>
1911    <x:h>Note:</x:h> The "no rewrite" rule prevents the proxy from changing the
1912    meaning of the request when the origin server is improperly using
1913    a non-reserved URI character for a reserved purpose.  Implementors
[969]1914    need to be aware that some pre-HTTP/1.1 proxies have been known to
[563]1915    rewrite the request-target.
1916  </t>
1920<section title="The Resource Identified by a Request" anchor="">
1922   The exact resource identified by an Internet request is determined by
[391]1923   examining both the request-target and the Host header field.
1926   An origin server that does not allow resources to differ by the
1927   requested host &MAY; ignore the Host header field value when
1928   determining the resource identified by an HTTP/1.1 request. (But see
1929   <xref target=""/>
1930   for other requirements on Host support in HTTP/1.1.)
1933   An origin server that does differentiate resources based on the host
1934   requested (sometimes referred to as virtual hosts or vanity host
1935   names) &MUST; use the following rules for determining the requested
1936   resource on an HTTP/1.1 request:
1937  <list style="numbers">
[391]1938    <t>If request-target is an absolute-URI, the host is part of the
1939     request-target. Any Host header field value in the request &MUST; be
[8]1940     ignored.</t>
[391]1941    <t>If the request-target is not an absolute-URI, and the request includes
[8]1942     a Host header field, the host is determined by the Host header
1943     field value.</t>
1944    <t>If the host as determined by rule 1 or 2 is not a valid host on
1945     the server, the response &MUST; be a 400 (Bad Request) error message.</t>
1946  </list>
1949   Recipients of an HTTP/1.0 request that lacks a Host header field &MAY;
1950   attempt to use heuristics (e.g., examination of the URI path for
1951   something unique to a particular host) in order to determine what
1952   exact resource is being requested.
[823]1956<section title="Effective Request URI" anchor="effective.request.uri">
[965]1957  <iref primary="true" item="effective request URI"/>
1958  <iref primary="true" item="target resource"/>
1960   HTTP requests often do not carry the absolute URI (<xref target="RFC3986" x:fmt="," x:sec="4.3"/>)
[965]1961   for the target resource; instead, the URI needs to be inferred from the
1962   request-target, Host header field, and connection context. The result of
1963   this process is called the "effective request URI".  The "target resource"
1964   is the resource identified by the effective request URI.
[965]1967   If the request-target is an absolute-URI, then the effective request URI is
[823]1968   the request-target.
[1480]1971   If the request-target uses the origin form or the asterisk form,
[992]1972   and the Host header field is present, then the effective request URI is
1973   constructed by concatenating
1976  <list style="symbols">
1977    <t>
1978      the scheme name: "http" if the request was received over an insecure
[953]1979      TCP connection, or "https" when received over a SSL/TLS-secured TCP
[823]1980      connection,
1981    </t>
1982    <t>
[1176]1983      the octet sequence "://",
[823]1984    </t>
1985    <t>
[991]1986      the authority component, as specified in the Host header field
[992]1987      (<xref target=""/>), and
[823]1988    </t>
1989    <t>
1990      the request-target obtained from the Request-Line, unless the
1991      request-target is just the asterisk "*".
1992    </t>
1993  </list>
[1480]1996   If the request-target uses the origin form or the asterisk form,
[992]1997   and the Host header field is not present, then the effective request URI is
1998   undefined.
[972]2001   Otherwise, when request-target uses the authority form, the effective
[983]2002   request URI is undefined.
[965]2006   Example 1: the effective request URI for the message
2008<artwork type="example" x:indent-with="  ">
2009GET /pub/WWW/TheProject.html HTTP/1.1
2013  (received over an insecure TCP connection) is "http", plus "://", plus the
2014  authority component "", plus the request-target
2015  "/pub/WWW/TheProject.html", thus
2016  "".
[965]2021   Example 2: the effective request URI for the message
2023<artwork type="example" x:indent-with="  ">
[1340]2024OPTIONS * HTTP/1.1
2028  (received over an SSL/TLS secured TCP connection) is "https", plus "://", plus the
2029  authority component "", thus "".
[965]2033   Effective request URIs are compared using the rules described in
[823]2034   <xref target="uri.comparison"/>, except that empty path components &MUST-NOT;
2035   be treated as equivalent to an absolute path of "/".
[623]2041<section title="Protocol Parameters" anchor="protocol.parameters">
2043<section title="Transfer Codings" anchor="transfer.codings">
2044  <x:anchor-alias value="transfer-coding"/>
2045  <x:anchor-alias value="transfer-extension"/>
2047   Transfer-coding values are used to indicate an encoding
[901]2048   transformation that has been, can be, or might need to be applied to a
[852]2049   payload body in order to ensure "safe transport" through the network.
[623]2050   This differs from a content coding in that the transfer-coding is a
[852]2051   property of the message rather than a property of the representation
2052   that is being transferred.
2054<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
[673]2055  <x:ref>transfer-coding</x:ref>         = "chunked" ; <xref target="chunked.encoding"/>
2056                          / "compress" ; <xref target="compress.coding"/>
2057                          / "deflate" ; <xref target="deflate.coding"/>
2058                          / "gzip" ; <xref target="gzip.coding"/>
2059                          / <x:ref>transfer-extension</x:ref>
[623]2060  <x:ref>transfer-extension</x:ref>      = <x:ref>token</x:ref> *( <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> <x:ref>transfer-parameter</x:ref> )
2062<t anchor="rule.parameter">
2063  <x:anchor-alias value="attribute"/>
2064  <x:anchor-alias value="transfer-parameter"/>
2065  <x:anchor-alias value="value"/>
[852]2066   Parameters are in the form of attribute/value pairs.
2068<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-parameter"/><iref primary="true" item="Grammar" subitem="attribute"/><iref primary="true" item="Grammar" subitem="value"/><iref primary="true" item="Grammar" subitem="date2"/><iref primary="true" item="Grammar" subitem="date3"/>
2069  <x:ref>transfer-parameter</x:ref>      = <x:ref>attribute</x:ref> <x:ref>BWS</x:ref> "=" <x:ref>BWS</x:ref> <x:ref>value</x:ref>
2070  <x:ref>attribute</x:ref>               = <x:ref>token</x:ref>
[810]2071  <x:ref>value</x:ref>                   = <x:ref>word</x:ref>
2074   All transfer-coding values are case-insensitive. HTTP/1.1 uses
2075   transfer-coding values in the TE header field (<xref target="header.te"/>) and in
2076   the Transfer-Encoding header field (<xref target="header.transfer-encoding"/>).
[641]2079   Transfer-codings are analogous to the Content-Transfer-Encoding values of
2080   MIME, which were designed to enable safe transport of binary data over a
2081   7-bit transport service (<xref target="RFC2045" x:fmt="," x:sec="6"/>).
2082   However, safe transport
[623]2083   has a different focus for an 8bit-clean transfer protocol. In HTTP,
2084   the only unsafe characteristic of message-bodies is the difficulty in
[864]2085   determining the exact message body length (<xref target="message.body"/>),
2086   or the desire to encrypt data over a shared transport.
[852]2089   A server that receives a request message with a transfer-coding it does
2090   not understand &SHOULD; respond with 501 (Not Implemented) and then
2091   close the connection. A server &MUST-NOT; send transfer-codings to an HTTP/1.0
[623]2092   client.
[673]2095<section title="Chunked Transfer Coding" anchor="chunked.encoding">
2096  <iref item="chunked (Coding Format)"/>
2097  <iref item="Coding Format" subitem="chunked"/>
[623]2098  <x:anchor-alias value="chunk"/>
2099  <x:anchor-alias value="Chunked-Body"/>
2100  <x:anchor-alias value="chunk-data"/>
2101  <x:anchor-alias value="chunk-ext"/>
2102  <x:anchor-alias value="chunk-ext-name"/>
2103  <x:anchor-alias value="chunk-ext-val"/>
2104  <x:anchor-alias value="chunk-size"/>
2105  <x:anchor-alias value="last-chunk"/>
2106  <x:anchor-alias value="trailer-part"/>
[707]2107  <x:anchor-alias value="quoted-str-nf"/>
2108  <x:anchor-alias value="qdtext-nf"/>
2110   The chunked encoding modifies the body of a message in order to
2111   transfer it as a series of chunks, each with its own size indicator,
[965]2112   followed by an &OPTIONAL; trailer containing header fields. This
[623]2113   allows dynamically produced content to be transferred along with the
2114   information necessary for the recipient to verify that it has
2115   received the full message.
