source: draft-ietf-httpbis/latest/p1-messaging.xml @ 2033

Last change on this file since 2033 was 2033, checked in by fielding@…, 7 years ago

Move new user agent requirement on rendering incomplete responses to
a suggestion in security considerations. Addresses #408 and #415

Consolidate more stuff on persistent connection reuse. Addresses #396

  • Property svn:eol-style set to native
  • Property svn:mime-type set to text/xml
File size: 221.8 KB
Line 
1<?xml version="1.0" encoding="utf-8"?>
2<?xml-stylesheet type='text/xsl' href='../myxml2rfc.xslt'?>
3<!DOCTYPE rfc [
4  <!ENTITY MAY "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>MAY</bcp14>">
5  <!ENTITY MUST "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>MUST</bcp14>">
6  <!ENTITY MUST-NOT "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>MUST NOT</bcp14>">
7  <!ENTITY OPTIONAL "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>OPTIONAL</bcp14>">
8  <!ENTITY RECOMMENDED "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>RECOMMENDED</bcp14>">
9  <!ENTITY REQUIRED "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>REQUIRED</bcp14>">
10  <!ENTITY SHALL "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHALL</bcp14>">
11  <!ENTITY SHALL-NOT "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHALL NOT</bcp14>">
12  <!ENTITY SHOULD "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHOULD</bcp14>">
13  <!ENTITY SHOULD-NOT "<bcp14 xmlns='http://purl.org/net/xml2rfc/ext'>SHOULD NOT</bcp14>">
14  <!ENTITY ID-VERSION "latest">
15  <!ENTITY ID-MONTH "December">
16  <!ENTITY ID-YEAR "2012">
17  <!ENTITY mdash "&#8212;">
18  <!ENTITY Note "<x:h xmlns:x='http://purl.org/net/xml2rfc/ext'>Note:</x:h>">
19  <!ENTITY caching-overview       "<xref target='Part6' x:rel='#caching.overview' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
20  <!ENTITY cache-incomplete       "<xref target='Part6' x:rel='#response.cacheability' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
21  <!ENTITY payload                "<xref target='Part2' x:rel='#payload' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
22  <!ENTITY media-type            "<xref target='Part2' x:rel='#media.type' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
23  <!ENTITY content-codings        "<xref target='Part2' x:rel='#content.codings' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
24  <!ENTITY CONNECT                "<xref target='Part2' x:rel='#CONNECT' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
25  <!ENTITY content.negotiation    "<xref target='Part2' x:rel='#content.negotiation' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
26  <!ENTITY diff-mime              "<xref target='Part2' x:rel='#differences.between.http.and.mime' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
27  <!ENTITY representation         "<xref target='Part2' x:rel='#representation' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
28  <!ENTITY HEAD                   "<xref target='Part2' x:rel='#HEAD' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
29  <!ENTITY header-allow           "<xref target='Part2' x:rel='#header.allow' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
30  <!ENTITY header-cache-control   "<xref target='Part6' x:rel='#header.cache-control' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
31  <!ENTITY header-content-encoding    "<xref target='Part2' x:rel='#header.content-encoding' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
32  <!ENTITY header-content-location    "<xref target='Part2' x:rel='#header.content-location' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
33  <!ENTITY header-content-range   "<xref target='Part5' x:rel='#header.content-range' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
34  <!ENTITY header-content-type    "<xref target='Part2' x:rel='#header.content-type' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
35  <!ENTITY header-date            "<xref target='Part2' x:rel='#header.date' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
36  <!ENTITY header-etag            "<xref target='Part4' x:rel='#header.etag' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
37  <!ENTITY header-expires         "<xref target='Part6' x:rel='#header.expires' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
38  <!ENTITY header-last-modified   "<xref target='Part4' x:rel='#header.last-modified' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
39  <!ENTITY header-mime-version    "<xref target='Part2' x:rel='#mime-version' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
40  <!ENTITY header-pragma          "<xref target='Part6' x:rel='#header.pragma' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
41  <!ENTITY header-proxy-authenticate  "<xref target='Part7' x:rel='#header.proxy-authenticate' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
42  <!ENTITY header-proxy-authorization "<xref target='Part7' x:rel='#header.proxy-authorization' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
43  <!ENTITY header-server          "<xref target='Part2' x:rel='#header.server' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
44  <!ENTITY header-warning         "<xref target='Part6' x:rel='#header.warning' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
45  <!ENTITY idempotent-methods     "<xref target='Part2' x:rel='#idempotent.methods' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
46  <!ENTITY methods                "<xref target='Part2' x:rel='#methods' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
47  <!ENTITY OPTIONS                "<xref target='Part2' x:rel='#OPTIONS' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
48  <!ENTITY qvalue                 "<xref target='Part2' x:rel='#quality.values' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
49  <!ENTITY resource               "<xref target='Part2' x:rel='#resource' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
50  <!ENTITY status-codes           "<xref target='Part2' x:rel='#status.codes' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
51  <!ENTITY status-1xx             "<xref target='Part2' x:rel='#status.1xx' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
52  <!ENTITY status-203             "<xref target='Part2' x:rel='#status.203' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
53  <!ENTITY status-3xx             "<xref target='Part2' x:rel='#status.3xx' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
54  <!ENTITY status-304             "<xref target='Part4' x:rel='#status.304' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
55  <!ENTITY status-4xx             "<xref target='Part2' x:rel='#status.4xx' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
56  <!ENTITY status-414             "<xref target='Part2' x:rel='#status.414' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
57  <!ENTITY iana-header-registry   "<xref target='Part2' x:rel='#header.field.registry' xmlns:x='http://purl.org/net/xml2rfc/ext'/>">
58]>
59<?rfc toc="yes" ?>
60<?rfc symrefs="yes" ?>
61<?rfc sortrefs="yes" ?>
62<?rfc compact="yes"?>
63<?rfc subcompact="no" ?>
64<?rfc linkmailto="no" ?>
65<?rfc editing="no" ?>
66<?rfc comments="yes"?>
67<?rfc inline="yes"?>
68<?rfc rfcedstyle="yes"?>
69<?rfc-ext allow-markup-in-artwork="yes" ?>
70<?rfc-ext include-references-in-index="yes" ?>
71<rfc obsoletes="2145,2616" updates="2817,2818" category="std" x:maturity-level="proposed"
72     ipr="pre5378Trust200902" docName="draft-ietf-httpbis-p1-messaging-&ID-VERSION;"
73     xmlns:x='http://purl.org/net/xml2rfc/ext'>
74<x:link rel="next" basename="p2-semantics"/>
75<x:feedback template="mailto:ietf-http-wg@w3.org?subject={docname},%20%22{section}%22&amp;body=&lt;{ref}&gt;:"/>
76<front>
77
78  <title abbrev="HTTP/1.1 Message Syntax and Routing">Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing</title>
79
80  <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
81    <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
82    <address>
83      <postal>
84        <street>345 Park Ave</street>
85        <city>San Jose</city>
86        <region>CA</region>
87        <code>95110</code>
88        <country>USA</country>
89      </postal>
90      <email>fielding@gbiv.com</email>
91      <uri>http://roy.gbiv.com/</uri>
92    </address>
93  </author>
94
95  <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
96    <organization abbrev="greenbytes">greenbytes GmbH</organization>
97    <address>
98      <postal>
99        <street>Hafenweg 16</street>
100        <city>Muenster</city><region>NW</region><code>48155</code>
101        <country>Germany</country>
102      </postal>
103      <email>julian.reschke@greenbytes.de</email>
104      <uri>http://greenbytes.de/tech/webdav/</uri>
105    </address>
106  </author>
107
108  <date month="&ID-MONTH;" year="&ID-YEAR;"/>
109  <workgroup>HTTPbis Working Group</workgroup>
110
111<abstract>
112<t>
113   The Hypertext Transfer Protocol (HTTP) is an application-level protocol for
114   distributed, collaborative, hypertext information systems. HTTP has been in
115   use by the World Wide Web global information initiative since 1990.
116   This document provides an overview of HTTP architecture and its associated
117   terminology, defines the "http" and "https" Uniform Resource Identifier
118   (URI) schemes, defines the HTTP/1.1 message syntax and parsing requirements,
119   and describes general security concerns for implementations.
120</t>   
121</abstract>
122
123<note title="Editorial Note (To be removed by RFC Editor)">
124  <t>
125    Discussion of this draft takes place on the HTTPBIS working group
126    mailing list (ietf-http-wg@w3.org), which is archived at
127    <eref target="http://lists.w3.org/Archives/Public/ietf-http-wg/"/>.
128  </t>
129  <t>
130    The current issues list is at
131    <eref target="http://tools.ietf.org/wg/httpbis/trac/report/3"/> and related
132    documents (including fancy diffs) can be found at
133    <eref target="http://tools.ietf.org/wg/httpbis/"/>.
134  </t>
135  <t>
136    The changes in this draft are summarized in <xref target="changes.since.21"/>.
137  </t>
138</note>
139</front>
140<middle>
141<section title="Introduction" anchor="introduction">
142<t>
143   The Hypertext Transfer Protocol (HTTP) is an application-level
144   request/response protocol that uses extensible semantics and MIME-like
145   message payloads for flexible interaction with network-based hypertext
146   information systems. This document is the first in a series of documents
147   that collectively form the HTTP/1.1 specification:
148   <list style="empty">
149    <t>RFC xxx1: Message Syntax and Routing</t>
150    <t><xref target="Part2" x:fmt="none">RFC xxx2</xref>: Semantics and Content</t>
151    <t><xref target="Part4" x:fmt="none">RFC xxx3</xref>: Conditional Requests</t>
152    <t><xref target="Part5" x:fmt="none">RFC xxx4</xref>: Range Requests</t>
153    <t><xref target="Part6" x:fmt="none">RFC xxx5</xref>: Caching</t>
154    <t><xref target="Part7" x:fmt="none">RFC xxx6</xref>: Authentication</t>
155   </list>
156</t>
157<t>
158   This HTTP/1.1 specification obsoletes and moves to historic status
159   <xref target="RFC2616" x:fmt="none">RFC 2616</xref>, its predecessor
160   <xref target="RFC2068" x:fmt="none">RFC 2068</xref>, and
161   <xref target="RFC2145" x:fmt="none">RFC 2145</xref> (on HTTP versioning).
162   This specification also updates the use of CONNECT to establish a tunnel,
163   previously defined in <xref target="RFC2817" x:fmt="none">RFC 2817</xref>,
164   and defines the "https" URI scheme that was described informally in
165   <xref target="RFC2818" x:fmt="none">RFC 2818</xref>.
166</t>
167<t>
168   HTTP is a generic interface protocol for information systems. It is
169   designed to hide the details of how a service is implemented by presenting
170   a uniform interface to clients that is independent of the types of
171   resources provided. Likewise, servers do not need to be aware of each
172   client's purpose: an HTTP request can be considered in isolation rather
173   than being associated with a specific type of client or a predetermined
174   sequence of application steps. The result is a protocol that can be used
175   effectively in many different contexts and for which implementations can
176   evolve independently over time.
177</t>
178<t>
179   HTTP is also designed for use as an intermediation protocol for translating
180   communication to and from non-HTTP information systems.
181   HTTP proxies and gateways can provide access to alternative information
182   services by translating their diverse protocols into a hypertext
183   format that can be viewed and manipulated by clients in the same way
184   as HTTP services.
185</t>
186<t>
187   One consequence of this flexibility is that the protocol cannot be
188   defined in terms of what occurs behind the interface. Instead, we
189   are limited to defining the syntax of communication, the intent
190   of received communication, and the expected behavior of recipients.
191   If the communication is considered in isolation, then successful
192   actions ought to be reflected in corresponding changes to the
193   observable interface provided by servers. However, since multiple
194   clients might act in parallel and perhaps at cross-purposes, we
195   cannot require that such changes be observable beyond the scope
196   of a single response.
197</t>
198<t>
199   This document describes the architectural elements that are used or
200   referred to in HTTP, defines the "http" and "https" URI schemes,
201   describes overall network operation and connection management,
202   and defines HTTP message framing and forwarding requirements.
203   Our goal is to define all of the mechanisms necessary for HTTP message
204   handling that are independent of message semantics, thereby defining the
205   complete set of requirements for message parsers and
206   message-forwarding intermediaries.
207</t>
208
209
210<section title="Requirement Notation" anchor="intro.requirements">
211<t>
212   The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT",
213   "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this
214   document are to be interpreted as described in <xref target="RFC2119"/>.
215</t>
216<t>
217   Conformance criteria and considerations regarding error handling
218   are defined in <xref target="conformance"/>.
219</t>
220</section>
221
222<section title="Syntax Notation" anchor="notation">
223<iref primary="true" item="Grammar" subitem="ALPHA"/>
224<iref primary="true" item="Grammar" subitem="CR"/>
225<iref primary="true" item="Grammar" subitem="CRLF"/>
226<iref primary="true" item="Grammar" subitem="CTL"/>
227<iref primary="true" item="Grammar" subitem="DIGIT"/>
228<iref primary="true" item="Grammar" subitem="DQUOTE"/>
229<iref primary="true" item="Grammar" subitem="HEXDIG"/>
230<iref primary="true" item="Grammar" subitem="HTAB"/>
231<iref primary="true" item="Grammar" subitem="LF"/>
232<iref primary="true" item="Grammar" subitem="OCTET"/>
233<iref primary="true" item="Grammar" subitem="SP"/>
234<iref primary="true" item="Grammar" subitem="VCHAR"/>
235<t>
236   This specification uses the Augmented Backus-Naur Form (ABNF) notation
237   of <xref target="RFC5234"/> with the list rule extension defined in
238   <xref target="abnf.extension"/><xref target="collected.abnf"/> shows
239   the collected ABNF with the list rule expanded.
240</t>
241<t anchor="core.rules">
242  <x:anchor-alias value="ALPHA"/>
243  <x:anchor-alias value="CTL"/>
244  <x:anchor-alias value="CR"/>
245  <x:anchor-alias value="CRLF"/>
246  <x:anchor-alias value="DIGIT"/>
247  <x:anchor-alias value="DQUOTE"/>
248  <x:anchor-alias value="HEXDIG"/>
249  <x:anchor-alias value="HTAB"/>
250  <x:anchor-alias value="LF"/>
251  <x:anchor-alias value="OCTET"/>
252  <x:anchor-alias value="SP"/>
253  <x:anchor-alias value="VCHAR"/>
254   The following core rules are included by
255   reference, as defined in <xref target="RFC5234" x:fmt="," x:sec="B.1"/>:
256   ALPHA (letters), CR (carriage return), CRLF (CR LF), CTL (controls),
257   DIGIT (decimal 0-9), DQUOTE (double quote),
258   HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line feed),
259   OCTET (any 8-bit sequence of data), SP (space), and
260   VCHAR (any visible <xref target="USASCII"/> character).
261</t>
262<t>
263   As a convention, ABNF rule names prefixed with "obs-" denote
264   "obsolete" grammar rules that appear for historical reasons.
265</t>
266</section>
267</section>
268
269<section title="Architecture" anchor="architecture">
270<t>
271   HTTP was created for the World Wide Web architecture
272   and has evolved over time to support the scalability needs of a worldwide
273   hypertext system. Much of that architecture is reflected in the terminology
274   and syntax productions used to define HTTP.
275</t>
276
277<section title="Client/Server Messaging" anchor="operation">
278<iref primary="true" item="client"/>
279<iref primary="true" item="server"/>
280<iref primary="true" item="connection"/>
281<t>
282   HTTP is a stateless request/response protocol that operates by exchanging
283   <x:dfn>messages</x:dfn> (<xref target="http.message"/>) across a reliable
284   transport or session-layer
285   "<x:dfn>connection</x:dfn>" (<xref target="connection.management"/>).
286   An HTTP "<x:dfn>client</x:dfn>" is a program that establishes a connection
287   to a server for the purpose of sending one or more HTTP requests.
288   An HTTP "<x:dfn>server</x:dfn>" is a program that accepts connections
289   in order to service HTTP requests by sending HTTP responses.
290</t>
291<iref primary="true" item="user agent"/>
292<iref primary="true" item="origin server"/>
293<iref primary="true" item="browser"/>
294<iref primary="true" item="spider"/>
295<iref primary="true" item="sender"/>
296<iref primary="true" item="recipient"/>
297<t>
298   The terms client and server refer only to the roles that
299   these programs perform for a particular connection.  The same program
300   might act as a client on some connections and a server on others.
301   We use the term "<x:dfn>user agent</x:dfn>" to refer to any of the various
302   client programs that initiate a request, including (but not limited to)
303   browsers, spiders (web-based robots), command-line tools, native
304   applications, and mobile apps.  The term "<x:dfn>origin server</x:dfn>" is
305   used to refer to the program that can originate authoritative responses to
306   a request. For general requirements, we use the terms
307   "<x:dfn>sender</x:dfn>" and "<x:dfn>recipient</x:dfn>" to refer to any
308   component that sends or receives, respectively, a given message.
309</t>
310<t>
311   HTTP relies upon the Uniform Resource Identifier (URI)
312   standard <xref target="RFC3986"/> to indicate the target resource
313   (<xref target="target-resource"/>) and relationships between resources.
314   Messages are passed in a format similar to that used by Internet mail
315   <xref target="RFC5322"/> and the Multipurpose Internet Mail Extensions
316   (MIME) <xref target="RFC2045"/> (see &diff-mime; for the differences
317   between HTTP and MIME messages).
318</t>
319<t>
320   Most HTTP communication consists of a retrieval request (GET) for
321   a representation of some resource identified by a URI.  In the
322   simplest case, this might be accomplished via a single bidirectional
323   connection (===) between the user agent (UA) and the origin server (O).
324</t>
325<figure><artwork type="drawing">
326         request   &gt;
327    <x:highlight>UA</x:highlight> ======================================= <x:highlight>O</x:highlight>
328                                &lt;   response
329</artwork></figure>
330<iref primary="true" item="message"/>
331<iref primary="true" item="request"/>
332<iref primary="true" item="response"/>
333<t>
334   A client sends an HTTP request to a server in the form of a <x:dfn>request</x:dfn>
335   message, beginning with a request-line that includes a method, URI, and
336   protocol version (<xref target="request.line"/>),
337   followed by header fields containing
338   request modifiers, client information, and representation metadata
339   (<xref target="header.fields"/>),
340   an empty line to indicate the end of the header section, and finally
341   a message body containing the payload body (if any,
342   <xref target="message.body"/>).
343</t>
344<t>
345   A server responds to a client's request by sending one or more HTTP
346   <x:dfn>response</x:dfn>
347   messages, each beginning with a status line that
348   includes the protocol version, a success or error code, and textual
349   reason phrase (<xref target="status.line"/>),
350   possibly followed by header fields containing server
351   information, resource metadata, and representation metadata
352   (<xref target="header.fields"/>),
353   an empty line to indicate the end of the header section, and finally
354   a message body containing the payload body (if any,
355   <xref target="message.body"/>).
356</t>
357<t>
358   A connection might be used for multiple request/response exchanges,
359   as defined in <xref target="persistent.connections"/>.
360</t>
361<t>
362   The following example illustrates a typical message exchange for a
363   GET request on the URI "http://www.example.com/hello.txt":
364</t>
365<figure><preamble>
366client request:
367</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
368GET /hello.txt HTTP/1.1
369User-Agent: curl/7.16.3 libcurl/7.16.3 OpenSSL/0.9.7l zlib/1.2.3
370Host: www.example.com
371Accept-Language: en, mi
372
373</artwork></figure>
374<figure><preamble>
375server response:
376</preamble><artwork type="message/http; msgtype=&#34;response&#34;" x:indent-with="  ">
377HTTP/1.1 200 OK
378Date: Mon, 27 Jul 2009 12:28:53 GMT
379Server: Apache
380Last-Modified: Wed, 22 Jul 2009 19:15:56 GMT
381ETag: "34aa387-d-1568eb00"
382Accept-Ranges: bytes
383Content-Length: <x:length-of target="exbody"/>
384Vary: Accept-Encoding
385Content-Type: text/plain
386
387<x:span anchor="exbody">Hello World!
388</x:span></artwork></figure>
389</section>
390
391<section title="Implementation Diversity" anchor="implementation-diversity">
392<t>
393   When considering the design of HTTP, it is easy to fall into a trap of
394   thinking that all user agents are general-purpose browsers and all origin
395   servers are large public websites. That is not the case in practice.
396   Common HTTP user agents include household appliances, stereos, scales,
397   firmware update scripts, command-line programs, mobile apps,
398   and communication devices in a multitude of shapes and sizes.  Likewise,
399   common HTTP origin servers include home automation units, configurable
400   networking components, office machines, autonomous robots, news feeds,
401   traffic cameras, ad selectors, and video delivery platforms.
402</t>
403<t>
404   The term "user agent" does not imply that there is a human user directly
405   interacting with the software agent at the time of a request. In many
406   cases, a user agent is installed or configured to run in the background
407   and save its results for later inspection (or save only a subset of those
408   results that might be interesting or erroneous). Spiders, for example, are
409   typically given a start URI and configured to follow certain behavior while
410   crawling the Web as a hypertext graph.
411</t>
412<t>
413   The implementation diversity of HTTP means that we cannot assume the
414   user agent can make interactive suggestions to a user or provide adequate
415   warning for security or privacy options.  In the few cases where this
416   specification requires reporting of errors to the user, it is acceptable
417   for such reporting to only be observable in an error console or log file.
418   Likewise, requirements that an automated action be confirmed by the user
419   before proceeding can be met via advance configuration choices,
420   run-time options, or simply not proceeding with the unsafe action.
421</t>
422</section>
423
424<section title="Intermediaries" anchor="intermediaries">
425<iref primary="true" item="intermediary"/>
426<t>
427   HTTP enables the use of intermediaries to satisfy requests through
428   a chain of connections.  There are three common forms of HTTP
429   <x:dfn>intermediary</x:dfn>: proxy, gateway, and tunnel.  In some cases,
430   a single intermediary might act as an origin server, proxy, gateway,
431   or tunnel, switching behavior based on the nature of each request.
432</t>
433<figure><artwork type="drawing">
434         &gt;             &gt;             &gt;             &gt;
435    <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>
436               &lt;             &lt;             &lt;             &lt;
437</artwork></figure>
438<t>
439   The figure above shows three intermediaries (A, B, and C) between the
440   user agent and origin server. A request or response message that
441   travels the whole chain will pass through four separate connections.
442   Some HTTP communication options
443   might apply only to the connection with the nearest, non-tunnel
444   neighbor, only to the end-points of the chain, or to all connections
445   along the chain. Although the diagram is linear, each participant might
446   be engaged in multiple, simultaneous communications. For example, B
447   might be receiving requests from many clients other than A, and/or
448   forwarding requests to servers other than C, at the same time that it
449   is handling A's request.
450</t>
451<t>
452<iref primary="true" item="upstream"/><iref primary="true" item="downstream"/>
453<iref primary="true" item="inbound"/><iref primary="true" item="outbound"/>
454   We use the terms "<x:dfn>upstream</x:dfn>" and "<x:dfn>downstream</x:dfn>"
455   to describe various requirements in relation to the directional flow of a
456   message: all messages flow from upstream to downstream.
457   Likewise, we use the terms inbound and outbound to refer to
458   directions in relation to the request path:
459   "<x:dfn>inbound</x:dfn>" means toward the origin server and
460   "<x:dfn>outbound</x:dfn>" means toward the user agent.
461</t>
462<t><iref primary="true" item="proxy"/>
463   A "<x:dfn>proxy</x:dfn>" is a message forwarding agent that is selected by the
464   client, usually via local configuration rules, to receive requests
465   for some type(s) of absolute URI and attempt to satisfy those
466   requests via translation through the HTTP interface.  Some translations
467   are minimal, such as for proxy requests for "http" URIs, whereas
468   other requests might require translation to and from entirely different
469   application-level protocols. Proxies are often used to group an
470   organization's HTTP requests through a common intermediary for the
471   sake of security, annotation services, or shared caching.
472</t>
473<t>
474<iref primary="true" item="transforming proxy"/>
475<iref primary="true" item="non-transforming proxy"/>
476   An HTTP-to-HTTP proxy is called a "<x:dfn>transforming proxy</x:dfn>" if it is designed
477   or configured to modify request or response messages in a semantically
478   meaningful way (i.e., modifications, beyond those required by normal
479   HTTP processing, that change the message in a way that would be
480   significant to the original sender or potentially significant to
481   downstream recipients).  For example, a transforming proxy might be
482   acting as a shared annotation server (modifying responses to include
483   references to a local annotation database), a malware filter, a
484   format transcoder, or an intranet-to-Internet privacy filter.  Such
485   transformations are presumed to be desired by the client (or client
486   organization) that selected the proxy and are beyond the scope of
487   this specification.  However, when a proxy is not intended to transform
488   a given message, we use the term "<x:dfn>non-transforming proxy</x:dfn>" to target
489   requirements that preserve HTTP message semantics. See &status-203; and
490   &header-warning; for status and warning codes related to transformations.
491</t>
492<t><iref primary="true" item="gateway"/><iref primary="true" item="reverse proxy"/>
493<iref primary="true" item="accelerator"/>
494   A "<x:dfn>gateway</x:dfn>" (a.k.a., "<x:dfn>reverse proxy</x:dfn>")
495   is a receiving agent that acts
496   as a layer above some other server(s) and translates the received
497   requests to the underlying server's protocol.  Gateways are often
498   used to encapsulate legacy or untrusted information services, to
499   improve server performance through "<x:dfn>accelerator</x:dfn>" caching, and to
500   enable partitioning or load-balancing of HTTP services across
501   multiple machines.
502</t>
503<t>
504   A gateway behaves as an origin server on its outbound connection and
505   as a user agent on its inbound connection.
506   All HTTP requirements applicable to an origin server
507   also apply to the outbound communication of a gateway.
508   A gateway communicates with inbound servers using any protocol that
509   it desires, including private extensions to HTTP that are outside
510   the scope of this specification.  However, an HTTP-to-HTTP gateway
511   that wishes to interoperate with third-party HTTP servers &MUST;
512   conform to HTTP user agent requirements on the gateway's inbound
513   connection and &MUST; implement the <x:ref>Connection</x:ref>
514   (<xref target="header.connection"/>) and <x:ref>Via</x:ref>
515   (<xref target="header.via"/>) header fields for both connections.
516</t>
517<t><iref primary="true" item="tunnel"/>
518   A "<x:dfn>tunnel</x:dfn>" acts as a blind relay between two connections
519   without changing the messages. Once active, a tunnel is not
520   considered a party to the HTTP communication, though the tunnel might
521   have been initiated by an HTTP request. A tunnel ceases to exist when
522   both ends of the relayed connection are closed. Tunnels are used to
523   extend a virtual connection through an intermediary, such as when
524   Transport Layer Security (TLS, <xref target="RFC5246"/>) is used to
525   establish confidential communication through a shared firewall proxy.
526</t>
527<t><iref primary="true" item="interception proxy"/>
528<iref primary="true" item="transparent proxy"/>
529<iref primary="true" item="captive portal"/>
530   The above categories for intermediary only consider those acting as
531   participants in the HTTP communication.  There are also intermediaries
532   that can act on lower layers of the network protocol stack, filtering or
533   redirecting HTTP traffic without the knowledge or permission of message
534   senders. Network intermediaries often introduce security flaws or
535   interoperability problems by violating HTTP semantics.  For example, an
536   "<x:dfn>interception proxy</x:dfn>" <xref target="RFC3040"/> (also commonly
537   known as a "<x:dfn>transparent proxy</x:dfn>" <xref target="RFC1919"/> or
538   "<x:dfn>captive portal</x:dfn>")
539   differs from an HTTP proxy because it is not selected by the client.
540   Instead, an interception proxy filters or redirects outgoing TCP port 80
541   packets (and occasionally other common port traffic).
542   Interception proxies are commonly found on public network access points,
543   as a means of enforcing account subscription prior to allowing use of
544   non-local Internet services, and within corporate firewalls to enforce
545   network usage policies.
