idnits 2.17.1 draft-ietf-httpbis-messaging-02.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The draft header indicates that this document obsoletes RFC7230, but the abstract doesn't seem to directly say this. It does mention RFC7230 though, so this could be OK. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document seems to contain a disclaimer for pre-RFC5378 work, but was first submitted on or after 10 November 2008. The disclaimer is usually necessary only for documents that revise or obsolete older RFCs, and that take significant amounts of text from those RFCs. If you can contact all authors of the source material and they are willing to grant the BCP78 rights to the IETF Trust, you can and should remove the disclaimer. Otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (July 2, 2018) is 2124 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'RFC3986' is defined on line 2005, but no explicit reference was found in the text == Unused Reference: 'RFC7231' is defined on line 2084, but no explicit reference was found in the text == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-cache-02 -- Possible downref: Normative reference to a draft: ref. 'Caching' ** Downref: Normative reference to an Informational RFC: RFC 1950 ** Downref: Normative reference to an Informational RFC: RFC 1951 ** Downref: Normative reference to an Informational RFC: RFC 1952 == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-semantics-02 -- Possible downref: Normative reference to a draft: ref. 'Semantics' -- Possible downref: Non-RFC (?) normative reference: ref. 'USASCII' -- Possible downref: Non-RFC (?) normative reference: ref. 'Welch' -- Obsolete informational reference (is this intentional?): RFC 7230 (ref. 'Err4667') (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 2068 (Obsoleted by RFC 2616) -- Duplicate reference: RFC7230, mentioned in 'RFC7230', was also mentioned in 'Err4667'. -- Obsolete informational reference (is this intentional?): RFC 7230 (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 7231 (Obsoleted by RFC 9110) Summary: 3 errors (**), 0 flaws (~~), 6 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 HTTP Working Group R. Fielding, Ed. 3 Internet-Draft Adobe 4 Obsoletes: 7230 (if approved) M. Nottingham, Ed. 5 Intended status: Standards Track Fastly 6 Expires: January 3, 2019 J. Reschke, Ed. 7 greenbytes 8 July 2, 2018 10 HTTP/1.1 Messaging 11 draft-ietf-httpbis-messaging-02 13 Abstract 15 The Hypertext Transfer Protocol (HTTP) is a stateless application- 16 level protocol for distributed, collaborative, hypertext information 17 systems. This document specifies the HTTP/1.1 message syntax, 18 message parsing, connection management, and related security 19 concerns. 21 This document obsoletes portions of RFC 7230. 23 Editorial Note 25 This note is to be removed before publishing as an RFC. 27 Discussion of this draft takes place on the HTTP working group 28 mailing list (ietf-http-wg@w3.org), which is archived at 29 . 31 Working Group information can be found at ; 32 source code and issues list for this draft can be found at 33 . 35 The changes in this draft are summarized in Appendix D.3. 37 Status of This Memo 39 This Internet-Draft is submitted in full conformance with the 40 provisions of BCP 78 and BCP 79. 42 Internet-Drafts are working documents of the Internet Engineering 43 Task Force (IETF). Note that other groups may also distribute 44 working documents as Internet-Drafts. The list of current Internet- 45 Drafts is at https://datatracker.ietf.org/drafts/current/. 47 Internet-Drafts are draft documents valid for a maximum of six months 48 and may be updated, replaced, or obsoleted by other documents at any 49 time. It is inappropriate to use Internet-Drafts as reference 50 material or to cite them other than as "work in progress." 52 This Internet-Draft will expire on January 3, 2019. 54 Copyright Notice 56 Copyright (c) 2018 IETF Trust and the persons identified as the 57 document authors. All rights reserved. 59 This document is subject to BCP 78 and the IETF Trust's Legal 60 Provisions Relating to IETF Documents 61 (https://trustee.ietf.org/license-info) in effect on the date of 62 publication of this document. Please review these documents 63 carefully, as they describe your rights and restrictions with respect 64 to this document. Code Components extracted from this document must 65 include Simplified BSD License text as described in Section 4.e of 66 the Trust Legal Provisions and are provided without warranty as 67 described in the Simplified BSD License. 69 This document may contain material from IETF Documents or IETF 70 Contributions published or made publicly available before November 71 10, 2008. The person(s) controlling the copyright in some of this 72 material may not have granted the IETF Trust the right to allow 73 modifications of such material outside the IETF Standards Process. 74 Without obtaining an adequate license from the person(s) controlling 75 the copyright in such materials, this document may not be modified 76 outside the IETF Standards Process, and derivative works of it may 77 not be created outside the IETF Standards Process, except to format 78 it for publication as an RFC or to translate it into languages other 79 than English. 81 Table of Contents 83 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 84 1.1. Requirements Notation . . . . . . . . . . . . . . . . . . 5 85 1.2. Syntax Notation . . . . . . . . . . . . . . . . . . . . . 5 86 2. Message . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 87 2.1. Message Format . . . . . . . . . . . . . . . . . . . . . 6 88 2.2. HTTP Version . . . . . . . . . . . . . . . . . . . . . . 6 89 2.3. Message Parsing . . . . . . . . . . . . . . . . . . . . . 7 90 3. Request Line . . . . . . . . . . . . . . . . . . . . . . . . 8 91 3.1. Method . . . . . . . . . . . . . . . . . . . . . . . . . 9 92 3.2. Request Target . . . . . . . . . . . . . . . . . . . . . 9 93 3.2.1. origin-form . . . . . . . . . . . . . . . . . . . . . 10 94 3.2.2. absolute-form . . . . . . . . . . . . . . . . . . . . 10 95 3.2.3. authority-form . . . . . . . . . . . . . . . . . . . 11 96 3.2.4. asterisk-form . . . . . . . . . . . . . . . . . . . . 11 98 3.3. Effective Request URI . . . . . . . . . . . . . . . . . . 12 99 4. Status Line . . . . . . . . . . . . . . . . . . . . . . . . . 13 100 5. Header Fields . . . . . . . . . . . . . . . . . . . . . . . . 14 101 5.1. Field Parsing . . . . . . . . . . . . . . . . . . . . . . 15 102 5.2. Obsolete Line Folding . . . . . . . . . . . . . . . . . . 15 103 6. Message Body . . . . . . . . . . . . . . . . . . . . . . . . 16 104 6.1. Transfer-Encoding . . . . . . . . . . . . . . . . . . . . 17 105 6.2. Content-Length . . . . . . . . . . . . . . . . . . . . . 18 106 6.3. Message Body Length . . . . . . . . . . . . . . . . . . . 19 107 7. Transfer Codings . . . . . . . . . . . . . . . . . . . . . . 21 108 7.1. Chunked Transfer Coding . . . . . . . . . . . . . . . . . 22 109 7.1.1. Chunk Extensions . . . . . . . . . . . . . . . . . . 23 110 7.1.2. Chunked Trailer Part . . . . . . . . . . . . . . . . 23 111 7.1.3. Decoding Chunked . . . . . . . . . . . . . . . . . . 24 112 7.2. Transfer Codings for Compression . . . . . . . . . . . . 25 113 7.3. Transfer Coding Registry . . . . . . . . . . . . . . . . 25 114 7.4. TE . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 115 8. Handling Incomplete Messages . . . . . . . . . . . . . . . . 27 116 9. Connection Management . . . . . . . . . . . . . . . . . . . . 28 117 9.1. Connection . . . . . . . . . . . . . . . . . . . . . . . 28 118 9.2. Establishment . . . . . . . . . . . . . . . . . . . . . . 30 119 9.3. Persistence . . . . . . . . . . . . . . . . . . . . . . . 30 120 9.3.1. Retrying Requests . . . . . . . . . . . . . . . . . . 31 121 9.3.2. Pipelining . . . . . . . . . . . . . . . . . . . . . 31 122 9.4. Concurrency . . . . . . . . . . . . . . . . . . . . . . . 32 123 9.5. Failures and Timeouts . . . . . . . . . . . . . . . . . . 33 124 9.6. Tear-down . . . . . . . . . . . . . . . . . . . . . . . . 33 125 9.7. Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . 34 126 9.7.1. Upgrade Protocol Names . . . . . . . . . . . . . . . 36 127 9.7.2. Upgrade Token Registry . . . . . . . . . . . . . . . 37 128 10. Enclosing Messages as Data . . . . . . . . . . . . . . . . . 37 129 10.1. Media Type message/http . . . . . . . . . . . . . . . . 38 130 10.2. Media Type application/http . . . . . . . . . . . . . . 39 131 11. Security Considerations . . . . . . . . . . . . . . . . . . . 40 132 11.1. Response Splitting . . . . . . . . . . . . . . . . . . . 40 133 11.2. Request Smuggling . . . . . . . . . . . . . . . . . . . 41 134 11.3. Message Integrity . . . . . . . . . . . . . . . . . . . 41 135 11.4. Message Confidentiality . . . . . . . . . . . . . . . . 42 136 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 42 137 12.1. Header Field Registration . . . . . . . . . . . . . . . 42 138 12.2. Media Type Registration . . . . . . . . . . . . . . . . 42 139 12.3. Transfer Coding Registration . . . . . . . . . . . . . . 42 140 12.4. Upgrade Token Registration . . . . . . . . . . . . . . . 43 141 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 43 142 13.1. Normative References . . . . . . . . . . . . . . . . . . 43 143 13.2. Informative References . . . . . . . . . . . . . . . . . 44 144 Appendix A. Collected ABNF . . . . . . . . . . . . . . . . . . . 46 145 Appendix B. Differences between HTTP and MIME . . . . . . . . . 47 146 B.1. MIME-Version . . . . . . . . . . . . . . . . . . . . . . 48 147 B.2. Conversion to Canonical Form . . . . . . . . . . . . . . 48 148 B.3. Conversion of Date Formats . . . . . . . . . . . . . . . 48 149 B.4. Conversion of Content-Encoding . . . . . . . . . . . . . 49 150 B.5. Conversion of Content-Transfer-Encoding . . . . . . . . . 49 151 B.6. MHTML and Line Length Limitations . . . . . . . . . . . . 49 152 Appendix C. HTTP Version History . . . . . . . . . . . . . . . . 49 153 C.1. Changes from HTTP/1.0 . . . . . . . . . . . . . . . . . . 50 154 C.1.1. Multihomed Web Servers . . . . . . . . . . . . . . . 50 155 C.1.2. Keep-Alive Connections . . . . . . . . . . . . . . . 51 156 C.1.3. Introduction of Transfer-Encoding . . . . . . . . . . 51 157 C.2. Changes from RFC 7230 . . . . . . . . . . . . . . . . . . 52 158 Appendix D. Change Log . . . . . . . . . . . . . . . . . . . . . 52 159 D.1. Between RFC7230 and draft 00 . . . . . . . . . . . . . . 52 160 D.2. Since draft-ietf-httpbis-messaging-00 . . . . . . . . . . 52 161 D.3. Since draft-ietf-httpbis-messaging-01 . . . . . . . . . . 53 162 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 54 163 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 56 164 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 56 166 1. Introduction 168 The Hypertext Transfer Protocol (HTTP) is a stateless application- 169 level request/response protocol that uses extensible semantics and 170 self-descriptive messages for flexible interaction with network-based 171 hypertext information systems. HTTP is defined by a series of 172 documents that collectively form the HTTP/1.1 specification: 174 o "HTTP Semantics" [Semantics] 176 o "HTTP Caching" [Caching] 178 o "HTTP/1.1 Messaging" (this document) 180 This document defines HTTP/1.1 message syntax and framing 181 requirements and their associated connection management. Our goal is 182 to define all of the mechanisms necessary for HTTP/1.1 message 183 handling that are independent of message semantics, thereby defining 184 the complete set of requirements for message parsers and message- 185 forwarding intermediaries. 187 This document obsoletes the portions of RFC 7230 related to HTTP/1.1 188 messaging and connection management, with the changes being 189 summarized in Appendix C.2. The other parts of RFC 7230 are 190 obsoleted by "HTTP Semantics" [Semantics]. 192 1.1. Requirements Notation 194 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 195 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 196 document are to be interpreted as described in [RFC2119]. 198 Conformance criteria and considerations regarding error handling are 199 defined in Section 3 of [Semantics]. 201 1.2. Syntax Notation 203 This specification uses the Augmented Backus-Naur Form (ABNF) 204 notation of [RFC5234] with a list extension, defined in Section 11 of 205 [Semantics], that allows for compact definition of comma-separated 206 lists using a '#' operator (similar to how the '*' operator indicates 207 repetition). Appendix A shows the collected grammar with all list 208 operators expanded to standard ABNF notation. 210 As a convention, ABNF rule names prefixed with "obs-" denote 211 "obsolete" grammar rules that appear for historical reasons. 213 The following core rules are included by reference, as defined in 214 [RFC5234], Appendix B.1: ALPHA (letters), CR (carriage return), CRLF 215 (CR LF), CTL (controls), DIGIT (decimal 0-9), DQUOTE (double quote), 216 HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line 217 feed), OCTET (any 8-bit sequence of data), SP (space), and VCHAR (any 218 visible [USASCII] character). 220 The rules below are defined in [Semantics]: 222 BWS = 223 OWS = 224 RWS = 225 absolute-URI = 226 absolute-path = 227 authority = 228 comment = 229 field-name = 230 field-value = 231 obs-text = 232 port = 233 query = 234 quoted-string = 235 token = 236 uri-host = 238 2. Message 240 2.1. Message Format 242 All HTTP/1.1 messages consist of a start-line followed by a sequence 243 of octets in a format similar to the Internet Message Format 244 [RFC5322]: zero or more header fields (collectively referred to as 245 the "headers" or the "header section"), an empty line indicating the 246 end of the header section, and an optional message body. 248 HTTP-message = start-line 249 *( header-field CRLF ) 250 CRLF 251 [ message-body ] 253 An HTTP message can be either a request from client to server or a 254 response from server to client. Syntactically, the two types of 255 message differ only in the start-line, which is either a request-line 256 (for requests) or a status-line (for responses), and in the algorithm 257 for determining the length of the message body (Section 6). 259 start-line = request-line / status-line 261 In theory, a client could receive requests and a server could receive 262 responses, distinguishing them by their different start-line formats. 