[707]2117<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Chunked-Body"/><iref primary="true" item="Grammar" subitem="chunk"/><iref primary="true" item="Grammar" subitem="chunk-size"/><iref primary="true" item="Grammar" subitem="last-chunk"/><iref primary="true" item="Grammar" subitem="chunk-ext"/><iref primary="true" item="Grammar" subitem="chunk-ext-name"/><iref primary="true" item="Grammar" subitem="chunk-ext-val"/><iref primary="true" item="Grammar" subitem="chunk-data"/><iref primary="true" item="Grammar" subitem="trailer-part"/><iref primary="true" item="Grammar" subitem="quoted-str-nf"/><iref primary="true" item="Grammar" subitem="qdtext-nf"/>
[623]2118  <x:ref>Chunked-Body</x:ref>   = *<x:ref>chunk</x:ref>
2119                   <x:ref>last-chunk</x:ref>
2120                   <x:ref>trailer-part</x:ref>
2121                   <x:ref>CRLF</x:ref>
[1425]2123  <x:ref>chunk</x:ref>          = <x:ref>chunk-size</x:ref> [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
[623]2124                   <x:ref>chunk-data</x:ref> <x:ref>CRLF</x:ref>
2125  <x:ref>chunk-size</x:ref>     = 1*<x:ref>HEXDIG</x:ref>
[1425]2126  <x:ref>last-chunk</x:ref>     = 1*("0") [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
[1425]2128  <x:ref>chunk-ext</x:ref>      = *( ";" <x:ref>chunk-ext-name</x:ref>
2129                      [ "=" <x:ref>chunk-ext-val</x:ref> ] )
[623]2130  <x:ref>chunk-ext-name</x:ref> = <x:ref>token</x:ref>
[707]2131  <x:ref>chunk-ext-val</x:ref>  = <x:ref>token</x:ref> / <x:ref>quoted-str-nf</x:ref>
[623]2132  <x:ref>chunk-data</x:ref>     = 1*<x:ref>OCTET</x:ref> ; a sequence of chunk-size octets
[965]2133  <x:ref>trailer-part</x:ref>   = *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
2135  <x:ref>quoted-str-nf</x:ref>  = <x:ref>DQUOTE</x:ref> *( <x:ref>qdtext-nf</x:ref> / <x:ref>quoted-pair</x:ref> ) <x:ref>DQUOTE</x:ref>
2136                 ; like <x:ref>quoted-string</x:ref>, but disallowing line folding
[1425]2137  <x:ref>qdtext-nf</x:ref>      = <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / %x21 / %x23-5B / %x5D-7E / <x:ref>obs-text</x:ref>
2140   The chunk-size field is a string of hex digits indicating the size of
2141   the chunk-data in octets. The chunked encoding is ended by any chunk whose size is
2142   zero, followed by the trailer, which is terminated by an empty line.
2145   The trailer allows the sender to include additional HTTP header
2146   fields at the end of the message. The Trailer header field can be
2147   used to indicate which header fields are included in a trailer (see
2148   <xref target="header.trailer"/>).
2151   A server using chunked transfer-coding in a response &MUST-NOT; use the
2152   trailer for any header fields unless at least one of the following is
2153   true:
2154  <list style="numbers">
2155    <t>the request included a TE header field that indicates "trailers" is
2156     acceptable in the transfer-coding of the  response, as described in
2157     <xref target="header.te"/>; or,</t>
2159    <t>the trailer fields consist entirely of optional metadata, and the
2160    recipient could use the message (in a manner acceptable to the server where
2161    the field originated) without receiving it. In other words, the server that
2162    generated the header (often but not always the origin server) is willing to
2163    accept the possibility that the trailer fields might be silently discarded
2164    along the path to the client.</t>
[623]2165  </list>
2168   This requirement prevents an interoperability failure when the
2169   message is being received by an HTTP/1.1 (or later) proxy and
2170   forwarded to an HTTP/1.0 recipient. It avoids a situation where
2171   compliance with the protocol would have necessitated a possibly
2172   infinite buffer on the proxy.
2175   A process for decoding the "chunked" transfer-coding
2176   can be represented in pseudo-code as:
2178<figure><artwork type="code">
2179  length := 0
2180  read chunk-size, chunk-ext (if any) and CRLF
2181  while (chunk-size &gt; 0) {
2182     read chunk-data and CRLF
[852]2183     append chunk-data to decoded-body
[623]2184     length := length + chunk-size
2185     read chunk-size and CRLF
2186  }
[852]2187  read header-field
2188  while (header-field not empty) {
2189     append header-field to existing header fields
2190     read header-field
[623]2191  }
2192  Content-Length := length
2193  Remove "chunked" from Transfer-Encoding
2196   All HTTP/1.1 applications &MUST; be able to receive and decode the
[852]2197   "chunked" transfer-coding and &MUST; ignore chunk-ext extensions
[623]2198   they do not understand.
2201   Since "chunked" is the only transfer-coding required to be understood
2202   by HTTP/1.1 recipients, it plays a crucial role in delimiting messages
2203   on a persistent connection.  Whenever a transfer-coding is applied to
2204   a payload body in a request, the final transfer-coding applied &MUST;
2205   be "chunked".  If a transfer-coding is applied to a response payload
2206   body, then either the final transfer-coding applied &MUST; be "chunked"
2207   or the message &MUST; be terminated by closing the connection. When the
2208   "chunked" transfer-coding is used, it &MUST; be the last transfer-coding
2209   applied to form the message-body. The "chunked" transfer-coding &MUST-NOT;
2210   be applied more than once in a message-body.
[673]2214<section title="Compression Codings" anchor="compression.codings">
2216   The codings defined below can be used to compress the payload of a
2217   message.
2220   <x:h>Note:</x:h> Use of program names for the identification of encoding formats
2221   is not desirable and is discouraged for future encodings. Their
2222   use here is representative of historical practice, not good
2223   design.
2226   <x:h>Note:</x:h> For compatibility with previous implementations of HTTP,
2227   applications &SHOULD; consider "x-gzip" and "x-compress" to be
2228   equivalent to "gzip" and "compress" respectively.
2231<section title="Compress Coding" anchor="compress.coding">
2232<iref item="compress (Coding Format)"/>
2233<iref item="Coding Format" subitem="compress"/>
2235   The "compress" format is produced by the common UNIX file compression
2236   program "compress". This format is an adaptive Lempel-Ziv-Welch
2237   coding (LZW).
2241<section title="Deflate Coding" anchor="deflate.coding">
2242<iref item="deflate (Coding Format)"/>
2243<iref item="Coding Format" subitem="deflate"/>
[801]2245   The "deflate" format is defined as the "deflate" compression mechanism
2246   (described in <xref target="RFC1951"/>) used inside the "zlib"
2247   data format (<xref target="RFC1950"/>).
2250  <t>
2251    <x:h>Note:</x:h> Some incorrect implementations send the "deflate"
2252    compressed data without the zlib wrapper.
2253   </t>
2257<section title="Gzip Coding" anchor="gzip.coding">
2258<iref item="gzip (Coding Format)"/>
2259<iref item="Coding Format" subitem="gzip"/>
2261   The "gzip" format is produced by the file compression program
2262   "gzip" (GNU zip), as described in <xref target="RFC1952"/>. This format is a
2263   Lempel-Ziv coding (LZ77) with a 32 bit CRC.
[670]2269<section title="Transfer Coding Registry" anchor="transfer.coding.registry">
2271   The HTTP Transfer Coding Registry defines the name space for the transfer
2272   coding names.
2275   Registrations &MUST; include the following fields:
2276   <list style="symbols">
2277     <t>Name</t>
2278     <t>Description</t>
2279     <t>Pointer to specification text</t>
2280   </list>
[808]2283   Names of transfer codings &MUST-NOT; overlap with names of content codings
2284   (&content-codings;), unless the encoding transformation is identical (as it
2285   is the case for the compression codings defined in
2286   <xref target="compression.codings"/>).
[942]2289   Values to be added to this name space require a specification
2290   (see "Specification Required" in <xref target="RFC5226" x:fmt="of" x:sec="4.1"/>), and &MUST;
[670]2291   conform to the purpose of transfer coding defined in this section.
2294   The registry itself is maintained at
2295   <eref target=""/>.