546   They are indistinguishable from a man-in-the-middle attack.
547</t>
548<t>
549   HTTP is defined as a stateless protocol, meaning that each request message
550   can be understood in isolation.  Many implementations depend on HTTP's
551   stateless design in order to reuse proxied connections or dynamically
552   load-balance requests across multiple servers.  Hence, servers &MUST-NOT;
553   assume that two requests on the same connection are from the same user
554   agent unless the connection is secured and specific to that agent.
555   Some non-standard HTTP extensions (e.g., <xref target="RFC4559"/>) have
556   been known to violate this requirement, resulting in security and
557   interoperability problems.
558</t>
559</section>
560
561<section title="Caches" anchor="caches">
562<iref primary="true" item="cache"/>
563<t>
564   A "<x:dfn>cache</x:dfn>" is a local store of previous response messages and the
565   subsystem that controls its message storage, retrieval, and deletion.
566   A cache stores cacheable responses in order to reduce the response
567   time and network bandwidth consumption on future, equivalent
568   requests. Any client or server &MAY; employ a cache, though a cache
569   cannot be used by a server while it is acting as a tunnel.
570</t>
571<t>
572   The effect of a cache is that the request/response chain is shortened
573   if one of the participants along the chain has a cached response
574   applicable to that request. The following illustrates the resulting
575   chain if B has a cached copy of an earlier response from O (via C)
576   for a request which has not been cached by UA or A.
577</t>
578<figure><artwork type="drawing">
579            &gt;             &gt;
580       <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>
581                  &lt;             &lt;
582</artwork></figure>
583<t><iref primary="true" item="cacheable"/>
584   A response is "<x:dfn>cacheable</x:dfn>" if a cache is allowed to store a copy of
585   the response message for use in answering subsequent requests.
586   Even when a response is cacheable, there might be additional
587   constraints placed by the client or by the origin server on when
588   that cached response can be used for a particular request. HTTP
589   requirements for cache behavior and cacheable responses are
590   defined in &caching-overview;
591</t>
592<t>
593   There are a wide variety of architectures and configurations
594   of caches deployed across the World Wide Web and
595   inside large organizations. These include national hierarchies
596   of proxy caches to save transoceanic bandwidth, collaborative systems that
597   broadcast or multicast cache entries, archives of pre-fetched cache
598   entries for use in off-line or high-latency environments, and so on.
599</t>
600</section>
601
602<section title="Conformance and Error Handling" anchor="conformance">
603<t>
604   This specification targets conformance criteria according to the role of
605   a participant in HTTP communication.  Hence, HTTP requirements are placed
606   on senders, recipients, clients, servers, user agents, intermediaries,
607   origin servers, proxies, gateways, or caches, depending on what behavior
608   is being constrained by the requirement. Additional (social) requirements
609   are placed on implementations, resource owners, and protocol element
610   registrations when they apply beyond the scope of a single communication.
611</t>
612<t>
613   The verb "generate" is used instead of "send" where a requirement
614   differentiates between creating a protocol element and merely forwarding a
615   received element downstream.
616</t>
617<t>
618   An implementation is considered conformant if it complies with all of the
619   requirements associated with the roles it partakes in HTTP. Note that
620   SHOULD-level requirements are relevant here, unless one of the documented
621   exceptions is applicable.
622</t>
623<t>
624   Conformance applies to both the syntax and semantics of HTTP protocol
625   elements. A sender &MUST-NOT; generate protocol elements that convey a
626   meaning that is known by that sender to be false. A sender &MUST-NOT;
627   generate protocol elements that do not match the grammar defined by the
628   ABNF rules for those protocol elements that are applicable to the sender's
629   role. If a received protocol element is processed, the recipient &MUST; be
630   able to parse any value that would match the ABNF rules for that protocol
631   element, excluding only those rules not applicable to the recipient's role.
632</t>
633<t>
634   Unless noted otherwise, a recipient &MAY; attempt to recover a usable
635   protocol element from an invalid construct.  HTTP does not define
636   specific error handling mechanisms except when they have a direct impact
637   on security, since different applications of the protocol require
638   different error handling strategies.  For example, a Web browser might
639   wish to transparently recover from a response where the
640   <x:ref>Location</x:ref> header field doesn't parse according to the ABNF,
641   whereas a systems control client might consider any form of error recovery
642   to be dangerous.
643</t>
644</section>
645
646<section title="Protocol Versioning" anchor="http.version">
647  <x:anchor-alias value="HTTP-version"/>
648  <x:anchor-alias value="HTTP-name"/>
649<t>
650   HTTP uses a "&lt;major&gt;.&lt;minor&gt;" numbering scheme to indicate
651   versions of the protocol. This specification defines version "1.1".
652   The protocol version as a whole indicates the sender's conformance
653   with the set of requirements laid out in that version's corresponding
654   specification of HTTP.
655</t>
656<t>
657   The version of an HTTP message is indicated by an HTTP-version field
658   in the first line of the message. HTTP-version is case-sensitive.
659</t>
660<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-version"/><iref primary="true" item="Grammar" subitem="HTTP-name"/>
661  <x:ref>HTTP-version</x:ref>  = <x:ref>HTTP-name</x:ref> "/" <x:ref>DIGIT</x:ref> "." <x:ref>DIGIT</x:ref>
662  <x:ref>HTTP-name</x:ref>     = <x:abnf-char-sequence>"HTTP"</x:abnf-char-sequence> ; "HTTP", case-sensitive
663</artwork></figure>
664<t>
665   The HTTP version number consists of two decimal digits separated by a "."
666   (period or decimal point).  The first digit ("major version") indicates the
667   HTTP messaging syntax, whereas the second digit ("minor version") indicates
668   the highest minor version to which the sender is
669   conformant and able to understand for future communication.  The minor
670   version advertises the sender's communication capabilities even when the
671   sender is only using a backwards-compatible subset of the protocol,
672   thereby letting the recipient know that more advanced features can
673   be used in response (by servers) or in future requests (by clients).
674</t>
675<t>
676   When an HTTP/1.1 message is sent to an HTTP/1.0 recipient
677   <xref target="RFC1945"/> or a recipient whose version is unknown,
678   the HTTP/1.1 message is constructed such that it can be interpreted
679   as a valid HTTP/1.0 message if all of the newer features are ignored.
680   This specification places recipient-version requirements on some
681   new features so that a conformant sender will only use compatible
682   features until it has determined, through configuration or the
683   receipt of a message, that the recipient supports HTTP/1.1.
684</t>
685<t>
686   The interpretation of a header field does not change between minor
687   versions of the same major HTTP version, though the default
688   behavior of a recipient in the absence of such a field can change.
689   Unless specified otherwise, header fields defined in HTTP/1.1 are
690   defined for all versions of HTTP/1.x.  In particular, the <x:ref>Host</x:ref>
691   and <x:ref>Connection</x:ref> header fields ought to be implemented by all
692   HTTP/1.x implementations whether or not they advertise conformance with
693   HTTP/1.1.
694</t>
695<t>
696   New header fields can be defined such that, when they are
697   understood by a recipient, they might override or enhance the
698   interpretation of previously defined header fields.  When an
699   implementation receives an unrecognized header field, the recipient
700   &MUST; ignore that header field for local processing regardless of
701   the message's HTTP version.  An unrecognized header field received
702   by a proxy &MUST; be forwarded downstream unless the header field's
703   field-name is listed in the message's <x:ref>Connection</x:ref> header field
704   (see <xref target="header.connection"/>).
705   These requirements allow HTTP's functionality to be enhanced without
706   requiring prior update of deployed intermediaries.
707</t>
708<t>
709   Intermediaries that process HTTP messages (i.e., all intermediaries
710   other than those acting as tunnels) &MUST; send their own HTTP-version
711   in forwarded messages.  In other words, they &MUST-NOT; blindly
712   forward the first line of an HTTP message without ensuring that the
713   protocol version in that message matches a version to which that
714   intermediary is conformant for both the receiving and
715   sending of messages.  Forwarding an HTTP message without rewriting
716   the HTTP-version might result in communication errors when downstream
717   recipients use the message sender's version to determine what features
718   are safe to use for later communication with that sender.
719</t>
720<t>
721   An HTTP client &SHOULD; send a request version equal to the highest
722   version to which the client is conformant and
723   whose major version is no higher than the highest version supported
724   by the server, if this is known.  An HTTP client &MUST-NOT; send a
725   version to which it is not conformant.
726</t>
727<t>
728   An HTTP client &MAY; send a lower request version if it is known that
729   the server incorrectly implements the HTTP specification, but only
730   after the client has attempted at least one normal request and determined
731   from the response status or header fields (e.g., <x:ref>Server</x:ref>) that
732   the server improperly handles higher request versions.
733</t>
734<t>
735   An HTTP server &SHOULD; send a response version equal to the highest
736   version to which the server is conformant and
737   whose major version is less than or equal to the one received in the
738   request.  An HTTP server &MUST-NOT; send a version to which it is not
739   conformant.  A server &MAY; send a <x:ref>505 (HTTP Version Not
740   Supported)</x:ref> response if it cannot send a response using the
741   major version used in the client's request.
742</t>
743<t>
744   An HTTP server &MAY; send an HTTP/1.0 response to an HTTP/1.0 request
745   if it is known or suspected that the client incorrectly implements the
746   HTTP specification and is incapable of correctly processing later
747   version responses, such as when a client fails to parse the version
748   number correctly or when an intermediary is known to blindly forward
749   the HTTP-version even when it doesn't conform to the given minor
750   version of the protocol. Such protocol downgrades &SHOULD-NOT; be
751   performed unless triggered by specific client attributes, such as when
752   one or more of the request header fields (e.g., <x:ref>User-Agent</x:ref>)
753   uniquely match the values sent by a client known to be in error.
754</t>
755<t>
756   The intention of HTTP's versioning design is that the major number
757   will only be incremented if an incompatible message syntax is
758   introduced, and that the minor number will only be incremented when
759   changes made to the protocol have the effect of adding to the message
760   semantics or implying additional capabilities of the sender.  However,
761   the minor version was not incremented for the changes introduced between
762   <xref target="RFC2068"/> and <xref target="RFC2616"/>, and this revision
763   has specifically avoiding any such changes to the protocol.
764</t>
765</section>
766
767<section title="Uniform Resource Identifiers" anchor="uri">
768<iref primary="true" item="resource"/>
769<t>
770   Uniform Resource Identifiers (URIs) <xref target="RFC3986"/> are used
771   throughout HTTP as the means for identifying resources (&resource;).
772   URI references are used to target requests, indicate redirects, and define
773   relationships.
774</t>
775  <x:anchor-alias value="URI-reference"/>
776  <x:anchor-alias value="absolute-URI"/>
777  <x:anchor-alias value="relative-part"/>
778  <x:anchor-alias value="authority"/>
779  <x:anchor-alias value="path-abempty"/>
780  <x:anchor-alias value="path-absolute"/>
781  <x:anchor-alias value="port"/>
782  <x:anchor-alias value="query"/>
783  <x:anchor-alias value="uri-host"/>
784  <x:anchor-alias value="partial-URI"/>
785<t>
786   This specification adopts the definitions of "URI-reference",
787   "absolute-URI", "relative-part", "port", "host",
788   "path-abempty", "path-absolute", "query", and "authority" from the
789   URI generic syntax.
790   In addition, we define a partial-URI rule for protocol elements
791   that allow a relative URI but not a fragment.
792</t>
793<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="URI-reference"><!--exported production--></iref><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"/><iref primary="true" item="Grammar" subitem="partial-URI"><!--exported production--></iref>
794  <x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in <xref target="RFC3986" x:fmt="," x:sec="4.1"/>&gt;
795  <x:ref>absolute-URI</x:ref>  = &lt;absolute-URI, defined in <xref target="RFC3986" x:fmt="," x:sec="4.3"/>&gt;
796  <x:ref>relative-part</x:ref> = &lt;relative-part, defined in <xref target="RFC3986" x:fmt="," x:sec="4.2"/>&gt;
797  <x:ref>authority</x:ref>     = &lt;authority, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2"/>&gt;
798  <x:ref>path-abempty</x:ref>  = &lt;path-abempty, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
799  <x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in <xref target="RFC3986" x:fmt="," x:sec="3.3"/>&gt;
800  <x:ref>port</x:ref>          = &lt;port, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.3"/>&gt;
801  <x:ref>query</x:ref>         = &lt;query, defined in <xref target="RFC3986" x:fmt="," x:sec="3.4"/>&gt;
802  <x:ref>uri-host</x:ref>      = &lt;host, defined in <xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>&gt;
803 
804  <x:ref>partial-URI</x:ref>   = relative-part [ "?" query ]
805</artwork></figure>
806<t>
807   Each protocol element in HTTP that allows a URI reference will indicate
808   in its ABNF production whether the element allows any form of reference
809   (URI-reference), only a URI in absolute form (absolute-URI), only the
810   path and optional query components, or some combination of the above.
811   Unless otherwise indicated, URI references are parsed
812   relative to the effective request URI
813   (<xref target="effective.request.uri"/>).
814</t>
815
816<section title="http URI scheme" anchor="http.uri">
817  <x:anchor-alias value="http-URI"/>
818  <iref item="http URI scheme" primary="true"/>
819  <iref item="URI scheme" subitem="http" primary="true"/>
820<t>
821   The "http" URI scheme is hereby defined for the purpose of minting
822   identifiers according to their association with the hierarchical
823   namespace governed by a potential HTTP origin server listening for
824   TCP connections on a given port.
825</t>
826<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="http-URI"><!--terminal production--></iref>
827  <x:ref>http-URI</x:ref> = "http:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
828</artwork></figure>
829<t>
830   The HTTP origin server is identified by the generic syntax's
831   <x:ref>authority</x:ref> component, which includes a host identifier
832   and optional TCP port (<xref target="RFC3986" x:fmt="," x:sec="3.2.2"/>).
833   The remainder of the URI, consisting of both the hierarchical path
834   component and optional query component, serves as an identifier for
835   a potential resource within that origin server's name space.
836</t>
837<t>
838   If the host identifier is provided as an IP address,
839   then the origin server is any listener on the indicated TCP port at
840   that IP address. If host is a registered name, then that name is
841   considered an indirect identifier and the recipient might use a name
842   resolution service, such as DNS, to find the address of a listener
843   for that host.
844   The host &MUST-NOT; be empty; if an "http" URI is received with an
845   empty host, then it &MUST; be rejected as invalid.
846   If the port subcomponent is empty or not given, then TCP port 80 is
847   assumed (the default reserved port for WWW services).
848</t>
849<t>
850   Regardless of the form of host identifier, access to that host is not
851   implied by the mere presence of its name or address. The host might or might
852   not exist and, even when it does exist, might or might not be running an
853   HTTP server or listening to the indicated port. The "http" URI scheme
854   makes use of the delegated nature of Internet names and addresses to
855   establish a naming authority (whatever entity has the ability to place
856   an HTTP server at that Internet name or address) and allows that
857   authority to determine which names are valid and how they might be used.
858</t>
859<t>
860   When an "http" URI is used within a context that calls for access to the
861   indicated resource, a client &MAY; attempt access by resolving
862   the host to an IP address, establishing a TCP connection to that address
863   on the indicated port, and sending an HTTP request message
864   (<xref target="http.message"/>) containing the URI's identifying data
865   (<xref target="message.routing"/>) to the server.
866   If the server responds to that request with a non-interim HTTP response
867   message, as described in &status-codes;, then that response
868   is considered an authoritative answer to the client's request.
869</t>
870<t>
871   Although HTTP is independent of the transport protocol, the "http"
872   scheme is specific to TCP-based services because the name delegation
873   process depends on TCP for establishing authority.
874   An HTTP service based on some other underlying connection protocol
875   would presumably be identified using a different URI scheme, just as
876   the "https" scheme (below) is used for resources that require an
877   end-to-end secured connection. Other protocols might also be used to
878   provide access to "http" identified resources &mdash; it is only the
879   authoritative interface used for mapping the namespace that is
880   specific to TCP.
881</t>
882<t>
883   The URI generic syntax for authority also includes a deprecated
884   userinfo subcomponent (<xref target="RFC3986" x:fmt="," x:sec="3.2.1"/>)
885   for including user authentication information in the URI.  Some
886   implementations make use of the userinfo component for internal
887   configuration of authentication information, such as within command
888   invocation options, configuration files, or bookmark lists, even
889   though such usage might expose a user identifier or password.
890   Senders &MUST; exclude the userinfo subcomponent (and its "@"
891   delimiter) when an "http" URI is transmitted within a message as a
892   request target or header field value.
893   Recipients of an "http" URI reference &SHOULD; parse for userinfo and
894   treat its presence as an error, since it is likely being used to obscure
895   the authority for the sake of phishing attacks.
896</t>
897</section>
898
899<section title="https URI scheme" anchor="https.uri">
900   <x:anchor-alias value="https-URI"/>
901   <iref item="https URI scheme"/>
902   <iref item="URI scheme" subitem="https"/>
903<t>
904   The "https" URI scheme is hereby defined for the purpose of minting
905   identifiers according to their association with the hierarchical
906   namespace governed by a potential HTTP origin server listening to a
907   given TCP port for TLS-secured connections <xref target="RFC5246"/>.
908</t>
909<t>
910   All of the requirements listed above for the "http" scheme are also
911   requirements for the "https" scheme, except that a default TCP port
912   of 443 is assumed if the port subcomponent is empty or not given,
913   and the TCP connection &MUST; be secured, end-to-end, through the
914   use of strong encryption prior to sending the first HTTP request.
915</t>
916<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="https-URI"><!--terminal production--></iref>
917  <x:ref>https-URI</x:ref> = "https:" "//" <x:ref>authority</x:ref> <x:ref>path-abempty</x:ref> [ "?" <x:ref>query</x:ref> ]
918</artwork></figure>
919<t>
920   Unlike the "http" scheme, responses to "https" identified requests
921   are never "public" and thus &MUST-NOT; be reused for shared caching.
922   They can, however, be reused in a private cache if the message is
923   cacheable by default in HTTP or specifically indicated as such by
924   the Cache-Control header field (&header-cache-control;).
925</t>
926<t>
927   Resources made available via the "https" scheme have no shared
928   identity with the "http" scheme even if their resource identifiers
929   indicate the same authority (the same host listening to the same
930   TCP port).  They are distinct name spaces and are considered to be
931   distinct origin servers.  However, an extension to HTTP that is
932   defined to apply to entire host domains, such as the Cookie protocol
933   <xref target="RFC6265"/>, can allow information
934   set by one service to impact communication with other services
935   within a matching group of host domains.
936</t>
937<t>
938   The process for authoritative access to an "https" identified
939   resource is defined in <xref target="RFC2818"/>.
940</t>
941</section>
942
943<section title="http and https URI Normalization and Comparison" anchor="uri.comparison">
944<t>
945   Since the "http" and "https" schemes conform to the URI generic syntax,
946   such URIs are normalized and compared according to the algorithm defined
947   in <xref target="RFC3986" x:fmt="," x:sec="6"/>, using the defaults
948   described above for each scheme.
949</t>
950<t>
951   If the port is equal to the default port for a scheme, the normal
952   form is to elide the port subcomponent. Likewise, an empty path
953   component is equivalent to an absolute path of "/", so the normal
954   form is to provide a path of "/" instead. The scheme and host
955   are case-insensitive and normally provided in lowercase; all
956   other components are compared in a case-sensitive manner.
957   Characters other than those in the "reserved" set are equivalent
958   to their percent-encoded octets (see <xref target="RFC3986"
959   x:fmt="," x:sec="2.1"/>): the normal form is to not encode them.
960</t>
961<t>
962   For example, the following three URIs are equivalent:
963</t>
964<figure><artwork type="example">
965   http://example.com:80/~smith/home.html
966   http://EXAMPLE.com/%7Esmith/home.html
967   http://EXAMPLE.com:/%7esmith/home.html
968</artwork></figure>
969</section>
970</section>
971</section>
972
973<section title="Message Format" anchor="http.message">
974<x:anchor-alias value="generic-message"/>
975<x:anchor-alias value="message.types"/>
976<x:anchor-alias value="HTTP-message"/>
977<x:anchor-alias value="start-line"/>
978<iref item="header section"/>
979<iref item="headers"/>
980<iref item="header field"/>
981<t>
982   All HTTP/1.1 messages consist of a start-line followed by a sequence of
983   octets in a format similar to the Internet Message Format
984   <xref target="RFC5322"/>: zero or more header fields (collectively
985   referred to as the "headers" or the "header section"), an empty line
986   indicating the end of the header section, and an optional message body.
987</t>
988<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="HTTP-message"><!--terminal production--></iref>
989  <x:ref>HTTP-message</x:ref>   = <x:ref>start-line</x:ref>
990                   *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
991                   <x:ref>CRLF</x:ref>
992                   [ <x:ref>message-body</x:ref> ]
993</artwork></figure>
994<t>
995   The normal procedure for parsing an HTTP message is to read the
996   start-line into a structure, read each header field into a hash
997   table by field name until the empty line, and then use the parsed
998   data to determine if a message body is expected.  If a message body
999   has been indicated, then it is read as a stream until an amount
1000   of octets equal to the message body length is read or the connection
1001   is closed.
1002</t>
1003<t>
1004   Recipients &MUST; parse an HTTP message as a sequence of octets in an
1005   encoding that is a superset of US-ASCII <xref target="USASCII"/>.
1006   Parsing an HTTP message as a stream of Unicode characters, without regard
1007   for the specific encoding, creates security vulnerabilities due to the
1008   varying ways that string processing libraries handle invalid multibyte
1009   character sequences that contain the octet LF (%x0A).  String-based
1010   parsers can only be safely used within protocol elements after the element
1011   has been extracted from the message, such as within a header field-value
1012   after message parsing has delineated the individual fields.
1013</t>
1014<t>
1015   An HTTP message can be parsed as a stream for incremental processing or
1016   forwarding downstream.  However, recipients cannot rely on incremental
1017   delivery of partial messages, since some implementations will buffer or
1018   delay message forwarding for the sake of network efficiency, security
1019   checks, or payload transformations.
1020</t>
1021
1022<section title="Start Line" anchor="start.line">
1023  <x:anchor-alias value="Start-Line"/>
1024<t>
1025   An HTTP message can either be a request from client to server or a
1026   response from server to client.  Syntactically, the two types of message
1027   differ only in the start-line, which is either a request-line (for requests)
1028   or a status-line (for responses), and in the algorithm for determining
1029   the length of the message body (<xref target="message.body"/>).
1030</t>
1031<t>
1032   In theory, a client could receive requests and a server could receive
1033   responses, distinguishing them by their different start-line formats,
1034   but in practice servers are implemented to only expect a request
1035   (a response is interpreted as an unknown or invalid request method)
1036   and clients are implemented to only expect a response.
1037</t>
1038<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="start-line"/>
1039  <x:ref>start-line</x:ref>     = <x:ref>request-line</x:ref> / <x:ref>status-line</x:ref>
1040</artwork></figure>
1041<t>
1042   A sender &MUST-NOT; send whitespace between the start-line and
1043   the first header field. The presence of such whitespace in a request
1044   might be an attempt to trick a server into ignoring that field or
1045   processing the line after it as a new request, either of which might
1046   result in a security vulnerability if other implementations within
1047   the request chain interpret the same message differently.
1048   Likewise, the presence of such whitespace in a response might be
1049   ignored by some clients or cause others to cease parsing.
1050</t>
1051<t>
1052   A recipient that receives whitespace between the start-line and
1053   the first header field &MUST; either reject the message as invalid or
1054   consume each whitespace-preceded line without further processing of it
1055   (i.e., ignore the entire line, along with any subsequent lines preceded
1056   by whitespace, until a properly formed header field is received or the
1057   header block is terminated).
1058</t>
1059
1060<section title="Request Line" anchor="request.line">
1061  <x:anchor-alias value="Request"/>
1062  <x:anchor-alias value="request-line"/>
1063<t>
1064   A request-line begins with a method token, followed by a single
1065   space (SP), the request-target, another single space (SP), the
1066   protocol version, and ending with CRLF.
1067</t>
1068<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-line"/>
1069  <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>
1070</artwork></figure>
1071<iref primary="true" item="method"/>
1072<t anchor="method">
1073   The method token indicates the request method to be performed on the
1074   target resource. The request method is case-sensitive.
1075</t>
1076<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="method"/>
1077  <x:ref>method</x:ref>         = <x:ref>token</x:ref>
1078</artwork></figure>
1079<t>
1080   The methods defined by this specification can be found in
1081   &methods;, along with information regarding the HTTP method registry
1082   and considerations for defining new methods.
1083</t>
1084<iref item="request-target"/>
1085<t>
1086   The request-target identifies the target resource upon which to apply
1087   the request, as defined in <xref target="request-target"/>.
1088</t>
1089<t>
1090   No whitespace is allowed inside the method, request-target, and
1091   protocol version.  Hence, recipients typically parse the request-line
1092   into its component parts by splitting on whitespace
1093   (see <xref target="message.robustness"/>).
1094</t>
1095<t>
1096   Unfortunately, some user agents fail to properly encode hypertext
1097   references that have embedded whitespace, sending the characters directly
1098   instead of properly encoding or excluding the disallowed characters.
1099   Recipients of an invalid request-line &SHOULD; respond with either a
1100   <x:ref>400 (Bad Request)</x:ref> error or a <x:ref>301 (Moved Permanently)</x:ref>
1101   redirect with the request-target properly encoded.  Recipients &SHOULD-NOT;
1102   attempt to autocorrect and then process the request without a redirect,
1103   since the invalid request-line might be deliberately crafted to bypass
1104   security filters along the request chain.
1105</t>
1106<t>
1107   HTTP does not place a pre-defined limit on the length of a request-line.
1108   A server that receives a method longer than any that it implements
1109   &SHOULD; respond with either a <x:ref>405 (Method Not Allowed)</x:ref>, if it is an origin
1110   server, or a <x:ref>501 (Not Implemented)</x:ref> status code.
1111   A server &MUST; be prepared to receive URIs of unbounded length and
1112   respond with the <x:ref>414 (URI Too Long)</x:ref> status code if the received
1113   request-target would be longer than the server wishes to handle
1114   (see &status-414;).
1115</t>
1116<t>
1117   Various ad-hoc limitations on request-line length are found in practice.
1118   It is &RECOMMENDED; that all HTTP senders and recipients support, at a
1119   minimum, request-line lengths of 8000 octets.
1120</t>
1121</section>
1122
1123<section title="Status Line" anchor="status.line">
1124  <x:anchor-alias value="response"/>
1125  <x:anchor-alias value="status-line"/>
1126  <x:anchor-alias value="status-code"/>
1127  <x:anchor-alias value="reason-phrase"/>
1128<t>
1129   The first line of a response message is the status-line, consisting
1130   of the protocol version, a space (SP), the status code, another space,
1131   a possibly-empty textual phrase describing the status code, and
1132   ending with CRLF.
1133</t>
1134<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-line"/>
1135  <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>
1136</artwork></figure>
1137<t>
1138   The status-code element is a 3-digit integer code describing the
1139   result of the server's attempt to understand and satisfy the client's
1140   corresponding request. The rest of the response message is to be
1141   interpreted in light of the semantics defined for that status code.
1142   See &status-codes; for information about the semantics of status codes,
1143   including the classes of status code (indicated by the first digit),
1144   the status codes defined by this specification, considerations for the
1145   definition of new status codes, and the IANA registry.
1146</t>
1147<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="status-code"/>
1148  <x:ref>status-code</x:ref>    = 3<x:ref>DIGIT</x:ref>
1149</artwork></figure>
1150<t>   
1151   The reason-phrase element exists for the sole purpose of providing a
1152   textual description associated with the numeric status code, mostly
1153   out of deference to earlier Internet application protocols that were more
1154   frequently used with interactive text clients. A client &SHOULD; ignore
1155   the reason-phrase content.
1156</t>
1157<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="reason-phrase"/>
1158  <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> )
1159</artwork></figure>
1160</section>
1161</section>
1162
1163<section title="Header Fields" anchor="header.fields">
1164  <x:anchor-alias value="header-field"/>
1165  <x:anchor-alias value="field-content"/>
1166  <x:anchor-alias value="field-name"/>
1167  <x:anchor-alias value="field-value"/>
1168  <x:anchor-alias value="obs-fold"/>
1169<t>
1170   Each HTTP header field consists of a case-insensitive field name
1171   followed by a colon (":"), optional whitespace, and the field value.