263 In practice, servers are implemented to only expect a request (a 264 response is interpreted as an unknown or invalid request method) and 265 clients are implemented to only expect a response. 267 Although HTTP makes use of some protocol elements similar to the 268 Multipurpose Internet Mail Extensions (MIME) [RFC2045], see 269 Appendix B for the differences between HTTP and MIME messages. 271 2.2. HTTP Version 273 HTTP uses a "." numbering scheme to indicate versions 274 of the protocol. This specification defines version "1.1". 275 Section 3.5 of [Semantics] specifies the semantics of HTTP version 276 numbers. 278 The version of an HTTP/1.x message is indicated by an HTTP-version 279 field in the start-line. HTTP-version is case-sensitive. 281 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 282 HTTP-name = %x48.54.54.50 ; "HTTP", case-sensitive 284 When an HTTP/1.1 message is sent to an HTTP/1.0 recipient [RFC1945] 285 or a recipient whose version is unknown, the HTTP/1.1 message is 286 constructed such that it can be interpreted as a valid HTTP/1.0 287 message if all of the newer features are ignored. This specification 288 places recipient-version requirements on some new features so that a 289 conformant sender will only use compatible features until it has 290 determined, through configuration or the receipt of a message, that 291 the recipient supports HTTP/1.1. 293 Intermediaries that process HTTP messages (i.e., all intermediaries 294 other than those acting as tunnels) MUST send their own HTTP-version 295 in forwarded messages. In other words, they are not allowed to 296 blindly forward the start-line without ensuring that the protocol 297 version in that message matches a version to which that intermediary 298 is conformant for both the receiving and sending of messages. 299 Forwarding an HTTP message without rewriting the HTTP-version might 300 result in communication errors when downstream recipients use the 301 message sender's version to determine what features are safe to use 302 for later communication with that sender. 304 A server MAY send an HTTP/1.0 response to an HTTP/1.1 request if it 305 is known or suspected that the client incorrectly implements the HTTP 306 specification and is incapable of correctly processing later version 307 responses, such as when a client fails to parse the version number 308 correctly or when an intermediary is known to blindly forward the 309 HTTP-version even when it doesn't conform to the given minor version 310 of the protocol. Such protocol downgrades SHOULD NOT be performed 311 unless triggered by specific client attributes, such as when one or 312 more of the request header fields (e.g., User-Agent) uniquely match 313 the values sent by a client known to be in error. 315 2.3. Message Parsing 317 The normal procedure for parsing an HTTP message is to read the 318 start-line into a structure, read each header field into a hash table 319 by field name until the empty line, and then use the parsed data to 320 determine if a message body is expected. If a message body has been 321 indicated, then it is read as a stream until an amount of octets 322 equal to the message body length is read or the connection is closed. 324 A recipient MUST parse an HTTP message as a sequence of octets in an 325 encoding that is a superset of US-ASCII [USASCII]. Parsing an HTTP 326 message as a stream of Unicode characters, without regard for the 327 specific encoding, creates security vulnerabilities due to the 328 varying ways that string processing libraries handle invalid 329 multibyte character sequences that contain the octet LF (%x0A). 330 String-based parsers can only be safely used within protocol elements 331 after the element has been extracted from the message, such as within 332 a header field-value after message parsing has delineated the 333 individual fields. 335 Although the line terminator for the start-line and header fields is 336 the sequence CRLF, a recipient MAY recognize a single LF as a line 337 terminator and ignore any preceding CR. 339 Older HTTP/1.0 user agent implementations might send an extra CRLF 340 after a POST request as a workaround for some early server 341 applications that failed to read message body content that was not 342 terminated by a line-ending. An HTTP/1.1 user agent MUST NOT preface 343 or follow a request with an extra CRLF. If terminating the request 344 message body with a line-ending is desired, then the user agent MUST 345 count the terminating CRLF octets as part of the message body length. 347 In the interest of robustness, a server that is expecting to receive 348 and parse a request-line SHOULD ignore at least one empty line (CRLF) 349 received prior to the request-line. 351 A sender MUST NOT send whitespace between the start-line and the 352 first header field. A recipient that receives whitespace between the 353 start-line and the first header field MUST either reject the message 354 as invalid or consume each whitespace-preceded line without further 355 processing of it (i.e., ignore the entire line, along with any 356 subsequent lines preceded by whitespace, until a properly formed 357 header field is received or the header section is terminated). 359 The presence of such whitespace in a request might be an attempt to 360 trick a server into ignoring that field or processing the line after 361 it as a new request, either of which might result in a security 362 vulnerability if other implementations within the request chain 363 interpret the same message differently. Likewise, the presence of 364 such whitespace in a response might be ignored by some clients or 365 cause others to cease parsing. 367 When a server listening only for HTTP request messages, or processing 368 what appears from the start-line to be an HTTP request message, 369 receives a sequence of octets that does not match the HTTP-message 370 grammar aside from the robustness exceptions listed above, the server 371 SHOULD respond with a 400 (Bad Request) response. 373 3. Request Line 375 A request-line begins with a method token, followed by a single space 376 (SP), the request-target, another single space (SP), the protocol 377 version, and ends with CRLF. 379 request-line = method SP request-target SP HTTP-version CRLF 381 Although the request-line grammar rule requires that each of the 382 component elements be separated by a single SP octet, recipients MAY 383 instead parse on whitespace-delimited word boundaries and, aside from 384 the CRLF terminator, treat any form of whitespace as the SP separator 385 while ignoring preceding or trailing whitespace; such whitespace 386 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 387 (%x0C), or bare CR. However, lenient parsing can result in request 388 smuggling security vulnerabilities if there are multiple recipients 389 of the message and each has its own unique interpretation of 390 robustness (see Section 11.2). 392 HTTP does not place a predefined limit on the length of a request- 393 line, as described in Section 3 of [Semantics]. A server that 394 receives a method longer than any that it implements SHOULD respond 395 with a 501 (Not Implemented) status code. A server that receives a 396 request-target longer than any URI it wishes to parse MUST respond 397 with a 414 (URI Too Long) status code (see Section 9.5.15 of 398 [Semantics]). 400 Various ad hoc limitations on request-line length are found in 401 practice. It is RECOMMENDED that all HTTP senders and recipients 402 support, at a minimum, request-line lengths of 8000 octets. 404 3.1. Method 406 The method token indicates the request method to be performed on the 407 target resource. The request method is case-sensitive. 409 method = token 411 The request methods defined by this specification can be found in 412 Section 7 of [Semantics], along with information regarding the HTTP 413 method registry and considerations for defining new methods. 415 3.2. Request Target 417 The request-target identifies the target resource upon which to apply 418 the request. The client derives a request-target from its desired 419 target URI. There are four distinct formats for the request-target, 420 depending on both the method being requested and whether the request 421 is to a proxy. 423 request-target = origin-form 424 / absolute-form 425 / authority-form 426 / asterisk-form 428 No whitespace is allowed in the request-target. Unfortunately, some 429 user agents fail to properly encode or exclude whitespace found in 430 hypertext references, resulting in those disallowed characters being 431 sent as the request-target in a malformed request-line. 433 Recipients of an invalid request-line SHOULD respond with either a 434 400 (Bad Request) error or a 301 (Moved Permanently) redirect with 435 the request-target properly encoded. A recipient SHOULD NOT attempt 436 to autocorrect and then process the request without a redirect, since 437 the invalid request-line might be deliberately crafted to bypass 438 security filters along the request chain. 440 3.2.1. origin-form 442 The most common form of request-target is the origin-form. 444 origin-form = absolute-path [ "?" query ] 446 When making a request directly to an origin server, other than a 447 CONNECT or server-wide OPTIONS request (as detailed below), a client 448 MUST send only the absolute path and query components of the target 449 URI as the request-target. If the target URI's path component is 450 empty, the client MUST send "/" as the path within the origin-form of 451 request-target. A Host header field is also sent, as defined in 452 Section 5.4 of [Semantics]. 454 For example, a client wishing to retrieve a representation of the 455 resource identified as 457 http://www.example.org/where?q=now 459 directly from the origin server would open (or reuse) a TCP 460 connection to port 80 of the host "www.example.org" and send the 461 lines: 463 GET /where?q=now HTTP/1.1 464 Host: www.example.org 466 followed by the remainder of the request message. 468 3.2.2. absolute-form 470 When making a request to a proxy, other than a CONNECT or server-wide 471 OPTIONS request (as detailed below), a client MUST send the target 472 URI in absolute-form as the request-target. 474 absolute-form = absolute-URI 476 The proxy is requested to either service that request from a valid 477 cache, if possible, or make the same request on the client's behalf 478 to either the next inbound proxy server or directly to the origin 479 server indicated by the request-target. Requirements on such 480 "forwarding" of messages are defined in Section 5.6 of [Semantics]. 482 An example absolute-form of request-line would be: 484 GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1 486 To allow for transition to the absolute-form for all requests in some 487 future version of HTTP, a server MUST accept the absolute-form in 488 requests, even though HTTP/1.1 clients will only send them in 489 requests to proxies. 491 3.2.3. authority-form 493 The authority-form of request-target is only used for CONNECT 494 requests (Section 7.3.6 of [Semantics]). 496 authority-form = authority 498 When making a CONNECT request to establish a tunnel through one or 499 more proxies, a client MUST send only the target URI's authority 500 component (excluding any userinfo and its "@" delimiter) as the 501 request-target. For example, 503 CONNECT www.example.com:80 HTTP/1.1 505 3.2.4. asterisk-form 507 The asterisk-form of request-target is only used for a server-wide 508 OPTIONS request (Section 7.3.7 of [Semantics]). 510 asterisk-form = "*" 512 When a client wishes to request OPTIONS for the server as a whole, as 513 opposed to a specific named resource of that server, the client MUST 514 send only "*" (%x2A) as the request-target. For example, 516 OPTIONS * HTTP/1.1 518 If a proxy receives an OPTIONS request with an absolute-form of 519 request-target in which the URI has an empty path and no query 520 component, then the last proxy on the request chain MUST send a 521 request-target of "*" when it forwards the request to the indicated 522 origin server. 524 For example, the request 526 OPTIONS http://www.example.org:8001 HTTP/1.1 528 would be forwarded by the final proxy as 530 OPTIONS * HTTP/1.1 531 Host: www.example.org:8001 533 after connecting to port 8001 of host "www.example.org". 535 3.3. Effective Request URI 537 Since the request-target often contains only part of the user agent's 538 target URI, a server reconstructs the intended target as an effective 539 request URI to properly service the request (Section 5.3 of 540 [Semantics]). 542 If the request-target is in absolute-form, the effective request URI 543 is the same as the request-target. Otherwise, the effective request 544 URI is constructed as follows: 546 If the server's configuration (or outbound gateway) provides a 547 fixed URI scheme, that scheme is used for the effective request 548 URI. Otherwise, if the request is received over a TLS-secured TCP 549 connection, the effective request URI's scheme is "https"; if not, 550 the scheme is "http". 552 If the server's configuration (or outbound gateway) provides a 553 fixed URI authority component, that authority is used for the 554 effective request URI. If not, then if the request-target is in 555 authority-form, the effective request URI's authority component is 556 the same as the request-target. If not, then if a Host header 557 field is supplied with a non-empty field-value, the authority 558 component is the same as the Host field-value. Otherwise, the 559 authority component is assigned the default name configured for 560 the server and, if the connection's incoming TCP port number 561 differs from the default port for the effective request URI's 562 scheme, then a colon (":") and the incoming port number (in 563 decimal form) are appended to the authority component. 565 If the request-target is in authority-form or asterisk-form, the 566 effective request URI's combined path and query component is 567 empty. Otherwise, the combined path and query component is the 568 same as the request-target. 570 The components of the effective request URI, once determined as 571 above, can be combined into absolute-URI form by concatenating the 572 scheme, "://", authority, and combined path and query component. 