2300<section title="Product Tokens" anchor="product.tokens">
2301  <x:anchor-alias value="product"/>
2302  <x:anchor-alias value="product-version"/>
2304   Product tokens are used to allow communicating applications to
2305   identify themselves by software name and version. Most fields using
2306   product tokens also allow sub-products which form a significant part
2307   of the application to be listed, separated by whitespace. By
2308   convention, the products are listed in order of their significance
2309   for identifying the application.
2311<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="product"/><iref primary="true" item="Grammar" subitem="product-version"/>
2312  <x:ref>product</x:ref>         = <x:ref>token</x:ref> ["/" <x:ref>product-version</x:ref>]
2313  <x:ref>product-version</x:ref> = <x:ref>token</x:ref>
2316   Examples:
2318<figure><artwork type="example">
2319  User-Agent: CERN-LineMode/2.15 libwww/2.17b3
2320  Server: Apache/0.8.4
2323   Product tokens &SHOULD; be short and to the point. They &MUST-NOT; be
2324   used for advertising or other non-essential information. Although any
[1176]2325   token octet &MAY; appear in a product-version, this token &SHOULD;
[623]2326   only be used for a version identifier (i.e., successive versions of
2327   the same product &SHOULD; only differ in the product-version portion of
2328   the product value).
2332<section title="Quality Values" anchor="quality.values">
2333  <x:anchor-alias value="qvalue"/>
[994]2335   Both transfer codings (TE request header field, <xref target="header.te"/>)
[623]2336   and content negotiation (&content.negotiation;) use short "floating point"
2337   numbers to indicate the relative importance ("weight") of various
2338   negotiable parameters.  A weight is normalized to a real number in
2339   the range 0 through 1, where 0 is the minimum and 1 the maximum
2340   value. If a parameter has a quality value of 0, then content with
[746]2341   this parameter is "not acceptable" for the client. HTTP/1.1
[623]2342   applications &MUST-NOT; generate more than three digits after the
2343   decimal point. User configuration of these values &SHOULD; also be
2344   limited in this fashion.
2346<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="qvalue"/>
2347  <x:ref>qvalue</x:ref>         = ( "0" [ "." 0*3<x:ref>DIGIT</x:ref> ] )
2348                 / ( "1" [ "." 0*3("0") ] )
2351  <t>
2352     <x:h>Note:</x:h> "Quality values" is a misnomer, since these values merely represent
2353     relative degradation in desired quality.
2354  </t>
[8]2360<section title="Connections" anchor="connections">
2362<section title="Persistent Connections" anchor="persistent.connections">
2364<section title="Purpose" anchor="persistent.purpose">
2366   Prior to persistent connections, a separate TCP connection was
[1173]2367   established for each request, increasing the load on HTTP servers
[8]2368   and causing congestion on the Internet. The use of inline images and
[761]2369   other associated data often requires a client to make multiple
[8]2370   requests of the same server in a short amount of time. Analysis of
2371   these performance problems and results from a prototype
2372   implementation are available <xref target="Pad1995"/> <xref target="Spe"/>. Implementation experience and
[578]2373   measurements of actual HTTP/1.1 implementations show good
[8]2374   results <xref target="Nie1997"/>. Alternatives have also been explored, for example,
2375   T/TCP <xref target="Tou1998"/>.
2378   Persistent HTTP connections have a number of advantages:
2379  <list style="symbols">
2380      <t>
2381        By opening and closing fewer TCP connections, CPU time is saved
2382        in routers and hosts (clients, servers, proxies, gateways,
2383        tunnels, or caches), and memory used for TCP protocol control
2384        blocks can be saved in hosts.
2385      </t>
2386      <t>
2387        HTTP requests and responses can be pipelined on a connection.
2388        Pipelining allows a client to make multiple requests without
2389        waiting for each response, allowing a single TCP connection to
2390        be used much more efficiently, with much lower elapsed time.
2391      </t>
2392      <t>
2393        Network congestion is reduced by reducing the number of packets
2394        caused by TCP opens, and by allowing TCP sufficient time to
2395        determine the congestion state of the network.
2396      </t>
2397      <t>
2398        Latency on subsequent requests is reduced since there is no time
2399        spent in TCP's connection opening handshake.
2400      </t>
2401      <t>
2402        HTTP can evolve more gracefully, since errors can be reported
2403        without the penalty of closing the TCP connection. Clients using
2404        future versions of HTTP might optimistically try a new feature,
2405        but if communicating with an older server, retry with old
2406        semantics after an error is reported.
2407      </t>
2408    </list>
2411   HTTP implementations &SHOULD; implement persistent connections.
2415<section title="Overall Operation" anchor="persistent.overall">
2417   A significant difference between HTTP/1.1 and earlier versions of
2418   HTTP is that persistent connections are the default behavior of any
2419   HTTP connection. That is, unless otherwise indicated, the client
2420   &SHOULD; assume that the server will maintain a persistent connection,
2421   even after error responses from the server.
2424   Persistent connections provide a mechanism by which a client and a
2425   server can signal the close of a TCP connection. This signaling takes
2426   place using the Connection header field (<xref target="header.connection"/>). Once a close
2427   has been signaled, the client &MUST-NOT; send any more requests on that
2428   connection.
2431<section title="Negotiation" anchor="persistent.negotiation">
2433   An HTTP/1.1 server &MAY; assume that a HTTP/1.1 client intends to
[994]2434   maintain a persistent connection unless a Connection header field including
[8]2435   the connection-token "close" was sent in the request. If the server
2436   chooses to close the connection immediately after sending the
[994]2437   response, it &SHOULD; send a Connection header field including the
[761]2438   connection-token "close".
2441   An HTTP/1.1 client &MAY; expect a connection to remain open, but would
2442   decide to keep it open based on whether the response from a server
[994]2443   contains a Connection header field with the connection-token close. In case
[8]2444   the client does not want to maintain a connection for more than that
[994]2445   request, it &SHOULD; send a Connection header field including the
[8]2446   connection-token close.
2449   If either the client or the server sends the close token in the
[994]2450   Connection header field, that request becomes the last one for the
[8]2451   connection.
2454   Clients and servers &SHOULD-NOT;  assume that a persistent connection is
2455   maintained for HTTP versions less than 1.1 unless it is explicitly
2456   signaled. See <xref target="compatibility.with.http.1.0.persistent.connections"/> for more information on backward
2457   compatibility with HTTP/1.0 clients.
2460   In order to remain persistent, all messages on the connection &MUST;
2461   have a self-defined message length (i.e., one not defined by closure
[864]2462   of the connection), as described in <xref target="message.body"/>.
2466<section title="Pipelining" anchor="pipelining">
2468   A client that supports persistent connections &MAY; "pipeline" its
2469   requests (i.e., send multiple requests without waiting for each
2470   response). A server &MUST; send its responses to those requests in the
2471   same order that the requests were received.
2474   Clients which assume persistent connections and pipeline immediately
2475   after connection establishment &SHOULD; be prepared to retry their
2476   connection if the first pipelined attempt fails. If a client does
2477   such a retry, it &MUST-NOT; pipeline before it knows the connection is
2478   persistent. Clients &MUST; also be prepared to resend their requests if
2479   the server closes the connection before sending all of the
2480   corresponding responses.
[1161]2483   Clients &SHOULD-NOT; pipeline requests using non-idempotent request methods or
2484   non-idempotent sequences of request methods (see &idempotent-methods;). Otherwise, a
[8]2485   premature termination of the transport connection could lead to
2486   indeterminate results. A client wishing to send a non-idempotent
2487   request &SHOULD; wait to send that request until it has received the
[919]2488   response status line for the previous request.
2493<section title="Proxy Servers" anchor="persistent.proxy">
2495   It is especially important that proxies correctly implement the
2496   properties of the Connection header field as specified in <xref target="header.connection"/>.
2499   The proxy server &MUST; signal persistent connections separately with
2500   its clients and the origin servers (or other proxy servers) that it
2501   connects to. Each persistent connection applies to only one transport
2502   link.
2505   A proxy server &MUST-NOT; establish a HTTP/1.1 persistent connection
[578]2506   with an HTTP/1.0 client (but see <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/>
[994]2507   for information and discussion of the problems with the Keep-Alive header field
[578]2508   implemented by many HTTP/1.0 clients).
[994]2511<section title="End-to-end and Hop-by-hop Header Fields" anchor="end-to-end.and.hop-by-hop.header-fields">
[769]2513  <cref anchor="TODO-end-to-end" source="jre">
2514    Restored from <eref target=""/>.
[839]2515    See also <eref target=""/>.
