idnits 2.17.1 draft-ietf-httpbis-messaging-18.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: ---------------------------------------------------------------------------- == There is 1 instance of lines with non-ascii characters in the document. 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 (18 August 2021) is 981 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(See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-cache-18 -- Possible downref: Normative reference to a draft: ref. 'CACHING' == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-semantics-18 -- Possible downref: Normative reference to a draft: ref. 'HTTP' ** 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 -- 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) Summary: 3 errors (**), 0 flaws (~~), 5 warnings (==), 10 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: 19 February 2022 J. Reschke, Ed. 7 greenbytes 8 18 August 2021 10 HTTP/1.1 11 draft-ietf-httpbis-messaging-18 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.19. 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 19 February 2022. 54 Copyright Notice 56 Copyright (c) 2021 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 (https://trustee.ietf.org/ 61 license-info) in effect on the date of publication of this document. 62 Please review these documents carefully, as they describe your rights 63 and restrictions with respect to this document. Code Components 64 extracted from this document must include Simplified BSD License text 65 as described in Section 4.e of the Trust Legal Provisions and are 66 provided without warranty as described in the Simplified BSD License. 68 This document may contain material from IETF Documents or IETF 69 Contributions published or made publicly available before November 70 10, 2008. The person(s) controlling the copyright in some of this 71 material may not have granted the IETF Trust the right to allow 72 modifications of such material outside the IETF Standards Process. 73 Without obtaining an adequate license from the person(s) controlling 74 the copyright in such materials, this document may not be modified 75 outside the IETF Standards Process, and derivative works of it may 76 not be created outside the IETF Standards Process, except to format 77 it for publication as an RFC or to translate it into languages other 78 than English. 80 Table of Contents 82 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 83 1.1. Requirements Notation . . . . . . . . . . . . . . . . . . 5 84 1.2. Syntax Notation . . . . . . . . . . . . . . . . . . . . . 5 85 2. Message . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 86 2.1. Message Format . . . . . . . . . . . . . . . . . . . . . 6 87 2.2. Message Parsing . . . . . . . . . . . . . . . . . . . . . 7 88 2.3. HTTP Version . . . . . . . . . . . . . . . . . . . . . . 8 89 3. Request Line . . . . . . . . . . . . . . . . . . . . . . . . 9 90 3.1. Method . . . . . . . . . . . . . . . . . . . . . . . . . 10 91 3.2. Request Target . . . . . . . . . . . . . . . . . . . . . 10 92 3.2.1. origin-form . . . . . . . . . . . . . . . . . . . . . 11 93 3.2.2. absolute-form . . . . . . . . . . . . . . . . . . . . 11 94 3.2.3. authority-form . . . . . . . . . . . . . . . . . . . 12 95 3.2.4. asterisk-form . . . . . . . . . . . . . . . . . . . . 12 96 3.3. Reconstructing the Target URI . . . . . . . . . . . . . . 13 97 4. Status Line . . . . . . . . . . . . . . . . . . . . . . . . . 15 98 5. Field Syntax . . . . . . . . . . . . . . . . . . . . . . . . 16 99 5.1. Field Line Parsing . . . . . . . . . . . . . . . . . . . 16 100 5.2. Obsolete Line Folding . . . . . . . . . . . . . . . . . . 17 101 6. Message Body . . . . . . . . . . . . . . . . . . . . . . . . 17 102 6.1. Transfer-Encoding . . . . . . . . . . . . . . . . . . . . 18 103 6.2. Content-Length . . . . . . . . . . . . . . . . . . . . . 20 104 6.3. Message Body Length . . . . . . . . . . . . . . . . . . . 20 105 7. Transfer Codings . . . . . . . . . . . . . . . . . . . . . . 23 106 7.1. Chunked Transfer Coding . . . . . . . . . . . . . . . . . 23 107 7.1.1. Chunk Extensions . . . . . . . . . . . . . . . . . . 24 108 7.1.2. Chunked Trailer Section . . . . . . . . . . . . . . . 25 109 7.1.3. Decoding Chunked . . . . . . . . . . . . . . . . . . 25 110 7.2. Transfer Codings for Compression . . . . . . . . . . . . 26 111 7.3. Transfer Coding Registry . . . . . . . . . . . . . . . . 26 112 7.4. Negotiating Transfer Codings . . . . . . . . . . . . . . 27 113 8. Handling Incomplete Messages . . . . . . . . . . . . . . . . 28 114 9. Connection Management . . . . . . . . . . . . . . . . . . . . 29 115 9.1. Establishment . . . . . . . . . . . . . . . . . . . . . . 29 116 9.2. Associating a Response to a Request . . . . . . . . . . . 29 117 9.3. Persistence . . . . . . . . . . . . . . . . . . . . . . . 30 118 9.3.1. Retrying Requests . . . . . . . . . . . . . . . . . . 