574 Example 1: the following message received over an insecure TCP 575 connection 577 GET /pub/WWW/TheProject.html HTTP/1.1 578 Host: www.example.org:8080 580 has an effective request URI of 582 http://www.example.org:8080/pub/WWW/TheProject.html 584 Example 2: the following message received over a TLS-secured TCP 585 connection 587 OPTIONS * HTTP/1.1 588 Host: www.example.org 590 has an effective request URI of 592 https://www.example.org 594 Recipients of an HTTP/1.0 request that lacks a Host header field 595 might need to use heuristics (e.g., examination of the URI path for 596 something unique to a particular host) in order to guess the 597 effective request URI's authority component. 599 4. Status Line 601 The first line of a response message is the status-line, consisting 602 of the protocol version, a space (SP), the status code, another 603 space, a possibly empty textual phrase describing the status code, 604 and ending with CRLF. 606 status-line = HTTP-version SP status-code SP reason-phrase CRLF 608 Although the status-line grammar rule requires that each of the 609 component elements be separated by a single SP octet, recipients MAY 610 instead parse on whitespace-delimited word boundaries and, aside from 611 the line terminator, treat any form of whitespace as the SP separator 612 while ignoring preceding or trailing whitespace; such whitespace 613 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 614 (%x0C), or bare CR. However, lenient parsing can result in response 615 splitting security vulnerabilities if there are multiple recipients 616 of the message and each has its own unique interpretation of 617 robustness (see Section 11.1). 619 The status-code element is a 3-digit integer code describing the 620 result of the server's attempt to understand and satisfy the client's 621 corresponding request. The rest of the response message is to be 622 interpreted in light of the semantics defined for that status code. 623 See Section 9 of [Semantics] for information about the semantics of 624 status codes, including the classes of status code (indicated by the 625 first digit), the status codes defined by this specification, 626 considerations for the definition of new status codes, and the IANA 627 registry. 629 status-code = 3DIGIT 631 The reason-phrase element exists for the sole purpose of providing a 632 textual description associated with the numeric status code, mostly 633 out of deference to earlier Internet application protocols that were 634 more frequently used with interactive text clients. A client SHOULD 635 ignore the reason-phrase content. 637 reason-phrase = *( HTAB / SP / VCHAR / obs-text ) 639 5. Header Fields 641 Each header field consists of a case-insensitive field name followed 642 by a colon (":"), optional leading whitespace, the field value, and 643 optional trailing whitespace. 645 header-field = field-name ":" OWS field-value OWS 647 Most HTTP field names and the rules for parsing within field values 648 are defined in Section 4 of [Semantics]. This section covers the 649 generic syntax for header field inclusion within, and extraction 650 from, HTTP/1.1 messages. In addition, the following header fields 651 are defined by this document because they are specific to HTTP/1.1 652 message processing: 654 +-------------------+----------+----------+---------------+ 655 | Header Field Name | Protocol | Status | Reference | 656 +-------------------+----------+----------+---------------+ 657 | Connection | http | standard | Section 9.1 | 658 | MIME-Version | http | standard | Appendix B.1 | 659 | TE | http | standard | Section 7.4 | 660 | Transfer-Encoding | http | standard | Section 6.1 | 661 | Upgrade | http | standard | Section 9.7 | 662 +-------------------+----------+----------+---------------+ 664 Furthermore, the field name "Close" is reserved, since using that 665 name as an HTTP header field might conflict with the "close" 666 connection option of the Connection header field (Section 9.1). 668 +-------------------+----------+----------+------------+ 669 | Header Field Name | Protocol | Status | Reference | 670 +-------------------+----------+----------+------------+ 671 | Close | http | reserved | Section 5 | 672 +-------------------+----------+----------+------------+ 674 5.1. Field Parsing 676 Messages are parsed using a generic algorithm, independent of the 677 individual header field names. The contents within a given field 678 value are not parsed until a later stage of message interpretation 679 (usually after the message's entire header section has been 680 processed). 682 No whitespace is allowed between the header field-name and colon. In 683 the past, differences in the handling of such whitespace have led to 684 security vulnerabilities in request routing and response handling. A 685 server MUST reject any received request message that contains 686 whitespace between a header field-name and colon with a response 687 status code of 400 (Bad Request). A proxy MUST remove any such 688 whitespace from a response message before forwarding the message 689 downstream. 691 A field value might be preceded and/or followed by optional 692 whitespace (OWS); a single SP preceding the field-value is preferred 693 for consistent readability by humans. The field value does not 694 include any leading or trailing whitespace: OWS occurring before the 695 first non-whitespace octet of the field value or after the last non- 696 whitespace octet of the field value ought to be excluded by parsers 697 when extracting the field value from a header field. 699 5.2. Obsolete Line Folding 701 Historically, HTTP header field values could be extended over 702 multiple lines by preceding each extra line with at least one space 703 or horizontal tab (obs-fold). This specification deprecates such 704 line folding except within the message/http media type 705 (Section 10.1). 707 obs-fold = CRLF 1*( SP / HTAB ) 708 ; obsolete line folding 710 A sender MUST NOT generate a message that includes line folding 711 (i.e., that has any field-value that contains a match to the obs-fold 712 rule) unless the message is intended for packaging within the 713 message/http media type. 715 A server that receives an obs-fold in a request message that is not 716 within a message/http container MUST either reject the message by 717 sending a 400 (Bad Request), preferably with a representation 718 explaining that obsolete line folding is unacceptable, or replace 719 each received obs-fold with one or more SP octets prior to 720 interpreting the field value or forwarding the message downstream. 722 A proxy or gateway that receives an obs-fold in a response message 723 that is not within a message/http container MUST either discard the 724 message and replace it with a 502 (Bad Gateway) response, preferably 725 with a representation explaining that unacceptable line folding was 726 received, or replace each received obs-fold with one or more SP 727 octets prior to interpreting the field value or forwarding the 728 message downstream. 730 A user agent that receives an obs-fold in a response message that is 731 not within a message/http container MUST replace each received obs- 732 fold with one or more SP octets prior to interpreting the field 733 value. 735 6. Message Body 737 The message body (if any) of an HTTP message is used to carry the 738 payload body of that request or response. The message body is 739 identical to the payload body unless a transfer coding has been 740 applied, as described in Section 6.1. 742 message-body = *OCTET 744 The rules for when a message body is allowed in a message differ for 745 requests and responses. 747 The presence of a message body in a request is signaled by a Content- 748 Length or Transfer-Encoding header field. Request message framing is 749 independent of method semantics, even if the method does not define 750 any use for a message body. 752 The presence of a message body in a response depends on both the 753 request method to which it is responding and the response status code 754 (Section 4). Responses to the HEAD request method (Section 7.3.2 of 755 [Semantics]) never include a message body because the associated 756 response header fields (e.g., Transfer-Encoding, Content-Length, 757 etc.), if present, indicate only what their values would have been if 758 the request method had been GET (Section 7.3.1 of [Semantics]). 2xx 759 (Successful) responses to a CONNECT request method (Section 7.3.6 of 760 [Semantics]) switch to tunnel mode instead of having a message body. 761 All 1xx (Informational), 204 (No Content), and 304 (Not Modified) 762 responses do not include a message body. All other responses do 763 include a message body, although the body might be of zero length. 765 6.1. Transfer-Encoding 767 The Transfer-Encoding header field lists the transfer coding names 768 corresponding to the sequence of transfer codings that have been (or 769 will be) applied to the payload body in order to form the message 770 body. Transfer codings are defined in Section 7. 772 Transfer-Encoding = 1#transfer-coding 774 Transfer-Encoding is analogous to the Content-Transfer-Encoding field 775 of MIME, which was designed to enable safe transport of binary data 776 over a 7-bit transport service ([RFC2045], Section 6). However, safe 777 transport has a different focus for an 8bit-clean transfer protocol. 778 In HTTP's case, Transfer-Encoding is primarily intended to accurately 779 delimit a dynamically generated payload and to distinguish payload 780 encodings that are only applied for transport efficiency or security 781 from those that are characteristics of the selected resource. 783 A recipient MUST be able to parse the chunked transfer coding 784 (Section 7.1) because it plays a crucial role in framing messages 785 when the payload body size is not known in advance. A sender MUST 786 NOT apply chunked more than once to a message body (i.e., chunking an 787 already chunked message is not allowed). If any transfer coding 788 other than chunked is applied to a request payload body, the sender 789 MUST apply chunked as the final transfer coding to ensure that the 790 message is properly framed. If any transfer coding other than 791 chunked is applied to a response payload body, the sender MUST either 792 apply chunked as the final transfer coding or terminate the message 793 by closing the connection. 795 For example, 797 Transfer-Encoding: gzip, chunked 799 indicates that the payload body has been compressed using the gzip 800 coding and then chunked using the chunked coding while forming the 801 message body. 803 Unlike Content-Encoding (Section 6.1.2 of [Semantics]), Transfer- 804 Encoding is a property of the message, not of the representation, and 805 any recipient along the request/response chain MAY decode the 806 received transfer coding(s) or apply additional transfer coding(s) to 807 the message body, assuming that corresponding changes are made to the 808 Transfer-Encoding field-value. Additional information about the 809 encoding parameters can be provided by other header fields not 810 defined by this specification. 812 Transfer-Encoding MAY be sent in a response to a HEAD request or in a 813 304 (Not Modified) response (Section 9.4.5 of [Semantics]) to a GET 814 request, neither of which includes a message body, to indicate that 815 the origin server would have applied a transfer coding to the message 816 body if the request had been an unconditional GET. This indication 817 is not required, however, because any recipient on the response chain 818 (including the origin server) can remove transfer codings when they 819 are not needed. 821 A server MUST NOT send a Transfer-Encoding header field in any 822 response with a status code of 1xx (Informational) or 204 (No 823 Content). A server MUST NOT send a Transfer-Encoding header field in 824 any 2xx (Successful) response to a CONNECT request (Section 7.3.6 of 825 [Semantics]). 827 Transfer-Encoding was added in HTTP/1.1. It is generally assumed 828 that implementations advertising only HTTP/1.0 support will not 829 understand how to process a transfer-encoded payload. A client MUST 830 NOT send a request containing Transfer-Encoding unless it knows the 831 server will handle HTTP/1.1 (or later) requests; such knowledge might 832 be in the form of specific user configuration or by remembering the 833 version of a prior received response. A server MUST NOT send a 834 response containing Transfer-Encoding unless the corresponding 835 request indicates HTTP/1.1 (or later). 837 A server that receives a request message with a transfer coding it 838 does not understand SHOULD respond with 501 (Not Implemented). 840 6.2. Content-Length 842 When a message does not have a Transfer-Encoding header field, a 843 Content-Length header field can provide the anticipated size, as a 844 decimal number of octets, for a potential payload body. For messages 845 that do include a payload body, the Content-Length field-value 846 provides the framing information necessary for determining where the 847 body (and message) ends. For messages that do not include a payload 848 body, the Content-Length indicates the size of the selected 849 representation (Section 6.2.4 of [Semantics]). 851 Note: HTTP's use of Content-Length for message framing differs 852 significantly from the same field's use in MIME, where it is an 853 optional field used only within the "message/external-body" media- 854 type. 856 6.3. Message Body Length 858 The length of a message body is determined by one of the following 859 (in order of precedence): 861 1. Any response to a HEAD request and any response with a 1xx 862 (Informational), 204 (No Content), or 304 (Not Modified) status 863 code is always terminated by the first empty line after the 864 header fields, regardless of the header fields present in the 865 message, and thus cannot contain a message body. 867 2. Any 2xx (Successful) response to a CONNECT request implies that 868 the connection will become a tunnel immediately after the empty 869 line that concludes the header fields. A client MUST ignore any 870 Content-Length or Transfer-Encoding header fields received in 871 such a message. 873 3. If a Transfer-Encoding header field is present and the chunked 874 transfer coding (Section 7.1) is the final encoding, the message 875 body length is determined by reading and decoding the chunked 876 data until the transfer coding indicates the data is complete. 