[769]2516  </cref>
2519   For the purpose of defining the behavior of caches and non-caching
[994]2520   proxies, we divide HTTP header fields into two categories:
[769]2521  <list style="symbols">
[994]2522      <t>End-to-end header fields, which are  transmitted to the ultimate
2523        recipient of a request or response. End-to-end header fields in
[769]2524        responses MUST be stored as part of a cache entry and &MUST; be
2525        transmitted in any response formed from a cache entry.</t>
[994]2527      <t>Hop-by-hop header fields, which are meaningful only for a single
[769]2528        transport-level connection, and are not stored by caches or
2529        forwarded by proxies.</t>
2530  </list>
[994]2533   The following HTTP/1.1 header fields are hop-by-hop header fields:
[769]2534  <list style="symbols">
2535      <t>Connection</t>
2536      <t>Keep-Alive</t>
2537      <t>Proxy-Authenticate</t>
2538      <t>Proxy-Authorization</t>
2539      <t>TE</t>
2540      <t>Trailer</t>
2541      <t>Transfer-Encoding</t>
2542      <t>Upgrade</t>
2543  </list>
[994]2546   All other header fields defined by HTTP/1.1 are end-to-end header fields.
[994]2549   Other hop-by-hop header fields &MUST; be listed in a Connection header field
[769]2550   (<xref target="header.connection"/>).
[994]2554<section title="Non-modifiable Header Fields" anchor="non-modifiable.header-fields">
[769]2556  <cref anchor="TODO-non-mod-headers" source="jre">
2557    Restored from <eref target=""/>.
[839]2558    See also <eref target=""/>.
[769]2559  </cref>
2562   Some features of HTTP/1.1, such as Digest Authentication, depend on the
[1107]2563   value of certain end-to-end header fields. A non-transforming proxy &SHOULD-NOT;
[994]2564   modify an end-to-end header field unless the definition of that header field requires
[769]2565   or specifically allows that.
[1107]2568   A non-transforming proxy &MUST-NOT; modify any of the following fields in a
[769]2569   request or response, and it &MUST-NOT; add any of these fields if not
2570   already present:
2571  <list style="symbols">
[1245]2572    <t>Allow</t>
2573    <t>Content-Location</t>
2574    <t>Content-MD5</t>
2575    <t>ETag</t>
2576    <t>Last-Modified</t>
2577    <t>Server</t>
[769]2578  </list>
[1107]2581   A non-transforming proxy &MUST-NOT; modify any of the following fields in a
[769]2582   response:
2583  <list style="symbols">
2584    <t>Expires</t>
2585  </list>
2588   but it &MAY; add any of these fields if not already present. If an
[994]2589   Expires header field is added, it &MUST; be given a field-value identical to
2590   that of the Date header field in that response.
2593   A proxy &MUST-NOT; modify or add any of the following fields in a
2594   message that contains the no-transform cache-control directive, or in
2595   any request:
2596  <list style="symbols">
2597    <t>Content-Encoding</t>
2598    <t>Content-Range</t>
2599    <t>Content-Type</t>
2600  </list>
[1107]2603   A transforming proxy &MAY; modify or add these fields to a message
[769]2604   that does not include no-transform, but if it does so, it &MUST; add a
2605   Warning 214 (Transformation applied) if one does not already appear
2606   in the message (see &header-warning;).
2609  <t>
[994]2610    <x:h>Warning:</x:h> Unnecessary modification of end-to-end header fields might
[769]2611    cause authentication failures if stronger authentication
2612    mechanisms are introduced in later versions of HTTP. Such
2613    authentication mechanisms &MAY; rely on the values of header fields
2614    not listed here.
2615  </t>
[1107]2618   A non-transforming proxy &MUST; preserve the message payload (&payload;),
[852]2619   though it &MAY; change the message-body through application or removal
2620   of a transfer-coding (<xref target="transfer.codings"/>).
[8]2626<section title="Practical Considerations" anchor="persistent.practical">
2628   Servers will usually have some time-out value beyond which they will
2629   no longer maintain an inactive connection. Proxy servers might make
2630   this a higher value since it is likely that the client will be making
2631   more connections through the same server. The use of persistent
2632   connections places no requirements on the length (or existence) of
2633   this time-out for either the client or the server.
2636   When a client or server wishes to time-out it &SHOULD; issue a graceful
2637   close on the transport connection. Clients and servers &SHOULD; both
2638   constantly watch for the other side of the transport close, and
2639   respond to it as appropriate. If a client or server does not detect
2640   the other side's close promptly it could cause unnecessary resource
2641   drain on the network.
2644   A client, server, or proxy &MAY; close the transport connection at any
2645   time. For example, a client might have started to send a new request
2646   at the same time that the server has decided to close the "idle"
2647   connection. From the server's point of view, the connection is being
2648   closed while it was idle, but from the client's point of view, a
2649   request is in progress.
[715]2652   Clients (including proxies) &SHOULD; limit the number of simultaneous
2653   connections that they maintain to a given server (including proxies).
2656   Previous revisions of HTTP gave a specific number of connections as a
2657   ceiling, but this was found to be impractical for many applications. As a
2658   result, this specification does not mandate a particular maximum number of
2659   connections, but instead encourages clients to be conservative when opening
2660   multiple connections.
2663   In particular, while using multiple connections avoids the "head-of-line
2664   blocking" problem (whereby a request that takes significant server-side
2665   processing and/or has a large payload can block subsequent requests on the
2666   same connection), each connection used consumes server resources (sometimes
2667   significantly), and furthermore using multiple connections can cause
2668   undesirable side effects in congested networks.
2671   Note that servers might reject traffic that they deem abusive, including an
2672   excessive number of connections from a client.
2676<section title="Retrying Requests" anchor="persistent.retrying.requests">
2678   Senders can close the transport connection at any time. Therefore,
2679   clients, servers, and proxies &MUST; be able to recover
2680   from asynchronous close events. Client software &MAY; reopen the
2681   transport connection and retransmit the aborted sequence of requests
2682   without user interaction so long as the request sequence is
2683   idempotent (see &idempotent-methods;). Non-idempotent request methods or sequences
2684   &MUST-NOT; be automatically retried, although user agents &MAY; offer a
2685   human operator the choice of retrying the request(s). Confirmation by
2686   user-agent software with semantic understanding of the application
2687   &MAY; substitute for user confirmation. The automatic retry &SHOULD-NOT; 
2688   be repeated if the second sequence of requests fails.
[8]2694<section title="Message Transmission Requirements" anchor="message.transmission.requirements">
2696<section title="Persistent Connections and Flow Control" anchor="persistent.flow">
2698   HTTP/1.1 servers &SHOULD; maintain persistent connections and use TCP's
2699   flow control mechanisms to resolve temporary overloads, rather than
2700   terminating connections with the expectation that clients will retry.
2701   The latter technique can exacerbate network congestion.
2705<section title="Monitoring Connections for Error Status Messages" anchor="persistent.monitor">
2707   An HTTP/1.1 (or later) client sending a message-body &SHOULD; monitor
[919]2708   the network connection for an error status code while it is transmitting
2709   the request. If the client sees an error status code, it &SHOULD;
[8]2710   immediately cease transmitting the body. If the body is being sent
2711   using a "chunked" encoding (<xref target="transfer.codings"/>), a zero length chunk and
2712   empty trailer &MAY; be used to prematurely mark the end of the message.
[994]2713   If the body was preceded by a Content-Length header field, the client &MUST;
[8]2714   close the connection.
2718<section title="Use of the 100 (Continue) Status" anchor="use.of.the.100.status">
[919]2720   The purpose of the 100 (Continue) status code (see &status-100;) is to
[8]2721   allow a client that is sending a request message with a request body
2722   to determine if the origin server is willing to accept the request
[994]2723   (based on the request header fields) before the client sends the request
[8]2724   body. In some cases, it might either be inappropriate or highly
2725   inefficient for the client to send the body if the server will reject
2726   the message without looking at the body.
2729   Requirements for HTTP/1.1 clients:
2730  <list style="symbols">
2731    <t>
2732        If a client will wait for a 100 (Continue) response before
[1163]2733        sending the request body, it &MUST; send an Expect header
[29]2734        field (&header-expect;) with the "100-continue" expectation.