1172</t>
1173<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"/><iref primary="true" item="Grammar" subitem="obs-fold"/>
1174  <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>
1175  <x:ref>field-name</x:ref>     = <x:ref>token</x:ref>
1176  <x:ref>field-value</x:ref>    = *( <x:ref>field-content</x:ref> / <x:ref>obs-fold</x:ref> )
1177  <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> )
1178  <x:ref>obs-fold</x:ref>       = <x:ref>CRLF</x:ref> ( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1179                 ; obsolete line folding
1180                 ; see <xref target="field.parsing"/>
1181</artwork></figure>
1182<t>
1183   The field-name token labels the corresponding field-value as having the
1184   semantics defined by that header field.  For example, the <x:ref>Date</x:ref>
1185   header field is defined in &header-date; as containing the origination
1186   timestamp for the message in which it appears.
1187</t>
1188
1189<section title="Field Extensibility" anchor="field.extensibility">
1190<t>
1191   HTTP header fields are fully extensible: there is no limit on the
1192   introduction of new field names, each presumably defining new semantics,
1193   nor on the number of header fields used in a given message.  Existing
1194   fields are defined in each part of this specification and in many other
1195   specifications outside the core standard.
1196   New header fields can be introduced without changing the protocol version
1197   if their defined semantics allow them to be safely ignored by recipients
1198   that do not recognize them.
1199</t>
1200<t>
1201   New HTTP header fields &SHOULD; be registered with IANA in the
1202   Message Header Field Registry, as described in &iana-header-registry;.
1203   Unrecognized header fields &MUST; be forwarded by a proxy unless the
1204   field-name is listed in the <x:ref>Connection</x:ref> header field
1205   (<xref target="header.connection"/>) or the proxy is specifically
1206   configured to block or otherwise transform such fields.
1207   Unrecognized header fields &SHOULD; be ignored by other recipients.
1208</t>
1209</section>
1210
1211<section title="Field Order" anchor="field.order">
1212<t>
1213   The order in which header fields with differing field names are
1214   received is not significant. However, it is "good practice" to send
1215   header fields that contain control data first, such as <x:ref>Host</x:ref>
1216   on requests and <x:ref>Date</x:ref> on responses, so that implementations
1217   can decide when not to handle a message as early as possible.  A server
1218   &MUST; wait until the entire header section is received before interpreting
1219   a request message, since later header fields might include conditionals,
1220   authentication credentials, or deliberately misleading duplicate
1221   header fields that would impact request processing.
1222</t>
1223<t>
1224   Multiple header fields with the same field name &MUST-NOT; be
1225   sent in a message unless the entire field value for that
1226   header field is defined as a comma-separated list [i.e., #(values)].
1227</t>
1228<t>
1229   Multiple header fields with the same field name can be combined into
1230   one "field-name: field-value" pair, without changing the semantics of the
1231   message, by appending each subsequent field value to the combined
1232   field value in order, separated by a comma. The order in which
1233   header fields with the same field name are received is therefore
1234   significant to the interpretation of the combined field value;
1235   a proxy &MUST-NOT; change the order of these field values when
1236   forwarding a message.
1237</t>
1238<x:note>
1239  <t>
1240   &Note; In practice, the "Set-Cookie" header field (<xref target="RFC6265"/>)
1241   often appears multiple times in a response message and does not use the
1242   list syntax, violating the above requirements on multiple header fields
1243   with the same name. Since it cannot be combined into a single field-value,
1244   recipients ought to handle "Set-Cookie" as a special case while processing
1245   header fields. (See Appendix A.2.3 of <xref target="Kri2001"/> for details.)
1246  </t>
1247</x:note>
1248</section>
1249
1250<section title="Whitespace" anchor="whitespace">
1251<t anchor="rule.LWS">
1252   This specification uses three rules to denote the use of linear
1253   whitespace: OWS (optional whitespace), RWS (required whitespace), and
1254   BWS ("bad" whitespace).
1255</t>
1256<t anchor="rule.OWS">
1257   The OWS rule is used where zero or more linear whitespace octets might
1258   appear. OWS &SHOULD; either not be generated or be generated as a single
1259   SP. Multiple OWS octets that occur within field-content &SHOULD; either
1260   be replaced with a single SP or transformed to all SP octets (each
1261   octet other than SP replaced with SP) before interpreting the field value
1262   or forwarding the message downstream.
1263</t>
1264<t anchor="rule.RWS">
1265   RWS is used when at least one linear whitespace octet is required to
1266   separate field tokens. RWS &SHOULD; be generated as a single SP.
1267   Multiple RWS octets that occur within field-content &SHOULD; either
1268   be replaced with a single SP or transformed to all SP octets before
1269   interpreting the field value or forwarding the message downstream.
1270</t>
1271<t anchor="rule.BWS">
1272   BWS is used where the grammar allows optional whitespace, for historical
1273   reasons, but senders &SHOULD-NOT; generate it in messages;
1274   recipients &MUST; accept such bad optional whitespace and remove it before
1275   interpreting the field value or forwarding the message downstream.
1276</t>
1277<t anchor="rule.whitespace">
1278  <x:anchor-alias value="BWS"/>
1279  <x:anchor-alias value="OWS"/>
1280  <x:anchor-alias value="RWS"/>
1281</t>
1282<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"/>
1283  <x:ref>OWS</x:ref>            = *( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1284                 ; optional whitespace
1285  <x:ref>RWS</x:ref>            = 1*( <x:ref>SP</x:ref> / <x:ref>HTAB</x:ref> )
1286                 ; required whitespace
1287  <x:ref>BWS</x:ref>            = <x:ref>OWS</x:ref>
1288                 ; "bad" whitespace
1289</artwork></figure>
1290</section>
1291
1292<section title="Field Parsing" anchor="field.parsing">
1293<t>
1294   No whitespace is allowed between the header field-name and colon.
1295   In the past, differences in the handling of such whitespace have led to
1296   security vulnerabilities in request routing and response handling.
1297   Any received request message that contains whitespace between a header
1298   field-name and colon &MUST; be rejected with a response code of 400
1299   (Bad Request).  A proxy &MUST; remove any such whitespace from a response
1300   message before forwarding the message downstream.
1301</t>
1302<t>
1303   A field value is preceded by optional whitespace (OWS); a single SP is
1304   preferred. The field value does not include any leading or trailing white
1305   space: OWS occurring before the first non-whitespace octet of the
1306   field value or after the last non-whitespace octet of the field value
1307   is ignored and &SHOULD; be removed before further processing (as this does
1308   not change the meaning of the header field).
1309</t>
1310<t>
1311   Historically, HTTP header field values could be extended over multiple
1312   lines by preceding each extra line with at least one space or horizontal
1313   tab (obs-fold). This specification deprecates such line
1314   folding except within the message/http media type
1315   (<xref target="internet.media.type.message.http"/>).
1316   HTTP senders &MUST-NOT; generate messages that include line folding
1317   (i.e., that contain any field-value that matches the obs-fold rule) unless
1318   the message is intended for packaging within the message/http media type.
1319   HTTP recipients &SHOULD; accept line folding and replace any embedded
1320   obs-fold whitespace with either a single SP or a matching number of SP
1321   octets (to avoid buffer copying) prior to interpreting the field value or
1322   forwarding the message downstream.
1323</t>
1324<t>
1325   Historically, HTTP has allowed field content with text in the ISO-8859-1
1326   <xref target="ISO-8859-1"/> charset, supporting other charsets only
1327   through use of <xref target="RFC2047"/> encoding.
1328   In practice, most HTTP header field values use only a subset of the
1329   US-ASCII charset <xref target="USASCII"/>. Newly defined
1330   header fields &SHOULD; limit their field values to US-ASCII octets.
1331   Recipients &SHOULD; treat other octets in field content (obs-text) as
1332   opaque data.
1333</t>
1334</section>
1335
1336<section title="Field Limits" anchor="field.limits">
1337<t>
1338   HTTP does not place a pre-defined limit on the length of each header field
1339   or on the length of the header block as a whole.  Various ad-hoc
1340   limitations on individual header field length are found in practice,
1341   often depending on the specific field semantics.
1342</t>
1343<t>
1344   A server &MUST; be prepared to receive request header fields of unbounded
1345   length and respond with an appropriate <x:ref>4xx (Client Error)</x:ref>
1346   status code if the received header field(s) are larger than the server
1347   wishes to process.
1348</t>
1349<t>
1350   A client &MUST; be prepared to receive response header fields of unbounded
1351   length. A client &MAY; discard or truncate received header fields that are
1352   larger than the client wishes to process if the field semantics are such
1353   that the dropped value(s) can be safely ignored without changing the
1354   response semantics.
1355</t>
1356</section>
1357
1358<section title="Field value components" anchor="field.components">
1359<t anchor="rule.token.separators">
1360  <x:anchor-alias value="tchar"/>
1361  <x:anchor-alias value="token"/>
1362  <x:anchor-alias value="special"/>
1363  <x:anchor-alias value="word"/>
1364   Many HTTP header field values consist of words (token or quoted-string)
1365   separated by whitespace or special characters. These special characters
1366   &MUST; be in a quoted string to be used within a parameter value (as defined
1367   in <xref target="transfer.codings"/>).
1368</t>
1369<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"><!--unused production--></iref>
1370  <x:ref>word</x:ref>           = <x:ref>token</x:ref> / <x:ref>quoted-string</x:ref>
1371
1372  <x:ref>token</x:ref>          = 1*<x:ref>tchar</x:ref>
1373<!--
1374  IMPORTANT: when editing "tchar" make sure that "special" is updated accordingly!!!
1375 -->
1376  <x:ref>tchar</x:ref>          = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*"
1377                 / "+" / "-" / "." / "^" / "_" / "`" / "|" / "~"
1378                 / <x:ref>DIGIT</x:ref> / <x:ref>ALPHA</x:ref>
1379                 ; any <x:ref>VCHAR</x:ref>, except <x:ref>special</x:ref>
1380
1381  <x:ref>special</x:ref>        = "(" / ")" / "&lt;" / ">" / "@" / ","
1382                 / ";" / ":" / "\" / DQUOTE / "/" / "["
1383                 / "]" / "?" / "=" / "{" / "}"
1384</artwork></figure>
1385<t anchor="rule.quoted-string">
1386  <x:anchor-alias value="quoted-string"/>
1387  <x:anchor-alias value="qdtext"/>
1388  <x:anchor-alias value="obs-text"/>
1389   A string of text is parsed as a single word if it is quoted using
1390   double-quote marks.
1391</t>
1392<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"/>
1393  <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>
1394  <x:ref>qdtext</x:ref>         = <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> /%x21 / %x23-5B / %x5D-7E / <x:ref>obs-text</x:ref>
1395  <x:ref>obs-text</x:ref>       = %x80-FF
1396</artwork></figure>
1397<t anchor="rule.quoted-pair">
1398  <x:anchor-alias value="quoted-pair"/>
1399   The backslash octet ("\") can be used as a single-octet
1400   quoting mechanism within quoted-string constructs:
1401</t>
1402<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-pair"/>
1403  <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> )
1404</artwork></figure>
1405<t>
1406   Recipients that process the value of a quoted-string &MUST; handle a
1407   quoted-pair as if it were replaced by the octet following the backslash.
1408</t>
1409<t>
1410   Senders &SHOULD-NOT; generate a quoted-pair in a quoted-string except where
1411   necessary to quote DQUOTE and backslash octets occurring within that string.
1412</t>
1413<t anchor="rule.comment">
1414  <x:anchor-alias value="comment"/>
1415  <x:anchor-alias value="ctext"/>
1416   Comments can be included in some HTTP header fields by surrounding
1417   the comment text with parentheses. Comments are only allowed in
1418   fields containing "comment" as part of their field value definition.
1419</t>
1420<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="comment"/><iref primary="true" item="Grammar" subitem="ctext"/>
1421  <x:ref>comment</x:ref>        = "(" *( <x:ref>ctext</x:ref> / <x:ref>quoted-cpair</x:ref> / <x:ref>comment</x:ref> ) ")"
1422  <x:ref>ctext</x:ref>          = <x:ref>HTAB</x:ref> / <x:ref>SP</x:ref> / %x21-27 / %x2A-5B / %x5D-7E / <x:ref>obs-text</x:ref>
1423</artwork></figure>
1424<t anchor="rule.quoted-cpair">
1425  <x:anchor-alias value="quoted-cpair"/>
1426   The backslash octet ("\") can be used as a single-octet
1427   quoting mechanism within comment constructs:
1428</t>
1429<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="quoted-cpair"/>
1430  <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> )
1431</artwork></figure>
1432<t>
1433   Senders &SHOULD-NOT; escape octets in comments that do not require escaping
1434   (i.e., other than the backslash octet "\" and the parentheses "(" and ")").
1435</t>
1436</section>
1437
1438</section>
1439
1440<section title="Message Body" anchor="message.body">
1441  <x:anchor-alias value="message-body"/>
1442<t>
1443   The message body (if any) of an HTTP message is used to carry the
1444   payload body of that request or response.  The message body is
1445   identical to the payload body unless a transfer coding has been
1446   applied, as described in <xref target="header.transfer-encoding"/>.
1447</t>
1448<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="message-body"/>
1449  <x:ref>message-body</x:ref> = *OCTET
1450</artwork></figure>
1451<t>
1452   The rules for when a message body is allowed in a message differ for
1453   requests and responses.
1454</t>
1455<t>
1456   The presence of a message body in a request is signaled by a
1457   <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1458   field. Request message framing is independent of method semantics,
1459   even if the method does not define any use for a message body.
1460</t>
1461<t>
1462   The presence of a message body in a response depends on both
1463   the request method to which it is responding and the response
1464   status code (<xref target="status.line"/>).
1465   Responses to the HEAD request method never include a message body
1466   because the associated response header fields (e.g.,
1467   <x:ref>Transfer-Encoding</x:ref>, <x:ref>Content-Length</x:ref>, etc.),
1468   if present, indicate only what their values would have been if the request
1469   method had been GET (&HEAD;).
1470   <x:ref>2xx (Successful)</x:ref> responses to CONNECT switch to tunnel
1471   mode instead of having a message body (&CONNECT;).
1472   All <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, and
1473   <x:ref>304 (Not Modified)</x:ref> responses &MUST-NOT; include a message body.
1474   All other responses do include a message body, although the body
1475   &MAY; be of zero length.
1476</t>
1477
1478<section title="Transfer-Encoding" anchor="header.transfer-encoding">
1479  <iref primary="true" item="Transfer-Encoding header field" x:for-anchor=""/>
1480  <iref item="chunked (Coding Format)"/>
1481  <x:anchor-alias value="Transfer-Encoding"/>
1482<t>
1483   The Transfer-Encoding header field lists the transfer coding names
1484   corresponding to the sequence of transfer codings that have been
1485   (or will be) applied to the payload body in order to form the message body.
1486   Transfer codings are defined in <xref target="transfer.codings"/>.
1487</t>
1488<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Transfer-Encoding"/>
1489  <x:ref>Transfer-Encoding</x:ref> = 1#<x:ref>transfer-coding</x:ref>
1490</artwork></figure>
1491<t>
1492   Transfer-Encoding is analogous to the Content-Transfer-Encoding field of
1493   MIME, which was designed to enable safe transport of binary data over a
1494   7-bit transport service (<xref target="RFC2045" x:fmt="," x:sec="6"/>).
1495   However, safe transport has a different focus for an 8bit-clean transfer
1496   protocol. In HTTP's case, Transfer-Encoding is primarily intended to
1497   accurately delimit a dynamically generated payload and to distinguish
1498   payload encodings that are only applied for transport efficiency or
1499   security from those that are characteristics of the selected resource.
1500</t>
1501<t>
1502   All HTTP/1.1 recipients &MUST; implement the chunked transfer coding
1503   (<xref target="chunked.encoding"/>) because it plays a crucial role in
1504   framing messages when the payload body size is not known in advance.
1505   If chunked is applied to a payload body, the sender &MUST-NOT; apply
1506   chunked more than once (i.e., chunking an already chunked message is not
1507   allowed).
1508   If any transfer coding is applied to a request payload body, the
1509   sender &MUST; apply chunked as the final transfer coding to ensure that
1510   the message is properly framed.
1511   If any transfer coding is applied to a response payload body, the
1512   sender &MUST; either apply chunked as the final transfer coding or
1513   terminate the message by closing the connection.
1514</t>
1515<figure><preamble>
1516   For example,
1517</preamble><artwork type="example">
1518  Transfer-Encoding: gzip, chunked
1519</artwork><postamble>
1520   indicates that the payload body has been compressed using the gzip
1521   coding and then chunked using the chunked coding while forming the
1522   message body.
1523</postamble></figure>
1524<t>
1525   Unlike <x:ref>Content-Encoding</x:ref> (&content-codings;),
1526   Transfer-Encoding is a property of the message, not of the payload, and
1527   any recipient along the request/response chain &MAY; decode the received
1528   transfer coding(s) or apply additional transfer coding(s) to the message
1529   body, assuming that corresponding changes are made to the Transfer-Encoding
1530   field-value. Additional information about the encoding parameters &MAY; be
1531   provided by other header fields not defined by this specification.
1532</t>
1533<t>
1534   Transfer-Encoding &MAY; be sent in a response to a HEAD request or in a
1535   <x:ref>304 (Not Modified)</x:ref> response (&status-304;) to a GET request,
1536   neither of which includes a message body,
1537   to indicate that the origin server would have applied a transfer coding
1538   to the message body if the request had been an unconditional GET.
1539   This indication is not required, however, because any recipient on
1540   the response chain (including the origin server) can remove transfer
1541   codings when they are not needed.
1542</t>
1543<t>
1544   Transfer-Encoding was added in HTTP/1.1.  It is generally assumed that
1545   implementations advertising only HTTP/1.0 support will not understand
1546   how to process a transfer-encoded payload.
1547   A client &MUST-NOT; send a request containing Transfer-Encoding unless it
1548   knows the server will handle HTTP/1.1 (or later) requests; such knowledge
1549   might be in the form of specific user configuration or by remembering the
1550   version of a prior received response.
1551   A server &MUST-NOT; send a response containing Transfer-Encoding unless
1552   the corresponding request indicates HTTP/1.1 (or later).
1553</t>
1554<t>
1555   A server that receives a request message with a transfer coding it does
1556   not understand &SHOULD; respond with <x:ref>501 (Not Implemented)</x:ref>.
1557</t>
1558</section>
1559
1560<section title="Content-Length" anchor="header.content-length">
1561  <iref primary="true" item="Content-Length header field" x:for-anchor=""/>
1562  <x:anchor-alias value="Content-Length"/>
1563<t>
1564   When a message is allowed to contain a message body, does not have a
1565   <x:ref>Transfer-Encoding</x:ref> header field, and has a payload body
1566   length that is known to the sender before the message header section has
1567   been sent, the sender &SHOULD; send a Content-Length header field to
1568   indicate the length of the payload body as a decimal number of octets.
1569</t>
1570<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Content-Length"/>
1571  <x:ref>Content-Length</x:ref> = 1*<x:ref>DIGIT</x:ref>
1572</artwork></figure>
1573<t>
1574   An example is
1575</t>
1576<figure><artwork type="example">
1577  Content-Length: 3495
1578</artwork></figure>
1579<t>
1580   A sender &MUST-NOT; send a Content-Length header field in any message that
1581   contains a <x:ref>Transfer-Encoding</x:ref> header field.
1582</t>
1583<t>
1584   A server &MAY; send a Content-Length header field in a response to a HEAD
1585   request (&HEAD;); a server &MUST-NOT; send Content-Length in such a
1586   response unless its field-value equals the decimal number of octets that
1587   would have been sent in the payload body of a response if the same
1588   request had used the GET method.
1589</t>
1590<t>
1591   A server &MAY; send a Content-Length header field in a
1592   <x:ref>304 (Not Modified)</x:ref> response to a conditional GET request
1593   (&status-304;); a server &MUST-NOT; send Content-Length in such a
1594   response unless its field-value equals the decimal number of octets that
1595   would have been sent in the payload body of a <x:ref>200 (OK)</x:ref>
1596   response to the same request.
1597</t>
1598<t>
1599   A server &MUST-NOT; send a Content-Length header field in any response
1600   with a status code of
1601   <x:ref>1xx (Informational)</x:ref> or <x:ref>204 (No Content)</x:ref>.
1602   A server &SHOULD-NOT; send a Content-Length header field in any
1603   <x:ref>2xx (Successful)</x:ref> response to a CONNECT request (&CONNECT;).
1604</t>
1605<t>
1606   Any Content-Length field value greater than or equal to zero is valid.
1607   Since there is no predefined limit to the length of an HTTP payload,
1608   recipients &SHOULD; anticipate potentially large decimal numerals and
1609   prevent parsing errors due to integer conversion overflows
1610   (<xref target="attack.protocol.element.size.overflows"/>).
1611</t>
1612<t>
1613   If a message is received that has multiple Content-Length header fields
1614   with field-values consisting of the same decimal value, or a single
1615   Content-Length header field with a field value containing a list of
1616   identical decimal values (e.g., "Content-Length: 42, 42"), indicating that
1617   duplicate Content-Length header fields have been generated or combined by an
1618   upstream message processor, then the recipient &MUST; either reject the
1619   message as invalid or replace the duplicated field-values with a single
1620   valid Content-Length field containing that decimal value prior to
1621   determining the message body length.
1622</t>
1623<x:note>
1624  <t>
1625   &Note; HTTP's use of Content-Length for message framing differs
1626   significantly from the same field's use in MIME, where it is an optional
1627   field used only within the "message/external-body" media-type.
1628  </t>
1629</x:note>
1630</section>
1631
1632<section title="Message Body Length" anchor="message.body.length">
1633  <iref item="chunked (Coding Format)"/>
1634<t>
1635   The length of a message body is determined by one of the following
1636   (in order of precedence):
1637</t>
1638<t>
1639  <list style="numbers">
1640    <x:lt><t>
1641     Any response to a HEAD request and any response with a
1642     <x:ref>1xx (Informational)</x:ref>, <x:ref>204 (No Content)</x:ref>, or
1643     <x:ref>304 (Not Modified)</x:ref> status code is always
1644     terminated by the first empty line after the header fields, regardless of
1645     the header fields present in the message, and thus cannot contain a
1646     message body.
1647    </t></x:lt>
1648    <x:lt><t>
1649     Any <x:ref>2xx (Successful)</x:ref> response to a CONNECT request implies that the
1650     connection will become a tunnel immediately after the empty line that
1651     concludes the header fields.  A client &MUST; ignore any
1652     <x:ref>Content-Length</x:ref> or <x:ref>Transfer-Encoding</x:ref> header
1653     fields received in such a message.
1654    </t></x:lt>
1655    <x:lt><t>
1656     If a <x:ref>Transfer-Encoding</x:ref> header field is present
1657     and the chunked transfer coding (<xref target="chunked.encoding"/>)
1658     is the final encoding, the message body length is determined by reading
1659     and decoding the chunked data until the transfer coding indicates the
1660     data is complete.
1661    </t>
1662    <t>
1663     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a
1664     response and the chunked transfer coding is not the final encoding, the
1665     message body length is determined by reading the connection until it is
1666     closed by the server.
1667     If a <x:ref>Transfer-Encoding</x:ref> header field is present in a request and the
1668     chunked transfer coding is not the final encoding, the message body
1669     length cannot be determined reliably; the server &MUST; respond with
1670     the <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1671    </t>
1672    <t>
1673     If a message is received with both a <x:ref>Transfer-Encoding</x:ref>
1674     and a <x:ref>Content-Length</x:ref> header field, the Transfer-Encoding
1675     overrides the Content-Length. Such a message might indicate an attempt
1676     to perform request or response smuggling (bypass of security-related
1677     checks on message routing or content) and thus ought to be handled as
1678     an error.  A sender &MUST; remove the received Content-Length field
1679     prior to forwarding such a message downstream.
1680    </t></x:lt>
1681    <x:lt><t>
1682     If a message is received without <x:ref>Transfer-Encoding</x:ref> and with
1683     either multiple <x:ref>Content-Length</x:ref> header fields having
1684     differing field-values or a single Content-Length header field having an
1685     invalid value, then the message framing is invalid and &MUST; be treated
1686     as an error to prevent request or response smuggling.
1687     If this is a request message, the server &MUST; respond with
1688     a <x:ref>400 (Bad Request)</x:ref> status code and then close the connection.
1689     If this is a response message received by a proxy, the proxy
1690     &MUST; discard the received response, send a <x:ref>502 (Bad Gateway)</x:ref>
1691     status code as its downstream response, and then close the connection.
1692     If this is a response message received by a user agent, it &MUST; be
1693     treated as an error by discarding the message and closing the connection.
1694    </t></x:lt>
1695    <x:lt><t>
1696     If a valid <x:ref>Content-Length</x:ref> header field is present without
1697     <x:ref>Transfer-Encoding</x:ref>, its decimal value defines the
1698     message body length in octets.  If the actual number of octets sent in
1699     the message is less than the indicated Content-Length, the recipient
1700     &MUST; consider the message to be incomplete and treat the connection
1701     as no longer usable.
1702     If the actual number of octets sent in the message is more than the indicated
1703     Content-Length, the recipient &MUST; only process the message body up to the
1704     field value's number of octets; the remainder of the message &MUST; either
1705     be discarded or treated as the next message in a pipeline.  For the sake of
1706     robustness, a user agent &MAY; attempt to detect and correct such an error
1707     in message framing if it is parsing the response to the last request on
1708     a connection and the connection has been closed by the server.
1709    </t></x:lt>
1710    <x:lt><t>
1711     If this is a request message and none of the above are true, then the
1712     message body length is zero (no message body is present).
1713    </t></x:lt>
1714    <x:lt><t>
1715     Otherwise, this is a response message without a declared message body
1716     length, so the message body length is determined by the number of octets
1717     received prior to the server closing the connection.
1718    </t></x:lt>
1719  </list>
1720</t>
1721<t>
1722   Since there is no way to distinguish a successfully completed,
1723   close-delimited message from a partially-received message interrupted
1724   by network failure, a server &SHOULD; use encoding or
1725   length-delimited messages whenever possible.  The close-delimiting
1726   feature exists primarily for backwards compatibility with HTTP/1.0.
1727</t>
1728<t>
1729   A server &MAY; reject a request that contains a message body but
1730   not a <x:ref>Content-Length</x:ref> by responding with
1731   <x:ref>411 (Length Required)</x:ref>.
1732</t>
1733<t>
1734   Unless a transfer coding other than chunked has been applied,
1735   a client that sends a request containing a message body &SHOULD;
1736   use a valid <x:ref>Content-Length</x:ref> header field if the message body
1737   length is known in advance, rather than the chunked transfer coding, since some
1738   existing services respond to chunked with a <x:ref>411 (Length Required)</x:ref>
1739   status code even though they understand the chunked transfer coding.  This
1740   is typically because such services are implemented via a gateway that
1741   requires a content-length in advance of being called and the server
1742   is unable or unwilling to buffer the entire request before processing.
1743</t>
1744<t>
1745   A client that sends a request containing a message body &MUST; include a
1746   valid <x:ref>Content-Length</x:ref> header field if it does not know the
1747   server will handle HTTP/1.1 (or later) requests; such knowledge can be in
1748   the form of specific user configuration or by remembering the version of a
1749   prior received response.
1750</t>
1751</section>
1752</section>
1753
1754<section anchor="incomplete.messages" title="Handling Incomplete Messages">
1755<t>
1756   A server that receives an incomplete request message, usually due to a
1757   canceled request or a triggered time-out exception, &MAY; send an error
1758   response prior to closing the connection.
1759</t>
1760<t>
1761   A client that receives an incomplete response message, which can occur
1762   when a connection is closed prematurely or when decoding a supposedly
1763   chunked transfer coding fails, &MUST; record the message as incomplete.
1764   Cache requirements for incomplete responses are defined in
1765   &cache-incomplete;.
1766</t>
1767<t>
1768   If a response terminates in the middle of the header block (before the
1769   empty line is received) and the status code might rely on header fields to
1770   convey the full meaning of the response, then the client cannot assume
1771   that meaning has been conveyed; the client might need to repeat the
1772   request in order to determine what action to take next.