878 If a Transfer-Encoding header field is present in a response and 879 the chunked transfer coding is not the final encoding, the 880 message body length is determined by reading the connection until 881 it is closed by the server. If a Transfer-Encoding header field 882 is present in a request and the chunked transfer coding is not 883 the final encoding, the message body length cannot be determined 884 reliably; the server MUST respond with the 400 (Bad Request) 885 status code and then close the connection. 887 If a message is received with both a Transfer-Encoding and a 888 Content-Length header field, the Transfer-Encoding overrides the 889 Content-Length. Such a message might indicate an attempt to 890 perform request smuggling (Section 11.2) or response splitting 891 (Section 11.1) and ought to be handled as an error. A sender 892 MUST remove the received Content-Length field prior to forwarding 893 such a message downstream. 895 4. If a message is received without Transfer-Encoding and with 896 either multiple Content-Length header fields having differing 897 field-values or a single Content-Length header field having an 898 invalid value, then the message framing is invalid and the 899 recipient MUST treat it as an unrecoverable error. If this is a 900 request message, the server MUST respond with a 400 (Bad Request) 901 status code and then close the connection. If this is a response 902 message received by a proxy, the proxy MUST close the connection 903 to the server, discard the received response, and send a 502 (Bad 904 Gateway) response to the client. If this is a response message 905 received by a user agent, the user agent MUST close the 906 connection to the server and discard the received response. 908 5. If a valid Content-Length header field is present without 909 Transfer-Encoding, its decimal value defines the expected message 910 body length in octets. If the sender closes the connection or 911 the recipient times out before the indicated number of octets are 912 received, the recipient MUST consider the message to be 913 incomplete and close the connection. 915 6. If this is a request message and none of the above are true, then 916 the message body length is zero (no message body is present). 918 7. Otherwise, this is a response message without a declared message 919 body length, so the message body length is determined by the 920 number of octets received prior to the server closing the 921 connection. 923 Since there is no way to distinguish a successfully completed, close- 924 delimited message from a partially received message interrupted by 925 network failure, a server SHOULD generate encoding or length- 926 delimited messages whenever possible. The close-delimiting feature 927 exists primarily for backwards compatibility with HTTP/1.0. 929 A server MAY reject a request that contains a message body but not a 930 Content-Length by responding with 411 (Length Required). 932 Unless a transfer coding other than chunked has been applied, a 933 client that sends a request containing a message body SHOULD use a 934 valid Content-Length header field if the message body length is known 935 in advance, rather than the chunked transfer coding, since some 936 existing services respond to chunked with a 411 (Length Required) 937 status code even though they understand the chunked transfer coding. 938 This is typically because such services are implemented via a gateway 939 that requires a content-length in advance of being called and the 940 server is unable or unwilling to buffer the entire request before 941 processing. 943 A user agent that sends a request containing a message body MUST send 944 a valid Content-Length header field if it does not know the server 945 will handle HTTP/1.1 (or later) requests; such knowledge can be in 946 the form of specific user configuration or by remembering the version 947 of a prior received response. 949 If the final response to the last request on a connection has been 950 completely received and there remains additional data to read, a user 951 agent MAY discard the remaining data or attempt to determine if that 952 data belongs as part of the prior response body, which might be the 953 case if the prior message's Content-Length value is incorrect. A 954 client MUST NOT process, cache, or forward such extra data as a 955 separate response, since such behavior would be vulnerable to cache 956 poisoning. 958 7. Transfer Codings 960 Transfer coding names are used to indicate an encoding transformation 961 that has been, can be, or might need to be applied to a payload body 962 in order to ensure "safe transport" through the network. This 963 differs from a content coding in that the transfer coding is a 964 property of the message rather than a property of the representation 965 that is being transferred. 967 transfer-coding = "chunked" ; Section 7.1 968 / "compress" ; [Semantics], Section 6.1.2.1 969 / "deflate" ; [Semantics], Section 6.1.2.2 970 / "gzip" ; [Semantics], Section 6.1.2.3 971 / transfer-extension 972 transfer-extension = token *( OWS ";" OWS transfer-parameter ) 974 Parameters are in the form of a name=value pair. 976 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 978 All transfer-coding names are case-insensitive and ought to be 979 registered within the HTTP Transfer Coding registry, as defined in 980 Section 7.3. They are used in the TE (Section 7.4) and Transfer- 981 Encoding (Section 6.1) header fields. 983 +------------+------------------------------------------+-----------+ 984 | Name | Description | Reference | 985 +------------+------------------------------------------+-----------+ 986 | chunked | Transfer in a series of chunks | Section 7 | 987 | | | .1 | 988 | compress | UNIX "compress" data format [Welch] | Section 7 | 989 | | | .2 | 990 | deflate | "deflate" compressed data ([RFC1951]) | Section 7 | 991 | | inside the "zlib" data format | .2 | 992 | | ([RFC1950]) | | 993 | gzip | GZIP file format [RFC1952] | Section 7 | 994 | | | .2 | 995 | trailers | (reserved) | Section 7 | 996 | x-compress | Deprecated (alias for compress) | Section 7 | 997 | | | .2 | 998 | x-gzip | Deprecated (alias for gzip) | Section 7 | 999 | | | .2 | 1000 +------------+------------------------------------------+-----------+ 1002 Note: the coding name "trailers" is reserved because it would 1003 clash with the use of the keyword "trailers" in the TE header 1004 field (Section 7.4). 1006 7.1. Chunked Transfer Coding 1008 The chunked transfer coding wraps the payload body in order to 1009 transfer it as a series of chunks, each with its own size indicator, 1010 followed by an OPTIONAL trailer containing header fields. Chunked 1011 enables content streams of unknown size to be transferred as a 1012 sequence of length-delimited buffers, which enables the sender to 1013 retain connection persistence and the recipient to know when it has 1014 received the entire message. 1016 chunked-body = *chunk 1017 last-chunk 1018 trailer-part 1019 CRLF 1021 chunk = chunk-size [ chunk-ext ] CRLF 1022 chunk-data CRLF 1023 chunk-size = 1*HEXDIG 1024 last-chunk = 1*("0") [ chunk-ext ] CRLF 1026 chunk-data = 1*OCTET ; a sequence of chunk-size octets 1028 The chunk-size field is a string of hex digits indicating the size of 1029 the chunk-data in octets. The chunked transfer coding is complete 1030 when a chunk with a chunk-size of zero is received, possibly followed 1031 by a trailer, and finally terminated by an empty line. 1033 A recipient MUST be able to parse and decode the chunked transfer 1034 coding. 1036 7.1.1. Chunk Extensions 1038 The chunked encoding allows each chunk to include zero or more chunk 1039 extensions, immediately following the chunk-size, for the sake of 1040 supplying per-chunk metadata (such as a signature or hash), mid- 1041 message control information, or randomization of message body size. 1043 chunk-ext = *( BWS ";" BWS chunk-ext-name 1044 [ BWS "=" BWS chunk-ext-val ] ) 1046 chunk-ext-name = token 1047 chunk-ext-val = token / quoted-string 1049 The chunked encoding is specific to each connection and is likely to 1050 be removed or recoded by each recipient (including intermediaries) 1051 before any higher-level application would have a chance to inspect 1052 the extensions. Hence, use of chunk extensions is generally limited 1053 to specialized HTTP services such as "long polling" (where client and 1054 server can have shared expectations regarding the use of chunk 1055 extensions) or for padding within an end-to-end secured connection. 1057 A recipient MUST ignore unrecognized chunk extensions. A server 1058 ought to limit the total length of chunk extensions received in a 1059 request to an amount reasonable for the services provided, in the 1060 same way that it applies length limitations and timeouts for other 1061 parts of a message, and generate an appropriate 4xx (Client Error) 1062 response if that amount is exceeded. 1064 7.1.2. Chunked Trailer Part 1066 A trailer allows the sender to include additional fields at the end 1067 of a chunked message in order to supply metadata that might be 1068 dynamically generated while the message body is sent, such as a 1069 message integrity check, digital signature, or post-processing 1070 status. The trailer fields are identical to header fields, except 1071 they are sent in a chunked trailer instead of the message's header 1072 section. 1074 trailer-part = *( header-field CRLF ) 1076 A sender MUST NOT generate a trailer that contains a field necessary 1077 for message framing (e.g., Transfer-Encoding and Content-Length), 1078 routing (e.g., Host), request modifiers (e.g., controls and 1079 conditionals in Section 8 of [Semantics]), authentication (e.g., see 1080 Section 8.5 of [Semantics] and [RFC6265]), response control data 1081 (e.g., see Section 10.1 of [Semantics]), or determining how to 1082 process the payload (e.g., Content-Encoding, Content-Type, Content- 1083 Range, and Trailer). 1085 When a chunked message containing a non-empty trailer is received, 1086 the recipient MAY process the fields (aside from those forbidden 1087 above) as if they were appended to the message's header section. A 1088 recipient MUST ignore (or consider as an error) any fields that are 1089 forbidden to be sent in a trailer, since processing them as if they 1090 were present in the header section might bypass external security 1091 filters. 1093 Unless the request includes a TE header field indicating "trailers" 1094 is acceptable, as described in Section 7.4, a server SHOULD NOT 1095 generate trailer fields that it believes are necessary for the user 1096 agent to receive. Without a TE containing "trailers", the server 1097 ought to assume that the trailer fields might be silently discarded 1098 along the path to the user agent. This requirement allows 1099 intermediaries to forward a de-chunked message to an HTTP/1.0 1100 recipient without buffering the entire response. 1102 When a message includes a message body encoded with the chunked 1103 transfer coding and the sender desires to send metadata in the form 1104 of trailer fields at the end of the message, the sender SHOULD 1105 generate a Trailer header field before the message body to indicate 1106 which fields will be present in the trailers. This allows the 1107 recipient to prepare for receipt of that metadata before it starts 1108 processing the body, which is useful if the message is being streamed 1109 and the recipient wishes to confirm an integrity check on the fly. 1111 7.1.3. Decoding Chunked 1113 A process for decoding the chunked transfer coding can be represented 1114 in pseudo-code as: 1116 length := 0 1117 read chunk-size, chunk-ext (if any), and CRLF 1118 while (chunk-size > 0) { 1119 read chunk-data and CRLF 1120 append chunk-data to decoded-body 1121 length := length + chunk-size 1122 read chunk-size, chunk-ext (if any), and CRLF 1123 } 1124 read trailer field 1125 while (trailer field is not empty) { 1126 if (trailer field is allowed to be sent in a trailer) { 1127 append trailer field to existing header fields 1128 } 1129 read trailer-field 1130 } 1131 Content-Length := length 1132 Remove "chunked" from Transfer-Encoding 1133 Remove Trailer from existing header fields 1135 7.2. Transfer Codings for Compression 1137 The following transfer coding names for compression are defined by 1138 the same algorithm as their corresponding content coding: 1140 compress (and x-compress) 1141 See Section 6.1.2.1 of [Semantics]. 1143 deflate 1144 See Section 6.1.2.2 of [Semantics]. 1146 gzip (and x-gzip) 1147 See Section 6.1.2.3 of [Semantics]. 1149 7.3. Transfer Coding Registry 1151 The "HTTP Transfer Coding Registry" defines the namespace for 1152 transfer coding names. It is maintained at 1153 . 1155 Registrations MUST include the following fields: 1157 o Name 1159 o Description 1161 o Pointer to specification text 1162 Names of transfer codings MUST NOT overlap with names of content 1163 codings (Section 6.1.2 of [Semantics]) unless the encoding 1164 transformation is identical, as is the case for the compression 1165 codings defined in Section 7.2. 1167 The TE header field (Section 7.4) uses a pseudo parameter named "q" 1168 as rank value when multiple transfer codings are acceptable. Future 1169 registrations of transfer codings SHOULD NOT define parameters called 1170 "q" (case-insensitively) in order to avoid ambiguities. 1172 Values to be added to this namespace require IETF Review (see 1173 Section 4.8 of [RFC8126]), and MUST conform to the purpose of 1174 transfer coding defined in this specification. 1176 Use of program names for the identification of encoding formats is 1177 not desirable and is discouraged for future encodings. 1179 7.4. TE 1181 The "TE" header field in a request indicates what transfer codings, 1182 besides chunked, the client is willing to accept in response, and 1183 whether or not the client is willing to accept trailer fields in a 1184 chunked transfer coding. 1186 The TE field-value consists of a comma-separated list of transfer 1187 coding names, each allowing for optional parameters (as described in 1188 Section 7), and/or the keyword "trailers". A client MUST NOT send 1189 the chunked transfer coding name in TE; chunked is always acceptable 1190 for HTTP/1.1 recipients. 1192 TE = #t-codings 1193 t-codings = "trailers" / ( transfer-coding [ t-ranking ] ) 1194 t-ranking = OWS ";" OWS "q=" rank 1195 rank = ( "0" [ "." 0*3DIGIT ] ) 1196 / ( "1" [ "." 0*3("0") ] ) 1198 Three examples of TE use are below. 