[8]2735    </t>
2736    <t>
[1163]2737        A client &MUST-NOT; send an Expect header field (&header-expect;)
[8]2738        with the "100-continue" expectation if it does not intend
2739        to send a request body.
2740    </t>
2741  </list>
2744   Because of the presence of older implementations, the protocol allows
[901]2745   ambiguous situations in which a client might send "Expect: 100-continue"
[919]2746   without receiving either a 417 (Expectation Failed)
2747   or a 100 (Continue) status code. Therefore, when a client sends this
[8]2748   header field to an origin server (possibly via a proxy) from which it
[919]2749   has never seen a 100 (Continue) status code, the client &SHOULD-NOT; 
2750   wait for an indefinite period before sending the request body.
2753   Requirements for HTTP/1.1 origin servers:
2754  <list style="symbols">
[1163]2755    <t> Upon receiving a request which includes an Expect header
[8]2756        field with the "100-continue" expectation, an origin server &MUST;
[919]2757        either respond with 100 (Continue) status code and continue to read
[8]2758        from the input stream, or respond with a final status code. The
2759        origin server &MUST-NOT; wait for the request body before sending
2760        the 100 (Continue) response. If it responds with a final status
2761        code, it &MAY; close the transport connection or it &MAY; continue
2762        to read and discard the rest of the request.  It &MUST-NOT;
[1161]2763        perform the request method if it returns a final status code.
[8]2764    </t>
2765    <t> An origin server &SHOULD-NOT;  send a 100 (Continue) response if
[1163]2766        the request message does not include an Expect header
[8]2767        field with the "100-continue" expectation, and &MUST-NOT; send a
2768        100 (Continue) response if such a request comes from an HTTP/1.0
2769        (or earlier) client. There is an exception to this rule: for
[97]2770        compatibility with <xref target="RFC2068"/>, a server &MAY; send a 100 (Continue)
[919]2771        status code in response to an HTTP/1.1 PUT or POST request that does
[1163]2772        not include an Expect header field with the "100-continue"
[8]2773        expectation. This exception, the purpose of which is
2774        to minimize any client processing delays associated with an
[919]2775        undeclared wait for 100 (Continue) status code, applies only to
[8]2776        HTTP/1.1 requests, and not to requests with any other HTTP-version
2777        value.
2778    </t>
2779    <t> An origin server &MAY; omit a 100 (Continue) response if it has
2780        already received some or all of the request body for the
2781        corresponding request.
2782    </t>
2783    <t> An origin server that sends a 100 (Continue) response &MUST;
2784    ultimately send a final status code, once the request body is
2785        received and processed, unless it terminates the transport
2786        connection prematurely.
2787    </t>
2788    <t> If an origin server receives a request that does not include an
[1163]2789        Expect header field with the "100-continue" expectation,
[8]2790        the request includes a request body, and the server responds
2791        with a final status code before reading the entire request body
2792        from the transport connection, then the server &SHOULD-NOT;  close
2793        the transport connection until it has read the entire request,
2794        or until the client closes the connection. Otherwise, the client
2795        might not reliably receive the response message. However, this
[1514]2796        requirement ought not be construed as preventing a server from
[8]2797        defending itself against denial-of-service attacks, or from
2798        badly broken client implementations.
2799      </t>
2800    </list>
2803   Requirements for HTTP/1.1 proxies:
2804  <list style="symbols">
[1163]2805    <t> If a proxy receives a request that includes an Expect header
[8]2806        field with the "100-continue" expectation, and the proxy
2807        either knows that the next-hop server complies with HTTP/1.1 or
2808        higher, or does not know the HTTP version of the next-hop
2809        server, it &MUST; forward the request, including the Expect header
2810        field.
2811    </t>
2812    <t> If the proxy knows that the version of the next-hop server is
2813        HTTP/1.0 or lower, it &MUST-NOT; forward the request, and it &MUST;
[919]2814        respond with a 417 (Expectation Failed) status code.
[8]2815    </t>
[1390]2816    <t> Proxies &SHOULD; maintain a record of the HTTP version
[8]2817        numbers received from recently-referenced next-hop servers.
2818    </t>
2819    <t> A proxy &MUST-NOT; forward a 100 (Continue) response if the
2820        request message was received from an HTTP/1.0 (or earlier)
[1163]2821        client and did not include an Expect header field with
[8]2822        the "100-continue" expectation. This requirement overrides the
[29]2823        general rule for forwarding of 1xx responses (see &status-1xx;).
[8]2824    </t>
2825  </list>
[901]2833<section title="Miscellaneous notes that might disappear" anchor="misc">
[651]2834<section title="Scheme aliases considered harmful" anchor="scheme.aliases">
[767]2836   <cref anchor="TBD-aliases-harmful">describe why aliases like webcal are harmful.</cref>
2840<section title="Use of HTTP for proxy communication" anchor="http.proxy">
[767]2842   <cref anchor="TBD-proxy-other">Configured to use HTTP to proxy HTTP or other protocols.</cref>
[651]2846<section title="Interception of HTTP for access control" anchor="http.intercept">
[767]2848   <cref anchor="TBD-intercept">Interception of HTTP traffic for initiating access control.</cref>
[651]2852<section title="Use of HTTP by other protocols" anchor="http.others">
[767]2854   <cref anchor="TBD-profiles">Profiles of HTTP defined by other protocol.
[651]2855   Extensions of HTTP like WebDAV.</cref>
2859<section title="Use of HTTP by media type specification" anchor="">
[767]2861   <cref anchor="TBD-hypertext">Instructions on composing HTTP requests via hypertext formats.</cref>
[647]2866<section title="Header Field Definitions" anchor="header.field.definitions">
[1175]2868   This section defines the syntax and semantics of HTTP header fields
[1431]2869   related to message origination, framing, and routing.
[1431]2871<texttable align="left">
2872  <ttcol>Header Field Name</ttcol>
2873  <ttcol>Defined in...</ttcol>
2875  <c>Connection</c> <c><xref target="header.connection"/></c>
2876  <c>Content-Length</c> <c><xref target="header.content-length"/></c>
2877  <c>Host</c> <c><xref target=""/></c>
2878  <c>TE</c> <c><xref target="header.te"/></c>
2879  <c>Trailer</c> <c><xref target="header.trailer"/></c>
2880  <c>Transfer-Encoding</c> <c><xref target="header.transfer-encoding"/></c>
2881  <c>Upgrade</c> <c><xref target="header.upgrade"/></c>
2882  <c>Via</c> <c><xref target="header.via"/></c>
2885<section title="Connection" anchor="header.connection">
[1120]2886  <iref primary="true" item="Connection header field" x:for-anchor=""/>
2887  <iref primary="true" item="Header Fields" subitem="Connection" x:for-anchor=""/>
[229]2888  <x:anchor-alias value="Connection"/>
2889  <x:anchor-alias value="connection-token"/>
[1163]2891   The "Connection" header field allows the sender to specify
[1175]2892   options that are desired only for that particular connection.
2893   Such connection options &MUST; be removed or replaced before the
2894   message can be forwarded downstream by a proxy or gateway.
2895   This mechanism also allows the sender to indicate which HTTP
2896   header fields used in the message are only intended for the
2897   immediate recipient ("hop-by-hop"), as opposed to all recipients
2898   on the chain ("end-to-end"), enabling the message to be
2899   self-descriptive and allowing future connection-specific extensions
2900   to be deployed in HTTP without fear that they will be blindly
2901   forwarded by previously deployed intermediaries.
[994]2904   The Connection header field's value has the following grammar:
[1236]2906<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-token"/>
2907  <x:ref>Connection</x:ref>       = 1#<x:ref>connection-token</x:ref>
[354]2908  <x:ref>connection-token</x:ref> = <x:ref>token</x:ref>
[1175]2911   A proxy or gateway &MUST; parse a received Connection
2912   header field before a message is forwarded and, for each
2913   connection-token in this field, remove any header field(s) from
2914   the message with the same name as the connection-token, and then
2915   remove the Connection header field itself or replace it with the
2916   sender's own connection options for the forwarded message.