1773</t>
1774<t>
1775   A message body that uses the chunked transfer coding is
1776   incomplete if the zero-sized chunk that terminates the encoding has not
1777   been received.  A message that uses a valid <x:ref>Content-Length</x:ref> is
1778   incomplete if the size of the message body received (in octets) is less than
1779   the value given by Content-Length.  A response that has neither chunked
1780   transfer coding nor Content-Length is terminated by closure of the
1781   connection, and thus is considered complete regardless of the number of
1782   message body octets received, provided that the header block was received
1783   intact.
1784</t>
1785</section>
1786
1787<section title="Message Parsing Robustness" anchor="message.robustness">
1788<t>
1789   Older HTTP/1.0 user agent implementations might send an extra CRLF
1790   after a POST request as a lame workaround for some early server
1791   applications that failed to read message body content that was
1792   not terminated by a line-ending. An HTTP/1.1 user agent &MUST-NOT;
1793   preface or follow a request with an extra CRLF.  If terminating
1794   the request message body with a line-ending is desired, then the
1795   user agent &MUST; include the terminating CRLF octets as part of the
1796   message body length.
1797</t>
1798<t>
1799   In the interest of robustness, servers &SHOULD; ignore at least one
1800   empty line received where a request-line is expected. In other words, if
1801   a server is reading the protocol stream at the beginning of a
1802   message and receives a CRLF first, the server &SHOULD; ignore the CRLF.
1803</t>
1804<t>
1805   Although the line terminator for the start-line and header
1806   fields is the sequence CRLF, recipients &MAY; recognize a
1807   single LF as a line terminator and ignore any preceding CR.
1808</t>
1809<t>
1810   Although the request-line and status-line grammar rules require that each
1811   of the component elements be separated by a single SP octet, recipients
1812   &MAY; instead parse on whitespace-delimited word boundaries and, aside
1813   from the CRLF terminator, treat any form of whitespace as the SP separator
1814   while ignoring preceding or trailing whitespace;
1815   such whitespace includes one or more of the following octets:
1816   SP, HTAB, VT (%x0B), FF (%x0C), or bare CR.
1817</t>
1818<t>
1819   When a server listening only for HTTP request messages, or processing
1820   what appears from the start-line to be an HTTP request message,
1821   receives a sequence of octets that does not match the HTTP-message
1822   grammar aside from the robustness exceptions listed above, the
1823   server &SHOULD; respond with a <x:ref>400 (Bad Request)</x:ref> response. 
1824</t>
1825</section>
1826</section>
1827
1828<section title="Transfer Codings" anchor="transfer.codings">
1829  <x:anchor-alias value="transfer-coding"/>
1830  <x:anchor-alias value="transfer-extension"/>
1831<t>
1832   Transfer coding names are used to indicate an encoding
1833   transformation that has been, can be, or might need to be applied to a
1834   payload body in order to ensure "safe transport" through the network.
1835   This differs from a content coding in that the transfer coding is a
1836   property of the message rather than a property of the representation
1837   that is being transferred.
1838</t>
1839<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="transfer-coding"/><iref primary="true" item="Grammar" subitem="transfer-extension"/>
1840  <x:ref>transfer-coding</x:ref>    = "chunked" ; <xref target="chunked.encoding"/>
1841                     / "compress" ; <xref target="compress.coding"/>
1842                     / "deflate" ; <xref target="deflate.coding"/>
1843                     / "gzip" ; <xref target="gzip.coding"/>
1844                     / <x:ref>transfer-extension</x:ref>
1845  <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> )
1846</artwork></figure>
1847<t anchor="rule.parameter">
1848  <x:anchor-alias value="attribute"/>
1849  <x:anchor-alias value="transfer-parameter"/>
1850  <x:anchor-alias value="value"/>
1851   Parameters are in the form of attribute/value pairs.
1852</t>
1853<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"/>
1854  <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>
1855  <x:ref>attribute</x:ref>          = <x:ref>token</x:ref>
1856  <x:ref>value</x:ref>              = <x:ref>word</x:ref>
1857</artwork></figure>
1858<t>
1859   All transfer-coding names are case-insensitive and &SHOULD; be registered
1860   within the HTTP Transfer Coding registry, as defined in
1861   <xref target="transfer.coding.registry"/>.
1862   They are used in the <x:ref>TE</x:ref> (<xref target="header.te"/>) and
1863   <x:ref>Transfer-Encoding</x:ref> (<xref target="header.transfer-encoding"/>)
1864   header fields.
1865</t>
1866
1867<section title="Chunked Transfer Coding" anchor="chunked.encoding">
1868  <iref primary="true" item="chunked (Coding Format)"/>
1869  <x:anchor-alias value="chunk"/>
1870  <x:anchor-alias value="chunked-body"/>
1871  <x:anchor-alias value="chunk-data"/>
1872  <x:anchor-alias value="chunk-ext"/>
1873  <x:anchor-alias value="chunk-ext-name"/>
1874  <x:anchor-alias value="chunk-ext-val"/>
1875  <x:anchor-alias value="chunk-size"/>
1876  <x:anchor-alias value="last-chunk"/>
1877  <x:anchor-alias value="trailer-part"/>
1878  <x:anchor-alias value="quoted-str-nf"/>
1879  <x:anchor-alias value="qdtext-nf"/>
1880<t>
1881   The chunked transfer coding modifies the body of a message in order to
1882   transfer it as a series of chunks, each with its own size indicator,
1883   followed by an &OPTIONAL; trailer containing header fields. This
1884   allows dynamically generated content to be transferred along with the
1885   information necessary for the recipient to verify that it has
1886   received the full message.
1887</t>
1888<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="chunked-body"><!--terminal production--></iref><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"/>
1889  <x:ref>chunked-body</x:ref>   = *<x:ref>chunk</x:ref>
1890                   <x:ref>last-chunk</x:ref>
1891                   <x:ref>trailer-part</x:ref>
1892                   <x:ref>CRLF</x:ref>
1893 
1894  <x:ref>chunk</x:ref>          = <x:ref>chunk-size</x:ref> [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1895                   <x:ref>chunk-data</x:ref> <x:ref>CRLF</x:ref>
1896  <x:ref>chunk-size</x:ref>     = 1*<x:ref>HEXDIG</x:ref>
1897  <x:ref>last-chunk</x:ref>     = 1*("0") [ <x:ref>chunk-ext</x:ref> ] <x:ref>CRLF</x:ref>
1898 
1899  <x:ref>chunk-ext</x:ref>      = *( ";" <x:ref>chunk-ext-name</x:ref> [ "=" <x:ref>chunk-ext-val</x:ref> ] )
1900  <x:ref>chunk-ext-name</x:ref> = <x:ref>token</x:ref>
1901  <x:ref>chunk-ext-val</x:ref>  = <x:ref>token</x:ref> / <x:ref>quoted-str-nf</x:ref>
1902  <x:ref>chunk-data</x:ref>     = 1*<x:ref>OCTET</x:ref> ; a sequence of chunk-size octets
1903  <x:ref>trailer-part</x:ref>   = *( <x:ref>header-field</x:ref> <x:ref>CRLF</x:ref> )
1904 
1905  <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>
1906                 ; like <x:ref>quoted-string</x:ref>, but disallowing line folding
1907  <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>
1908</artwork></figure>
1909<t>
1910   Chunk extensions within the chunked transfer coding are deprecated.
1911   Senders &SHOULD-NOT; send chunk-ext.
1912   Definition of new chunk extensions is discouraged.
1913</t>
1914<t>
1915   The chunk-size field is a string of hex digits indicating the size of
1916   the chunk-data in octets. The chunked transfer coding is complete when a
1917   chunk with a chunk-size of zero is received, possibly followed by a
1918   trailer, and finally terminated by an empty line.
1919</t>
1920
1921<section title="Trailer" anchor="header.trailer">
1922  <iref primary="true" item="Trailer header field" x:for-anchor=""/>
1923  <x:anchor-alias value="Trailer"/>
1924<t>
1925   A trailer allows the sender to include additional fields at the end of a
1926   chunked message in order to supply metadata that might be dynamically
1927   generated while the message body is sent, such as a message integrity
1928   check, digital signature, or post-processing status.
1929   The trailer &MUST-NOT; contain fields that need to be known before a
1930   recipient processes the body, such as <x:ref>Transfer-Encoding</x:ref>,
1931   <x:ref>Content-Length</x:ref>, and <x:ref>Trailer</x:ref>.
1932</t>
1933<t>
1934   When a message includes a message body encoded with the chunked
1935   transfer coding and the sender desires to send metadata in the form of
1936   trailer fields at the end of the message, the sender &SHOULD; send a
1937   <x:ref>Trailer</x:ref> header field before the message body to indicate
1938   which fields will be present in the trailers. This allows the recipient
1939   to prepare for receipt of that metadata before it starts processing the body,
1940   which is useful if the message is being streamed and the recipient wishes
1941   to confirm an integrity check on the fly.
1942</t>
1943<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Trailer"/>
1944  <x:ref>Trailer</x:ref> = 1#<x:ref>field-name</x:ref>
1945</artwork></figure>
1946<t>
1947   If no <x:ref>Trailer</x:ref> header field is present, the sender of a
1948   chunked message body &SHOULD; send an empty trailer.
1949</t>
1950<t>
1951   A server &MUST; send an empty trailer with the chunked transfer coding
1952   unless at least one of the following is true:
1953  <list style="numbers">
1954    <t>the request included a <x:ref>TE</x:ref> header field that indicates
1955    "trailers" is acceptable in the transfer coding of the response, as
1956    described in <xref target="header.te"/>; or,</t>
1957     
1958    <t>the trailer fields consist entirely of optional metadata and the
1959    recipient could use the message (in a manner acceptable to the server where
1960    the field originated) without receiving that metadata. In other words,
1961    the server that generated the header field is willing to accept the
1962    possibility that the trailer fields might be silently discarded along
1963    the path to the client.</t>
1964  </list>
1965</t>
1966<t>
1967   The above requirement prevents the need for an infinite buffer when a
1968   message is being received by an HTTP/1.1 (or later) proxy and forwarded to
1969   an HTTP/1.0 recipient.
1970</t>
1971</section>
1972
1973<section title="Decoding chunked" anchor="decoding.chunked">
1974<t>
1975   A process for decoding the chunked transfer coding
1976   can be represented in pseudo-code as:
1977</t>
1978<figure><artwork type="code">
1979  length := 0
1980  read chunk-size, chunk-ext (if any) and CRLF
1981  while (chunk-size &gt; 0) {
1982     read chunk-data and CRLF
1983     append chunk-data to decoded-body
1984     length := length + chunk-size
1985     read chunk-size and CRLF
1986  }
1987  read header-field
1988  while (header-field not empty) {
1989     append header-field to existing header fields
1990     read header-field
1991  }
1992  Content-Length := length
1993  Remove "chunked" from Transfer-Encoding
1994  Remove Trailer from existing header fields
1995</artwork></figure>
1996<t>
1997   All recipients &MUST; be able to receive and decode the
1998   chunked transfer coding and &MUST; ignore chunk-ext extensions
1999   they do not understand.
2000</t>
2001</section>
2002</section>
2003
2004<section title="Compression Codings" anchor="compression.codings">
2005<t>
2006   The codings defined below can be used to compress the payload of a
2007   message.
2008</t>
2009
2010<section title="Compress Coding" anchor="compress.coding">
2011<iref item="compress (Coding Format)"/>
2012<t>
2013   The "compress" format is produced by the common UNIX file compression
2014   program "compress". This format is an adaptive Lempel-Ziv-Welch
2015   coding (LZW). Recipients &SHOULD; consider "x-compress" to be
2016   equivalent to "compress".
2017</t>
2018</section>
2019
2020<section title="Deflate Coding" anchor="deflate.coding">
2021<iref item="deflate (Coding Format)"/>
2022<t>
2023   The "deflate" format is defined as the "deflate" compression mechanism
2024   (described in <xref target="RFC1951"/>) used inside the "zlib"
2025   data format (<xref target="RFC1950"/>).
2026</t>
2027<x:note>
2028  <t>
2029    &Note; Some incorrect implementations send the "deflate"
2030    compressed data without the zlib wrapper.
2031   </t>
2032</x:note>
2033</section>
2034
2035<section title="Gzip Coding" anchor="gzip.coding">
2036<iref item="gzip (Coding Format)"/>
2037<t>
2038   The "gzip" format is produced by the file compression program
2039   "gzip" (GNU zip), as described in <xref target="RFC1952"/>. This format is a
2040   Lempel-Ziv coding (LZ77) with a 32 bit CRC.
2041   Recipients &SHOULD; consider "x-gzip" to be equivalent to "gzip".
2042</t>
2043</section>
2044
2045</section>
2046
2047<section title="TE" anchor="header.te">
2048  <iref primary="true" item="TE header field" x:for-anchor=""/>
2049  <x:anchor-alias value="TE"/>
2050  <x:anchor-alias value="t-codings"/>
2051  <x:anchor-alias value="t-ranking"/>
2052  <x:anchor-alias value="rank"/>
2053<t>
2054   The "TE" header field in a request indicates what transfer codings,
2055   besides chunked, the client is willing to accept in response, and
2056   whether or not the client is willing to accept trailer fields in a
2057   chunked transfer coding.
2058</t>
2059<t>
2060   The TE field-value consists of a comma-separated list of transfer coding
2061   names, each allowing for optional parameters (as described in
2062   <xref target="transfer.codings"/>), and/or the keyword "trailers".
2063   Clients &MUST-NOT; send the chunked transfer coding name in TE;
2064   chunked is always acceptable for HTTP/1.1 recipients.
2065</t>
2066<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="t-ranking"/><iref primary="true" item="Grammar" subitem="rank"/>
2067  <x:ref>TE</x:ref>        = #<x:ref>t-codings</x:ref>
2068  <x:ref>t-codings</x:ref> = "trailers" / ( <x:ref>transfer-coding</x:ref> [ <x:ref>t-ranking</x:ref> ] )
2069  <x:ref>t-ranking</x:ref> = <x:ref>OWS</x:ref> ";" <x:ref>OWS</x:ref> "q=" <x:ref>rank</x:ref>
2070  <x:ref>rank</x:ref>      = ( "0" [ "." 0*3<x:ref>DIGIT</x:ref> ] )
2071             / ( "1" [ "." 0*3("0") ] )
2072</artwork></figure>
2073<t>
2074   Three examples of TE use are below.
2075</t>
2076<figure><artwork type="example">
2077  TE: deflate
2078  TE:
2079  TE: trailers, deflate;q=0.5
2080</artwork></figure>
2081<t>
2082   The presence of the keyword "trailers" indicates that the client is
2083   willing to accept trailer fields in a chunked transfer coding,
2084   as defined in <xref target="chunked.encoding"/>, on behalf of itself and
2085   any downstream clients. For chained requests, this implies that either:
2086   (a) all downstream clients are willing to accept trailer fields in the
2087   forwarded response; or,
2088   (b) the client will attempt to buffer the response on behalf of downstream
2089   recipients.
2090   Note that HTTP/1.1 does not define any means to limit the size of a
2091   chunked response such that a client can be assured of buffering the
2092   entire response.
2093</t>
2094<t>
2095   When multiple transfer codings are acceptable, the client &MAY; rank the
2096   codings by preference using a case-insensitive "q" parameter (similar to
2097   the qvalues used in content negotiation fields, &qvalue;). The rank value
2098   is a real number in the range 0 through 1, where 0.001 is the least
2099   preferred and 1 is the most preferred; a value of 0 means "not acceptable".
2100</t>
2101<t>
2102   If the TE field-value is empty or if no TE field is present, the only
2103   acceptable transfer coding is chunked. A message with no transfer coding
2104   is always acceptable.
2105</t>
2106<t>
2107   Since the TE header field only applies to the immediate connection,
2108   a sender of TE &MUST; also send a "TE" connection option within the
2109   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
2110   in order to prevent the TE field from being forwarded by intermediaries
2111   that do not support its semantics.
2112</t>
2113</section>
2114</section>
2115
2116<section title="Message Routing" anchor="message.routing">
2117<t>
2118   HTTP request message routing is determined by each client based on the
2119   target resource, the client's proxy configuration, and
2120   establishment or reuse of an inbound connection.  The corresponding
2121   response routing follows the same connection chain back to the client.
2122</t>
2123
2124<section title="Identifying a Target Resource" anchor="target-resource">
2125  <iref primary="true" item="target resource"/>
2126  <iref primary="true" item="target URI"/>
2127  <x:anchor-alias value="target resource"/>
2128  <x:anchor-alias value="target URI"/>
2129<t>
2130   HTTP is used in a wide variety of applications, ranging from
2131   general-purpose computers to home appliances.  In some cases,
2132   communication options are hard-coded in a client's configuration.
2133   However, most HTTP clients rely on the same resource identification
2134   mechanism and configuration techniques as general-purpose Web browsers.
2135</t>
2136<t>
2137   HTTP communication is initiated by a user agent for some purpose.
2138   The purpose is a combination of request semantics, which are defined in
2139   <xref target="Part2"/>, and a target resource upon which to apply those
2140   semantics.  A URI reference (<xref target="uri"/>) is typically used as
2141   an identifier for the "<x:dfn>target resource</x:dfn>", which a user agent
2142   would resolve to its absolute form in order to obtain the
2143   "<x:dfn>target URI</x:dfn>".  The target URI
2144   excludes the reference's fragment identifier component, if any,
2145   since fragment identifiers are reserved for client-side processing
2146   (<xref target="RFC3986" x:fmt="," x:sec="3.5"/>).
2147</t>
2148</section>
2149
2150<section title="Connecting Inbound" anchor="connecting.inbound">
2151<t>
2152   Once the target URI is determined, a client needs to decide whether
2153   a network request is necessary to accomplish the desired semantics and,
2154   if so, where that request is to be directed.
2155</t>
2156<t>
2157   If the client has a response cache and the request semantics can be
2158   satisfied by a cache (<xref target="Part6"/>), then the request is
2159   usually directed to the cache first.
2160</t>
2161<t>
2162   If the request is not satisfied by a cache, then a typical client will
2163   check its configuration to determine whether a proxy is to be used to
2164   satisfy the request.  Proxy configuration is implementation-dependent,
2165   but is often based on URI prefix matching, selective authority matching,
2166   or both, and the proxy itself is usually identified by an "http" or
2167   "https" URI.  If a proxy is applicable, the client connects inbound by
2168   establishing (or reusing) a connection to that proxy.
2169</t>
2170<t>
2171   If no proxy is applicable, a typical client will invoke a handler routine,
2172   usually specific to the target URI's scheme, to connect directly
2173   to an authority for the target resource.  How that is accomplished is
2174   dependent on the target URI scheme and defined by its associated
2175   specification, similar to how this specification defines origin server
2176   access for resolution of the "http" (<xref target="http.uri"/>) and
2177   "https" (<xref target="https.uri"/>) schemes.
2178</t>
2179<t>
2180   HTTP requirements regarding connection management are defined in
2181   <xref target="connection.management"/>.
2182</t>
2183</section>
2184
2185<section title="Request Target" anchor="request-target">
2186<t>
2187   Once an inbound connection is obtained,
2188   the client sends an HTTP request message (<xref target="http.message"/>)
2189   with a request-target derived from the target URI.
2190   There are four distinct formats for the request-target, depending on both
2191   the method being requested and whether the request is to a proxy.
2192</t>   
2193<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="request-target"/><iref primary="true" item="Grammar" subitem="origin-form"/><iref primary="true" item="Grammar" subitem="absolute-form"/><iref primary="true" item="Grammar" subitem="authority-form"/><iref primary="true" item="Grammar" subitem="asterisk-form"/>
2194  <x:ref>request-target</x:ref> = <x:ref>origin-form</x:ref>
2195                 / <x:ref>absolute-form</x:ref>
2196                 / <x:ref>authority-form</x:ref>
2197                 / <x:ref>asterisk-form</x:ref>
2198
2199  <x:ref>origin-form</x:ref>    = <x:ref>path-absolute</x:ref> [ "?" <x:ref>query</x:ref> ]
2200  <x:ref>absolute-form</x:ref>  = <x:ref>absolute-URI</x:ref>
2201  <x:ref>authority-form</x:ref> = <x:ref>authority</x:ref>
2202  <x:ref>asterisk-form</x:ref>  = "*"
2203</artwork></figure>
2204<t anchor="origin-form"><iref item="origin-form (of request-target)"/>
2205   The most common form of request-target is the origin-form.
2206   When making a request directly to an origin server, other than a CONNECT
2207   or server-wide OPTIONS request (as detailed below),
2208   a client &MUST; send only the absolute path and query components of
2209   the target URI as the request-target.
2210   If the target URI's path component is empty, then the client &MUST; send
2211   "/" as the path within the origin-form of request-target.
2212   A <x:ref>Host</x:ref> header field is also sent, as defined in
2213   <xref target="header.host"/>, containing the target URI's
2214   authority component (excluding any userinfo).
2215</t>
2216<t>
2217   For example, a client wishing to retrieve a representation of the resource
2218   identified as
2219</t>
2220<figure><artwork x:indent-with="  " type="example">
2221http://www.example.org/where?q=now
2222</artwork></figure>
2223<t>
2224   directly from the origin server would open (or reuse) a TCP connection
2225   to port 80 of the host "www.example.org" and send the lines:
2226</t>
2227<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2228GET /where?q=now HTTP/1.1
2229Host: www.example.org
2230</artwork></figure>
2231<t>
2232   followed by the remainder of the request message.
2233</t>
2234<t anchor="absolute-form"><iref item="absolute-form (of request-target)"/>
2235   When making a request to a proxy, other than a CONNECT or server-wide
2236   OPTIONS request (as detailed below), a client &MUST; send the target URI
2237   in absolute-form as the request-target.
2238   The proxy is requested to either service that request from a valid cache,
2239   if possible, or make the same request on the client's behalf to either
2240   the next inbound proxy server or directly to the origin server indicated
2241   by the request-target.  Requirements on such "forwarding" of messages are
2242   defined in <xref target="message.forwarding"/>.
2243</t>
2244<t>
2245   An example absolute-form of request-line would be:
2246</t>
2247<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2248GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1
2249</artwork></figure>
2250<t>
2251   To allow for transition to the absolute-form for all requests in some
2252   future version of HTTP, HTTP/1.1 servers &MUST; accept the absolute-form
2253   in requests, even though HTTP/1.1 clients will only send them in requests
2254   to proxies.
2255</t>
2256<t anchor="authority-form"><iref item="authority-form (of request-target)"/>
2257   The authority-form of request-target is only used for CONNECT requests
2258   (&CONNECT;).  When making a CONNECT request to establish a tunnel through
2259   one or more proxies, a client &MUST; send only the target URI's
2260   authority component (excluding any userinfo) as the request-target.
2261   For example,
2262</t>
2263<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2264CONNECT www.example.com:80 HTTP/1.1
2265</artwork></figure>
2266<t anchor="asterisk-form"><iref item="asterisk-form (of request-target)"/>
2267   The asterisk-form of request-target is only used for a server-wide
2268   OPTIONS request (&OPTIONS;).  When a client wishes to request OPTIONS
2269   for the server as a whole, as opposed to a specific named resource of
2270   that server, the client &MUST; send only "*" (%x2A) as the request-target.
2271   For example,
2272</t>
2273<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2274OPTIONS * HTTP/1.1
2275</artwork></figure>
2276<t>
2277   If a proxy receives an OPTIONS request with an absolute-form of
2278   request-target in which the URI has an empty path and no query component,
2279   then the last proxy on the request chain &MUST; send a request-target
2280   of "*" when it forwards the request to the indicated origin server.
2281</t>
2282<figure><preamble>   
2283   For example, the request
2284</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2285OPTIONS http://www.example.org:8001 HTTP/1.1
2286</artwork></figure>
2287<figure><preamble>   
2288  would be forwarded by the final proxy as
2289</preamble><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2290OPTIONS * HTTP/1.1
2291Host: www.example.org:8001
2292</artwork>
2293<postamble>
2294   after connecting to port 8001 of host "www.example.org".
2295</postamble>
2296</figure>
2297</section>
2298
2299<section title="Host" anchor="header.host">
2300  <iref primary="true" item="Host header field" x:for-anchor=""/>
2301  <x:anchor-alias value="Host"/>
2302<t>
2303   The "Host" header field in a request provides the host and port
2304   information from the target URI, enabling the origin
2305   server to distinguish among resources while servicing requests
2306   for multiple host names on a single IP address.  Since the Host
2307   field-value is critical information for handling a request, it
2308   &SHOULD; be sent as the first header field following the request-line.
2309</t>
2310<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Host"/>
2311  <x:ref>Host</x:ref> = <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ; <xref target="http.uri"/>
2312</artwork></figure>
2313<t>
2314   A client &MUST; send a Host header field in all HTTP/1.1 request
2315   messages.  If the target URI includes an authority component, then
2316   the Host field-value &MUST; be identical to that authority component
2317   after excluding any userinfo (<xref target="http.uri"/>).
2318   If the authority component is missing or undefined for the target URI,
2319   then the Host header field &MUST; be sent with an empty field-value.
2320</t>
2321<t>
2322   For example, a GET request to the origin server for
2323   &lt;http://www.example.org/pub/WWW/&gt; would begin with:
2324</t>
2325<figure><artwork type="message/http; msgtype=&#34;request&#34;" x:indent-with="  ">
2326GET /pub/WWW/ HTTP/1.1
2327Host: www.example.org
2328</artwork></figure>
2329<t>
2330   The Host header field &MUST; be sent in an HTTP/1.1 request even
2331   if the request-target is in the absolute-form, since this
2332   allows the Host information to be forwarded through ancient HTTP/1.0
2333   proxies that might not have implemented Host.
2334</t>
2335<t>
2336   When a proxy receives a request with an absolute-form of
2337   request-target, the proxy &MUST; ignore the received
2338   Host header field (if any) and instead replace it with the host
2339   information of the request-target.  If the proxy forwards the request,
2340   it &MUST; generate a new Host field-value based on the received
2341   request-target rather than forward the received Host field-value.
2342</t>
2343<t>
2344   Since the Host header field acts as an application-level routing
2345   mechanism, it is a frequent target for malware seeking to poison
2346   a shared cache or redirect a request to an unintended server.
2347   An interception proxy is particularly vulnerable if it relies on
2348   the Host field-value for redirecting requests to internal
2349   servers, or for use as a cache key in a shared cache, without
2350   first verifying that the intercepted connection is targeting a
2351   valid IP address for that host.
2352</t>
2353<t>
2354   A server &MUST; respond with a <x:ref>400 (Bad Request)</x:ref> status code
2355   to any HTTP/1.1 request message that lacks a Host header field and
2356   to any request message that contains more than one Host header field
2357   or a Host header field with an invalid field-value.
2358</t>
2359</section>
2360
2361<section title="Effective Request URI" anchor="effective.request.uri">
2362  <iref primary="true" item="effective request URI"/>
2363<t>
2364   A server that receives an HTTP request message &MUST; reconstruct
2365   the user agent's original target URI, based on the pieces of information
2366   learned from the request-target, <x:ref>Host</x:ref> header field, and
2367   connection context, in order to identify the intended target resource and
2368   properly service the request. The URI derived from this reconstruction
2369   process is referred to as the "<x:dfn>effective request URI</x:dfn>".
2370</t>
2371<t>
2372   For a user agent, the effective request URI is the target URI.
2373</t>
2374<t>
2375   If the request-target is in absolute-form, then the effective request URI
2376   is the same as the request-target.  Otherwise, the effective request URI
2377   is constructed as follows.
2378</t>
2379<t>
2380   If the request is received over a TLS-secured TCP connection,
2381   then the effective request URI's scheme is "https"; otherwise, the
2382   scheme is "http".
2383</t>
2384<t>
2385   If the request-target is in authority-form, then the effective
2386   request URI's authority component is the same as the request-target.
2387   Otherwise, if a <x:ref>Host</x:ref> header field is supplied with a
2388   non-empty field-value, then the authority component is the same as the
2389   Host field-value. Otherwise, the authority component is the concatenation of
2390   the default host name configured for the server, a colon (":"), and the
2391   connection's incoming TCP port number in decimal form.