1200 TE: deflate 1201 TE: 1202 TE: trailers, deflate;q=0.5 1204 The presence of the keyword "trailers" indicates that the client is 1205 willing to accept trailer fields in a chunked transfer coding, as 1206 defined in Section 7.1.2, on behalf of itself and any downstream 1207 clients. For requests from an intermediary, this implies that 1208 either: (a) all downstream clients are willing to accept trailer 1209 fields in the forwarded response; or, (b) the intermediary will 1210 attempt to buffer the response on behalf of downstream recipients. 1211 Note that HTTP/1.1 does not define any means to limit the size of a 1212 chunked response such that an intermediary can be assured of 1213 buffering the entire response. 1215 When multiple transfer codings are acceptable, the client MAY rank 1216 the codings by preference using a case-insensitive "q" parameter 1217 (similar to the qvalues used in content negotiation fields, 1218 Section 8.4.1 of [Semantics]). The rank value is a real number in 1219 the range 0 through 1, where 0.001 is the least preferred and 1 is 1220 the most preferred; a value of 0 means "not acceptable". 1222 If the TE field-value is empty or if no TE field is present, the only 1223 acceptable transfer coding is chunked. A message with no transfer 1224 coding is always acceptable. 1226 Since the TE header field only applies to the immediate connection, a 1227 sender of TE MUST also send a "TE" connection option within the 1228 Connection header field (Section 9.1) in order to prevent the TE 1229 field from being forwarded by intermediaries that do not support its 1230 semantics. 1232 8. Handling Incomplete Messages 1234 A server that receives an incomplete request message, usually due to 1235 a canceled request or a triggered timeout exception, MAY send an 1236 error response prior to closing the connection. 1238 A client that receives an incomplete response message, which can 1239 occur when a connection is closed prematurely or when decoding a 1240 supposedly chunked transfer coding fails, MUST record the message as 1241 incomplete. Cache requirements for incomplete responses are defined 1242 in Section 3 of [Caching]. 1244 If a response terminates in the middle of the header section (before 1245 the empty line is received) and the status code might rely on header 1246 fields to convey the full meaning of the response, then the client 1247 cannot assume that meaning has been conveyed; the client might need 1248 to repeat the request in order to determine what action to take next. 1250 A message body that uses the chunked transfer coding is incomplete if 1251 the zero-sized chunk that terminates the encoding has not been 1252 received. A message that uses a valid Content-Length is incomplete 1253 if the size of the message body received (in octets) is less than the 1254 value given by Content-Length. A response that has neither chunked 1255 transfer coding nor Content-Length is terminated by closure of the 1256 connection and, thus, is considered complete regardless of the number 1257 of message body octets received, provided that the header section was 1258 received intact. 1260 9. Connection Management 1262 HTTP messaging is independent of the underlying transport- or 1263 session-layer connection protocol(s). HTTP only presumes a reliable 1264 transport with in-order delivery of requests and the corresponding 1265 in-order delivery of responses. The mapping of HTTP request and 1266 response structures onto the data units of an underlying transport 1267 protocol is outside the scope of this specification. 1269 As described in Section 5.2 of [Semantics], the specific connection 1270 protocols to be used for an HTTP interaction are determined by client 1271 configuration and the target URI. For example, the "http" URI scheme 1272 (Section 2.5.1 of [Semantics]) indicates a default connection of TCP 1273 over IP, with a default TCP port of 80, but the client might be 1274 configured to use a proxy via some other connection, port, or 1275 protocol. 1277 HTTP implementations are expected to engage in connection management, 1278 which includes maintaining the state of current connections, 1279 establishing a new connection or reusing an existing connection, 1280 processing messages received on a connection, detecting connection 1281 failures, and closing each connection. Most clients maintain 1282 multiple connections in parallel, including more than one connection 1283 per server endpoint. Most servers are designed to maintain thousands 1284 of concurrent connections, while controlling request queues to enable 1285 fair use and detect denial-of-service attacks. 1287 9.1. Connection 1289 The "Connection" header field allows the sender to indicate desired 1290 control options for the current connection. In order to avoid 1291 confusing downstream recipients, a proxy or gateway MUST remove or 1292 replace any received connection options before forwarding the 1293 message. 1295 When a header field aside from Connection is used to supply control 1296 information for or about the current connection, the sender MUST list 1297 the corresponding field-name within the Connection header field. A 1298 proxy or gateway MUST parse a received Connection header field before 1299 a message is forwarded and, for each connection-option in this field, 1300 remove any header field(s) from the message with the same name as the 1301 connection-option, and then remove the Connection header field itself 1302 (or replace it with the intermediary's own connection options for the 1303 forwarded message). 1305 Hence, the Connection header field provides a declarative way of 1306 distinguishing header fields that are only intended for the immediate 1307 recipient ("hop-by-hop") from those fields that are intended for all 1308 recipients on the chain ("end-to-end"), enabling the message to be 1309 self-descriptive and allowing future connection-specific extensions 1310 to be deployed without fear that they will be blindly forwarded by 1311 older intermediaries. 1313 The Connection header field's value has the following grammar: 1315 Connection = 1#connection-option 1316 connection-option = token 1318 Connection options are case-insensitive. 1320 A sender MUST NOT send a connection option corresponding to a header 1321 field that is intended for all recipients of the payload. For 1322 example, Cache-Control is never appropriate as a connection option 1323 (Section 5.2 of [Caching]). 1325 The connection options do not always correspond to a header field 1326 present in the message, since a connection-specific header field 1327 might not be needed if there are no parameters associated with a 1328 connection option. In contrast, a connection-specific header field 1329 that is received without a corresponding connection option usually 1330 indicates that the field has been improperly forwarded by an 1331 intermediary and ought to be ignored by the recipient. 1333 When defining new connection options, specification authors ought to 1334 survey existing header field names and ensure that the new connection 1335 option does not share the same name as an already deployed header 1336 field. Defining a new connection option essentially reserves that 1337 potential field-name for carrying additional information related to 1338 the connection option, since it would be unwise for senders to use 1339 that field-name for anything else. 1341 The "close" connection option is defined for a sender to signal that 1342 this connection will be closed after completion of the response. For 1343 example, 1345 Connection: close 1347 in either the request or the response header fields indicates that 1348 the sender is going to close the connection after the current 1349 request/response is complete (Section 9.6). 1351 A client that does not support persistent connections MUST send the 1352 "close" connection option in every request message. 1354 A server that does not support persistent connections MUST send the 1355 "close" connection option in every response message that does not 1356 have a 1xx (Informational) status code. 1358 9.2. Establishment 1360 It is beyond the scope of this specification to describe how 1361 connections are established via various transport- or session-layer 1362 protocols. Each connection applies to only one transport link. 1364 9.3. Persistence 1366 HTTP/1.1 defaults to the use of "persistent connections", allowing 1367 multiple requests and responses to be carried over a single 1368 connection. The "close" connection option is used to signal that a 1369 connection will not persist after the current request/response. HTTP 1370 implementations SHOULD support persistent connections. 1372 A recipient determines whether a connection is persistent or not 1373 based on the most recently received message's protocol version and 1374 Connection header field (if any): 1376 o If the "close" connection option is present, the connection will 1377 not persist after the current response; else, 1379 o If the received protocol is HTTP/1.1 (or later), the connection 1380 will persist after the current response; else, 1382 o If the received protocol is HTTP/1.0, the "keep-alive" connection 1383 option is present, either the recipient is not a proxy or the 1384 message is a response, and the recipient wishes to honor the 1385 HTTP/1.0 "keep-alive" mechanism, the connection will persist after 1386 the current response; otherwise, 1388 o The connection will close after the current response. 1390 A client MAY send additional requests on a persistent connection 1391 until it sends or receives a "close" connection option or receives an 1392 HTTP/1.0 response without a "keep-alive" connection option. 1394 In order to remain persistent, all messages on a connection need to 1395 have a self-defined message length (i.e., one not defined by closure 1396 of the connection), as described in Section 6. A server MUST read 1397 the entire request message body or close the connection after sending 1398 its response, since otherwise the remaining data on a persistent 1399 connection would be misinterpreted as the next request. Likewise, a 1400 client MUST read the entire response message body if it intends to 1401 reuse the same connection for a subsequent request. 1403 A proxy server MUST NOT maintain a persistent connection with an 1404 HTTP/1.0 client (see Section 19.7.1 of [RFC2068] for information and 1405 discussion of the problems with the Keep-Alive header field 1406 implemented by many HTTP/1.0 clients). 1408 See Appendix C.1.2 for more information on backwards compatibility 1409 with HTTP/1.0 clients. 1411 9.3.1. Retrying Requests 1413 Connections can be closed at any time, with or without intention. 1414 Implementations ought to anticipate the need to recover from 1415 asynchronous close events. 1417 When an inbound connection is closed prematurely, a client MAY open a 1418 new connection and automatically retransmit an aborted sequence of 1419 requests if all of those requests have idempotent methods 1420 (Section 7.2.2 of [Semantics]). A proxy MUST NOT automatically retry 1421 non-idempotent requests. 1423 A user agent MUST NOT automatically retry a request with a non- 1424 idempotent method unless it has some means to know that the request 1425 semantics are actually idempotent, regardless of the method, or some 1426 means to detect that the original request was never applied. For 1427 example, a user agent that knows (through design or configuration) 1428 that a POST request to a given resource is safe can repeat that 1429 request automatically. Likewise, a user agent designed specifically 1430 to operate on a version control repository might be able to recover 1431 from partial failure conditions by checking the target resource 1432 revision(s) after a failed connection, reverting or fixing any 1433 changes that were partially applied, and then automatically retrying 1434 the requests that failed. 1436 A client SHOULD NOT automatically retry a failed automatic retry. 1438 9.3.2. Pipelining 1440 A client that supports persistent connections MAY "pipeline" its 1441 requests (i.e., send multiple requests without waiting for each 1442 response). A server MAY process a sequence of pipelined requests in 1443 parallel if they all have safe methods (Section 7.2.1 of 1444 [Semantics]), but it MUST send the corresponding responses in the 1445 same order that the requests were received. 1447 A client that pipelines requests SHOULD retry unanswered requests if 1448 the connection closes before it receives all of the corresponding 1449 responses. When retrying pipelined requests after a failed 1450 connection (a connection not explicitly closed by the server in its 1451 last complete response), a client MUST NOT pipeline immediately after 1452 connection establishment, since the first remaining request in the 1453 prior pipeline might have caused an error response that can be lost 1454 again if multiple requests are sent on a prematurely closed 1455 connection (see the TCP reset problem described in Section 9.6). 1457 Idempotent methods (Section 7.2.2 of [Semantics]) are significant to 1458 pipelining because they can be automatically retried after a 1459 connection failure. A user agent SHOULD NOT pipeline requests after 1460 a non-idempotent method, until the final response status code for 1461 that method has been received, unless the user agent has a means to 1462 detect and recover from partial failure conditions involving the 1463 pipelined sequence. 1465 An intermediary that receives pipelined requests MAY pipeline those 1466 requests when forwarding them inbound, since it can rely on the 1467 outbound user agent(s) to determine what requests can be safely 1468 pipelined. If the inbound connection fails before receiving a 1469 response, the pipelining intermediary MAY attempt to retry a sequence 1470 of requests that have yet to receive a response if the requests all 1471 have idempotent methods; otherwise, the pipelining intermediary 1472 SHOULD forward any received responses and then close the 1473 corresponding outbound connection(s) so that the outbound user 1474 agent(s) can recover accordingly. 1476 9.4. Concurrency 1478 A client ought to limit the number of simultaneous open connections 1479 that it maintains to a given server. 1481 Previous revisions of HTTP gave a specific number of connections as a 1482 ceiling, but this was found to be impractical for many applications. 1483 As a result, this specification does not mandate a particular maximum 1484 number of connections but, instead, encourages clients to be 1485 conservative when opening multiple connections. 