[1175]2919   A sender &MUST-NOT; include field-names in the Connection header
2920   field-value for fields that are defined as expressing constraints
2921   for all recipients in the request or response chain, such as the
2922   Cache-Control header field (&header-cache-control;).
[1175]2925   The connection options do not have to correspond to a header field
2926   present in the message, since a connection-specific header field
2927   might not be needed if there are no parameters associated with that
2928   connection option.  Recipients that trigger certain connection
2929   behavior based on the presence of connection options &MUST; do so
2930   based on the presence of the connection-token rather than only the
2931   presence of the optional header field.  In other words, if the
2932   connection option is received as a header field but not indicated
2933   within the Connection field-value, then the recipient &MUST; ignore
2934   the connection-specific header field because it has likely been
2935   forwarded by an intermediary that is only partially compliant.
2938   When defining new connection options, specifications ought to
2939   carefully consider existing deployed header fields and ensure
2940   that the new connection-token does not share the same name as
2941   an unrelated header field that might already be deployed.
2942   Defining a new connection-token essentially reserves that potential
2943   field-name for carrying additional information related to the
2944   connection option, since it would be unwise for senders to use
2945   that field-name for anything else.
[8]2948   HTTP/1.1 defines the "close" connection option for the sender to
2949   signal that the connection will be closed after completion of the
2950   response. For example,
2952<figure><artwork type="example">
[354]2953  Connection: close
2956   in either the request or the response header fields indicates that
[746]2957   the connection &SHOULD-NOT;  be considered "persistent" (<xref target="persistent.connections"/>)
[8]2958   after the current request/response is complete.
[86]2961   An HTTP/1.1 client that does not support persistent connections &MUST;
2962   include the "close" connection option in every request message.
[86]2965   An HTTP/1.1 server that does not support persistent connections &MUST;
2966   include the "close" connection option in every response message that
[753]2967   does not have a 1xx (Informational) status code.
2971<section title="Content-Length" anchor="header.content-length">
[1120]2972  <iref primary="true" item="Content-Length header field" x:for-anchor=""/>
2973  <iref primary="true" item="Header Fields" subitem="Content-Length" x:for-anchor=""/>
[229]2974  <x:anchor-alias value="Content-Length"/>
[852]2976   The "Content-Length" header field indicates the size of the
2977   message-body, in decimal number of octets, for any message other than
[1161]2978   a response to a HEAD request or a response with a status code of 304.
2979   In the case of a response to a HEAD request, Content-Length indicates
2980   the size of the payload body (not including any potential transfer-coding)
2981   that would have been sent had the request been a GET.
2982   In the case of a 304 (Not Modified) response to a GET request,
2983   Content-Length indicates the size of the payload body (not including
2984   any potential transfer-coding) that would have been sent in a 200 (OK)
2985   response.
[1236]2987<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
2988  <x:ref>Content-Length</x:ref> = 1*<x:ref>DIGIT</x:ref>
2991   An example is
2993<figure><artwork type="example">
[354]2994  Content-Length: 3495
[852]2997   Implementations &SHOULD; use this field to indicate the message-body
2998   length when no transfer-coding is being applied and the
2999   payload's body length can be determined prior to being transferred.
[864]3000   <xref target="message.body"/> describes how recipients determine the length
3001   of a message-body.
3004   Any Content-Length greater than or equal to zero is a valid value.
[864]3007   Note that the use of this field in HTTP is significantly different from
[8]3008   the corresponding definition in MIME, where it is an optional field
[852]3009   used within the "message/external-body" content-type.
3013<section title="Host" anchor="">
[1120]3014  <iref primary="true" item="Host header field" x:for-anchor=""/>
3015  <iref primary="true" item="Header Fields" subitem="Host" x:for-anchor=""/>
[229]3016  <x:anchor-alias value="Host"/>
[1173]3018   The "Host" header field in a request provides the host and port
3019   information from the target resource's URI, enabling the origin
3020   server to distinguish between resources while servicing requests
3021   for multiple host names on a single IP address.  Since the Host
3022   field-value is critical information for handling a request, it
3023   &SHOULD; be sent as the first header field following the Request-Line.
[1236]3025<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
3026  <x:ref>Host</x:ref> = <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ; <xref target="http.uri"/>
[1173]3029   A client &MUST; send a Host header field in all HTTP/1.1 request
3030   messages.  If the target resource's URI includes an authority
3031   component, then the Host field-value &MUST; be identical to that
3032   authority component after excluding any userinfo (<xref target="http.uri"/>).
3033   If the authority component is missing or undefined for the target
3034   resource's URI, then the Host header field &MUST; be sent with an
3035   empty field-value.
3038   For example, a GET request to the origin server for
3039   &lt;; would begin with:
[803]3041<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
3042GET /pub/WWW/ HTTP/1.1
[1173]3046   The Host header field &MUST; be sent in an HTTP/1.1 request even
3047   if the request-target is in the form of an absolute-URI, since this
3048   allows the Host information to be forwarded through ancient HTTP/1.0
3049   proxies that might not have implemented Host.
[1173]3052   When an HTTP/1.1 proxy receives a request with a request-target in
3053   the form of an absolute-URI, the proxy &MUST; ignore the received
3054   Host header field (if any) and instead replace it with the host
3055   information of the request-target.  When a proxy forwards a request,
3056   it &MUST; generate the Host header field based on the received
3057   absolute-URI rather than the received Host.
3060   Since the Host header field acts as an application-level routing
3061   mechanism, it is a frequent target for malware seeking to poison
3062   a shared cache or redirect a request to an unintended server.
3063   An interception proxy is particularly vulnerable if it relies on
3064   the Host header field value for redirecting requests to internal
3065   servers, or for use as a cache key in a shared cache, without
3066   first verifying that the intercepted connection is targeting a
3067   valid IP address for that host.
3070   A server &MUST; respond with a 400 (Bad Request) status code to
3071   any HTTP/1.1 request message that lacks a Host header field and
3072   to any request message that contains more than one Host header field
3073   or a Host header field with an invalid field-value.
[97]3076   See Sections <xref target="" format="counter"/>
[8]3077   and <xref target="" format="counter"/>
3078   for other requirements relating to Host.
3082<section title="TE" anchor="header.te">
[1120]3083  <iref primary="true" item="TE header field" x:for-anchor=""/>
3084  <iref primary="true" item="Header Fields" subitem="TE" x:for-anchor=""/>
[229]3085  <x:anchor-alias value="TE"/>
3086  <x:anchor-alias value="t-codings"/>
[457]3087  <x:anchor-alias value="te-params"/>
3088  <x:anchor-alias value="te-ext"/>
[1163]3090   The "TE" header field indicates what extension transfer-codings
[1514]3091   the client is willing to accept in the response, and whether or not it is
[698]3092   willing to accept trailer fields in a chunked transfer-coding.
[975]3095   Its value consists of the keyword "trailers" and/or a comma-separated
[8]3096   list of extension transfer-coding names with optional accept
3097   parameters (as described in <xref target="transfer.codings"/>).
[1236]3099<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="TE"/><iref primary="true" item="Grammar" subitem="t-codings"/><iref primary="true" item="Grammar" subitem="te-params"/><iref primary="true" item="Grammar" subitem="te-ext"/>
3100  <x:ref>TE</x:ref>        = #<x:ref>t-codings</x:ref>
[457]3101  <x:ref>t-codings</x:ref> = "trailers" / ( <x:ref>transfer-extension</x:ref> [ <x:ref>te-params</x:ref> ] )
3102  <x:ref>te-params</x:ref> = <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> "q=" <x:ref>qvalue</x:ref> *( <x:ref>te-ext</x:ref> )
[810]3103  <x:ref>te-ext</x:ref>    = <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> <x:ref>token</x:ref> [ "=" <x:ref>word</x:ref> ]
3106   The presence of the keyword "trailers" indicates that the client is
3107   willing to accept trailer fields in a chunked transfer-coding, as
[673]3108   defined in <xref target="chunked.encoding"/>. This keyword is reserved for use with
[8]3109   transfer-coding values even though it does not itself represent a
3110   transfer-coding.
3113   Examples of its use are:
3115<figure><artwork type="example">
[354]3116  TE: deflate
3117  TE:
3118  TE: trailers, deflate;q=0.5
3121   The TE header field only applies to the immediate connection.
3122   Therefore, the keyword &MUST; be supplied within a Connection header
3123   field (<xref target="header.connection"/>) whenever TE is present in an HTTP/1.1 message.
3126   A server tests whether a transfer-coding is acceptable, according to
3127   a TE field, using these rules:
3128  <list style="numbers">
3129    <x:lt>
3130      <t>The "chunked" transfer-coding is always acceptable. If the
3131         keyword "trailers" is listed, the client indicates that it is
3132         willing to accept trailer fields in the chunked response on
3133         behalf of itself and any downstream clients. The implication is
3134         that, if given, the client is stating that either all
3135         downstream clients are willing to accept trailer fields in the
3136         forwarded response, or that it will attempt to buffer the
3137         response on behalf of downstream recipients.