2392</t>
2393<t>
2394   If the request-target is in authority-form or asterisk-form, then the
2395   effective request URI's combined path and query component is empty.
2396   Otherwise, the combined path and query component is the same as the
2397   request-target.
2398</t>
2399<t>
2400   The components of the effective request URI, once determined as above,
2401   can be combined into absolute-URI form by concatenating the scheme,
2402   "://", authority, and combined path and query component.
2403</t>
2404<figure>
2405<preamble>
2406   Example 1: the following message received over an insecure TCP connection
2407</preamble> 
2408<artwork type="example" x:indent-with="  ">
2409GET /pub/WWW/TheProject.html HTTP/1.1
2410Host: www.example.org:8080
2411</artwork>
2412</figure>
2413<figure>
2414<preamble>
2415  has an effective request URI of
2416</preamble>
2417<artwork type="example" x:indent-with="  ">
2418http://www.example.org:8080/pub/WWW/TheProject.html
2419</artwork>
2420</figure>
2421<figure>
2422<preamble>
2423   Example 2: the following message received over a TLS-secured TCP connection
2424</preamble> 
2425<artwork type="example" x:indent-with="  ">
2426OPTIONS * HTTP/1.1
2427Host: www.example.org
2428</artwork>
2429</figure>
2430<figure>
2431<preamble>
2432  has an effective request URI of
2433</preamble>
2434<artwork type="example" x:indent-with="  ">
2435https://www.example.org
2436</artwork>
2437</figure>
2438<t>
2439   An origin server that does not allow resources to differ by requested
2440   host &MAY; ignore the <x:ref>Host</x:ref> field-value and instead replace it
2441   with a configured server name when constructing the effective request URI.
2442</t>
2443<t>
2444   Recipients of an HTTP/1.0 request that lacks a <x:ref>Host</x:ref> header
2445   field &MAY; attempt to use heuristics (e.g., examination of the URI path for
2446   something unique to a particular host) in order to guess the
2447   effective request URI's authority component.
2448</t>
2449</section>
2450
2451<section title="Associating a Response to a Request" anchor="associating.response.to.request">
2452<t>
2453   HTTP does not include a request identifier for associating a given
2454   request message with its corresponding one or more response messages.
2455   Hence, it relies on the order of response arrival to correspond exactly
2456   to the order in which requests are made on the same connection.
2457   More than one response message per request only occurs when one or more
2458   informational responses (<x:ref>1xx</x:ref>, see &status-1xx;) precede a
2459   final response to the same request.
2460</t>
2461<t>
2462   A client that has more than one outstanding request on a connection &MUST;
2463   maintain a list of outstanding requests in the order sent and &MUST;
2464   associate each received response message on that connection to the highest
2465   ordered request that has not yet received a final (non-<x:ref>1xx</x:ref>)
2466   response.
2467</t>
2468</section>
2469
2470<section title="Message Forwarding" anchor="message.forwarding">
2471<t>
2472   As described in <xref target="intermediaries"/>, intermediaries can serve
2473   a variety of roles in the processing of HTTP requests and responses.
2474   Some intermediaries are used to improve performance or availability.
2475   Others are used for access control or to filter content.
2476   Since an HTTP stream has characteristics similar to a pipe-and-filter
2477   architecture, there are no inherent limits to the extent an intermediary
2478   can enhance (or interfere) with either direction of the stream.
2479</t>
2480<t>
2481   Intermediaries that forward a message &MUST; implement the
2482   <x:ref>Connection</x:ref> header field, as specified in
2483   <xref target="header.connection"/>, to exclude fields that are only
2484   intended for the incoming connection.
2485</t>
2486<t>
2487   In order to avoid request loops, a proxy that forwards requests to other
2488   proxies &MUST; be able to recognize and exclude all of its own server
2489   names, including any aliases, local variations, or literal IP addresses.
2490</t>
2491
2492<section title="Via" anchor="header.via">
2493  <iref primary="true" item="Via header field" x:for-anchor=""/>
2494  <x:anchor-alias value="pseudonym"/>
2495  <x:anchor-alias value="received-by"/>
2496  <x:anchor-alias value="received-protocol"/>
2497  <x:anchor-alias value="Via"/>
2498<t>
2499   The "Via" header field &MUST; be sent by a proxy or gateway in forwarded
2500   messages to indicate the intermediate protocols and recipients between the
2501   user agent and the server on requests, and between the origin server and
2502   the client on responses. It is analogous to the "Received" field
2503   used by email systems (<xref target="RFC5322" x:fmt="of" x:sec="3.6.7"/>).
2504   Via is used in HTTP for tracking message forwards,
2505   avoiding request loops, and identifying the protocol capabilities of
2506   all senders along the request/response chain.
2507</t>
2508<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"/>
2509  <x:ref>Via</x:ref>               = 1#( <x:ref>received-protocol</x:ref> <x:ref>RWS</x:ref> <x:ref>received-by</x:ref>
2510                          [ <x:ref>RWS</x:ref> <x:ref>comment</x:ref> ] )
2511  <x:ref>received-protocol</x:ref> = [ <x:ref>protocol-name</x:ref> "/" ] <x:ref>protocol-version</x:ref>
2512                      ; see <xref target="header.upgrade"/>
2513  <x:ref>received-by</x:ref>       = ( <x:ref>uri-host</x:ref> [ ":" <x:ref>port</x:ref> ] ) / <x:ref>pseudonym</x:ref>
2514  <x:ref>pseudonym</x:ref>         = <x:ref>token</x:ref>
2515</artwork></figure>
2516<t>
2517   The received-protocol indicates the protocol version of the message
2518   received by the server or client along each segment of the
2519   request/response chain. The received-protocol version is appended to
2520   the Via field value when the message is forwarded so that information
2521   about the protocol capabilities of upstream applications remains
2522   visible to all recipients.
2523</t>
2524<t>
2525   The protocol-name is excluded if and only if it would be "HTTP". The
2526   received-by field is normally the host and optional port number of a
2527   recipient server or client that subsequently forwarded the message.
2528   However, if the real host is considered to be sensitive information,
2529   it &MAY; be replaced by a pseudonym. If the port is not given, it &MAY;
2530   be assumed to be the default port of the received-protocol.
2531</t>
2532<t>
2533   Multiple Via field values represent each proxy or gateway that has
2534   forwarded the message. Each recipient &MUST; append its information
2535   such that the end result is ordered according to the sequence of
2536   forwarding applications.
2537</t>
2538<t>
2539   Comments &MAY; be used in the Via header field to identify the software
2540   of each recipient, analogous to the <x:ref>User-Agent</x:ref> and
2541   <x:ref>Server</x:ref> header fields. However, all comments in the Via field
2542   are optional and &MAY; be removed by any recipient prior to forwarding the
2543   message.
2544</t>
2545<t>
2546   For example, a request message could be sent from an HTTP/1.0 user
2547   agent to an internal proxy code-named "fred", which uses HTTP/1.1 to
2548   forward the request to a public proxy at p.example.net, which completes
2549   the request by forwarding it to the origin server at www.example.com.
2550   The request received by www.example.com would then have the following
2551   Via header field:
2552</t>
2553<figure><artwork type="example">
2554  Via: 1.0 fred, 1.1 p.example.net (Apache/1.1)
2555</artwork></figure>
2556<t>
2557   A proxy or gateway used as a portal through a network firewall
2558   &SHOULD-NOT; forward the names and ports of hosts within the firewall
2559   region unless it is explicitly enabled to do so. If not enabled, the
2560   received-by host of any host behind the firewall &SHOULD; be replaced
2561   by an appropriate pseudonym for that host.
2562</t>
2563<t>
2564   A proxy or gateway &MAY; combine an ordered subsequence of Via header
2565   field entries into a single such entry if the entries have identical
2566   received-protocol values. For example,
2567</t>
2568<figure><artwork type="example">
2569  Via: 1.0 ricky, 1.1 ethel, 1.1 fred, 1.0 lucy
2570</artwork></figure>
2571<t>
2572  could be collapsed to
2573</t>
2574<figure><artwork type="example">
2575  Via: 1.0 ricky, 1.1 mertz, 1.0 lucy
2576</artwork></figure>
2577<t>
2578   Senders &SHOULD-NOT; combine multiple entries unless they are all
2579   under the same organizational control and the hosts have already been
2580   replaced by pseudonyms. Senders &MUST-NOT; combine entries which
2581   have different received-protocol values.
2582</t>
2583</section>
2584
2585<section title="Transformation" anchor="message.transformation">
2586<t>
2587   If a proxy receives a request-target with a host name that is not a
2588   fully qualified domain name, it &MAY; add its own domain to the host name
2589   it received when forwarding the request.  A proxy &MUST-NOT; change the
2590   host name if it is a fully qualified domain name.
2591</t>
2592<t>
2593   A non-transforming proxy &MUST-NOT; modify the "path-absolute" and "query"
2594   parts of the received request-target when forwarding it to the next inbound
2595   server, except as noted above to replace an empty path with "/" or "*".
2596</t>
2597<t>
2598   A non-transforming proxy &MUST; preserve the message payload (&payload;),
2599   though it &MAY; change the message body through application or removal
2600   of a transfer coding (<xref target="transfer.codings"/>).
2601</t>
2602<t>
2603   A non-transforming proxy &SHOULD-NOT; modify header fields that provide
2604   information about the end points of the communication chain, the resource
2605   state, or the selected representation.
2606</t>
2607<t>
2608   A non-transforming proxy &MUST-NOT; modify any of the following fields in a
2609   request or response, and it &MUST-NOT; add any of these fields if not
2610   already present:
2611  <list style="symbols">
2612    <t><x:ref>Allow</x:ref> (&header-allow;)</t>
2613    <t><x:ref>Content-Location</x:ref> (&header-content-location;)</t>
2614    <t>Content-MD5 (<xref target="RFC2616" x:fmt="of" x:sec="14.15"/>)</t>
2615    <t><x:ref>ETag</x:ref> (&header-etag;)</t>
2616    <t><x:ref>Last-Modified</x:ref> (&header-last-modified;)</t>
2617    <t><x:ref>Server</x:ref> (&header-server;)</t>
2618  </list>
2619</t>
2620<t>
2621   A non-transforming proxy &MUST-NOT; modify an <x:ref>Expires</x:ref>
2622   header field (&header-expires;) if already present in a response, but
2623   it &MAY; add an <x:ref>Expires</x:ref> header field with a field-value
2624   identical to that of the <x:ref>Date</x:ref> header field.
2625</t>
2626<t>
2627   A proxy &MUST-NOT; modify or add any of the following fields in a
2628   message that contains the no-transform cache-control directive:
2629  <list style="symbols">
2630    <t><x:ref>Content-Encoding</x:ref> (&header-content-encoding;)</t>
2631    <t><x:ref>Content-Range</x:ref> (&header-content-range;)</t>
2632    <t><x:ref>Content-Type</x:ref> (&header-content-type;)</t>
2633  </list>
2634</t>
2635<t>
2636   A transforming proxy &MAY; modify or add these fields to a message
2637   that does not include no-transform, but if it does so, it &MUST; add a
2638   Warning 214 (Transformation applied) if one does not already appear
2639   in the message (see &header-warning;).
2640</t>
2641<x:note>
2642  <t>
2643    <x:h>Warning:</x:h> Unnecessary modification of header fields might
2644    cause authentication failures if stronger authentication
2645    mechanisms are introduced in later versions of HTTP. Such
2646    authentication mechanisms &MAY; rely on the values of header fields
2647    not listed here.
2648  </t>
2649</x:note>
2650</section>
2651</section>
2652</section>
2653
2654<section title="Connection Management" anchor="connection.management">
2655<t>
2656   HTTP messaging is independent of the underlying transport or
2657   session-layer connection protocol(s).  HTTP only presumes a reliable
2658   transport with in-order delivery of requests and the corresponding
2659   in-order delivery of responses.  The mapping of HTTP request and
2660   response structures onto the data units of an underlying transport
2661   protocol is outside the scope of this specification.
2662</t>
2663<t>
2664   As described in <xref target="connecting.inbound"/>, the specific
2665   connection protocols to be used for an HTTP interaction are determined by
2666   client configuration and the <x:ref>target URI</x:ref>.
2667   For example, the "http" URI scheme
2668   (<xref target="http.uri"/>) indicates a default connection of TCP
2669   over IP, with a default TCP port of 80, but the client might be
2670   configured to use a proxy via some other connection, port, or protocol.
2671</t>
2672<t>
2673   HTTP implementations are expected to engage in connection management,
2674   which includes maintaining the state of current connections,
2675   establishing a new connection or reusing an existing connection,
2676   processing messages received on a connection, detecting connection
2677   failures, and closing each connection.
2678   Most clients maintain multiple connections in parallel, including
2679   more than one connection per server endpoint.
2680   Most servers are designed to maintain thousands of concurrent connections,
2681   while controlling request queues to enable fair use and detect
2682   denial of service attacks.
2683</t>
2684
2685<section title="Connection" anchor="header.connection">
2686  <iref primary="true" item="Connection header field" x:for-anchor=""/>
2687  <iref primary="true" item="close" x:for-anchor=""/>
2688  <x:anchor-alias value="Connection"/>
2689  <x:anchor-alias value="connection-option"/>
2690  <x:anchor-alias value="close"/>
2691<t>
2692   The "Connection" header field allows the sender to indicate desired
2693   control options for the current connection.  In order to avoid confusing
2694   downstream recipients, a proxy or gateway &MUST; remove or replace any
2695   received connection options before forwarding the message.
2696</t>
2697<t>
2698   When a header field is used to supply control information for or about
2699   the current connection, the sender &SHOULD; list the corresponding
2700   field-name within the "Connection" header field.
2701   A proxy or gateway &MUST; parse a received Connection
2702   header field before a message is forwarded and, for each
2703   connection-option in this field, remove any header field(s) from
2704   the message with the same name as the connection-option, and then
2705   remove the Connection header field itself (or replace it with the
2706   intermediary's own connection options for the forwarded message).
2707</t>
2708<t>
2709   Hence, the Connection header field provides a declarative way of
2710   distinguishing header fields that are only intended for the
2711   immediate recipient ("hop-by-hop") from those fields that are
2712   intended for all recipients on the chain ("end-to-end"), enabling the
2713   message to be self-descriptive and allowing future connection-specific
2714   extensions to be deployed without fear that they will be blindly
2715   forwarded by older intermediaries.
2716</t>
2717<t>
2718   The Connection header field's value has the following grammar:
2719</t>
2720<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Connection"/><iref primary="true" item="Grammar" subitem="connection-option"/>
2721  <x:ref>Connection</x:ref>        = 1#<x:ref>connection-option</x:ref>
2722  <x:ref>connection-option</x:ref> = <x:ref>token</x:ref>
2723</artwork></figure>
2724<t>
2725   Connection options are case-insensitive.
2726</t>
2727<t>
2728   A sender &MUST-NOT; include field-names in the Connection header
2729   field-value for fields that are defined as expressing constraints
2730   for all recipients in the request or response chain, such as the
2731   Cache-Control header field (&header-cache-control;).
2732</t>
2733<t>
2734   The connection options do not have to correspond to a header field
2735   present in the message, since a connection-specific header field
2736   might not be needed if there are no parameters associated with that
2737   connection option.  Recipients that trigger certain connection
2738   behavior based on the presence of connection options &MUST; do so
2739   based on the presence of the connection-option rather than only the
2740   presence of the optional header field.  In other words, if the
2741   connection option is received as a header field but not indicated
2742   within the Connection field-value, then the recipient &MUST; ignore
2743   the connection-specific header field because it has likely been
2744   forwarded by an intermediary that is only partially conformant.
2745</t>
2746<t>
2747   When defining new connection options, specifications ought to
2748   carefully consider existing deployed header fields and ensure
2749   that the new connection option does not share the same name as
2750   an unrelated header field that might already be deployed.
2751   Defining a new connection option essentially reserves that potential
2752   field-name for carrying additional information related to the
2753   connection option, since it would be unwise for senders to use
2754   that field-name for anything else.
2755</t>
2756<t>
2757   The "<x:dfn>close</x:dfn>" connection option is defined for a
2758   sender to signal that this connection will be closed after completion of
2759   the response. For example,
2760</t>
2761<figure><artwork type="example">
2762  Connection: close
2763</artwork></figure>
2764<t>
2765   in either the request or the response header fields indicates that
2766   the connection &SHOULD; be closed after the current request/response
2767   is complete (<xref target="persistent.tear-down"/>).
2768</t>
2769<t>
2770   A client that does not support <x:ref>persistent connections</x:ref> &MUST;
2771   send the "close" connection option in every request message.
2772</t>
2773<t>
2774   A server that does not support <x:ref>persistent connections</x:ref> &MUST;
2775   send the "close" connection option in every response message that
2776   does not have a <x:ref>1xx (Informational)</x:ref> status code.
2777</t>
2778</section>
2779
2780<section title="Establishment" anchor="persistent.establishment">
2781<t>
2782   It is beyond the scope of this specification to describe how connections
2783   are established via various transport or session-layer protocols.
2784   Each connection applies to only one transport link.
2785</t>
2786</section>
2787
2788<section title="Persistence" anchor="persistent.connections">
2789   <x:anchor-alias value="persistent connections"/>
2790<t>
2791   HTTP/1.1 defaults to the use of "<x:dfn>persistent connections</x:dfn>",
2792   which allow multiple requests and responses to be carried over a single
2793   connection. The "<x:ref>close</x:ref>" connection-option is used to signal
2794   that a connection will not persist after the current request/response.
2795   HTTP implementations &SHOULD; support persistent connections.
2796</t>
2797<t>
2798   A recipient determines whether a connection is persistent or not based on
2799   the most recently received message's protocol version and
2800   <x:ref>Connection</x:ref> header field (if any):
2801   <list style="symbols">
2802     <t>If the <x:ref>close</x:ref> connection option is present, the
2803        connection will not persist after the current response; else,</t>
2804     <t>If the received protocol is HTTP/1.1 (or later), the connection will
2805        persist after the current response; else,</t>
2806     <t>If the received protocol is HTTP/1.0, the "keep-alive"
2807        connection option is present, the recipient is not a proxy, and
2808        the recipient wishes to honor the HTTP/1.0 "keep-alive" mechanism,
2809        the connection will persist after the current response; otherwise,</t>
2810     <t>The connection will close after the current response.</t>
2811   </list>
2812</t>
2813<t>
2814   A server &MAY; assume that an HTTP/1.1 client intends to maintain a
2815   persistent connection until a <x:ref>close</x:ref> connection option
2816   is received in a request.
2817</t>
2818<t>
2819   A client &MAY; reuse a persistent connection until it sends or receives
2820   a <x:ref>close</x:ref> connection option or receives an HTTP/1.0 response
2821   without a "keep-alive" connection option.
2822</t>
2823<t>
2824   In order to remain persistent, all messages on a connection &MUST;
2825   have a self-defined message length (i.e., one not defined by closure
2826   of the connection), as described in <xref target="message.body"/>.
2827   A server &MUST; read the entire request message body or close
2828   the connection after sending its response, since otherwise the
2829   remaining data on a persistent connection would be misinterpreted
2830   as the next request.  Likewise,
2831   a client &MUST; read the entire response message body if it intends
2832   to reuse the same connection for a subsequent request.
2833</t>
2834<t>
2835   A proxy server &MUST-NOT; maintain a persistent connection with an
2836   HTTP/1.0 client (see <xref x:sec="19.7.1" x:fmt="of" target="RFC2068"/> for
2837   information and discussion of the problems with the Keep-Alive header field
2838   implemented by many HTTP/1.0 clients).
2839</t>
2840<t>
2841   Clients and servers &SHOULD-NOT; assume that a persistent connection is
2842   maintained for HTTP versions less than 1.1 unless it is explicitly
2843   signaled.
2844   See <xref target="compatibility.with.http.1.0.persistent.connections"/>
2845   for more information on backward compatibility with HTTP/1.0 clients.
2846</t>
2847
2848<section title="Pipelining" anchor="pipelining">
2849<t>
2850   A client that supports persistent connections &MAY; "pipeline" its
2851   requests (i.e., send multiple requests without waiting for each
2852   response). A server &MUST; send its responses to those requests in the
2853   same order that the requests were received.
2854</t>
2855<t>
2856   Clients which assume persistent connections and pipeline immediately
2857   after connection establishment &SHOULD; be prepared to retry their
2858   connection if the first pipelined attempt fails. If a client does
2859   such a retry, it &MUST-NOT; pipeline before it knows the connection is
2860   persistent. Clients &MUST; also be prepared to resend their requests if
2861   the server closes the connection before sending all of the
2862   corresponding responses.
2863</t>
2864<t>
2865   Clients &SHOULD-NOT; pipeline requests using non-idempotent request methods
2866   or non-idempotent sequences of request methods (see &idempotent-methods;).
2867   Otherwise, a premature termination of the transport connection could lead
2868   to indeterminate results. A client wishing to send a non-idempotent
2869   request &SHOULD; wait to send that request until it has received the
2870   response status line for the previous request.
2871</t>
2872</section>
2873
2874<section title="Retrying Requests" anchor="persistent.retrying.requests">
2875<t>
2876   Connections can be closed at any time, with or without intention.
2877   Implementations ought to anticipate the need to recover
2878   from asynchronous close events.
2879   A client &MAY; open a new connection and retransmit an aborted sequence
2880   of requests without user interaction so long as the request sequence is
2881   idempotent (see &idempotent-methods;).
2882   A client &MUST-NOT; automatically retry non-idempotent request sequences,
2883   although user agents &MAY; offer a human operator the choice of retrying
2884   the request(s). Confirmation by
2885   user agent software with semantic understanding of the application
2886   &MAY; substitute for user confirmation. An automatic retry &SHOULD-NOT; 
2887   be repeated if a second sequence of requests fails.
2888</t>
2889</section>
2890</section>
2891   
2892<section title="Concurrency" anchor="persistent.concurrency">
2893<t>
2894   Clients &SHOULD; limit the number of simultaneous
2895   connections that they maintain to a given server.
2896</t>
2897<t>
2898   Previous revisions of HTTP gave a specific number of connections as a
2899   ceiling, but this was found to be impractical for many applications. As a
2900   result, this specification does not mandate a particular maximum number of
2901   connections, but instead encourages clients to be conservative when opening
2902   multiple connections.
2903</t>
2904<t>
2905   Multiple connections are typically used to avoid the "head-of-line
2906   blocking" problem, wherein a request that takes significant server-side
2907   processing and/or has a large payload blocks subsequent requests on the
2908   same connection. However, each connection consumes server resources.
2909   Furthermore, using multiple connections can cause undesirable side effects
2910   in congested networks.
2911</t>
2912<t>
2913   Note that servers might reject traffic that they deem abusive, including an
2914   excessive number of connections from a client.
2915</t>
2916</section>
2917
2918<section title="Failures and Time-outs" anchor="persistent.failures">
2919<t>
2920   Servers will usually have some time-out value beyond which they will
2921   no longer maintain an inactive connection. Proxy servers might make
2922   this a higher value since it is likely that the client will be making
2923   more connections through the same server. The use of persistent
2924   connections places no requirements on the length (or existence) of
2925   this time-out for either the client or the server.
2926</t>
2927<t>
2928   When a client or server wishes to time-out it &SHOULD; issue a graceful
2929   close on the transport connection. Clients and servers &SHOULD; both
2930   constantly watch for the other side of the transport close, and
2931   respond to it as appropriate. If a client or server does not detect
2932   the other side's close promptly it could cause unnecessary resource
2933   drain on the network.
2934</t>
2935<t>
2936   A client, server, or proxy &MAY; close the transport connection at any
2937   time. For example, a client might have started to send a new request
2938   at the same time that the server has decided to close the "idle"
2939   connection. From the server's point of view, the connection is being
2940   closed while it was idle, but from the client's point of view, a
2941   request is in progress.
2942</t>
2943<t>
2944   Servers &SHOULD; maintain persistent connections and allow the underlying
2945   transport's flow control mechanisms to resolve temporary overloads, rather
2946   than terminate connections with the expectation that clients will retry.
2947   The latter technique can exacerbate network congestion.
2948</t>
2949<t>
2950   A client sending a message body &SHOULD; monitor
2951   the network connection for an error status code while it is transmitting
2952   the request. If the client sees an error status code, it &SHOULD;
2953   immediately cease transmitting the body and close the connection.
2954</t>
2955</section>
2956   
2957<section title="Tear-down" anchor="persistent.tear-down">
2958  <iref primary="false" item="Connection header field" x:for-anchor=""/>
2959  <iref primary="false" item="close" x:for-anchor=""/>
2960<t>
2961   The <x:ref>Connection</x:ref> header field
2962   (<xref target="header.connection"/>) provides a "<x:ref>close</x:ref>"
2963   connection option that a sender &SHOULD; send when it wishes to close
2964   the connection after the current request/response pair.
2965</t>
2966<t>
2967   A client that sends a <x:ref>close</x:ref> connection option &MUST-NOT;
2968   send further requests on that connection (after the one containing
2969   <x:ref>close</x:ref>) and &MUST; close the connection after reading the
2970   final response message corresponding to this request.
2971</t>
2972<t>
2973   A server that receives a <x:ref>close</x:ref> connection option &MUST;
2974   initiate a lingering close (see below) of the connection after it sends the
2975   final response to the request that contained <x:ref>close</x:ref>.
2976   The server &SHOULD; include a <x:ref>close</x:ref> connection option
2977   in its final response on that connection. The server &MUST-NOT; process
2978   any further requests received on that connection.
2979</t>
2980<t>
2981   A server that sends a <x:ref>close</x:ref> connection option &MUST;
2982   initiate a lingering close of the connection after it sends the
2983   response containing <x:ref>close</x:ref>. The server &MUST-NOT; process
2984   any further requests received on that connection.
2985</t>
2986<t>
2987   A client that receives a <x:ref>close</x:ref> connection option &MUST;
2988   cease sending requests on that connection and close the connection
2989   after reading the response message containing the close; if additional
2990   pipelined requests had been sent on the connection, the client &SHOULD;
2991   assume that they will not be processed by the server.
2992</t>
2993<t>
2994   If a server performs an immediate close of a TCP connection, there is a
2995   significant risk that the client will not be able to read the last HTTP
2996   response.  If the server receives additional data from the client on a
2997   fully-closed connection, such as another request that was sent by the
2998   client before receiving the server's response, the server's TCP stack will
2999   send a reset packet to the client; unfortunately, the reset packet might
3000   erase the client's unacknowledged input buffers before they can be read
3001   and interpreted by the client's HTTP parser.
3002</t>
3003<t>
3004   To avoid the TCP reset problem, a server can perform a lingering close on a
3005   connection by closing only the write side of the read/write connection
3006   (a half-close) and continuing to read from the connection until the
3007   connection is closed by the client or the server is reasonably certain
3008   that its own TCP stack has received the client's acknowledgement of the
3009   packet(s) containing the server's last response. It is then safe for the
3010   server to fully close the connection.
3011</t>
3012<t>
3013   It is unknown whether the reset problem is exclusive to TCP or might also
3014   be found in other transport connection protocols.
3015</t>
3016</section>
3017
3018<section title="Upgrade" anchor="header.upgrade">
3019  <iref primary="true" item="Upgrade header field" x:for-anchor=""/>
3020  <x:anchor-alias value="Upgrade"/>
3021  <x:anchor-alias value="protocol"/>
3022  <x:anchor-alias value="protocol-name"/>
3023  <x:anchor-alias value="protocol-version"/>
3024<t>
3025   The "Upgrade" header field is intended to provide a simple mechanism
3026   for transitioning from HTTP/1.1 to some other protocol on the same
3027   connection.  A client &MAY; send a list of protocols in the Upgrade
3028   header field of a request to invite the server to switch to one or
3029   more of those protocols before sending the final response.
3030   A server &MUST; send an Upgrade header field in <x:ref>101 (Switching
3031   Protocols)</x:ref> responses to indicate which protocol(s) are being
3032   switched to, and &MUST; send it in <x:ref>426 (Upgrade Required)</x:ref>
3033   responses to indicate acceptable protocols.
3034   A server &MAY; send an Upgrade header field in any other response to
3035   indicate that they might be willing to upgrade to one of the
3036   specified protocols for a future request.