1487 Multiple connections are typically used to avoid the "head-of-line 1488 blocking" problem, wherein a request that takes significant server- 1489 side processing and/or has a large payload blocks subsequent requests 1490 on the same connection. However, each connection consumes server 1491 resources. Furthermore, using multiple connections can cause 1492 undesirable side effects in congested networks. 1494 Note that a server might reject traffic that it deems abusive or 1495 characteristic of a denial-of-service attack, such as an excessive 1496 number of open connections from a single client. 1498 9.5. Failures and Timeouts 1500 Servers will usually have some timeout value beyond which they will 1501 no longer maintain an inactive connection. Proxy servers might make 1502 this a higher value since it is likely that the client will be making 1503 more connections through the same proxy server. The use of 1504 persistent connections places no requirements on the length (or 1505 existence) of this timeout for either the client or the server. 1507 A client or server that wishes to time out SHOULD issue a graceful 1508 close on the connection. Implementations SHOULD constantly monitor 1509 open connections for a received closure signal and respond to it as 1510 appropriate, since prompt closure of both sides of a connection 1511 enables allocated system resources to be reclaimed. 1513 A client, server, or proxy MAY close the transport connection at any 1514 time. For example, a client might have started to send a new request 1515 at the same time that the server has decided to close the "idle" 1516 connection. From the server's point of view, the connection is being 1517 closed while it was idle, but from the client's point of view, a 1518 request is in progress. 1520 A server SHOULD sustain persistent connections, when possible, and 1521 allow the underlying transport's flow-control mechanisms to resolve 1522 temporary overloads, rather than terminate connections with the 1523 expectation that clients will retry. The latter technique can 1524 exacerbate network congestion. 1526 A client sending a message body SHOULD monitor the network connection 1527 for an error response while it is transmitting the request. If the 1528 client sees a response that indicates the server does not wish to 1529 receive the message body and is closing the connection, the client 1530 SHOULD immediately cease transmitting the body and close its side of 1531 the connection. 1533 9.6. Tear-down 1535 The Connection header field (Section 9.1) provides a "close" 1536 connection option that a sender SHOULD send when it wishes to close 1537 the connection after the current request/response pair. 1539 A client that sends a "close" connection option MUST NOT send further 1540 requests on that connection (after the one containing "close") and 1541 MUST close the connection after reading the final response message 1542 corresponding to this request. 1544 A server that receives a "close" connection option MUST initiate a 1545 close of the connection (see below) after it sends the final response 1546 to the request that contained "close". The server SHOULD send a 1547 "close" connection option in its final response on that connection. 1548 The server MUST NOT process any further requests received on that 1549 connection. 1551 A server that sends a "close" connection option MUST initiate a close 1552 of the connection (see below) after it sends the response containing 1553 "close". The server MUST NOT process any further requests received 1554 on that connection. 1556 A client that receives a "close" connection option MUST cease sending 1557 requests on that connection and close the connection after reading 1558 the response message containing the "close"; if additional pipelined 1559 requests had been sent on the connection, the client SHOULD NOT 1560 assume that they will be processed by the server. 1562 If a server performs an immediate close of a TCP connection, there is 1563 a significant risk that the client will not be able to read the last 1564 HTTP response. If the server receives additional data from the 1565 client on a fully closed connection, such as another request that was 1566 sent by the client before receiving the server's response, the 1567 server's TCP stack will send a reset packet to the client; 1568 unfortunately, the reset packet might erase the client's 1569 unacknowledged input buffers before they can be read and interpreted 1570 by the client's HTTP parser. 1572 To avoid the TCP reset problem, servers typically close a connection 1573 in stages. First, the server performs a half-close by closing only 1574 the write side of the read/write connection. The server then 1575 continues to read from the connection until it receives a 1576 corresponding close by the client, or until the server is reasonably 1577 certain that its own TCP stack has received the client's 1578 acknowledgement of the packet(s) containing the server's last 1579 response. Finally, the server fully closes the connection. 1581 It is unknown whether the reset problem is exclusive to TCP or might 1582 also be found in other transport connection protocols. 1584 9.7. Upgrade 1586 The "Upgrade" header field is intended to provide a simple mechanism 1587 for transitioning from HTTP/1.1 to some other protocol on the same 1588 connection. A client MAY send a list of protocols in the Upgrade 1589 header field of a request to invite the server to switch to one or 1590 more of those protocols, in order of descending preference, before 1591 sending the final response. A server MAY ignore a received Upgrade 1592 header field if it wishes to continue using the current protocol on 1593 that connection. Upgrade cannot be used to insist on a protocol 1594 change. 1596 Upgrade = 1#protocol 1598 protocol = protocol-name ["/" protocol-version] 1599 protocol-name = token 1600 protocol-version = token 1602 A server that sends a 101 (Switching Protocols) response MUST send an 1603 Upgrade header field to indicate the new protocol(s) to which the 1604 connection is being switched; if multiple protocol layers are being 1605 switched, the sender MUST list the protocols in layer-ascending 1606 order. A server MUST NOT switch to a protocol that was not indicated 1607 by the client in the corresponding request's Upgrade header field. A 1608 server MAY choose to ignore the order of preference indicated by the 1609 client and select the new protocol(s) based on other factors, such as 1610 the nature of the request or the current load on the server. 1612 A server that sends a 426 (Upgrade Required) response MUST send an 1613 Upgrade header field to indicate the acceptable protocols, in order 1614 of descending preference. 1616 A server MAY send an Upgrade header field in any other response to 1617 advertise that it implements support for upgrading to the listed 1618 protocols, in order of descending preference, when appropriate for a 1619 future request. 1621 The following is a hypothetical example sent by a client: 1623 GET /hello.txt HTTP/1.1 1624 Host: www.example.com 1625 Connection: upgrade 1626 Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11 1628 The capabilities and nature of the application-level communication 1629 after the protocol change is entirely dependent upon the new 1630 protocol(s) chosen. However, immediately after sending the 101 1631 (Switching Protocols) response, the server is expected to continue 1632 responding to the original request as if it had received its 1633 equivalent within the new protocol (i.e., the server still has an 1634 outstanding request to satisfy after the protocol has been changed, 1635 and is expected to do so without requiring the request to be 1636 repeated). 1638 For example, if the Upgrade header field is received in a GET request 1639 and the server decides to switch protocols, it first responds with a 1640 101 (Switching Protocols) message in HTTP/1.1 and then immediately 1641 follows that with the new protocol's equivalent of a response to a 1642 GET on the target resource. This allows a connection to be upgraded 1643 to protocols with the same semantics as HTTP without the latency cost 1644 of an additional round trip. A server MUST NOT switch protocols 1645 unless the received message semantics can be honored by the new 1646 protocol; an OPTIONS request can be honored by any protocol. 1648 The following is an example response to the above hypothetical 1649 request: 1651 HTTP/1.1 101 Switching Protocols 1652 Connection: upgrade 1653 Upgrade: HTTP/2.0 1655 [... data stream switches to HTTP/2.0 with an appropriate response 1656 (as defined by new protocol) to the "GET /hello.txt" request ...] 1658 When Upgrade is sent, the sender MUST also send a Connection header 1659 field (Section 9.1) that contains an "upgrade" connection option, in 1660 order to prevent Upgrade from being accidentally forwarded by 1661 intermediaries that might not implement the listed protocols. A 1662 server MUST ignore an Upgrade header field that is received in an 1663 HTTP/1.0 request. 1665 A client cannot begin using an upgraded protocol on the connection 1666 until it has completely sent the request message (i.e., the client 1667 can't change the protocol it is sending in the middle of a message). 1668 If a server receives both an Upgrade and an Expect header field with 1669 the "100-continue" expectation (Section 8.1.1 of [Semantics]), the 1670 server MUST send a 100 (Continue) response before sending a 101 1671 (Switching Protocols) response. 1673 The Upgrade header field only applies to switching protocols on top 1674 of the existing connection; it cannot be used to switch the 1675 underlying connection (transport) protocol, nor to switch the 1676 existing communication to a different connection. For those 1677 purposes, it is more appropriate to use a 3xx (Redirection) response 1678 (Section 9.4 of [Semantics]). 1680 9.7.1. Upgrade Protocol Names 1682 This specification only defines the protocol name "HTTP" for use by 1683 the family of Hypertext Transfer Protocols, as defined by the HTTP 1684 version rules of Section 3.5 of [Semantics] and future updates to 1685 this specification. Additional protocol names ought to be registered 1686 using the registration procedure defined in Section 9.7.2. 1688 +------+-------------------+--------------------+-------------------+ 1689 | Name | Description | Expected Version | Reference | 1690 | | | Tokens | | 1691 +------+-------------------+--------------------+-------------------+ 1692 | HTTP | Hypertext | any DIGIT.DIGIT | Section 3.5 of | 1693 | | Transfer Protocol | (e.g, "2.0") | [Semantics] | 1694 +------+-------------------+--------------------+-------------------+ 1696 9.7.2. Upgrade Token Registry 1698 The "Hypertext Transfer Protocol (HTTP) Upgrade Token Registry" 1699 defines the namespace for protocol-name tokens used to identify 1700 protocols in the Upgrade header field. The registry is maintained at 1701 . 1703 Each registered protocol name is associated with contact information 1704 and an optional set of specifications that details how the connection 1705 will be processed after it has been upgraded. 1707 Registrations happen on a "First Come First Served" basis (see 1708 Section 4.4 of [RFC8126]) and are subject to the following rules: 1710 1. A protocol-name token, once registered, stays registered forever. 1712 2. The registration MUST name a responsible party for the 1713 registration. 1715 3. The registration MUST name a point of contact. 1717 4. The registration MAY name a set of specifications associated with 1718 that token. Such specifications need not be publicly available. 1720 5. The registration SHOULD name a set of expected "protocol-version" 1721 tokens associated with that token at the time of registration. 1723 6. The responsible party MAY change the registration at any time. 1724 The IANA will keep a record of all such changes, and make them 1725 available upon request. 1727 7. The IESG MAY reassign responsibility for a protocol token. This 1728 will normally only be used in the case when a responsible party 1729 cannot be contacted. 1731 10. Enclosing Messages as Data 1732 10.1. Media Type message/http 1734 The message/http media type can be used to enclose a single HTTP 1735 request or response message, provided that it obeys the MIME 1736 restrictions for all "message" types regarding line length and 1737 encodings. 1739 Type name: message 1741 Subtype name: http 1743 Required parameters: N/A 1745 Optional parameters: version, msgtype 1747 version: The HTTP-version number of the enclosed message (e.g., 1748 "1.1"). If not present, the version can be determined from the 1749 first line of the body. 1751 msgtype: The message type -- "request" or "response". If not 1752 present, the type can be determined from the first line of the 1753 body. 1755 Encoding considerations: only "7bit", "8bit", or "binary" are 1756 permitted 1758 Security considerations: see Section 11 1760 Interoperability considerations: N/A 1762 Published specification: This specification (see Section 10.1). 1764 Applications that use this media type: N/A 1766 Fragment identifier considerations: N/A 1768 Additional information: 1770 Magic number(s): N/A 1772 Deprecated alias names for this type: N/A 1774 File extension(s): N/A 1776 Macintosh file type code(s): N/A 1778 Person and email address to contact for further information: 1779 See Authors' Addresses section. 1781 Intended usage: COMMON 1783 Restrictions on usage: N/A 1785 Author: See Authors' Addresses section. 1787 Change controller: IESG 1789 10.2. Media Type application/http 1791 The application/http media type can be used to enclose a pipeline of 1792 one or more HTTP request or response messages (not intermixed). 1794 Type name: application 1796 Subtype name: http 1798 Required parameters: N/A 1800 Optional parameters: version, msgtype 1802 version: The HTTP-version number of the enclosed messages (e.g., 1803 "1.1"). If not present, the version can be determined from the 1804 first line of the body. 1806 msgtype: The message type -- "request" or "response". If not 1807 present, the type can be determined from the first line of the 1808 body. 1810 Encoding considerations: HTTP messages enclosed by this type are in 1811 "binary" format; use of an appropriate Content-Transfer-Encoding 1812 is required when transmitted via email. 1814 Security considerations: see Section 11 1816 Interoperability considerations: N/A 1818 Published specification: This specification (see Section 10.2). 