3138      </t><t>
3139         <x:h>Note:</x:h> HTTP/1.1 does not define any means to limit the size of a
3140         chunked response such that a client can be assured of buffering
3141         the entire response.</t>
3142    </x:lt>
3143    <x:lt>
3144      <t>If the transfer-coding being tested is one of the transfer-codings
3145         listed in the TE field, then it is acceptable unless it
[457]3146         is accompanied by a qvalue of 0. (As defined in <xref target="quality.values"/>, a
[879]3147         qvalue of 0 means "not acceptable".)</t>
[8]3148    </x:lt>
3149    <x:lt>
3150      <t>If multiple transfer-codings are acceptable, then the
3151         acceptable transfer-coding with the highest non-zero qvalue is
3152         preferred.  The "chunked" transfer-coding always has a qvalue
3153         of 1.</t>
3154    </x:lt>
3155  </list>
3158   If the TE field-value is empty or if no TE field is present, the only
[1514]3159   acceptable transfer-coding is "chunked". A message with no transfer-coding is
[8]3160   always acceptable.
3164<section title="Trailer" anchor="header.trailer">
[1120]3165  <iref primary="true" item="Trailer header field" x:for-anchor=""/>
3166  <iref primary="true" item="Header Fields" subitem="Trailer" x:for-anchor=""/>
[229]3167  <x:anchor-alias value="Trailer"/>
[1163]3169   The "Trailer" header field indicates that the given set of
[8]3170   header fields is present in the trailer of a message encoded with
3171   chunked transfer-coding.
[1236]3173<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
3174  <x:ref>Trailer</x:ref> = 1#<x:ref>field-name</x:ref>
3177   An HTTP/1.1 message &SHOULD; include a Trailer header field in a
3178   message using chunked transfer-coding with a non-empty trailer. Doing
3179   so allows the recipient to know which header fields to expect in the
3180   trailer.
3183   If no Trailer header field is present, the trailer &SHOULD-NOT;  include
[673]3184   any header fields. See <xref target="chunked.encoding"/> for restrictions on the use of
[8]3185   trailer fields in a "chunked" transfer-coding.
3188   Message header fields listed in the Trailer header field &MUST-NOT;
3189   include the following header fields:
3190  <list style="symbols">
3191    <t>Transfer-Encoding</t>
3192    <t>Content-Length</t>
3193    <t>Trailer</t>
3194  </list>
3198<section title="Transfer-Encoding" anchor="header.transfer-encoding">
[1120]3199  <iref primary="true" item="Transfer-Encoding header field" x:for-anchor=""/>
3200  <iref primary="true" item="Header Fields" subitem="Transfer-Encoding" x:for-anchor=""/>
[229]3201  <x:anchor-alias value="Transfer-Encoding"/>
[1163]3203   The "Transfer-Encoding" header field indicates what transfer-codings
[698]3204   (if any) have been applied to the message body. It differs from
3205   Content-Encoding (&content-codings;) in that transfer-codings are a property
3206   of the message (and therefore are removed by intermediaries), whereas
3207   content-codings are not.
[1236]3209<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
3210  <x:ref>Transfer-Encoding</x:ref> = 1#<x:ref>transfer-coding</x:ref>
3213   Transfer-codings are defined in <xref target="transfer.codings"/>. An example is:
3215<figure><artwork type="example">
3216  Transfer-Encoding: chunked
[866]3219   If multiple encodings have been applied to a representation, the transfer-codings
[8]3220   &MUST; be listed in the order in which they were applied.
3221   Additional information about the encoding parameters &MAY; be provided
[965]3222   by other header fields not defined by this specification.
3225   Many older HTTP/1.0 applications do not understand the Transfer-Encoding
[994]3226   header field.
3230<section title="Upgrade" anchor="header.upgrade">
[1120]3231  <iref primary="true" item="Upgrade header field" x:for-anchor=""/>
3232  <iref primary="true" item="Header Fields" subitem="Upgrade" x:for-anchor=""/>
[229]3233  <x:anchor-alias value="Upgrade"/>
[1163]3235   The "Upgrade" header field allows the client to specify what
[698]3236   additional communication protocols it would like to use, if the server
[1071]3237   chooses to switch protocols. Servers can use it to indicate what protocols
3238   they are willing to switch to.
[1236]3240<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
3241  <x:ref>Upgrade</x:ref> = 1#<x:ref>product</x:ref>
3244   For example,
3246<figure><artwork type="example">
[354]3247  Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
3250   The Upgrade header field is intended to provide a simple mechanism
[1514]3251   for transitioning from HTTP/1.1 to some other, incompatible protocol. It
[8]3252   does so by allowing the client to advertise its desire to use another
3253   protocol, such as a later version of HTTP with a higher major version
3254   number, even though the current request has been made using HTTP/1.1.
3255   This eases the difficult transition between incompatible protocols by
3256   allowing the client to initiate a request in the more commonly
3257   supported protocol while indicating to the server that it would like
3258   to use a "better" protocol if available (where "better" is determined
[1161]3259   by the server, possibly according to the nature of the request method
3260   or target resource).
3263   The Upgrade header field only applies to switching application-layer
3264   protocols upon the existing transport-layer connection. Upgrade
3265   cannot be used to insist on a protocol change; its acceptance and use
3266   by the server is optional. The capabilities and nature of the
3267   application-layer communication after the protocol change is entirely
3268   dependent upon the new protocol chosen, although the first action
3269   after changing the protocol &MUST; be a response to the initial HTTP
3270   request containing the Upgrade header field.
3273   The Upgrade header field only applies to the immediate connection.
3274   Therefore, the upgrade keyword &MUST; be supplied within a Connection
3275   header field (<xref target="header.connection"/>) whenever Upgrade is present in an
3276   HTTP/1.1 message.
3279   The Upgrade header field cannot be used to indicate a switch to a
3280   protocol on a different connection. For that purpose, it is more
[1069]3281   appropriate to use a 3xx redirection response (&status-3xx;).
[1071]3284   Servers &MUST; include the "Upgrade" header field in 101 (Switching
3285   Protocols) responses to indicate which protocol(s) are being switched to,
3286   and &MUST; include it in 426 (Upgrade Required) responses to indicate
3287   acceptable protocols to upgrade to. Servers &MAY; include it in any other
3288   response to indicate that they are willing to upgrade to one of the
3289   specified protocols.
[8]3292   This specification only defines the protocol name "HTTP" for use by
3293   the family of Hypertext Transfer Protocols, as defined by the HTTP
3294   version rules of <xref target="http.version"/> and future updates to this
[684]3295   specification. Additional tokens can be registered with IANA using the
3296   registration procedure defined below. 
3299<section title="Upgrade Token Registry" anchor="upgrade.token.registry">
3301   The HTTP Upgrade Token Registry defines the name space for product
3302   tokens used to identify protocols in the Upgrade header field.
[969]3303   Each registered token is associated with contact information and
3304   an optional set of specifications that details how the connection
3305   will be processed after it has been upgraded.
[969]3308   Registrations are allowed on a First Come First Served basis as
3309   described in <xref target="RFC5226" x:sec="4.1" x:fmt="of"/>. The
3310   specifications need not be IETF documents or be subject to IESG review.
3311   Registrations are subject to the following rules:
[684]3312  <list style="numbers">
3313    <t>A token, once registered, stays registered forever.</t>
3314    <t>The registration &MUST; name a responsible party for the
3315       registration.</t>
3316    <t>The registration &MUST; name a point of contact.</t>
[969]3317    <t>The registration &MAY; name a set of specifications associated with that
3318       token. Such specifications need not be publicly available.</t>
[684]3319    <t>The responsible party &MAY; change the registration at any time.