3037</t>
3038<figure><artwork type="abnf2616"><iref primary="true" item="Grammar" subitem="Upgrade"/>
3039  <x:ref>Upgrade</x:ref>          = 1#<x:ref>protocol</x:ref>
3040
3041  <x:ref>protocol</x:ref>         = <x:ref>protocol-name</x:ref> ["/" <x:ref>protocol-version</x:ref>]
3042  <x:ref>protocol-name</x:ref>    = <x:ref>token</x:ref>
3043  <x:ref>protocol-version</x:ref> = <x:ref>token</x:ref>
3044</artwork></figure>
3045<t>
3046   For example,
3047</t>
3048<figure><artwork type="example">
3049  Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11
3050</artwork></figure>
3051<t>
3052   Upgrade eases the difficult transition between incompatible protocols by
3053   allowing the client to initiate a request in the more commonly
3054   supported protocol while indicating to the server that it would like
3055   to use a "better" protocol if available (where "better" is determined
3056   by the server, possibly according to the nature of the request method
3057   or target resource).
3058</t>
3059<t>
3060   Upgrade cannot be used to insist on a protocol change; its acceptance and
3061   use by the server is optional. The capabilities and nature of the
3062   application-level communication after the protocol change is entirely
3063   dependent upon the new protocol chosen, although the first action
3064   after changing the protocol &MUST; be a response to the initial HTTP
3065   request that contained the Upgrade header field.
3066</t>
3067<t>
3068   For example, if the Upgrade header field is received in a GET request
3069   and the server decides to switch protocols, then it &MUST; first respond
3070   with a <x:ref>101 (Switching Protocols)</x:ref> message in HTTP/1.1 and
3071   then immediately follow that with the new protocol's equivalent of a
3072   response to a GET on the target resource.  This allows a connection to be
3073   upgraded to protocols with the same semantics as HTTP without the
3074   latency cost of an additional round-trip.  A server &MUST-NOT; switch
3075   protocols unless the received message semantics can be honored by the new
3076   protocol; an OPTIONS request can be honored by any protocol.
3077</t>
3078<t>
3079   When Upgrade is sent, a sender &MUST; also send a
3080   <x:ref>Connection</x:ref> header field (<xref target="header.connection"/>)
3081   that contains the "upgrade" connection option, in order to prevent Upgrade
3082   from being accidentally forwarded by intermediaries that might not implement
3083   the listed protocols.  A server &MUST; ignore an Upgrade header field that
3084   is received in an HTTP/1.0 request.
3085</t>
3086<t>
3087   The Upgrade header field only applies to switching application-level
3088   protocols on the existing connection; it cannot be used
3089   to switch to a protocol on a different connection. For that purpose, it is
3090   more appropriate to use a <x:ref>3xx (Redirection)</x:ref> response
3091   (&status-3xx;).
3092</t>
3093<t>
3094   This specification only defines the protocol name "HTTP" for use by
3095   the family of Hypertext Transfer Protocols, as defined by the HTTP
3096   version rules of <xref target="http.version"/> and future updates to this
3097   specification. Additional tokens can be registered with IANA using the
3098   registration procedure defined in <xref target="upgrade.token.registry"/>.
3099</t>
3100</section>
3101</section>
3102
3103<section title="IANA Considerations" anchor="IANA.considerations">
3104
3105<section title="Header Field Registration" anchor="header.field.registration">
3106<t>
3107   HTTP header fields are registered within the Message Header Field Registry
3108   <xref target="RFC3864"/> maintained by IANA at
3109   <eref target="http://www.iana.org/assignments/message-headers/message-header-index.html"/>.
3110</t>
3111<t>
3112   This document defines the following HTTP header fields, so their
3113   associated registry entries shall be updated according to the permanent
3114   registrations below:
3115</t>
3116<?BEGININC p1-messaging.iana-headers ?>
3117<!--AUTOGENERATED FROM extract-header-defs.xslt, do not edit manually-->
3118<texttable align="left" suppress-title="true" anchor="iana.header.registration.table">
3119   <ttcol>Header Field Name</ttcol>
3120   <ttcol>Protocol</ttcol>
3121   <ttcol>Status</ttcol>
3122   <ttcol>Reference</ttcol>
3123
3124   <c>Connection</c>
3125   <c>http</c>
3126   <c>standard</c>
3127   <c>
3128      <xref target="header.connection"/>
3129   </c>
3130   <c>Content-Length</c>
3131   <c>http</c>
3132   <c>standard</c>
3133   <c>
3134      <xref target="header.content-length"/>
3135   </c>
3136   <c>Host</c>
3137   <c>http</c>
3138   <c>standard</c>
3139   <c>
3140      <xref target="header.host"/>
3141   </c>
3142   <c>TE</c>
3143   <c>http</c>
3144   <c>standard</c>
3145   <c>
3146      <xref target="header.te"/>
3147   </c>
3148   <c>Trailer</c>
3149   <c>http</c>
3150   <c>standard</c>
3151   <c>
3152      <xref target="header.trailer"/>
3153   </c>
3154   <c>Transfer-Encoding</c>
3155   <c>http</c>
3156   <c>standard</c>
3157   <c>
3158      <xref target="header.transfer-encoding"/>
3159   </c>
3160   <c>Upgrade</c>
3161   <c>http</c>
3162   <c>standard</c>
3163   <c>
3164      <xref target="header.upgrade"/>
3165   </c>
3166   <c>Via</c>
3167   <c>http</c>
3168   <c>standard</c>
3169   <c>
3170      <xref target="header.via"/>
3171   </c>
3172</texttable>
3173<!--(END)-->
3174<?ENDINC p1-messaging.iana-headers ?>
3175<t>
3176   Furthermore, the header field-name "Close" shall be registered as
3177   "reserved", since using that name as an HTTP header field might
3178   conflict with the "close" connection option of the "<x:ref>Connection</x:ref>"
3179   header field (<xref target="header.connection"/>).
3180</t>
3181<texttable align="left" suppress-title="true">
3182   <ttcol>Header Field Name</ttcol>
3183   <ttcol>Protocol</ttcol>
3184   <ttcol>Status</ttcol>
3185   <ttcol>Reference</ttcol>
3186
3187   <c>Close</c>
3188   <c>http</c>
3189   <c>reserved</c>
3190   <c>
3191      <xref target="header.field.registration"/>
3192   </c>
3193</texttable>
3194<t>
3195   The change controller is: "IETF (iesg@ietf.org) - Internet Engineering Task Force".
3196</t>
3197</section>
3198
3199<section title="URI Scheme Registration" anchor="uri.scheme.registration">
3200<t>
3201   IANA maintains the registry of URI Schemes <xref target="RFC4395"/> at
3202   <eref target="http://www.iana.org/assignments/uri-schemes.html"/>.
3203</t>
3204<t>
3205   This document defines the following URI schemes, so their
3206   associated registry entries shall be updated according to the permanent
3207   registrations below:
3208</t>
3209<texttable align="left" suppress-title="true">
3210   <ttcol>URI Scheme</ttcol>
3211   <ttcol>Description</ttcol>
3212   <ttcol>Reference</ttcol>
3213
3214   <c>http</c>
3215   <c>Hypertext Transfer Protocol</c>
3216   <c><xref target="http.uri"/></c>
3217
3218   <c>https</c>
3219   <c>Hypertext Transfer Protocol Secure</c>
3220   <c><xref target="https.uri"/></c>
3221</texttable>
3222</section>
3223
3224<section title="Internet Media Type Registrations" anchor="internet.media.type.http">
3225<t>
3226   This document serves as the specification for the Internet media types
3227   "message/http" and "application/http". The following is to be registered with
3228   IANA (see <xref target="RFC4288"/>).
3229</t>
3230<section title="Internet Media Type message/http" anchor="internet.media.type.message.http">
3231<iref item="Media Type" subitem="message/http" primary="true"/>
3232<iref item="message/http Media Type" primary="true"/>
3233<t>
3234   The message/http type can be used to enclose a single HTTP request or
3235   response message, provided that it obeys the MIME restrictions for all
3236   "message" types regarding line length and encodings.
3237</t>
3238<t>
3239  <list style="hanging" x:indent="12em">
3240    <t hangText="Type name:">
3241      message
3242    </t>
3243    <t hangText="Subtype name:">
3244      http
3245    </t>
3246    <t hangText="Required parameters:">
3247      none
3248    </t>
3249    <t hangText="Optional parameters:">
3250      version, msgtype
3251      <list style="hanging">
3252        <t hangText="version:">
3253          The HTTP-version number of the enclosed message
3254          (e.g., "1.1"). If not present, the version can be
3255          determined from the first line of the body.
3256        </t>
3257        <t hangText="msgtype:">
3258          The message type &mdash; "request" or "response". If not
3259          present, the type can be determined from the first
3260          line of the body.
3261        </t>
3262      </list>
3263    </t>
3264    <t hangText="Encoding considerations:">
3265      only "7bit", "8bit", or "binary" are permitted
3266    </t>
3267    <t hangText="Security considerations:">
3268      none
3269    </t>
3270    <t hangText="Interoperability considerations:">
3271      none
3272    </t>
3273    <t hangText="Published specification:">
3274      This specification (see <xref target="internet.media.type.message.http"/>).
3275    </t>
3276    <t hangText="Applications that use this media type:">
3277    </t>
3278    <t hangText="Additional information:">
3279      <list style="hanging">
3280        <t hangText="Magic number(s):">none</t>
3281        <t hangText="File extension(s):">none</t>
3282        <t hangText="Macintosh file type code(s):">none</t>
3283      </list>
3284    </t>
3285    <t hangText="Person and email address to contact for further information:">
3286      See Authors Section.
3287    </t>
3288    <t hangText="Intended usage:">
3289      COMMON
3290    </t>
3291    <t hangText="Restrictions on usage:">
3292      none
3293    </t>
3294    <t hangText="Author/Change controller:">
3295      IESG
3296    </t>
3297  </list>
3298</t>
3299</section>
3300<section title="Internet Media Type application/http" anchor="internet.media.type.application.http">
3301<iref item="Media Type" subitem="application/http" primary="true"/>
3302<iref item="application/http Media Type" primary="true"/>
3303<t>
3304   The application/http type can be used to enclose a pipeline of one or more
3305   HTTP request or response messages (not intermixed).
3306</t>
3307<t>
3308  <list style="hanging" x:indent="12em">
3309    <t hangText="Type name:">
3310      application
3311    </t>
3312    <t hangText="Subtype name:">
3313      http
3314    </t>
3315    <t hangText="Required parameters:">
3316      none
3317    </t>
3318    <t hangText="Optional parameters:">
3319      version, msgtype
3320      <list style="hanging">
3321        <t hangText="version:">
3322          The HTTP-version number of the enclosed messages
3323          (e.g., "1.1"). If not present, the version can be
3324          determined from the first line of the body.
3325        </t>
3326        <t hangText="msgtype:">
3327          The message type &mdash; "request" or "response". If not
3328          present, the type can be determined from the first
3329          line of the body.
3330        </t>
3331      </list>
3332    </t>
3333    <t hangText="Encoding considerations:">
3334      HTTP messages enclosed by this type
3335      are in "binary" format; use of an appropriate
3336      Content-Transfer-Encoding is required when
3337      transmitted via E-mail.
3338    </t>
3339    <t hangText="Security considerations:">
3340      none
3341    </t>
3342    <t hangText="Interoperability considerations:">
3343      none
3344    </t>
3345    <t hangText="Published specification:">
3346      This specification (see <xref target="internet.media.type.application.http"/>).
3347    </t>
3348    <t hangText="Applications that use this media type:">
3349    </t>
3350    <t hangText="Additional information:">
3351      <list style="hanging">
3352        <t hangText="Magic number(s):">none</t>
3353        <t hangText="File extension(s):">none</t>
3354        <t hangText="Macintosh file type code(s):">none</t>
3355      </list>
3356    </t>
3357    <t hangText="Person and email address to contact for further information:">
3358      See Authors Section.
3359    </t>
3360    <t hangText="Intended usage:">
3361      COMMON
3362    </t>
3363    <t hangText="Restrictions on usage:">
3364      none
3365    </t>
3366    <t hangText="Author/Change controller:">
3367      IESG
3368    </t>
3369  </list>
3370</t>
3371</section>
3372</section>
3373
3374<section title="Transfer Coding Registry" anchor="transfer.coding.registry">
3375<t>
3376   The HTTP Transfer Coding Registry defines the name space for transfer
3377   coding names.
3378</t>
3379<t>
3380   Registrations &MUST; include the following fields:
3381   <list style="symbols">
3382     <t>Name</t>
3383     <t>Description</t>
3384     <t>Pointer to specification text</t>
3385   </list>
3386</t>
3387<t>
3388   Names of transfer codings &MUST-NOT; overlap with names of content codings
3389   (&content-codings;) unless the encoding transformation is identical, as
3390   is the case for the compression codings defined in
3391   <xref target="compression.codings"/>.
3392</t>
3393<t>
3394   Values to be added to this name space require IETF Review (see
3395   <xref target="RFC5226" x:fmt="of" x:sec="4.1"/>), and &MUST;
3396   conform to the purpose of transfer coding defined in this section.
3397   Use of program names for the identification of encoding formats
3398   is not desirable and is discouraged for future encodings.
3399</t>
3400<t>
3401   The registry itself is maintained at
3402   <eref target="http://www.iana.org/assignments/http-parameters"/>.
3403</t>
3404</section>
3405
3406<section title="Transfer Coding Registrations" anchor="transfer.coding.registration">
3407<t>
3408   The HTTP Transfer Coding Registry shall be updated with the registrations
3409   below:
3410</t>
3411<texttable align="left" suppress-title="true" anchor="iana.transfer.coding.registration.table">
3412   <ttcol>Name</ttcol>
3413   <ttcol>Description</ttcol>
3414   <ttcol>Reference</ttcol>
3415   <c>chunked</c>
3416   <c>Transfer in a series of chunks</c>
3417   <c>
3418      <xref target="chunked.encoding"/>
3419   </c>
3420   <c>compress</c>
3421   <c>UNIX "compress" program method</c>
3422   <c>
3423      <xref target="compress.coding"/>
3424   </c>
3425   <c>deflate</c>
3426   <c>"deflate" compression mechanism (<xref target="RFC1951"/>) used inside
3427   the "zlib" data format (<xref target="RFC1950"/>)
3428   </c>
3429   <c>
3430      <xref target="deflate.coding"/>
3431   </c>
3432   <c>gzip</c>
3433   <c>Same as GNU zip <xref target="RFC1952"/></c>
3434   <c>
3435      <xref target="gzip.coding"/>
3436   </c>
3437</texttable>
3438</section>
3439
3440<section title="Upgrade Token Registry" anchor="upgrade.token.registry">
3441<t>
3442   The HTTP Upgrade Token Registry defines the name space for protocol-name
3443   tokens used to identify protocols in the <x:ref>Upgrade</x:ref> header
3444   field. Each registered protocol name is associated with contact information
3445   and an optional set of specifications that details how the connection
3446   will be processed after it has been upgraded.
3447</t>
3448<t>
3449   Registrations happen on a "First Come First Served" basis (see
3450   <xref target="RFC5226" x:sec="4.1" x:fmt="of"/>) and are subject to the
3451   following rules:
3452  <list style="numbers">
3453    <t>A protocol-name token, once registered, stays registered forever.</t>
3454    <t>The registration &MUST; name a responsible party for the
3455       registration.</t>
3456    <t>The registration &MUST; name a point of contact.</t>
3457    <t>The registration &MAY; name a set of specifications associated with
3458       that token. Such specifications need not be publicly available.</t>
3459    <t>The registration &SHOULD; name a set of expected "protocol-version"
3460       tokens associated with that token at the time of registration.</t>
3461    <t>The responsible party &MAY; change the registration at any time.
3462       The IANA will keep a record of all such changes, and make them
3463       available upon request.</t>
3464    <t>The IESG &MAY; reassign responsibility for a protocol token.
3465       This will normally only be used in the case when a
3466       responsible party cannot be contacted.</t>
3467  </list>
3468</t>
3469<t>
3470   This registration procedure for HTTP Upgrade Tokens replaces that
3471   previously defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
3472</t>
3473</section>
3474
3475<section title="Upgrade Token Registration" anchor="upgrade.token.registration">
3476<t>
3477   The HTTP Upgrade Token Registry shall be updated with the registration
3478   below:
3479</t>
3480<texttable align="left" suppress-title="true">
3481   <ttcol>Value</ttcol>
3482   <ttcol>Description</ttcol>
3483   <ttcol>Expected Version Tokens</ttcol>
3484   <ttcol>Reference</ttcol>
3485
3486   <c>HTTP</c>
3487   <c>Hypertext Transfer Protocol</c>
3488   <c>any DIGIT.DIGIT (e.g, "2.0")</c>
3489   <c><xref target="http.version"/></c>
3490</texttable>
3491<t>
3492   The responsible party is: "IETF (iesg@ietf.org) - Internet Engineering Task Force".
3493</t>
3494</section>
3495
3496</section>
3497
3498<section title="Security Considerations" anchor="security.considerations">
3499<t>
3500   This section is meant to inform application developers, information
3501   providers, and users of the security limitations in HTTP/1.1 as
3502   described by this document. The discussion does not include
3503   definitive solutions to the problems revealed, though it does make
3504   some suggestions for reducing security risks.
3505</t>
3506
3507<section title="Personal Information" anchor="personal.information">
3508<t>
3509   HTTP clients are often privy to large amounts of personal information,
3510   including both information provided by the user to interact with resources
3511   (e.g., the user's name, location, mail address, passwords, encryption
3512   keys, etc.) and information about the user's browsing activity over
3513   time (e.g., history, bookmarks, etc.). HTTP implementations need to
3514   prevent unintentional leakage of this information.
3515</t>
3516</section>
3517
3518<section title="Abuse of Server Log Information" anchor="abuse.of.server.log.information">
3519<t>
3520   A server is in the position to save personal data about a user's
3521   requests which might identify their reading patterns or subjects of
3522   interest.  In particular, log information gathered at an intermediary
3523   often contains a history of user agent interaction, across a multitude
3524   of sites, that can be traced to individual users.
3525</t>
3526<t>
3527   HTTP log information is confidential in nature; its handling is often
3528   constrained by laws and regulations.  Log information needs to be securely
3529   stored and appropriate guidelines followed for its analysis.
3530   Anonymization of personal information within individual entries helps,
3531   but is generally not sufficient to prevent real log traces from being
3532   re-identified based on correlation with other access characteristics.
3533   As such, access traces that are keyed to a specific client should not
3534   be published even if the key is pseudonymous.
3535</t>
3536<t>
3537   To minimize the risk of theft or accidental publication, log information
3538   should be purged of personally identifiable information, including
3539   user identifiers, IP addresses, and user-provided query parameters,
3540   as soon as that information is no longer necessary to support operational
3541   needs for security, auditing, or fraud control.
3542</t>
3543</section>
3544
3545<section title="Attacks Based On File and Path Names" anchor="attack.pathname">
3546<t>
3547   Origin servers &SHOULD; be careful to restrict
3548   the documents returned by HTTP requests to be only those that were
3549   intended by the server administrators. If an HTTP server translates
3550   HTTP URIs directly into file system calls, the server &MUST; take
3551   special care not to serve files that were not intended to be
3552   delivered to HTTP clients. For example, UNIX, Microsoft Windows, and
3553   other operating systems use ".." as a path component to indicate a
3554   directory level above the current one. On such a system, an HTTP
3555   server &MUST; disallow any such construct in the request-target if it
3556   would otherwise allow access to a resource outside those intended to
3557   be accessible via the HTTP server. Similarly, files intended for
3558   reference only internally to the server (such as access control
3559   files, configuration files, and script code) &MUST; be protected from
3560   inappropriate retrieval, since they might contain sensitive
3561   information.
3562</t>
3563</section>
3564
3565<section title="DNS-related Attacks" anchor="dns.related.attacks">
3566<t>
3567   HTTP clients rely heavily on the Domain Name Service (DNS), and are thus
3568   generally prone to security attacks based on the deliberate misassociation
3569   of IP addresses and DNS names not protected by DNSSEC. Clients need to be
3570   cautious in assuming the validity of an IP number/DNS name association unless
3571   the response is protected by DNSSEC (<xref target="RFC4033"/>).
3572</t>
3573</section>
3574
3575<section title="Intermediaries and Caching" anchor="attack.intermediaries">
3576<t>
3577   By their very nature, HTTP intermediaries are men-in-the-middle, and
3578   represent an opportunity for man-in-the-middle attacks. Compromise of
3579   the systems on which the intermediaries run can result in serious security
3580   and privacy problems. Intermediaries have access to security-related
3581   information, personal information about individual users and
3582   organizations, and proprietary information belonging to users and
3583   content providers. A compromised intermediary, or an intermediary
3584   implemented or configured without regard to security and privacy
3585   considerations, might be used in the commission of a wide range of
3586   potential attacks.
3587</t>
3588<t>
3589   Intermediaries that contain a shared cache are especially vulnerable
3590   to cache poisoning attacks.
3591</t>
3592<t>
3593   Implementers need to consider the privacy and security
3594   implications of their design and coding decisions, and of the
3595   configuration options they provide to operators (especially the
3596   default configuration).
3597</t>
3598<t>
3599   Users need to be aware that intermediaries are no more trustworthy than
3600   the people who run them; HTTP itself cannot solve this problem.
3601</t>
3602</section>
3603
3604<section title="Buffer Overflows" anchor="attack.protocol.element.size.overflows">
3605<t>
3606   Because HTTP uses mostly textual, character-delimited fields, attackers can
3607   overflow buffers in implementations, and/or perform a Denial of Service
3608   against implementations that accept fields with unlimited lengths.
3609</t>
3610<t>
3611   To promote interoperability, this specification makes specific
3612   recommendations for minimum size limits on request-line
3613   (<xref target="request.line"/>)
3614   and blocks of header fields (<xref target="header.fields"/>). These are
3615   minimum recommendations, chosen to be supportable even by implementations
3616   with limited resources; it is expected that most implementations will
3617   choose substantially higher limits.
3618</t>
3619<t>
3620   This specification also provides a way for servers to reject messages that
3621   have request-targets that are too long (&status-414;) or request entities
3622   that are too large (&status-4xx;).
3623</t>
3624<t>
3625   Recipients &SHOULD; carefully limit the extent to which they read other
3626   fields, including (but not limited to) request methods, response status
3627   phrases, header field-names, and body chunks, so as to avoid denial of
3628   service attacks without impeding interoperability.
3629</t>
3630</section>
3631
3632<section title="Message Integrity" anchor="message.integrity">
3633<t>
3634   HTTP does not define a specific mechanism for ensuring message integrity,
3635   instead relying on the error-detection ability of underlying transport
3636   protocols and the use of length or chunk-delimited framing to detect
3637   completeness. Additional integrity mechanisms, such as hash functions or
3638   digital signatures applied to the content, can be selectively added to
3639   messages via extensible metadata header fields. Historically, the lack of
3640   a single integrity mechanism has been justified by the informal nature of
3641   most HTTP communication.  However, the prevalence of HTTP as an information
3642   access mechanism has resulted in its increasing use within environments
3643   where verification of message integrity is crucial.
3644</t>
3645<t>
3646   User agents are encouraged to implement configurable means for detecting
3647   and reporting failures of message integrity such that those means can be
3648   enabled within environments for which integrity is necessary. For example,
3649   a browser being used to view medical history or drug interaction
3650   information needs to indicate to the user when such information is detected
3651   by the protocol to be incomplete, expired, or corrupted during transfer.
3652   Such mechanisms might be selectively enabled via user agent extensions or
3653   the presence of message integrity metadata in a response.
3654   At a minimum, user agents ought to provide some indication that allows a
3655   user to distinguish between a complete and incomplete response message
3656   (<xref target="incomplete.messages"/>) when such verification is desired.
3657</t>
3658</section>
3659</section>
3660
3661<section title="Acknowledgments" anchor="acks">
3662<t>
3663   This edition of HTTP/1.1 builds on the many contributions that went into
3664   <xref target="RFC1945" format="none">RFC 1945</xref>,
3665   <xref target="RFC2068" format="none">RFC 2068</xref>,
3666   <xref target="RFC2145" format="none">RFC 2145</xref>, and
3667   <xref target="RFC2616" format="none">RFC 2616</xref>, including
3668   substantial contributions made by the previous authors, editors, and
3669   working group chairs: Tim Berners-Lee, Ari Luotonen, Roy T. Fielding,
3670   Henrik Frystyk Nielsen, Jim Gettys, Jeffrey C. Mogul, Larry Masinter,
3671   and Paul J. Leach. Mark Nottingham oversaw this effort as working group chair.
3672</t>
3673<t>
3674   Since 1999, the following contributors have helped improve the HTTP
3675   specification by reporting bugs, asking smart questions, drafting or
3676   reviewing text, and evaluating open issues:
3677</t>
3678<?BEGININC acks ?>
3679<t>Adam Barth,
3680Adam Roach,
3681Addison Phillips,
3682Adrian Chadd,
3683Adrien W. de Croy,
3684Alan Ford,
3685Alan Ruttenberg,
3686Albert Lunde,
3687Alek Storm,
3688Alex Rousskov,
3689Alexandre Morgaut,
3690Alexey Melnikov,
3691Alisha Smith,
3692Amichai Rothman,
3693Amit Klein,
3694Amos Jeffries,
3695Andreas Maier,
3696Andreas Petersson,
3697Anil Sharma,
3698Anne van Kesteren,
3699Anthony Bryan,
3700Asbjorn Ulsberg,
3701Ashok Kumar,
3702Balachander Krishnamurthy,
3703Barry Leiba,
3704Ben Laurie,
3705Benjamin Niven-Jenkins,
3706Bil Corry,
3707Bill Burke,
3708Bjoern Hoehrmann,
3709Bob Scheifler,
3710Boris Zbarsky,
3711Brett Slatkin,
3712Brian Kell,
3713Brian McBarron,
3714Brian Pane,
3715Brian Smith,
3716Bryce Nesbitt,
3717Cameron Heavon-Jones,
3718Carl Kugler,
3719Carsten Bormann,
3720Charles Fry,
3721Chris Newman,
3722Cyrus Daboo,
3723Dale Robert Anderson,
3724Dan Wing,
3725Dan Winship,
3726Daniel Stenberg,
3727Dave Cridland,
3728Dave Crocker,
3729Dave Kristol,
3730David Booth,
3731David Singer,
3732David W. Morris,
3733Diwakar Shetty,
3734Dmitry Kurochkin,
3735Drummond Reed,
3736Duane Wessels,
3737Edward Lee,
3738Eliot Lear,
3739Eran Hammer-Lahav,
3740Eric D. Williams,
3741Eric J. Bowman,
3742Eric Lawrence,
3743Eric Rescorla,
3744Erik Aronesty,
3745Evan Prodromou,
3746Florian Weimer,
3747Frank Ellermann,
3748Fred Bohle,
3749Gabriel Montenegro,
3750Geoffrey Sneddon,
3751Gervase Markham,
3752Grahame Grieve,
3753Greg Wilkins,
3754Harald Tveit Alvestrand,
3755Harry Halpin,
3756Helge Hess,
3757Henrik Nordstrom,
3758Henry S. Thompson,
3759Henry Story,
3760Herbert van de Sompel,
3761Howard Melman,
3762Hugo Haas,
3763Ian Fette,
3764Ian Hickson,
3765Ido Safruti,
3766Ilya Grigorik,
3767Ingo Struck,
3768J. Ross Nicoll,
3769James H. Manger,
3770James Lacey,
3771James M. Snell,
3772Jamie Lokier,
3773Jan Algermissen,
3774Jeff Hodges (who came up with the term 'effective Request-URI'),
3775Jeff Walden,
3776Jim Luther,
3777Joe D. Williams,
3778Joe Gregorio,
3779Joe Orton,
3780John C. Klensin,
3781John C. Mallery,
3782John Cowan,
3783John Kemp,
3784John Panzer,
3785John Schneider,
3786John Stracke,
3787John Sullivan,
3788Jonas Sicking,
3789Jonathan Billington,
3790Jonathan Moore,
3791Jonathan Rees,
3792Jonathan Silvera,
3793Jordi Ros,
3794Joris Dobbelsteen,
3795Josh Cohen,
3796Julien Pierre,
3797Jungshik Shin,
3798Justin Chapweske,
3799Justin Erenkrantz,
3800Justin James,
3801Kalvinder Singh,
3802Karl Dubost,
3803Keith Hoffman,
3804Keith Moore,
3805Ken Murchison,
3806Koen Holtman,
3807Konstantin Voronkov,
3808Kris Zyp,
3809Lisa Dusseault,
3810Maciej Stachowiak,
3811Marc Schneider,
3812Marc Slemko,
3813Mark Baker,
3814Mark Pauley,
3815Mark Watson,
3816Markus Isomaki,
3817Markus Lanthaler,
3818Martin J. Duerst,
3819Martin Musatov,
3820Martin Nilsson,
3821Martin Thomson,
3822Matt Lynch,
3823Matthew Cox,
3824Max Clark,
3825Michael Burrows,
3826Michael Hausenblas,
3827Mike Amundsen,
3828Mike Belshe,
3829Mike Kelly,
3830Mike Schinkel,
3831Miles Sabin,
3832Murray S. Kucherawy,
3833Mykyta Yevstifeyev,
3834Nathan Rixham,
3835Nicholas Shanks,
3836Nico Williams,
3837Nicolas Alvarez,
3838Nicolas Mailhot,
3839Noah Slater,
3840Pablo Castro,
3841Pat Hayes,
3842Patrick R. McManus,
3843Paul E. Jones,
3844Paul Hoffman,
3845Paul Marquess,
3846Peter Lepeska,
3847Peter Saint-Andre,
3848Peter Watkins,
3849Phil Archer,
3850Philippe Mougin,
3851Phillip Hallam-Baker,
3852Poul-Henning Kamp,
3853Preethi Natarajan,
3854Rajeev Bector,
3855Ray Polk,
3856Reto Bachmann-Gmuer,
3857Richard Cyganiak,
3858Robert Brewer,
3859Robert Collins,
3860Robert O'Callahan,
3861Robert Olofsson,
3862Robert Sayre,
3863Robert Siemer,
3864Robert de Wilde,
3865Roberto Javier Godoy,
3866Roberto Peon,
3867Roland Zink,
3868Ronny Widjaja,
3869S. Mike Dierken,
3870Salvatore Loreto,
3871Sam Johnston,
3872Sam Ruby,
3873Scott Lawrence (who maintained the original issues list),
3874Sean B. Palmer,
3875Shane McCarron,
3876Stefan Eissing,
3877Stefan Tilkov,
3878Stefanos Harhalakis,
3879Stephane Bortzmeyer,
3880Stephen Farrell,
3881Stephen Ludin,
3882Stuart Williams,
3883Subbu Allamaraju,
3884Sylvain Hellegouarch,
3885Tapan Divekar,
3886Tatsuya Hayashi,
3887Ted Hardie,
3888Thomas Broyer,
3889Thomas Fossati,
3890Thomas Nordin,
3891Thomas Roessler,
3892Tim Bray,
3893Tim Morgan,
3894Tim Olsen,
3895Tom Zhou,
3896Travis Snoozy,
3897Tyler Close,
3898Vincent Murphy,
3899Wenbo Zhu,
3900Werner Baumann,
3901Wilbur Streett,
3902Wilfredo Sanchez Vega,
3903William A. Rowe Jr.,
3904William Chan,
3905Willy Tarreau,
3906Xiaoshu Wang,
3907Yaron Goland,
3908Yngve Nysaeter Pettersen,
3909Yoav Nir,
3910Yogesh Bang,
3911Yutaka Oiwa,
3912Yves Lafon (long-time member of the editor team),
3913Zed A. Shaw, and
3914Zhong Yu.