31 119 9.3.2. Pipelining . . . . . . . . . . . . . . . . . . . . . 31 120 9.4. Concurrency . . . . . . . . . . . . . . . . . . . . . . . 32 121 9.5. Failures and Timeouts . . . . . . . . . . . . . . . . . . 32 122 9.6. Tear-down . . . . . . . . . . . . . . . . . . . . . . . . 33 123 9.7. TLS Connection Initiation . . . . . . . . . . . . . . . . 35 124 9.8. TLS Connection Closure . . . . . . . . . . . . . . . . . 35 125 10. Enclosing Messages as Data . . . . . . . . . . . . . . . . . 36 126 10.1. Media Type message/http . . . . . . . . . . . . . . . . 36 127 10.2. Media Type application/http . . . . . . . . . . . . . . 37 128 11. Security Considerations . . . . . . . . . . . . . . . . . . . 38 129 11.1. Response Splitting . . . . . . . . . . . . . . . . . . . 38 130 11.2. Request Smuggling . . . . . . . . . . . . . . . . . . . 39 131 11.3. Message Integrity . . . . . . . . . . . . . . . . . . . 40 132 11.4. Message Confidentiality . . . . . . . . . . . . . . . . 40 133 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 41 134 12.1. Field Name Registration . . . . . . . . . . . . . . . . 41 135 12.2. Media Type Registration . . . . . . . . . . . . . . . . 41 136 12.3. Transfer Coding Registration . . . . . . . . . . . . . . 41 137 12.4. ALPN Protocol ID Registration . . . . . . . . . . . . . 42 138 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 43 139 13.1. Normative References . . . . . . . . . . . . . . . . . . 43 140 13.2. Informative References . . . . . . . . . . . . . . . . . 44 141 Appendix A. Collected ABNF . . . . . . . . . . . . . . . . . . . 45 142 Appendix B. Differences between HTTP and MIME . . . . . . . . . 47 143 B.1. MIME-Version . . . . . . . . . . . . . . . . . . . . . . 47 144 B.2. Conversion to Canonical Form . . . . . . . . . . . . . . 47 145 B.3. Conversion of Date Formats . . . . . . . . . . . . . . . 48 146 B.4. Conversion of Content-Encoding . . . . . . . . . . . . . 48 147 B.5. Conversion of Content-Transfer-Encoding . . . . . . . . . 48 148 B.6. MHTML and Line Length Limitations . . . . . . . . . . . . 48 149 Appendix C. Changes from previous RFCs . . . . . . . . . . . . . 49 150 C.1. Changes from HTTP/0.9 . . . . . . . . . . . . . . . . . . 49 151 C.2. Changes from HTTP/1.0 . . . . . . . . . . . . . . . . . . 49 152 C.2.1. Multihomed Web Servers . . . . . . . . . . . . . . . 49 153 C.2.2. Keep-Alive Connections . . . . . . . . . . . . . . . 49 154 C.2.3. Introduction of Transfer-Encoding . . . . . . . . . . 50 155 C.3. Changes from RFC 7230 . . . . . . . . . . . . . . . . . . 50 156 Appendix D. Change Log . . . . . . . . . . . . . . . . . . . . . 51 157 D.1. Between RFC7230 and draft 00 . . . . . . . . . . . . . . 51 158 D.2. Since draft-ietf-httpbis-messaging-00 . . . . . . . . . . 51 159 D.3. Since draft-ietf-httpbis-messaging-01 . . . . . . . . . . 52 160 D.4. Since draft-ietf-httpbis-messaging-02 . . . . . . . . . . 52 161 D.5. Since draft-ietf-httpbis-messaging-03 . . . . . . . . . . 53 162 D.6. Since draft-ietf-httpbis-messaging-04 . . . . . . . . . . 53 163 D.7. Since draft-ietf-httpbis-messaging-05 . . . . . . . . . . 53 164 D.8. Since draft-ietf-httpbis-messaging-06 . . . . . . . . . . 54 165 D.9. Since draft-ietf-httpbis-messaging-07 . . . . . . . . . . 54 166 D.10. Since draft-ietf-httpbis-messaging-08 . . . . . . . . . . 54 167 D.11. Since draft-ietf-httpbis-messaging-09 . . . . . . . . . . 55 168 D.12. Since draft-ietf-httpbis-messaging-10 . . . . . . . . . . 55 169 D.13. Since draft-ietf-httpbis-messaging-11 . . . . . . . . . . 55 170 D.14. Since draft-ietf-httpbis-messaging-12 . . . . . . . . . . 55 171 D.15. Since draft-ietf-httpbis-messaging-13 . . . . . . . . . . 56 172 D.16. Since draft-ietf-httpbis-messaging-14 . . . . . . . . . . 56 173 D.17. Since draft-ietf-httpbis-messaging-15 . . . . . . . . . . 57 174 D.18. Since draft-ietf-httpbis-messaging-16 . . . . . . . . . . 57 175 D.19. Since draft-ietf-httpbis-messaging-17 . . . . . . . . . . 57 176 Acknowledgements . . . . . . . . . . . . . . . . . . . . . . . . 57 177 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 57 178 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 60 180 1. Introduction 182 The Hypertext Transfer Protocol (HTTP) is a stateless application- 183 level request/response protocol that uses extensible semantics and 184 self-descriptive messages for flexible interaction with network-based 185 hypertext information systems. HTTP/1.1 is defined by: 187 * This document 189 * "HTTP Semantics" [HTTP] 190 * "HTTP Caching" [CACHING] 192 This document specifies how HTTP semantics are conveyed using the 193 HTTP/1.1 message syntax, framing and connection management 194 mechanisms. Its goal is to define the complete set of requirements 195 for HTTP/1.1 message parsers and message-forwarding intermediaries. 