1820 Applications that use this media type: N/A 1822 Fragment identifier considerations: N/A 1824 Additional information: 1826 Deprecated alias names for this type: N/A 1828 Magic number(s): N/A 1829 File extension(s): N/A 1831 Macintosh file type code(s): N/A 1833 Person and email address to contact for further information: 1834 See Authors' Addresses section. 1836 Intended usage: COMMON 1838 Restrictions on usage: N/A 1840 Author: See Authors' Addresses section. 1842 Change controller: IESG 1844 11. Security Considerations 1846 This section is meant to inform developers, information providers, 1847 and users of known security considerations relevant to HTTP message 1848 syntax, parsing, and routing. Security considerations about HTTP 1849 semantics and payloads are addressed in [Semantics]. 1851 11.1. Response Splitting 1853 Response splitting (a.k.a, CRLF injection) is a common technique, 1854 used in various attacks on Web usage, that exploits the line-based 1855 nature of HTTP message framing and the ordered association of 1856 requests to responses on persistent connections [Klein]. This 1857 technique can be particularly damaging when the requests pass through 1858 a shared cache. 1860 Response splitting exploits a vulnerability in servers (usually 1861 within an application server) where an attacker can send encoded data 1862 within some parameter of the request that is later decoded and echoed 1863 within any of the response header fields of the response. If the 1864 decoded data is crafted to look like the response has ended and a 1865 subsequent response has begun, the response has been split and the 1866 content within the apparent second response is controlled by the 1867 attacker. The attacker can then make any other request on the same 1868 persistent connection and trick the recipients (including 1869 intermediaries) into believing that the second half of the split is 1870 an authoritative answer to the second request. 1872 For example, a parameter within the request-target might be read by 1873 an application server and reused within a redirect, resulting in the 1874 same parameter being echoed in the Location header field of the 1875 response. If the parameter is decoded by the application and not 1876 properly encoded when placed in the response field, the attacker can 1877 send encoded CRLF octets and other content that will make the 1878 application's single response look like two or more responses. 1880 A common defense against response splitting is to filter requests for 1881 data that looks like encoded CR and LF (e.g., "%0D" and "%0A"). 1882 However, that assumes the application server is only performing URI 1883 decoding, rather than more obscure data transformations like charset 1884 transcoding, XML entity translation, base64 decoding, sprintf 1885 reformatting, etc. A more effective mitigation is to prevent 1886 anything other than the server's core protocol libraries from sending 1887 a CR or LF within the header section, which means restricting the 1888 output of header fields to APIs that filter for bad octets and not 1889 allowing application servers to write directly to the protocol 1890 stream. 1892 11.2. Request Smuggling 1894 Request smuggling ([Linhart]) is a technique that exploits 1895 differences in protocol parsing among various recipients to hide 1896 additional requests (which might otherwise be blocked or disabled by 1897 policy) within an apparently harmless request. Like response 1898 splitting, request smuggling can lead to a variety of attacks on HTTP 1899 usage. 1901 This specification has introduced new requirements on request 1902 parsing, particularly with regard to message framing in Section 6.3, 1903 to reduce the effectiveness of request smuggling. 1905 11.3. Message Integrity 1907 HTTP does not define a specific mechanism for ensuring message 1908 integrity, instead relying on the error-detection ability of 1909 underlying transport protocols and the use of length or chunk- 1910 delimited framing to detect completeness. Additional integrity 1911 mechanisms, such as hash functions or digital signatures applied to 1912 the content, can be selectively added to messages via extensible 1913 metadata header fields. Historically, the lack of a single integrity 1914 mechanism has been justified by the informal nature of most HTTP 1915 communication. However, the prevalence of HTTP as an information 1916 access mechanism has resulted in its increasing use within 1917 environments where verification of message integrity is crucial. 1919 User agents are encouraged to implement configurable means for 1920 detecting and reporting failures of message integrity such that those 1921 means can be enabled within environments for which integrity is 1922 necessary. For example, a browser being used to view medical history 1923 or drug interaction information needs to indicate to the user when 1924 such information is detected by the protocol to be incomplete, 1925 expired, or corrupted during transfer. Such mechanisms might be 1926 selectively enabled via user agent extensions or the presence of 1927 message integrity metadata in a response. At a minimum, user agents 1928 ought to provide some indication that allows a user to distinguish 1929 between a complete and incomplete response message (Section 8) when 1930 such verification is desired. 1932 11.4. Message Confidentiality 1934 HTTP relies on underlying transport protocols to provide message 1935 confidentiality when that is desired. HTTP has been specifically 1936 designed to be independent of the transport protocol, such that it 1937 can be used over many different forms of encrypted connection, with 1938 the selection of such transports being identified by the choice of 1939 URI scheme or within user agent configuration. 1941 The "https" scheme can be used to identify resources that require a 1942 confidential connection, as described in Section 2.5.2 of 1943 [Semantics]. 1945 12. IANA Considerations 1947 The change controller for the following registrations is: "IETF 1948 (iesg@ietf.org) - Internet Engineering Task Force". 1950 12.1. Header Field Registration 1952 Please update the "Message Headers" registry of "Permanent Message 1953 Header Field Names" at with the header field names listed in the two tables of 1955 Section 5. 1957 12.2. Media Type Registration 1959 Please update the "Media Types" registry at 1960 with the registration 1961 information in Section 10.1 and Section 10.2 for the media types 1962 "message/http" and "application/http", respectively. 1964 12.3. Transfer Coding Registration 1966 Please update the "HTTP Transfer Coding Registry" at 1967 with the 1968 registration procedure of Section 7.3 and the content coding names 1969 summarized in the table of Section 7. 1971 12.4. Upgrade Token Registration 1973 Please update the "Hypertext Transfer Protocol (HTTP) Upgrade Token 1974 Registry" at 1975 with the registration procedure of Section 9.7.2 and the upgrade 1976 token names summarized in the table of Section 9.7.1. 1978 13. References 1980 13.1. Normative References 1982 [Caching] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 1983 Ed., "HTTP Caching", draft-ietf-httpbis-cache-02 (work in 1984 progress), July 2018. 1986 [RFC1950] Deutsch, L. and J-L. Gailly, "ZLIB Compressed Data Format 1987 Specification version 3.3", RFC 1950, 1988 DOI 10.17487/RFC1950, May 1996, 1989 . 1991 [RFC1951] Deutsch, P., "DEFLATE Compressed Data Format Specification 1992 version 1.3", RFC 1951, DOI 10.17487/RFC1951, May 1996, 1993 . 1995 [RFC1952] Deutsch, P., Gailly, J-L., Adler, M., Deutsch, L., and G. 1996 Randers-Pehrson, "GZIP file format specification version 1997 4.3", RFC 1952, DOI 10.17487/RFC1952, May 1996, 1998 . 2000 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2001 Requirement Levels", BCP 14, RFC 2119, 2002 DOI 10.17487/RFC2119, March 1997, 2003 . 2005 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 2006 Resource Identifier (URI): Generic Syntax", STD 66, 2007 RFC 3986, DOI 10.17487/RFC3986, January 2005, 2008 . 2010 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 2011 Specifications: ABNF", STD 68, RFC 5234, 2012 DOI 10.17487/RFC5234, January 2008, 2013 . 2015 [Semantics] 2016 Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 2017 Ed., "HTTP Semantics", draft-ietf-httpbis-semantics-02 2018 (work in progress), July 2018. 2020 [USASCII] American National Standards Institute, "Coded Character 2021 Set -- 7-bit American Standard Code for Information 2022 Interchange", ANSI X3.4, 1986. 2024 [Welch] Welch, T., "A Technique for High-Performance Data 2025 Compression", IEEE Computer 17(6), June 1984. 2027 13.2. Informative References 2029 [Err4667] RFC Errata, Erratum ID 4667, RFC 7230, 2030 . 2032 [Klein] Klein, A., "Divide and Conquer - HTTP Response Splitting, 2033 Web Cache Poisoning Attacks, and Related Topics", March 2034 2004, . 2037 [Linhart] Linhart, C., Klein, A., Heled, R., and S. Orrin, "HTTP 2038 Request Smuggling", June 2005, 2039 . 2041 [RFC1945] Berners-Lee, T., Fielding, R., and H. Nielsen, "Hypertext 2042 Transfer Protocol -- HTTP/1.0", RFC 1945, 2043 DOI 10.17487/RFC1945, May 1996, 2044 . 2046 [RFC2045] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2047 Extensions (MIME) Part One: Format of Internet Message 2048 Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996, 2049 . 2051 [RFC2046] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2052 Extensions (MIME) Part Two: Media Types", RFC 2046, 2053 DOI 10.17487/RFC2046, November 1996, 2054 . 2056 [RFC2049] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2057 Extensions (MIME) Part Five: Conformance Criteria and 2058 Examples", RFC 2049, DOI 10.17487/RFC2049, November 1996, 2059 . 2061 [RFC2068] Fielding, R., Gettys, J., Mogul, J., Nielsen, H., and T. 2062 Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", 2063 RFC 2068, DOI 10.17487/RFC2068, January 1997, 2064 . 2066 [RFC2557] Palme, F., Hopmann, A., Shelness, N., and E. Stefferud, 2067 "MIME Encapsulation of Aggregate Documents, such as HTML 2068 (MHTML)", RFC 2557, DOI 10.17487/RFC2557, March 1999, 2069 . 2071 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 2072 DOI 10.17487/RFC5322, October 2008, 2073 . 2075 [RFC6265] Barth, A., "HTTP State Management Mechanism", RFC 6265, 2076 DOI 10.17487/RFC6265, April 2011, 2077 . 2079 [RFC7230] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2080 Protocol (HTTP/1.1): Message Syntax and Routing", 2081 RFC 7230, DOI 10.17487/RFC7230, June 2014, 2082 . 2084 [RFC7231] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2085 Protocol (HTTP/1.1): Semantics and Content", RFC 7231, 2086 DOI 10.17487/RFC7231, June 2014, 2087 . 2089 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2090 Writing an IANA Considerations Section in RFCs", BCP 26, 2091 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2092 . 2094 Appendix A. Collected ABNF 2096 In the collected ABNF below, list rules are expanded as per 2097 Section 11 of [Semantics]. 2099 BWS = 2101 Connection = *( "," OWS ) connection-option *( OWS "," [ OWS 2102 connection-option ] ) 2104 HTTP-message = start-line *( header-field CRLF ) CRLF [ message-body 2105 ] 2106 HTTP-name = %x48.54.54.50 ; HTTP 2107 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 2109 OWS = 2111 RWS = 2113 TE = [ ( "," / t-codings ) *( OWS "," [ OWS t-codings ] ) ] 2114 Transfer-Encoding = *( "," OWS ) transfer-coding *( OWS "," [ OWS 2115 transfer-coding ] ) 2117 Upgrade = *( "," OWS ) protocol *( OWS "," [ OWS protocol ] ) 2119 absolute-URI = 2120 absolute-form = absolute-URI 2121 absolute-path = 2122 asterisk-form = "*" 2123 authority = 2124 authority-form = authority 2126 chunk = chunk-size [ chunk-ext ] CRLF chunk-data CRLF 2127 chunk-data = 1*OCTET 2128 chunk-ext = *( BWS ";" BWS chunk-ext-name [ BWS "=" BWS chunk-ext-val 2129 ] ) 2130 chunk-ext-name = token 2131 chunk-ext-val = token / quoted-string 2132 chunk-size = 1*HEXDIG 2133 chunked-body = *chunk last-chunk trailer-part CRLF 2134 comment = 2135 connection-option = token 2137 field-name = 2138 field-value = 2140 header-field = field-name ":" OWS field-value OWS 2141 last-chunk = 1*"0" [ chunk-ext ] CRLF 2143 message-body = *OCTET 2144 method = token 2146 obs-fold = CRLF 1*( SP / HTAB ) 2147 obs-text = 2148 origin-form = absolute-path [ "?" query ] 2150 port = 2151 protocol = protocol-name [ "/" protocol-version ] 2152 protocol-name = token 2153 protocol-version = token 2155 query = 2156 quoted-string = 2158 rank = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] ) 2159 reason-phrase = *( HTAB / SP / VCHAR / obs-text ) 2160 request-line = method SP request-target SP HTTP-version CRLF 2161 request-target = origin-form / absolute-form / authority-form / 2162 asterisk-form 2164 start-line = request-line / status-line 2165 status-code = 3DIGIT 2166 status-line = HTTP-version SP status-code SP reason-phrase CRLF 2168 t-codings = "trailers" / ( transfer-coding [ t-ranking ] ) 2169 t-ranking = OWS ";" OWS "q=" rank 2170 token = 2171 trailer-part = *( header-field CRLF ) 2172 transfer-coding = "chunked" / "compress" / "deflate" / "gzip" / 2173 transfer-extension 2174 transfer-extension = token *( OWS ";" OWS transfer-parameter ) 2175 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 2177 uri-host = 2179 Appendix B. Differences between HTTP and MIME 2181 HTTP/1.1 uses many of the constructs defined for the Internet Message 2182 Format [RFC5322] and the Multipurpose Internet Mail Extensions (MIME) 2183 [RFC2045] to allow a message body to be transmitted in an open 2184 variety of representations and with extensible header fields. 2185 However, RFC 2045 is focused only on email; applications of HTTP have 2186 many characteristics that differ from email; hence, HTTP has features 2187 that differ from MIME. These differences were carefully chosen to 2188 optimize performance over binary connections, to allow greater 2189 freedom in the use of new media types, to make date comparisons 2190 easier, and to acknowledge the practice of some early HTTP servers 2191 and clients. 2193 This appendix describes specific areas where HTTP differs from MIME. 2194 Proxies and gateways to and from strict MIME environments need to be 2195 aware of these differences and provide the appropriate conversions 2196 where necessary. 