3915</t>
3916<?ENDINC acks ?>
3917<t>
3918   See <xref target="RFC2616" x:fmt="of" x:sec="16"/> for additional
3919   acknowledgements from prior revisions.
3920</t>
3921</section>
3922
3923</middle>
3924<back>
3925
3926<references title="Normative References">
3927
3928<reference anchor="Part2">
3929  <front>
3930    <title>Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content</title>
3931    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
3932      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3933      <address><email>fielding@gbiv.com</email></address>
3934    </author>
3935    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
3936      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3937      <address><email>julian.reschke@greenbytes.de</email></address>
3938    </author>
3939    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
3940  </front>
3941  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p2-semantics-&ID-VERSION;"/>
3942  <x:source href="p2-semantics.xml" basename="p2-semantics">
3943    <x:defines>1xx (Informational)</x:defines>
3944    <x:defines>1xx</x:defines>
3945    <x:defines>100 (Continue)</x:defines>
3946    <x:defines>101 (Switching Protocols)</x:defines>
3947    <x:defines>2xx (Successful)</x:defines>
3948    <x:defines>2xx</x:defines>
3949    <x:defines>200 (OK)</x:defines>
3950    <x:defines>204 (No Content)</x:defines>
3951    <x:defines>3xx (Redirection)</x:defines>
3952    <x:defines>3xx</x:defines>
3953    <x:defines>301 (Moved Permanently)</x:defines>
3954    <x:defines>4xx (Client Error)</x:defines>
3955    <x:defines>4xx</x:defines>
3956    <x:defines>400 (Bad Request)</x:defines>
3957    <x:defines>405 (Method Not Allowed)</x:defines>
3958    <x:defines>411 (Length Required)</x:defines>
3959    <x:defines>414 (URI Too Long)</x:defines>
3960    <x:defines>417 (Expectation Failed)</x:defines>
3961    <x:defines>426 (Upgrade Required)</x:defines>
3962    <x:defines>501 (Not Implemented)</x:defines>
3963    <x:defines>502 (Bad Gateway)</x:defines>
3964    <x:defines>505 (HTTP Version Not Supported)</x:defines>
3965    <x:defines>Allow</x:defines>
3966    <x:defines>Content-Encoding</x:defines>
3967    <x:defines>Content-Location</x:defines>
3968    <x:defines>Content-Type</x:defines>
3969    <x:defines>Date</x:defines>
3970    <x:defines>Expect</x:defines>
3971    <x:defines>Location</x:defines>
3972    <x:defines>Server</x:defines>
3973    <x:defines>User-Agent</x:defines>
3974  </x:source>
3975</reference>
3976
3977<reference anchor="Part4">
3978  <front>
3979    <title>Hypertext Transfer Protocol (HTTP/1.1): Conditional Requests</title>
3980    <author fullname="Roy T. Fielding" initials="R." role="editor" surname="Fielding">
3981      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
3982      <address><email>fielding@gbiv.com</email></address>
3983    </author>
3984    <author fullname="Julian F. Reschke" initials="J. F." role="editor" surname="Reschke">
3985      <organization abbrev="greenbytes">greenbytes GmbH</organization>
3986      <address><email>julian.reschke@greenbytes.de</email></address>
3987    </author>
3988    <date month="&ID-MONTH;" year="&ID-YEAR;" />
3989  </front>
3990  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p4-conditional-&ID-VERSION;" />
3991  <x:source basename="p4-conditional" href="p4-conditional.xml">
3992    <x:defines>304 (Not Modified)</x:defines>
3993    <x:defines>ETag</x:defines>
3994    <x:defines>Last-Modified</x:defines>
3995  </x:source>
3996</reference>
3997
3998<reference anchor="Part5">
3999  <front>
4000    <title>Hypertext Transfer Protocol (HTTP/1.1): Range Requests</title>
4001    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4002      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4003      <address><email>fielding@gbiv.com</email></address>
4004    </author>
4005    <author initials="Y." surname="Lafon" fullname="Yves Lafon" role="editor">
4006      <organization abbrev="W3C">World Wide Web Consortium</organization>
4007      <address><email>ylafon@w3.org</email></address>
4008    </author>
4009    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4010      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4011      <address><email>julian.reschke@greenbytes.de</email></address>
4012    </author>
4013    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4014  </front>
4015  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p5-range-&ID-VERSION;"/>
4016  <x:source href="p5-range.xml" basename="p5-range">
4017    <x:defines>Content-Range</x:defines>
4018  </x:source>
4019</reference>
4020
4021<reference anchor="Part6">
4022  <front>
4023    <title>Hypertext Transfer Protocol (HTTP/1.1): Caching</title>
4024    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4025      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4026      <address><email>fielding@gbiv.com</email></address>
4027    </author>
4028    <author initials="M." surname="Nottingham" fullname="Mark Nottingham" role="editor">
4029      <organization>Akamai</organization>
4030      <address><email>mnot@mnot.net</email></address>
4031    </author>
4032    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4033      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4034      <address><email>julian.reschke@greenbytes.de</email></address>
4035    </author>
4036    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4037  </front>
4038  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p6-cache-&ID-VERSION;"/>
4039  <x:source href="p6-cache.xml" basename="p6-cache">
4040    <x:defines>Expires</x:defines>
4041  </x:source>
4042</reference>
4043
4044<reference anchor="Part7">
4045  <front>
4046    <title>Hypertext Transfer Protocol (HTTP/1.1): Authentication</title>
4047    <author initials="R." surname="Fielding" fullname="Roy T. Fielding" role="editor">
4048      <organization abbrev="Adobe">Adobe Systems Incorporated</organization>
4049      <address><email>fielding@gbiv.com</email></address>
4050    </author>
4051    <author initials="J. F." surname="Reschke" fullname="Julian F. Reschke" role="editor">
4052      <organization abbrev="greenbytes">greenbytes GmbH</organization>
4053      <address><email>julian.reschke@greenbytes.de</email></address>
4054    </author>
4055    <date month="&ID-MONTH;" year="&ID-YEAR;"/>
4056  </front>
4057  <seriesInfo name="Internet-Draft" value="draft-ietf-httpbis-p7-auth-&ID-VERSION;"/>
4058  <x:source href="p7-auth.xml" basename="p7-auth">
4059    <x:defines>Proxy-Authenticate</x:defines>
4060    <x:defines>Proxy-Authorization</x:defines>
4061  </x:source>
4062</reference>
4063
4064<reference anchor="RFC5234">
4065  <front>
4066    <title abbrev="ABNF for Syntax Specifications">Augmented BNF for Syntax Specifications: ABNF</title>
4067    <author initials="D." surname="Crocker" fullname="Dave Crocker" role="editor">
4068      <organization>Brandenburg InternetWorking</organization>
4069      <address>
4070        <email>dcrocker@bbiw.net</email>
4071      </address> 
4072    </author>
4073    <author initials="P." surname="Overell" fullname="Paul Overell">
4074      <organization>THUS plc.</organization>
4075      <address>
4076        <email>paul.overell@thus.net</email>
4077      </address>
4078    </author>
4079    <date month="January" year="2008"/>
4080  </front>
4081  <seriesInfo name="STD" value="68"/>
4082  <seriesInfo name="RFC" value="5234"/>
4083</reference>
4084
4085<reference anchor="RFC2119">
4086  <front>
4087    <title>Key words for use in RFCs to Indicate Requirement Levels</title>
4088    <author initials="S." surname="Bradner" fullname="Scott Bradner">
4089      <organization>Harvard University</organization>
4090      <address><email>sob@harvard.edu</email></address>
4091    </author>
4092    <date month="March" year="1997"/>
4093  </front>
4094  <seriesInfo name="BCP" value="14"/>
4095  <seriesInfo name="RFC" value="2119"/>
4096</reference>
4097
4098<reference anchor="RFC3986">
4099 <front>
4100  <title abbrev='URI Generic Syntax'>Uniform Resource Identifier (URI): Generic Syntax</title>
4101  <author initials='T.' surname='Berners-Lee' fullname='Tim Berners-Lee'>
4102    <organization abbrev="W3C/MIT">World Wide Web Consortium</organization>
4103    <address>
4104       <email>timbl@w3.org</email>
4105       <uri>http://www.w3.org/People/Berners-Lee/</uri>
4106    </address>
4107  </author>
4108  <author initials='R.' surname='Fielding' fullname='Roy T. Fielding'>
4109    <organization abbrev="Day Software">Day Software</organization>
4110    <address>
4111      <email>fielding@gbiv.com</email>
4112      <uri>http://roy.gbiv.com/</uri>
4113    </address>
4114  </author>
4115  <author initials='L.' surname='Masinter' fullname='Larry Masinter'>
4116    <organization abbrev="Adobe Systems">Adobe Systems Incorporated</organization>
4117    <address>
4118      <email>LMM@acm.org</email>
4119      <uri>http://larry.masinter.net/</uri>
4120    </address>
4121  </author>
4122  <date month='January' year='2005'></date>
4123 </front>
4124 <seriesInfo name="STD" value="66"/>
4125 <seriesInfo name="RFC" value="3986"/>
4126</reference>
4127
4128<reference anchor="USASCII">
4129  <front>
4130    <title>Coded Character Set -- 7-bit American Standard Code for Information Interchange</title>
4131    <author>
4132      <organization>American National Standards Institute</organization>
4133    </author>
4134    <date year="1986"/>
4135  </front>
4136  <seriesInfo name="ANSI" value="X3.4"/>
4137</reference>
4138
4139<reference anchor="RFC1950">
4140  <front>
4141    <title>ZLIB Compressed Data Format Specification version 3.3</title>
4142    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4143      <organization>Aladdin Enterprises</organization>
4144      <address><email>ghost@aladdin.com</email></address>
4145    </author>
4146    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly"/>
4147    <date month="May" year="1996"/>
4148  </front>
4149  <seriesInfo name="RFC" value="1950"/>
4150  <!--<annotation>
4151    RFC 1950 is an Informational RFC, thus it might be less stable than
4152    this specification. On the other hand, this downward reference was
4153    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4154    therefore it is unlikely to cause problems in practice. See also
4155    <xref target="BCP97"/>.
4156  </annotation>-->
4157</reference>
4158
4159<reference anchor="RFC1951">
4160  <front>
4161    <title>DEFLATE Compressed Data Format Specification version 1.3</title>
4162    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4163      <organization>Aladdin Enterprises</organization>
4164      <address><email>ghost@aladdin.com</email></address>
4165    </author>
4166    <date month="May" year="1996"/>
4167  </front>
4168  <seriesInfo name="RFC" value="1951"/>
4169  <!--<annotation>
4170    RFC 1951 is an Informational RFC, thus it might be less stable than
4171    this specification. On the other hand, this downward reference was
4172    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4173    therefore it is unlikely to cause problems in practice. See also
4174    <xref target="BCP97"/>.
4175  </annotation>-->
4176</reference>
4177
4178<reference anchor="RFC1952">
4179  <front>
4180    <title>GZIP file format specification version 4.3</title>
4181    <author initials="P." surname="Deutsch" fullname="L. Peter Deutsch">
4182      <organization>Aladdin Enterprises</organization>
4183      <address><email>ghost@aladdin.com</email></address>
4184    </author>
4185    <author initials="J-L." surname="Gailly" fullname="Jean-Loup Gailly">
4186      <address><email>gzip@prep.ai.mit.edu</email></address>
4187    </author>
4188    <author initials="M." surname="Adler" fullname="Mark Adler">
4189      <address><email>madler@alumni.caltech.edu</email></address>
4190    </author>
4191    <author initials="L.P." surname="Deutsch" fullname="L. Peter Deutsch">
4192      <address><email>ghost@aladdin.com</email></address>
4193    </author>
4194    <author initials="G." surname="Randers-Pehrson" fullname="Glenn Randers-Pehrson">
4195      <address><email>randeg@alumni.rpi.edu</email></address>
4196    </author>
4197    <date month="May" year="1996"/>
4198  </front>
4199  <seriesInfo name="RFC" value="1952"/>
4200  <!--<annotation>
4201    RFC 1952 is an Informational RFC, thus it might be less stable than
4202    this specification. On the other hand, this downward reference was
4203    present since the publication of <xref target="RFC2068" x:fmt="none">RFC 2068</xref> in 1997,
4204    therefore it is unlikely to cause problems in practice. See also
4205    <xref target="BCP97"/>.
4206  </annotation>-->
4207</reference>
4208
4209</references>
4210
4211<references title="Informative References">
4212
4213<reference anchor="ISO-8859-1">
4214  <front>
4215    <title>
4216     Information technology -- 8-bit single-byte coded graphic character sets -- Part 1: Latin alphabet No. 1
4217    </title>
4218    <author>
4219      <organization>International Organization for Standardization</organization>
4220    </author>
4221    <date year="1998"/>
4222  </front>
4223  <seriesInfo name="ISO/IEC" value="8859-1:1998"/>
4224</reference>
4225
4226<reference anchor='RFC1919'>
4227  <front>
4228    <title>Classical versus Transparent IP Proxies</title>
4229    <author initials='M.' surname='Chatel' fullname='Marc Chatel'>
4230      <address><email>mchatel@pax.eunet.ch</email></address>
4231    </author>
4232    <date year='1996' month='March' />
4233  </front>
4234  <seriesInfo name='RFC' value='1919' />
4235</reference>
4236
4237<reference anchor="RFC1945">
4238  <front>
4239    <title abbrev="HTTP/1.0">Hypertext Transfer Protocol -- HTTP/1.0</title>
4240    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4241      <organization>MIT, Laboratory for Computer Science</organization>
4242      <address><email>timbl@w3.org</email></address>
4243    </author>
4244    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4245      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4246      <address><email>fielding@ics.uci.edu</email></address>
4247    </author>
4248    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4249      <organization>W3 Consortium, MIT Laboratory for Computer Science</organization>
4250      <address><email>frystyk@w3.org</email></address>
4251    </author>
4252    <date month="May" year="1996"/>
4253  </front>
4254  <seriesInfo name="RFC" value="1945"/>
4255</reference>
4256
4257<reference anchor="RFC2045">
4258  <front>
4259    <title abbrev="Internet Message Bodies">Multipurpose Internet Mail Extensions (MIME) Part One: Format of Internet Message Bodies</title>
4260    <author initials="N." surname="Freed" fullname="Ned Freed">
4261      <organization>Innosoft International, Inc.</organization>
4262      <address><email>ned@innosoft.com</email></address>
4263    </author>
4264    <author initials="N.S." surname="Borenstein" fullname="Nathaniel S. Borenstein">
4265      <organization>First Virtual Holdings</organization>
4266      <address><email>nsb@nsb.fv.com</email></address>
4267    </author>
4268    <date month="November" year="1996"/>
4269  </front>
4270  <seriesInfo name="RFC" value="2045"/>
4271</reference>
4272
4273<reference anchor="RFC2047">
4274  <front>
4275    <title abbrev="Message Header Extensions">MIME (Multipurpose Internet Mail Extensions) Part Three: Message Header Extensions for Non-ASCII Text</title>
4276    <author initials="K." surname="Moore" fullname="Keith Moore">
4277      <organization>University of Tennessee</organization>
4278      <address><email>moore@cs.utk.edu</email></address>
4279    </author>
4280    <date month="November" year="1996"/>
4281  </front>
4282  <seriesInfo name="RFC" value="2047"/>
4283</reference>
4284
4285<reference anchor="RFC2068">
4286  <front>
4287    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4288    <author initials="R." surname="Fielding" fullname="Roy T. Fielding">
4289      <organization>University of California, Irvine, Department of Information and Computer Science</organization>
4290      <address><email>fielding@ics.uci.edu</email></address>
4291    </author>
4292    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4293      <organization>MIT Laboratory for Computer Science</organization>
4294      <address><email>jg@w3.org</email></address>
4295    </author>
4296    <author initials="J." surname="Mogul" fullname="Jeffrey C. Mogul">
4297      <organization>Digital Equipment Corporation, Western Research Laboratory</organization>
4298      <address><email>mogul@wrl.dec.com</email></address>
4299    </author>
4300    <author initials="H." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4301      <organization>MIT Laboratory for Computer Science</organization>
4302      <address><email>frystyk@w3.org</email></address>
4303    </author>
4304    <author initials="T." surname="Berners-Lee" fullname="Tim Berners-Lee">
4305      <organization>MIT Laboratory for Computer Science</organization>
4306      <address><email>timbl@w3.org</email></address>
4307    </author>
4308    <date month="January" year="1997"/>
4309  </front>
4310  <seriesInfo name="RFC" value="2068"/>
4311</reference>
4312
4313<reference anchor="RFC2145">
4314  <front>
4315    <title abbrev="HTTP Version Numbers">Use and Interpretation of HTTP Version Numbers</title>
4316    <author initials="J.C." surname="Mogul" fullname="Jeffrey C. Mogul">
4317      <organization>Western Research Laboratory</organization>
4318      <address><email>mogul@wrl.dec.com</email></address>
4319    </author>
4320    <author initials="R.T." surname="Fielding" fullname="Roy T. Fielding">
4321      <organization>Department of Information and Computer Science</organization>
4322      <address><email>fielding@ics.uci.edu</email></address>
4323    </author>
4324    <author initials="J." surname="Gettys" fullname="Jim Gettys">
4325      <organization>MIT Laboratory for Computer Science</organization>
4326      <address><email>jg@w3.org</email></address>
4327    </author>
4328    <author initials="H.F." surname="Nielsen" fullname="Henrik Frystyk Nielsen">
4329      <organization>W3 Consortium</organization>
4330      <address><email>frystyk@w3.org</email></address>
4331    </author>
4332    <date month="May" year="1997"/>
4333  </front>
4334  <seriesInfo name="RFC" value="2145"/>
4335</reference>
4336
4337<reference anchor="RFC2616">
4338  <front>
4339    <title>Hypertext Transfer Protocol -- HTTP/1.1</title>
4340    <author initials="R." surname="Fielding" fullname="R. Fielding">
4341      <organization>University of California, Irvine</organization>
4342      <address><email>fielding@ics.uci.edu</email></address>
4343    </author>
4344    <author initials="J." surname="Gettys" fullname="J. Gettys">
4345      <organization>W3C</organization>
4346      <address><email>jg@w3.org</email></address>
4347    </author>
4348    <author initials="J." surname="Mogul" fullname="J. Mogul">
4349      <organization>Compaq Computer Corporation</organization>
4350      <address><email>mogul@wrl.dec.com</email></address>
4351    </author>
4352    <author initials="H." surname="Frystyk" fullname="H. Frystyk">
4353      <organization>MIT Laboratory for Computer Science</organization>
4354      <address><email>frystyk@w3.org</email></address>
4355    </author>
4356    <author initials="L." surname="Masinter" fullname="L. Masinter">
4357      <organization>Xerox Corporation</organization>
4358      <address><email>masinter@parc.xerox.com</email></address>
4359    </author>
4360    <author initials="P." surname="Leach" fullname="P. Leach">
4361      <organization>Microsoft Corporation</organization>
4362      <address><email>paulle@microsoft.com</email></address>
4363    </author>
4364    <author initials="T." surname="Berners-Lee" fullname="T. Berners-Lee">
4365      <organization>W3C</organization>
4366      <address><email>timbl@w3.org</email></address>
4367    </author>
4368    <date month="June" year="1999"/>
4369  </front>
4370  <seriesInfo name="RFC" value="2616"/>
4371</reference>
4372
4373<reference anchor='RFC2817'>
4374  <front>
4375    <title>Upgrading to TLS Within HTTP/1.1</title>
4376    <author initials='R.' surname='Khare' fullname='R. Khare'>
4377      <organization>4K Associates / UC Irvine</organization>
4378      <address><email>rohit@4K-associates.com</email></address>
4379    </author>
4380    <author initials='S.' surname='Lawrence' fullname='S. Lawrence'>
4381      <organization>Agranat Systems, Inc.</organization>
4382      <address><email>lawrence@agranat.com</email></address>
4383    </author>
4384    <date year='2000' month='May' />
4385  </front>
4386  <seriesInfo name='RFC' value='2817' />
4387</reference>
4388
4389<reference anchor='RFC2818'>
4390  <front>
4391    <title>HTTP Over TLS</title>
4392    <author initials='E.' surname='Rescorla' fullname='Eric Rescorla'>
4393      <organization>RTFM, Inc.</organization>
4394      <address><email>ekr@rtfm.com</email></address>
4395    </author>
4396    <date year='2000' month='May' />
4397  </front>
4398  <seriesInfo name='RFC' value='2818' />
4399</reference>
4400
4401<reference anchor='RFC3040'>
4402  <front>
4403    <title>Internet Web Replication and Caching Taxonomy</title>
4404    <author initials='I.' surname='Cooper' fullname='I. Cooper'>
4405      <organization>Equinix, Inc.</organization>
4406    </author>
4407    <author initials='I.' surname='Melve' fullname='I. Melve'>
4408      <organization>UNINETT</organization>
4409    </author>
4410    <author initials='G.' surname='Tomlinson' fullname='G. Tomlinson'>
4411      <organization>CacheFlow Inc.</organization>
4412    </author>
4413    <date year='2001' month='January' />
4414  </front>
4415  <seriesInfo name='RFC' value='3040' />
4416</reference>
4417
4418<reference anchor='RFC3864'>
4419  <front>
4420    <title>Registration Procedures for Message Header Fields</title>
4421    <author initials='G.' surname='Klyne' fullname='G. Klyne'>
4422      <organization>Nine by Nine</organization>
4423      <address><email>GK-IETF@ninebynine.org</email></address>
4424    </author>
4425    <author initials='M.' surname='Nottingham' fullname='M. Nottingham'>
4426      <organization>BEA Systems</organization>
4427      <address><email>mnot@pobox.com</email></address>
4428    </author>
4429    <author initials='J.' surname='Mogul' fullname='J. Mogul'>
4430      <organization>HP Labs</organization>
4431      <address><email>JeffMogul@acm.org</email></address>
4432    </author>
4433    <date year='2004' month='September' />
4434  </front>
4435  <seriesInfo name='BCP' value='90' />
4436  <seriesInfo name='RFC' value='3864' />
4437</reference>
4438
4439<reference anchor='RFC4033'>
4440  <front>
4441    <title>DNS Security Introduction and Requirements</title>
4442    <author initials='R.' surname='Arends' fullname='R. Arends'/>
4443    <author initials='R.' surname='Austein' fullname='R. Austein'/>
4444    <author initials='M.' surname='Larson' fullname='M. Larson'/>
4445    <author initials='D.' surname='Massey' fullname='D. Massey'/>
4446    <author initials='S.' surname='Rose' fullname='S. Rose'/>
4447    <date year='2005' month='March' />
4448  </front>
4449  <seriesInfo name='RFC' value='4033' />
4450</reference>
4451
4452<reference anchor="RFC4288">
4453  <front>
4454    <title>Media Type Specifications and Registration Procedures</title>
4455    <author initials="N." surname="Freed" fullname="N. Freed">
4456      <organization>Sun Microsystems</organization>
4457      <address>
4458        <email>ned.freed@mrochek.com</email>
4459      </address>
4460    </author>
4461    <author initials="J." surname="Klensin" fullname="J. Klensin">
4462      <address>
4463        <email>klensin+ietf@jck.com</email>
4464      </address>
4465    </author>
4466    <date year="2005" month="December"/>
4467  </front>
4468  <seriesInfo name="BCP" value="13"/>
4469  <seriesInfo name="RFC" value="4288"/>
4470</reference>
4471
4472<reference anchor='RFC4395'>
4473  <front>
4474    <title>Guidelines and Registration Procedures for New URI Schemes</title>
4475    <author initials='T.' surname='Hansen' fullname='T. Hansen'>
4476      <organization>AT&amp;T Laboratories</organization>
4477      <address>
4478        <email>tony+urireg@maillennium.att.com</email>
4479      </address>
4480    </author>
4481    <author initials='T.' surname='Hardie' fullname='T. Hardie'>
4482      <organization>Qualcomm, Inc.</organization>
4483      <address>
4484        <email>hardie@qualcomm.com</email>
4485      </address>
4486    </author>
4487    <author initials='L.' surname='Masinter' fullname='L. Masinter'>
4488      <organization>Adobe Systems</organization>
4489      <address>
4490        <email>LMM@acm.org</email>
4491      </address>
4492    </author>
4493    <date year='2006' month='February' />
4494  </front>
4495  <seriesInfo name='BCP' value='115' />
4496  <seriesInfo name='RFC' value='4395' />
4497</reference>
4498
4499<reference anchor='RFC4559'>
4500  <front>
4501    <title>SPNEGO-based Kerberos and NTLM HTTP Authentication in Microsoft Windows</title>
4502    <author initials='K.' surname='Jaganathan' fullname='K. Jaganathan'/>
4503    <author initials='L.' surname='Zhu' fullname='L. Zhu'/>
4504    <author initials='J.' surname='Brezak' fullname='J. Brezak'/>
4505    <date year='2006' month='June' />
4506  </front>
4507  <seriesInfo name='RFC' value='4559' />
4508</reference>
4509
4510<reference anchor='RFC5226'>
4511  <front>
4512    <title>Guidelines for Writing an IANA Considerations Section in RFCs</title>
4513    <author initials='T.' surname='Narten' fullname='T. Narten'>
4514      <organization>IBM</organization>
4515      <address><email>narten@us.ibm.com</email></address>
4516    </author>
4517    <author initials='H.' surname='Alvestrand' fullname='H. Alvestrand'>
4518      <organization>Google</organization>
4519      <address><email>Harald@Alvestrand.no</email></address>
4520    </author>
4521    <date year='2008' month='May' />
4522  </front>
4523  <seriesInfo name='BCP' value='26' />
4524  <seriesInfo name='RFC' value='5226' />
4525</reference>
4526
4527<reference anchor='RFC5246'>
4528   <front>
4529      <title>The Transport Layer Security (TLS) Protocol Version 1.2</title>
4530      <author initials='T.' surname='Dierks' fullname='T. Dierks'>
4531         <organization />
4532      </author>
4533      <author initials='E.' surname='Rescorla' fullname='E. Rescorla'>
4534         <organization>RTFM, Inc.</organization>
4535      </author>
4536      <date year='2008' month='August' />
4537   </front>
4538   <seriesInfo name='RFC' value='5246' />
4539</reference>
4540
4541<reference anchor="RFC5322">
4542  <front>
4543    <title>Internet Message Format</title>
4544    <author initials="P." surname="Resnick" fullname="P. Resnick">
4545      <organization>Qualcomm Incorporated</organization>
4546    </author>
4547    <date year="2008" month="October"/>
4548  </front> 
4549  <seriesInfo name="RFC" value="5322"/>
4550</reference>
4551
4552<reference anchor="RFC6265">
4553  <front>
4554    <title>HTTP State Management Mechanism</title>
4555    <author initials="A." surname="Barth" fullname="Adam Barth">
4556      <organization abbrev="U.C. Berkeley">
4557        University of California, Berkeley
4558      </organization>
4559      <address><email>abarth@eecs.berkeley.edu</email></address>
4560    </author>
4561    <date year="2011" month="April" />
4562  </front>
4563  <seriesInfo name="RFC" value="6265"/>
4564</reference>
4565
4566<!--<reference anchor='BCP97'>
4567  <front>
4568    <title>Handling Normative References to Standards-Track Documents</title>
4569    <author initials='J.' surname='Klensin' fullname='J. Klensin'>
4570      <address>
4571        <email>klensin+ietf@jck.com</email>
4572      </address>
4573    </author>
4574    <author initials='S.' surname='Hartman' fullname='S. Hartman'>
4575      <organization>MIT</organization>
4576      <address>
4577        <email>hartmans-ietf@mit.edu</email>
4578      </address>
4579    </author>
4580    <date year='2007' month='June' />
4581  </front>
4582  <seriesInfo name='BCP' value='97' />
4583  <seriesInfo name='RFC' value='4897' />
4584</reference>-->
4585
4586<reference anchor="Kri2001" target="http://arxiv.org/abs/cs.SE/0105018">
4587  <front>
4588    <title>HTTP Cookies: Standards, Privacy, and Politics</title>
4589    <author initials="D." surname="Kristol" fullname="David M. Kristol"/>
4590    <date year="2001" month="November"/>
4591  </front>
4592  <seriesInfo name="ACM Transactions on Internet Technology" value="Vol. 1, #2"/>
4593</reference>
4594
4595</references>
4596
4597
4598<section title="HTTP Version History" anchor="compatibility">
4599<t>
4600   HTTP has been in use by the World-Wide Web global information initiative
4601   since 1990. The first version of HTTP, later referred to as HTTP/0.9,
4602   was a simple protocol for hypertext data transfer across the Internet
4603   with only a single request method (GET) and no metadata.