197 This document obsoletes the portions of RFC 7230 related to HTTP/1.1 198 messaging and connection management, with the changes being 199 summarized in Appendix C.3. The other parts of RFC 7230 are 200 obsoleted by "HTTP Semantics" [HTTP]. 202 1.1. Requirements Notation 204 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 205 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 206 "OPTIONAL" in this document are to be interpreted as described in BCP 207 14 [RFC2119] [RFC8174] when, and only when, they appear in all 208 capitals, as shown here. 210 Conformance criteria and considerations regarding error handling are 211 defined in Section 2 of [HTTP]. 213 1.2. Syntax Notation 215 This specification uses the Augmented Backus-Naur Form (ABNF) 216 notation of [RFC5234], extended with the notation for case- 217 sensitivity in strings defined in [RFC7405]. 219 It also uses a list extension, defined in Section 5.6.1 of [HTTP], 220 that allows for compact definition of comma-separated lists using a 221 '#' operator (similar to how the '*' operator indicates repetition). 222 Appendix A shows the collected grammar with all list operators 223 expanded to standard ABNF notation. 225 As a convention, ABNF rule names prefixed with "obs-" denote 226 "obsolete" grammar rules that appear for historical reasons. 228 The following core rules are included by reference, as defined in 229 [RFC5234], Appendix B.1: ALPHA (letters), CR (carriage return), CRLF 230 (CR LF), CTL (controls), DIGIT (decimal 0-9), DQUOTE (double quote), 231 HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line 232 feed), OCTET (any 8-bit sequence of data), SP (space), and VCHAR (any 233 visible [USASCII] character). 235 The rules below are defined in [HTTP]: 237 BWS = 238 OWS = 239 RWS = 240 absolute-path = 241 field-name = 242 field-value = 243 obs-text = 244 quoted-string = 245 token = 246 transfer-coding = 247 249 The rules below are defined in [URI]: 251 absolute-URI = 252 authority = 253 uri-host = 254 port = 255 query = 257 2. Message 259 HTTP/1.1 clients and servers communicate by sending messages. See 260 Section 3 of [HTTP] for the general terminology and core concepts of 261 HTTP. 263 2.1. Message Format 265 An HTTP/1.1 message consists of a start-line followed by a CRLF and a 266 sequence of octets in a format similar to the Internet Message Format 267 [RFC5322]: zero or more header field lines (collectively referred to 268 as the "headers" or the "header section"), an empty line indicating 269 the end of the header section, and an optional message body. 271 HTTP-message = start-line CRLF 272 *( field-line CRLF ) 273 CRLF 274 [ message-body ] 276 A message can be either a request from client to server or a response 277 from server to client. Syntactically, the two types of message 278 differ only in the start-line, which is either a request-line (for 279 requests) or a status-line (for responses), and in the algorithm for 280 determining the length of the message body (Section 6). 282 start-line = request-line / status-line 284 In theory, a client could receive requests and a server could receive 285 responses, distinguishing them by their different start-line formats. 286 In practice, servers are implemented to only expect a request (a 287 response is interpreted as an unknown or invalid request method) and 288 clients are implemented to only expect a response. 290 HTTP makes use of some protocol elements similar to the Multipurpose 291 Internet Mail Extensions (MIME) [RFC2045]. See Appendix B for the 292 differences between HTTP and MIME messages. 294 2.2. Message Parsing 296 The normal procedure for parsing an HTTP message is to read the 297 start-line into a structure, read each header field line into a hash 298 table by field name until the empty line, and then use the parsed 299 data to determine if a message body is expected. If a message body 300 has been indicated, then it is read as a stream until an amount of 301 octets equal to the message body length is read or the connection is 302 closed. 304 A recipient MUST parse an HTTP message as a sequence of octets in an 305 encoding that is a superset of US-ASCII [USASCII]. Parsing an HTTP 306 message as a stream of Unicode characters, without regard for the 307 specific encoding, creates security vulnerabilities due to the 308 varying ways that string processing libraries handle invalid 309 multibyte character sequences that contain the octet LF (%x0A). 310 String-based parsers can only be safely used within protocol elements 311 after the element has been extracted from the message, such as within 312 a header field line value after message parsing has delineated the 313 individual field lines. 315 Although the line terminator for the start-line and fields is the 316 sequence CRLF, a recipient MAY recognize a single LF as a line 317 terminator and ignore any preceding CR. 319 A sender MUST NOT generate a bare CR (a CR character not immediately 320 followed by LF) within any protocol elements other than the content. 