2198 B.1. MIME-Version 2200 HTTP is not a MIME-compliant protocol. However, messages can include 2201 a single MIME-Version header field to indicate what version of the 2202 MIME protocol was used to construct the message. Use of the MIME- 2203 Version header field indicates that the message is in full 2204 conformance with the MIME protocol (as defined in [RFC2045]). 2205 Senders are responsible for ensuring full conformance (where 2206 possible) when exporting HTTP messages to strict MIME environments. 2208 B.2. Conversion to Canonical Form 2210 MIME requires that an Internet mail body part be converted to 2211 canonical form prior to being transferred, as described in Section 4 2212 of [RFC2049]. Section 6.1.1.2 of [Semantics] describes the forms 2213 allowed for subtypes of the "text" media type when transmitted over 2214 HTTP. [RFC2046] requires that content with a type of "text" 2215 represent line breaks as CRLF and forbids the use of CR or LF outside 2216 of line break sequences. HTTP allows CRLF, bare CR, and bare LF to 2217 indicate a line break within text content. 2219 A proxy or gateway from HTTP to a strict MIME environment ought to 2220 translate all line breaks within text media types to the RFC 2049 2221 canonical form of CRLF. Note, however, this might be complicated by 2222 the presence of a Content-Encoding and by the fact that HTTP allows 2223 the use of some charsets that do not use octets 13 and 10 to 2224 represent CR and LF, respectively. 2226 Conversion will break any cryptographic checksums applied to the 2227 original content unless the original content is already in canonical 2228 form. Therefore, the canonical form is recommended for any content 2229 that uses such checksums in HTTP. 2231 B.3. Conversion of Date Formats 2233 HTTP/1.1 uses a restricted set of date formats (Section 10.1.1.1 of 2234 [Semantics]) to simplify the process of date comparison. Proxies and 2235 gateways from other protocols ought to ensure that any Date header 2236 field present in a message conforms to one of the HTTP/1.1 formats 2237 and rewrite the date if necessary. 2239 B.4. Conversion of Content-Encoding 2241 MIME does not include any concept equivalent to HTTP/1.1's Content- 2242 Encoding header field. Since this acts as a modifier on the media 2243 type, proxies and gateways from HTTP to MIME-compliant protocols 2244 ought to either change the value of the Content-Type header field or 2245 decode the representation before forwarding the message. (Some 2246 experimental applications of Content-Type for Internet mail have used 2247 a media-type parameter of ";conversions=" to perform 2248 a function equivalent to Content-Encoding. However, this parameter 2249 is not part of the MIME standards). 2251 B.5. Conversion of Content-Transfer-Encoding 2253 HTTP does not use the Content-Transfer-Encoding field of MIME. 2254 Proxies and gateways from MIME-compliant protocols to HTTP need to 2255 remove any Content-Transfer-Encoding prior to delivering the response 2256 message to an HTTP client. 2258 Proxies and gateways from HTTP to MIME-compliant protocols are 2259 responsible for ensuring that the message is in the correct format 2260 and encoding for safe transport on that protocol, where "safe 2261 transport" is defined by the limitations of the protocol being used. 2262 Such a proxy or gateway ought to transform and label the data with an 2263 appropriate Content-Transfer-Encoding if doing so will improve the 2264 likelihood of safe transport over the destination protocol. 2266 B.6. MHTML and Line Length Limitations 2268 HTTP implementations that share code with MHTML [RFC2557] 2269 implementations need to be aware of MIME line length limitations. 2270 Since HTTP does not have this limitation, HTTP does not fold long 2271 lines. MHTML messages being transported by HTTP follow all 2272 conventions of MHTML, including line length limitations and folding, 2273 canonicalization, etc., since HTTP transfers message-bodies as 2274 payload and, aside from the "multipart/byteranges" type 2275 (Section 6.3.4 of [Semantics]), does not interpret the content or any 2276 MIME header lines that might be contained therein. 2278 Appendix C. HTTP Version History 2280 HTTP has been in use since 1990. The first version, later referred 2281 to as HTTP/0.9, was a simple protocol for hypertext data transfer 2282 across the Internet, using only a single request method (GET) and no 2283 metadata. HTTP/1.0, as defined by [RFC1945], added a range of 2284 request methods and MIME-like messaging, allowing for metadata to be 2285 transferred and modifiers placed on the request/response semantics. 2286 However, HTTP/1.0 did not sufficiently take into consideration the 2287 effects of hierarchical proxies, caching, the need for persistent 2288 connections, or name-based virtual hosts. The proliferation of 2289 incompletely implemented applications calling themselves "HTTP/1.0" 2290 further necessitated a protocol version change in order for two 2291 communicating applications to determine each other's true 2292 capabilities. 2294 HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent 2295 requirements that enable reliable implementations, adding only those 2296 features that can either be safely ignored by an HTTP/1.0 recipient 2297 or only be sent when communicating with a party advertising 2298 conformance with HTTP/1.1. 2300 HTTP/1.1 has been designed to make supporting previous versions easy. 2301 A general-purpose HTTP/1.1 server ought to be able to understand any 2302 valid request in the format of HTTP/1.0, responding appropriately 2303 with an HTTP/1.1 message that only uses features understood (or 2304 safely ignored) by HTTP/1.0 clients. Likewise, an HTTP/1.1 client 2305 can be expected to understand any valid HTTP/1.0 response. 2307 Since HTTP/0.9 did not support header fields in a request, there is 2308 no mechanism for it to support name-based virtual hosts (selection of 2309 resource by inspection of the Host header field). Any server that 2310 implements name-based virtual hosts ought to disable support for 2311 HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in fact, 2312 badly constructed HTTP/1.x requests caused by a client failing to 2313 properly encode the request-target. 2315 C.1. Changes from HTTP/1.0 2317 This section summarizes major differences between versions HTTP/1.0 2318 and HTTP/1.1. 2320 C.1.1. Multihomed Web Servers 2322 The requirements that clients and servers support the Host header 2323 field (Section 5.4 of [Semantics]), report an error if it is missing 2324 from an HTTP/1.1 request, and accept absolute URIs (Section 3.2) are 2325 among the most important changes defined by HTTP/1.1. 2327 Older HTTP/1.0 clients assumed a one-to-one relationship of IP 2328 addresses and servers; there was no other established mechanism for 2329 distinguishing the intended server of a request than the IP address 2330 to which that request was directed. The Host header field was 2331 introduced during the development of HTTP/1.1 and, though it was 2332 quickly implemented by most HTTP/1.0 browsers, additional 2333 requirements were placed on all HTTP/1.1 requests in order to ensure 2334 complete adoption. At the time of this writing, most HTTP-based 2335 services are dependent upon the Host header field for targeting 2336 requests. 2338 C.1.2. Keep-Alive Connections 2340 In HTTP/1.0, each connection is established by the client prior to 2341 the request and closed by the server after sending the response. 2342 However, some implementations implement the explicitly negotiated 2343 ("Keep-Alive") version of persistent connections described in 2344 Section 19.7.1 of [RFC2068]. 2346 Some clients and servers might wish to be compatible with these 2347 previous approaches to persistent connections, by explicitly 2348 negotiating for them with a "Connection: keep-alive" request header 2349 field. However, some experimental implementations of HTTP/1.0 2350 persistent connections are faulty; for example, if an HTTP/1.0 proxy 2351 server doesn't understand Connection, it will erroneously forward 2352 that header field to the next inbound server, which would result in a 2353 hung connection. 2355 One attempted solution was the introduction of a Proxy-Connection 2356 header field, targeted specifically at proxies. In practice, this 2357 was also unworkable, because proxies are often deployed in multiple 2358 layers, bringing about the same problem discussed above. 2360 As a result, clients are encouraged not to send the Proxy-Connection 2361 header field in any requests. 2363 Clients are also encouraged to consider the use of Connection: keep- 2364 alive in requests carefully; while they can enable persistent 2365 connections with HTTP/1.0 servers, clients using them will need to 2366 monitor the connection for "hung" requests (which indicate that the 2367 client ought stop sending the header field), and this mechanism ought 2368 not be used by clients at all when a proxy is being used. 2370 C.1.3. Introduction of Transfer-Encoding 2372 HTTP/1.1 introduces the Transfer-Encoding header field (Section 6.1). 2373 Transfer codings need to be decoded prior to forwarding an HTTP 2374 message over a MIME-compliant protocol. 2376 C.2. Changes from RFC 7230 2378 Most of the sections introducing HTTP's design goals, history, 2379 architecture, conformance criteria, protocol versioning, URIs, 2380 message routing, and header field values have been moved to 2381 [Semantics]. This document has been reduced to just the messaging 2382 syntax and connection management requirements specific to HTTP/1.1. 2384 Furthermore: 2386 In the ABNF for chunked extensions, re-introduce (bad) whitespace 2387 around ";" and "=". Whitespace was removed in [RFC7230], but later 2388 this change was found to break existing implementations (see 2389 [Err4667]). (Section 7.1.1) 2391 Disallow transfer coding parameters called "q" in order to avoid 2392 conflicts with the use of ranks in the TE header field. 2393 (Section 7.3) 2395 Appendix D. Change Log 2397 This section is to be removed before publishing as an RFC. 2399 D.1. Between RFC7230 and draft 00 2401 The changes were purely editorial: 2403 o Change boilerplate and abstract to indicate the "draft" status, 2404 and update references to ancestor specifications. 2406 o Adjust historical notes. 2408 o Update links to sibling specifications. 2410 o Replace sections listing changes from RFC 2616 by new empty 2411 sections referring to RFC 723x. 2413 o Remove acknowledgements specific to RFC 723x. 2415 o Move "Acknowledgements" to the very end and make them unnumbered. 2417 D.2. Since draft-ietf-httpbis-messaging-00 2419 The changes in this draft are editorial, with respect to HTTP as a 2420 whole, to move all core HTTP semantics into [Semantics]: 2422 o Moved introduction, architecture, conformance, and ABNF extensions 2423 from RFC 7230 (Messaging) to semantics [Semantics]. 2425 o Moved discussion of MIME differences from RFC 7231 (Semantics) to 2426 Appendix B since they mostly cover transforming 1.1 messages. 2428 o Moved all extensibility tips, registration procedures, and 2429 registry tables from the IANA considerations to normative 2430 sections, reducing the IANA considerations to just instructions 2431 that will be removed prior to publication as an RFC. 2433 D.3. Since draft-ietf-httpbis-messaging-01 2435 o Cite RFC 8126 instead of RFC 5226 () 2438 o Resolved erratum 4779, no change needed here 2439 (, 2440 ) 2442 o In Section 7, fixed prose claiming transfer parameters allow bare 2443 names (, 2444 ) 2446 o Resolved erratum 4225, no change needed here 2447 (, 2448 ) 2450 o Replace "response code" with "response status code" 2451 (, 2452 ) 2454 o In Section 9.3, clarify statement about HTTP/1.0 keep-alive 2455 (, 2456 ) 2458 o In Section 7.1.1, re-introduce (bad) whitespace around ";" and "=" 2459 (, 2460 , ) 2463 o In Section 7.3, state that transfer codings should not use 2464 parameters named "q" (, ) 2467 o In Section 7, mark coding name "trailers" as reserved in the IANA 2468 registry () 2470 Index 2472 A 2473 absolute-form (of request-target) 10 2474 application/http Media Type 39 2475 asterisk-form (of request-target) 11 2476 authority-form (of request-target) 11 2478 C 2479 Connection header field 28, 33 2480 Content-Length header field 18 2481 Content-Transfer-Encoding header field 49 2482 chunked (Coding Format) 17, 19 2483 chunked (transfer coding) 22 2484 close 28, 33 2485 compress (transfer coding) 25 2487 D 2488 deflate (transfer coding) 25 2490 E 2491 effective request URI 12 2493 G 2494 Grammar 2495 absolute-form 9-10 2496 ALPHA 5 2497 asterisk-form 9, 11 2498 authority-form 9, 11 2499 chunk 22 2500 chunk-data 22 2501 chunk-ext 22-23 2502 chunk-ext-name 23 2503 chunk-ext-val 23 2504 chunk-size 22 2505 chunked-body 22-23 2506 Connection 29 2507 connection-option 29 2508 CR 5 2509 CRLF 5 2510 CTL 5 2511 DIGIT 5 2512 DQUOTE 5 2513 field-name 14 2514 field-value 14 2515 header-field 14, 23 2516 HEXDIG 5 2517 HTAB 5 2518 HTTP-message 6 2519 HTTP-name 6 2520 HTTP-version 6 2521 last-chunk 22 2522 LF 5 2523 message-body 16 2524 method 9 2525 obs-fold 15 2526 OCTET 5 2527 origin-form 9-10 2528 rank 26 2529 reason-phrase 14 2530 request-line 8 2531 request-target 9 2532 SP 5 2533 start-line 6 2534 status-code 14 2535 status-line 13 2536 t-codings 26 2537 t-ranking 26 2538 TE 26 2539 trailer-part 22-23 2540 transfer-coding 21 2541 Transfer-Encoding 17 2542 transfer-extension 21 2543 transfer-parameter 21 2544 Upgrade 35 2545 VCHAR 5 2546 gzip (transfer coding) 25 2548 H 2549 header field 6 2550 header section 6 2551 headers 6 2553 M 2554 MIME-Version header field 48 2555 Media Type 2556 application/http 39 2557 message/http 38 2558 message/http Media Type 38 2559 method 9 2561 O 2562 origin-form (of request-target) 10 2564 R 2565 request-target 9 2567 T 2568 TE header field 26 2569 Transfer-Encoding header field 17 2571 U 2572 Upgrade header field 34 2574 X 2575 x-compress (transfer coding) 25 2576 x-gzip (transfer coding) 25 2578 Acknowledgments 2580 See Appendix "Acknowledgments" of [Semantics]. 2582 Authors' Addresses 2584 Roy T. Fielding (editor) 2585 Adobe 2586 345 Park Ave 2587 San Jose, CA 95110 2588 USA 2590 EMail: fielding@gbiv.com 2591 URI: https://roy.gbiv.com/ 2593 Mark Nottingham (editor) 2594 Fastly 2596 EMail: mnot@mnot.net 2597 URI: https://www.mnot.net/ 2599 Julian F. Reschke (editor) 2600 greenbytes GmbH 2601 Hafenweg 16 2602 Muenster, NW 48155 2603 Germany 2605 EMail: julian.reschke@greenbytes.de 2606 URI: https://greenbytes.de/tech/webdav/