4604   HTTP/1.0, as defined by <xref target="RFC1945"/>, added a range of request
4605   methods and MIME-like messaging that could include metadata about the data
4606   transferred and modifiers on the request/response semantics. However,
4607   HTTP/1.0 did not sufficiently take into consideration the effects of
4608   hierarchical proxies, caching, the need for persistent connections, or
4609   name-based virtual hosts. The proliferation of incompletely-implemented
4610   applications calling themselves "HTTP/1.0" further necessitated a
4611   protocol version change in order for two communicating applications
4612   to determine each other's true capabilities.
4613</t>
4614<t>
4615   HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent
4616   requirements that enable reliable implementations, adding only
4617   those new features that will either be safely ignored by an HTTP/1.0
4618   recipient or only sent when communicating with a party advertising
4619   conformance with HTTP/1.1.
4620</t>
4621<t>
4622   It is beyond the scope of a protocol specification to mandate
4623   conformance with previous versions. HTTP/1.1 was deliberately
4624   designed, however, to make supporting previous versions easy.
4625   We would expect a general-purpose HTTP/1.1 server to understand
4626   any valid request in the format of HTTP/1.0 and respond appropriately
4627   with an HTTP/1.1 message that only uses features understood (or
4628   safely ignored) by HTTP/1.0 clients.  Likewise, we would expect
4629   an HTTP/1.1 client to understand any valid HTTP/1.0 response.
4630</t>
4631<t>
4632   Since HTTP/0.9 did not support header fields in a request,
4633   there is no mechanism for it to support name-based virtual
4634   hosts (selection of resource by inspection of the <x:ref>Host</x:ref> header
4635   field).  Any server that implements name-based virtual hosts
4636   ought to disable support for HTTP/0.9.  Most requests that
4637   appear to be HTTP/0.9 are, in fact, badly constructed HTTP/1.x
4638   requests wherein a buggy client failed to properly encode
4639   linear whitespace found in a URI reference and placed in
4640   the request-target.
4641</t>
4642
4643<section title="Changes from HTTP/1.0" anchor="changes.from.1.0">
4644<t>
4645   This section summarizes major differences between versions HTTP/1.0
4646   and HTTP/1.1.
4647</t>
4648
4649<section title="Multi-homed Web Servers" anchor="changes.to.simplify.multi-homed.web.servers.and.conserve.ip.addresses">
4650<t>
4651   The requirements that clients and servers support the <x:ref>Host</x:ref>
4652   header field (<xref target="header.host"/>), report an error if it is
4653   missing from an HTTP/1.1 request, and accept absolute URIs (<xref target="request-target"/>)
4654   are among the most important changes defined by HTTP/1.1.
4655</t>
4656<t>
4657   Older HTTP/1.0 clients assumed a one-to-one relationship of IP
4658   addresses and servers; there was no other established mechanism for
4659   distinguishing the intended server of a request than the IP address
4660   to which that request was directed. The <x:ref>Host</x:ref> header field was
4661   introduced during the development of HTTP/1.1 and, though it was
4662   quickly implemented by most HTTP/1.0 browsers, additional requirements
4663   were placed on all HTTP/1.1 requests in order to ensure complete
4664   adoption.  At the time of this writing, most HTTP-based services
4665   are dependent upon the Host header field for targeting requests.
4666</t>
4667</section>
4668
4669<section title="Keep-Alive Connections" anchor="compatibility.with.http.1.0.persistent.connections">
4670<t>
4671   In HTTP/1.0, each connection is established by the client prior to the
4672   request and closed by the server after sending the response. However, some
4673   implementations implement the explicitly negotiated ("Keep-Alive") version
4674   of persistent connections described in <xref x:sec="19.7.1" x:fmt="of"
4675   target="RFC2068"/>.
4676</t>
4677<t>
4678   Some clients and servers might wish to be compatible with these previous
4679   approaches to persistent connections, by explicitly negotiating for them
4680   with a "Connection: keep-alive" request header field. However, some
4681   experimental implementations of HTTP/1.0 persistent connections are faulty;
4682   for example, if an HTTP/1.0 proxy server doesn't understand
4683   <x:ref>Connection</x:ref>, it will erroneously forward that header field
4684   to the next inbound server, which would result in a hung connection.
4685</t>
4686<t>
4687   One attempted solution was the introduction of a Proxy-Connection header
4688   field, targeted specifically at proxies. In practice, this was also
4689   unworkable, because proxies are often deployed in multiple layers, bringing
4690   about the same problem discussed above.
4691</t>
4692<t>
4693   As a result, clients are encouraged not to send the Proxy-Connection header
4694   field in any requests.
4695</t>
4696<t>
4697   Clients are also encouraged to consider the use of Connection: keep-alive
4698   in requests carefully; while they can enable persistent connections with
4699   HTTP/1.0 servers, clients using them need will need to monitor the
4700   connection for "hung" requests (which indicate that the client ought stop
4701   sending the header field), and this mechanism ought not be used by clients
4702   at all when a proxy is being used.
4703</t>
4704</section>
4705
4706<section title="Introduction of Transfer-Encoding" anchor="introduction.of.transfer-encoding">
4707<t>
4708   HTTP/1.1 introduces the <x:ref>Transfer-Encoding</x:ref> header field
4709   (<xref target="header.transfer-encoding"/>). Proxies/gateways &MUST; remove
4710   any transfer coding prior to forwarding a message via a MIME-compliant
4711   protocol.
4712</t>
4713</section>
4714
4715</section>
4716
4717<section title="Changes from RFC 2616" anchor="changes.from.rfc.2616">
4718<t>
4719  HTTP's approach to error handling has been explained.
4720  (<xref target="conformance"/>)
4721</t>
4722<t>
4723  The expectation to support HTTP/0.9 requests has been removed.
4724</t>
4725<t>
4726  The term "Effective Request URI" has been introduced.
4727  (<xref target="effective.request.uri" />)
4728</t>
4729<t>
4730  HTTP messages can be (and often are) buffered by implementations; despite
4731  it sometimes being available as a stream, HTTP is fundamentally a
4732  message-oriented protocol.
4733  (<xref target="http.message" />)
4734</t>
4735<t>
4736  Minimum supported sizes for various protocol elements have been
4737  suggested, to improve interoperability.
4738</t>
4739<t>
4740  Header fields that span multiple lines ("line folding") are deprecated.
4741  (<xref target="field.parsing" />)
4742</t>
4743<t>
4744  The HTTP-version ABNF production has been clarified to be case-sensitive.
4745  Additionally, version numbers has been restricted to single digits, due
4746  to the fact that implementations are known to handle multi-digit version
4747  numbers incorrectly.
4748  (<xref target="http.version"/>)
4749</t>
4750<t>
4751  The HTTPS URI scheme is now defined by this specification; previously,
4752  it was done in  <xref target="RFC2818" x:fmt="of" x:sec="2.4"/>.
4753  (<xref target="https.uri"/>)
4754</t>
4755<t>
4756  The HTTPS URI scheme implies end-to-end security.
4757  (<xref target="https.uri"/>)
4758</t>
4759<t>
4760  Userinfo (i.e., username and password) are now disallowed in HTTP and
4761  HTTPS URIs, because of security issues related to their transmission on the
4762  wire.
4763  (<xref target="http.uri" />)
4764</t>
4765<t>
4766  Invalid whitespace around field-names is now required to be rejected,
4767  because accepting it represents a security vulnerability.
4768  (<xref target="header.fields"/>)
4769</t>
4770<t>
4771  The ABNF productions defining header fields now only list the field value.
4772  (<xref target="header.fields"/>)
4773</t>
4774<t>
4775  Rules about implicit linear whitespace between certain grammar productions
4776  have been removed; now whitespace is only allowed where specifically
4777  defined in the ABNF.
4778  (<xref target="whitespace"/>)
4779</t>
4780<t> 
4781  The NUL octet is no longer allowed in comment and quoted-string text, and
4782  handling of backslash-escaping in them has been clarified.
4783  (<xref target="field.components"/>)
4784</t> 
4785<t>
4786  The quoted-pair rule no longer allows escaping control characters other than
4787  HTAB.
4788  (<xref target="field.components"/>)
4789</t>
4790<t>
4791  Non-ASCII content in header fields and the reason phrase has been obsoleted
4792  and made opaque (the TEXT rule was removed).
4793  (<xref target="field.components"/>)
4794</t>
4795<t>
4796  Bogus "<x:ref>Content-Length</x:ref>" header fields are now required to be
4797  handled as errors by recipients.
4798  (<xref target="header.content-length"/>)
4799</t>
4800<t>
4801  The "identity" transfer coding token has been removed.
4802  (Sections <xref format="counter" target="message.body"/> and
4803  <xref format="counter" target="transfer.codings"/>)
4804</t>
4805<t>
4806  The algorithm for determining the message body length has been clarified
4807  to indicate all of the special cases (e.g., driven by methods or status
4808  codes) that affect it, and that new protocol elements cannot define such
4809  special cases.
4810  (<xref target="message.body.length"/>)
4811</t>
4812<t>
4813  "multipart/byteranges" is no longer a way of determining message body length
4814  detection.
4815  (<xref target="message.body.length"/>)
4816</t>
4817<t>
4818  CONNECT is a new, special case in determining message body length.
4819  (<xref target="message.body.length"/>)
4820</t>
4821<t>
4822  Chunk length does not include the count of the octets in the
4823  chunk header and trailer.
4824  (<xref target="chunked.encoding"/>)
4825</t>
4826<t>
4827  Use of chunk extensions is deprecated, and line folding in them is
4828  disallowed.
4829  (<xref target="chunked.encoding"/>)
4830</t>
4831<t>
4832  The path-absolute + query components of RFC3986 have been used to define the
4833  request-target, instead of abs_path from RFC 1808.
4834  (<xref target="request-target"/>)
4835</t>
4836<t>
4837  The asterisk form of the request-target is only allowed in the OPTIONS
4838  method.
4839  (<xref target="request-target"/>)
4840</t>
4841<t>
4842  Exactly when "close" connection options have to be sent has been clarified.
4843  (<xref target="header.connection"/>)
4844</t>
4845<t> 
4846  "hop-by-hop" header fields are required to appear in the Connection header
4847  field; just because they're defined as hop-by-hop in this specification
4848  doesn't exempt them.
4849  (<xref target="header.connection"/>)
4850</t>
4851<t>
4852  The limit of two connections per server has been removed.
4853  (<xref target="persistent.connections"/>)
4854</t>
4855<t>
4856  An idempotent sequence of requests is no longer required to be retried.
4857  (<xref target="persistent.connections"/>)
4858</t>
4859<t>
4860  The requirement to retry requests under certain circumstances when the
4861  server prematurely closes the connection has been removed.
4862  (<xref target="persistent.connections"/>)
4863</t>
4864<t>
4865  Some extraneous requirements about when servers are allowed to close
4866  connections prematurely have been removed.
4867  (<xref target="persistent.connections"/>)
4868</t>
4869<t>
4870  The semantics of the <x:ref>Upgrade</x:ref> header field is now defined in
4871  responses other than 101 (this was incorporated from <xref
4872  target="RFC2817"/>).
4873  (<xref target="header.upgrade"/>)
4874</t>
4875<t>
4876  Registration of Transfer Codings now requires IETF Review
4877  (<xref target="transfer.coding.registry"/>)
4878</t>
4879<t>
4880  The meaning of the "deflate" content coding has been clarified.
4881  (<xref target="deflate.coding" />)
4882</t>
4883<t>
4884  This specification now defines the Upgrade Token Registry, previously
4885  defined in <xref target="RFC2817" x:fmt="of" x:sec="7.2"/>.
4886  (<xref target="upgrade.token.registry"/>)
4887</t>
4888<t>
4889  Empty list elements in list productions (e.g., a list header containing
4890  ", ,") have been deprecated.
4891  (<xref target="abnf.extension"/>)
4892</t>
4893<t>
4894  Issues with the Keep-Alive and Proxy-Connection headers in requests
4895  are pointed out, with use of the latter being discouraged altogether.
4896  (<xref target="compatibility.with.http.1.0.persistent.connections" />)
4897</t>
4898</section>
4899</section>
4900
4901<section title="ABNF list extension: #rule" anchor="abnf.extension">
4902<t>
4903  A #rule extension to the ABNF rules of <xref target="RFC5234"/> is used to
4904  improve readability in the definitions of some header field values.
4905</t>
4906<t>
4907  A construct "#" is defined, similar to "*", for defining comma-delimited
4908  lists of elements. The full form is "&lt;n&gt;#&lt;m&gt;element" indicating
4909  at least &lt;n&gt; and at most &lt;m&gt; elements, each separated by a single
4910  comma (",") and optional whitespace (OWS).   
4911</t>
4912<figure><preamble>
4913  Thus,
4914</preamble><artwork type="example">
4915  1#element =&gt; element *( OWS "," OWS element )
4916</artwork></figure>
4917<figure><preamble>
4918  and:
4919</preamble><artwork type="example">
4920  #element =&gt; [ 1#element ]
4921</artwork></figure>
4922<figure><preamble>
4923  and for n &gt;= 1 and m &gt; 1:
4924</preamble><artwork type="example">
4925  &lt;n&gt;#&lt;m&gt;element =&gt; element &lt;n-1&gt;*&lt;m-1&gt;( OWS "," OWS element )
4926</artwork></figure>
4927<t>
4928  For compatibility with legacy list rules, recipients &SHOULD; accept empty
4929  list elements. In other words, consumers would follow the list productions:
4930</t>
4931<figure><artwork type="example">
4932  #element =&gt; [ ( "," / element ) *( OWS "," [ OWS element ] ) ]
4933 
4934  1#element =&gt; *( "," OWS ) element *( OWS "," [ OWS element ] )
4935</artwork></figure>
4936<t>
4937  Note that empty elements do not contribute to the count of elements present,
4938  though.
4939</t>
4940<t>
4941  For example, given these ABNF productions:
4942</t>
4943<figure><artwork type="example">
4944  example-list      = 1#example-list-elmt
4945  example-list-elmt = token ; see <xref target="field.components"/> 
4946</artwork></figure>
4947<t>
4948  Then these are valid values for example-list (not including the double
4949  quotes, which are present for delimitation only):
4950</t>
4951<figure><artwork type="example">
4952  "foo,bar"
4953  "foo ,bar,"
4954  "foo , ,bar,charlie   "
4955</artwork></figure>
4956<t>
4957  But these values would be invalid, as at least one non-empty element is
4958  required:
4959</t>
4960<figure><artwork type="example">
4961  ""
4962  ","
4963  ",   ,"
4964</artwork></figure>
4965<t>
4966  <xref target="collected.abnf"/> shows the collected ABNF, with the list rules
4967  expanded as explained above.
4968</t>
4969</section>
4970
4971<?BEGININC p1-messaging.abnf-appendix ?>
4972<section xmlns:x="http://purl.org/net/xml2rfc/ext" title="Collected ABNF" anchor="collected.abnf">
4973<figure>
4974<artwork type="abnf" name="p1-messaging.parsed-abnf">
4975<x:ref>BWS</x:ref> = OWS
4976
4977<x:ref>Connection</x:ref> = *( "," OWS ) connection-option *( OWS "," [ OWS
4978 connection-option ] )
4979<x:ref>Content-Length</x:ref> = 1*DIGIT
4980
4981<x:ref>HTTP-message</x:ref> = start-line *( header-field CRLF ) CRLF [ message-body
4982 ]
4983<x:ref>HTTP-name</x:ref> = %x48.54.54.50 ; HTTP
4984<x:ref>HTTP-version</x:ref> = HTTP-name "/" DIGIT "." DIGIT
4985<x:ref>Host</x:ref> = uri-host [ ":" port ]
4986
4987<x:ref>OWS</x:ref> = *( SP / HTAB )
4988
4989<x:ref>RWS</x:ref> = 1*( SP / HTAB )
4990
4991<x:ref>TE</x:ref> = [ ( "," / t-codings ) *( OWS "," [ OWS t-codings ] ) ]
4992<x:ref>Trailer</x:ref> = *( "," OWS ) field-name *( OWS "," [ OWS field-name ] )
4993<x:ref>Transfer-Encoding</x:ref> = *( "," OWS ) transfer-coding *( OWS "," [ OWS
4994 transfer-coding ] )
4995
4996<x:ref>URI-reference</x:ref> = &lt;URI-reference, defined in [RFC3986], Section 4.1&gt;
4997<x:ref>Upgrade</x:ref> = *( "," OWS ) protocol *( OWS "," [ OWS protocol ] )
4998
4999<x:ref>Via</x:ref> = *( "," OWS ) ( received-protocol RWS received-by [ RWS comment
5000 ] ) *( OWS "," [ OWS ( received-protocol RWS received-by [ RWS
5001 comment ] ) ] )
5002
5003<x:ref>absolute-URI</x:ref> = &lt;absolute-URI, defined in [RFC3986], Section 4.3&gt;
5004<x:ref>absolute-form</x:ref> = absolute-URI
5005<x:ref>asterisk-form</x:ref> = "*"
5006<x:ref>attribute</x:ref> = token
5007<x:ref>authority</x:ref> = &lt;authority, defined in [RFC3986], Section 3.2&gt;
5008<x:ref>authority-form</x:ref> = authority
5009
5010<x:ref>chunk</x:ref> = chunk-size [ chunk-ext ] CRLF chunk-data CRLF
5011<x:ref>chunk-data</x:ref> = 1*OCTET
5012<x:ref>chunk-ext</x:ref> = *( ";" chunk-ext-name [ "=" chunk-ext-val ] )
5013<x:ref>chunk-ext-name</x:ref> = token
5014<x:ref>chunk-ext-val</x:ref> = token / quoted-str-nf
5015<x:ref>chunk-size</x:ref> = 1*HEXDIG
5016<x:ref>chunked-body</x:ref> = *chunk last-chunk trailer-part CRLF
5017<x:ref>comment</x:ref> = "(" *( ctext / quoted-cpair / comment ) ")"
5018<x:ref>connection-option</x:ref> = token
5019<x:ref>ctext</x:ref> = HTAB / SP / %x21-27 ; '!'-'''
5020 / %x2A-5B ; '*'-'['
5021 / %x5D-7E ; ']'-'~'
5022 / obs-text
5023
5024<x:ref>field-content</x:ref> = *( HTAB / SP / VCHAR / obs-text )
5025<x:ref>field-name</x:ref> = token
5026<x:ref>field-value</x:ref> = *( field-content / obs-fold )
5027
5028<x:ref>header-field</x:ref> = field-name ":" OWS field-value BWS
5029<x:ref>http-URI</x:ref> = "http://" authority path-abempty [ "?" query ]
5030<x:ref>https-URI</x:ref> = "https://" authority path-abempty [ "?" query ]
5031
5032<x:ref>last-chunk</x:ref> = 1*"0" [ chunk-ext ] CRLF
5033
5034<x:ref>message-body</x:ref> = *OCTET
5035<x:ref>method</x:ref> = token
5036
5037<x:ref>obs-fold</x:ref> = CRLF ( SP / HTAB )
5038<x:ref>obs-text</x:ref> = %x80-FF
5039<x:ref>origin-form</x:ref> = path-absolute [ "?" query ]
5040
5041<x:ref>partial-URI</x:ref> = relative-part [ "?" query ]
5042<x:ref>path-abempty</x:ref> = &lt;path-abempty, defined in [RFC3986], Section 3.3&gt;
5043<x:ref>path-absolute</x:ref> = &lt;path-absolute, defined in [RFC3986], Section 3.3&gt;
5044<x:ref>port</x:ref> = &lt;port, defined in [RFC3986], Section 3.2.3&gt;
5045<x:ref>protocol</x:ref> = protocol-name [ "/" protocol-version ]
5046<x:ref>protocol-name</x:ref> = token
5047<x:ref>protocol-version</x:ref> = token
5048<x:ref>pseudonym</x:ref> = token
5049
5050<x:ref>qdtext</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5051 / %x5D-7E ; ']'-'~'
5052 / obs-text
5053<x:ref>qdtext-nf</x:ref> = HTAB / SP / "!" / %x23-5B ; '#'-'['
5054 / %x5D-7E ; ']'-'~'
5055 / obs-text
5056<x:ref>query</x:ref> = &lt;query, defined in [RFC3986], Section 3.4&gt;
5057<x:ref>quoted-cpair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5058<x:ref>quoted-pair</x:ref> = "\" ( HTAB / SP / VCHAR / obs-text )
5059<x:ref>quoted-str-nf</x:ref> = DQUOTE *( qdtext-nf / quoted-pair ) DQUOTE
5060<x:ref>quoted-string</x:ref> = DQUOTE *( qdtext / quoted-pair ) DQUOTE
5061
5062<x:ref>rank</x:ref> = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] )
5063<x:ref>reason-phrase</x:ref> = *( HTAB / SP / VCHAR / obs-text )
5064<x:ref>received-by</x:ref> = ( uri-host [ ":" port ] ) / pseudonym
5065<x:ref>received-protocol</x:ref> = [ protocol-name "/" ] protocol-version
5066<x:ref>relative-part</x:ref> = &lt;relative-part, defined in [RFC3986], Section 4.2&gt;
5067<x:ref>request-line</x:ref> = method SP request-target SP HTTP-version CRLF
5068<x:ref>request-target</x:ref> = origin-form / absolute-form / authority-form /
5069 asterisk-form
5070
5071<x:ref>special</x:ref> = "(" / ")" / "&lt;" / "&gt;" / "@" / "," / ";" / ":" / "\" /
5072 DQUOTE / "/" / "[" / "]" / "?" / "=" / "{" / "}"
5073<x:ref>start-line</x:ref> = request-line / status-line
5074<x:ref>status-code</x:ref> = 3DIGIT
5075<x:ref>status-line</x:ref> = HTTP-version SP status-code SP reason-phrase CRLF
5076
5077<x:ref>t-codings</x:ref> = "trailers" / ( transfer-coding [ t-ranking ] )
5078<x:ref>t-ranking</x:ref> = OWS ";" OWS "q=" rank
5079<x:ref>tchar</x:ref> = "!" / "#" / "$" / "%" / "&amp;" / "'" / "*" / "+" / "-" / "." /
5080 "^" / "_" / "`" / "|" / "~" / DIGIT / ALPHA
5081<x:ref>token</x:ref> = 1*tchar
5082<x:ref>trailer-part</x:ref> = *( header-field CRLF )
5083<x:ref>transfer-coding</x:ref> = "chunked" / "compress" / "deflate" / "gzip" /
5084 transfer-extension
5085<x:ref>transfer-extension</x:ref> = token *( OWS ";" OWS transfer-parameter )
5086<x:ref>transfer-parameter</x:ref> = attribute BWS "=" BWS value
5087
5088<x:ref>uri-host</x:ref> = &lt;host, defined in [RFC3986], Section 3.2.2&gt;
5089
5090<x:ref>value</x:ref> = word
5091
5092<x:ref>word</x:ref> = token / quoted-string
5093</artwork>
5094</figure>
5095</section>
5096<?ENDINC p1-messaging.abnf-appendix ?>
5097
5098<section title="Change Log (to be removed by RFC Editor before publication)" anchor="change.log">
5099
5100<section title="Since RFC 2616">
5101<t>
5102  Changes up to the first Working Group Last Call draft are summarized
5103  in <eref target="http://trac.tools.ietf.org/html/draft-ietf-httpbis-p1-messaging-21#appendix-D"/>.
5104</t>
5105</section>
5106
5107<section title="Since draft-ietf-httpbis-p1-messaging-21" anchor="changes.since.21">
5108<t>
5109  Closed issues:
5110  <list style="symbols">
5111    <t>
5112      <eref target="http://tools.ietf.org/wg/httpbis/trac/ticket/128"/>:
5113      "Cite HTTPS URI scheme definition" (the spec now includes the HTTPs
5114      scheme definition and thus updates RFC 2818)
5115    </t>
5116    <t>
5117      <eref target="http://tools.ietf.org/wg/httpbis/trac/ticket/389"/>:
5118      "mention of 'proxies' in section about caches"
5119    </t>
5120  </list>
5121</t>
5122</section>
5123
5124</section>
5125
5126</back>
5127</rfc>
Note: See TracBrowser for help on using the repository browser.