3320       The IANA will keep a record of all such changes, and make them
3321       available upon request.</t>
3322    <t>The responsible party for the first registration of a "product"
3323       token &MUST; approve later registrations of a "version" token
3324       together with that "product" token before they can be registered.</t>
3325    <t>If absolutely required, the IESG &MAY; reassign the responsibility
3326       for a token. This will normally only be used in the case when a
3327       responsible party cannot be contacted.</t>
3328  </list>
[8]3335<section title="Via" anchor="header.via">
[1120]3336  <iref primary="true" item="Via header field" x:for-anchor=""/>
3337  <iref primary="true" item="Header Fields" subitem="Via" x:for-anchor=""/>
[229]3338  <x:anchor-alias value="protocol-name"/>
3339  <x:anchor-alias value="protocol-version"/>
3340  <x:anchor-alias value="pseudonym"/>
3341  <x:anchor-alias value="received-by"/>
3342  <x:anchor-alias value="received-protocol"/>
3343  <x:anchor-alias value="Via"/>
[1175]3345   The "Via" header field &MUST; be sent by a proxy or gateway to
[8]3346   indicate the intermediate protocols and recipients between the user
3347   agent and the server on requests, and between the origin server and
[1175]3348   the client on responses. It is analogous to the "Received" field
3349   used by email systems (<xref target="RFC5322" x:fmt="of" x:sec="3.6.7"/>)
3350   and is intended to be used for tracking message forwards,
[8]3351   avoiding request loops, and identifying the protocol capabilities of
3352   all senders along the request/response chain.
[1236]3354<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Via"/><iref primary="true" item="Grammar" subitem="received-protocol"/><iref primary="true" item="Grammar" subitem="protocol-name"/><iref primary="true" item="Grammar" subitem="protocol-version"/><iref primary="true" item="Grammar" subitem="received-by"/><iref primary="true" item="Grammar" subitem="pseudonym"/>
3355  <x:ref>Via</x:ref>               = 1#( <x:ref>received-protocol</x:ref> <x:ref>RWS</x:ref> <x:ref>received-by</x:ref>
[376]3356                          [ <x:ref>RWS</x:ref> <x:ref>comment</x:ref> ] )
[229]3357  <x:ref>received-protocol</x:ref> = [ <x:ref>protocol-name</x:ref> "/" ] <x:ref>protocol-version</x:ref>
3358  <x:ref>protocol-name</x:ref>     = <x:ref>token</x:ref>
3359  <x:ref>protocol-version</x:ref>  = <x:ref>token</x:ref>
[334]3360  <x:ref>received-by</x:ref>       = ( <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ) / <x:ref>pseudonym</x:ref>
[229]3361  <x:ref>pseudonym</x:ref>         = <x:ref>token</x:ref>
3364   The received-protocol indicates the protocol version of the message
3365   received by the server or client along each segment of the
3366   request/response chain. The received-protocol version is appended to
3367   the Via field value when the message is forwarded so that information
3368   about the protocol capabilities of upstream applications remains
3369   visible to all recipients.
[1175]3372   The protocol-name is excluded if and only if it would be "HTTP". The
[8]3373   received-by field is normally the host and optional port number of a
3374   recipient server or client that subsequently forwarded the message.
3375   However, if the real host is considered to be sensitive information,
3376   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
3377   be assumed to be the default port of the received-protocol.
[1175]3380   Multiple Via field values represent each proxy or gateway that has
[8]3381   forwarded the message. Each recipient &MUST; append its information
3382   such that the end result is ordered according to the sequence of
3383   forwarding applications.
3386   Comments &MAY; be used in the Via header field to identify the software
[1175]3387   of each recipient, analogous to the User-Agent and Server header fields.
3388   However, all comments in the Via field are optional and &MAY; be removed
3389   by any recipient prior to forwarding the message.
3392   For example, a request message could be sent from an HTTP/1.0 user
3393   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
[90]3394   forward the request to a public proxy at, which completes
3395   the request by forwarding it to the origin server at
3396   The request received by would then have the following
[8]3397   Via header field:
3399<figure><artwork type="example">
[354]3400  Via: 1.0 fred, 1.1 (Apache/1.1)
[1175]3403   A proxy or gateway used as a portal through a network firewall
3404   &SHOULD-NOT; forward the names and ports of hosts within the firewall
3405   region unless it is explicitly enabled to do so. If not enabled, the
3406   received-by host of any host behind the firewall &SHOULD; be replaced
3407   by an appropriate pseudonym for that host.
3410   For organizations that have strong privacy requirements for hiding
[1175]3411   internal structures, a proxy or gateway &MAY; combine an ordered
3412   subsequence of Via header field entries with identical received-protocol
3413   values into a single such entry. For example,
3415<figure><artwork type="example">
[354]3416  Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
[803]3419  could be collapsed to
3421<figure><artwork type="example">
[354]3422  Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
[1175]3425   Senders &SHOULD-NOT; combine multiple entries unless they are all
[8]3426   under the same organizational control and the hosts have already been
[1175]3427   replaced by pseudonyms. Senders &MUST-NOT; combine entries which
[8]3428   have different received-protocol values.
[29]3434<section title="IANA Considerations" anchor="IANA.considerations">
[921]3436<section title="Header Field Registration" anchor="header.field.registration">
[969]3438   The Message Header Field Registry located at <eref target=""/> shall be updated
[290]3439   with the permanent registrations below (see <xref target="RFC3864"/>):
[680]3441<?BEGININC p1-messaging.iana-headers ?>
[290]3442<!--AUTOGENERATED FROM extract-header-defs.xslt, do not edit manually-->
3443<texttable align="left" suppress-title="true" anchor="iana.header.registration.table">
[253]3444   <ttcol>Header Field Name</ttcol>
3445   <ttcol>Protocol</ttcol>
3446   <ttcol>Status</ttcol>
3447   <ttcol>Reference</ttcol>
3449   <c>Connection</c>
3450   <c>http</c>
3451   <c>standard</c>
3452   <c>
3453      <xref target="header.connection"/>
3454   </c>
3455   <c>Content-Length</c>
3456   <c>http</c>
3457   <c>standard</c>
3458   <c>
3459      <xref target="header.content-length"/>
3460   </c>
3461   <c>Host</c>
3462   <c>http</c>
3463   <c>standard</c>
3464   <c>
3465      <xref target=""/>
3466   </c>
3467   <c>TE</c>
3468   <c>http</c>
3469   <c>standard</c>
3470   <c>
3471      <xref target="header.te"/>
3472   </c>
3473   <c>Trailer</c>
3474   <c>http</c>
3475   <c>standard</c>
3476   <c>
3477      <xref target="header.trailer"/>
3478   </c>
3479   <c>Transfer-Encoding</c>
3480   <c>http</c>
3481   <c>standard</c>
3482   <c>
3483      <xref target="header.transfer-encoding"/>
3484   </c>
3485   <c>Upgrade</c>
3486   <c>http</c>
3487   <c>standard</c>
3488   <c>
3489      <xref target="header.upgrade"/>
3490   </c>
3491   <c>Via</c>
3492   <c>http</c>
3493   <c>standard</c>
3494   <c>
3495      <xref target="header.via"/>
3496   </c>
[680]3499<?ENDINC p1-messaging.iana-headers ?>
[1338]3501   Furthermore, the header field name "Close" shall be registered as "reserved", as its use as
3502   HTTP header field would be in conflict with the use of the "close" connection
3503   option for the "Connection" header field (<xref target="header.connection"/>).
3505<texttable align="left" suppress-title="true">
3506   <ttcol>Header Field Name</ttcol>
3507   <ttcol>Protocol</ttcol>
3508   <ttcol>Status</ttcol>
3509   <ttcol>Reference</ttcol>
3511   <c>Close</c>
3512   <c>http</c>
3513   <c>reserved</c>
3514   <c>
3515      <xref target="header.field.registration"/>
3516   </c>
[290]3519   The change controller is: "IETF ( - Internet Engineering Task Force".
3523<section title="URI Scheme Registration" anchor="uri.scheme.registration">
[646]3525   The entries for the "http" and "https" URI Schemes in the registry located at
[307]3526   <eref target=""/>
[969]3527   shall be updated to point to Sections <xref target="http.uri" format="counter"/>
[646]3528   and <xref target="https.uri" format="counter"/> of this document
[307]3529   (see <xref target="RFC4395"/>).
[296]3533<section title="Internet Media Type Registrations" anchor="">
3535   This document serves as the specification for the Internet media types
3536   "message/http" and "application/http". The following is to be registered with
3537   IANA (see <xref target="RFC4288"/>).
3539<section title="Internet Media Type message/http" anchor="">
3540<iref item="Media Type" subitem="message/http" primary="true"/>
3541<iref item="message/http Media Type" primary="true"/>
3543   The message/http type can be used to enclose a single HTTP request or
3544   response message, provided that it obeys the MIME restrictions for all