321 A recipient of such a bare CR MUST consider that element to be 322 invalid or replace each bare CR with SP before processing the element 323 or forwarding the message. 325 Older HTTP/1.0 user agent implementations might send an extra CRLF 326 after a POST request as a workaround for some early server 327 applications that failed to read message body content that was not 328 terminated by a line-ending. An HTTP/1.1 user agent MUST NOT preface 329 or follow a request with an extra CRLF. If terminating the request 330 message body with a line-ending is desired, then the user agent MUST 331 count the terminating CRLF octets as part of the message body length. 333 In the interest of robustness, a server that is expecting to receive 334 and parse a request-line SHOULD ignore at least one empty line (CRLF) 335 received prior to the request-line. 337 A sender MUST NOT send whitespace between the start-line and the 338 first header field. 340 A recipient that receives whitespace between the start-line and the 341 first header field MUST either reject the message as invalid or 342 consume each whitespace-preceded line without further processing of 343 it (i.e., ignore the entire line, along with any subsequent lines 344 preceded by whitespace, until a properly formed header field is 345 received or the header section is terminated). Rejection or removal 346 of invalid whitespace-preceded lines is necessary to prevent their 347 misinterpretation by downstream recipients that might be vulnerable 348 to request smuggling (Section 11.2) or response splitting 349 (Section 11.1) attacks. 351 When a server listening only for HTTP request messages, or processing 352 what appears from the start-line to be an HTTP request message, 353 receives a sequence of octets that does not match the HTTP-message 354 grammar aside from the robustness exceptions listed above, the server 355 SHOULD respond with a 400 (Bad Request) response and close the 356 connection. 358 2.3. HTTP Version 360 HTTP uses a "." numbering scheme to indicate versions 361 of the protocol. This specification defines version "1.1". 362 Section 2.5 of [HTTP] specifies the semantics of HTTP version 363 numbers. 365 The version of an HTTP/1.x message is indicated by an HTTP-version 366 field in the start-line. HTTP-version is case-sensitive. 368 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 369 HTTP-name = %s"HTTP" 371 When an HTTP/1.1 message is sent to an HTTP/1.0 recipient [HTTP/1.0] 372 or a recipient whose version is unknown, the HTTP/1.1 message is 373 constructed such that it can be interpreted as a valid HTTP/1.0 374 message if all of the newer features are ignored. This specification 375 places recipient-version requirements on some new features so that a 376 conformant sender will only use compatible features until it has 377 determined, through configuration or the receipt of a message, that 378 the recipient supports HTTP/1.1. 380 Intermediaries that process HTTP messages (i.e., all intermediaries 381 other than those acting as tunnels) MUST send their own HTTP-version 382 in forwarded messages, unless it is purposefully downgraded as a 383 workaround for an upstream issue. In other words, an intermediary is 384 not allowed to blindly forward the start-line without ensuring that 385 the protocol version in that message matches a version to which that 386 intermediary is conformant for both the receiving and sending of 387 messages. Forwarding an HTTP message without rewriting the HTTP- 388 version might result in communication errors when downstream 389 recipients use the message sender's version to determine what 390 features are safe to use for later communication with that sender. 392 A server MAY send an HTTP/1.0 response to an HTTP/1.1 request if it 393 is known or suspected that the client incorrectly implements the HTTP 394 specification and is incapable of correctly processing later version 395 responses, such as when a client fails to parse the version number 396 correctly or when an intermediary is known to blindly forward the 397 HTTP-version even when it doesn't conform to the given minor version 398 of the protocol. Such protocol downgrades SHOULD NOT be performed 399 unless triggered by specific client attributes, such as when one or 400 more of the request header fields (e.g., User-Agent) uniquely match 401 the values sent by a client known to be in error. 403 3. Request Line 405 A request-line begins with a method token, followed by a single space 406 (SP), the request-target, another single space (SP), and ends with 407 the protocol version. 409 request-line = method SP request-target SP HTTP-version 411 Although the request-line grammar rule requires that each of the 412 component elements be separated by a single SP octet, recipients MAY 413 instead parse on whitespace-delimited word boundaries and, aside from 414 the CRLF terminator, treat any form of whitespace as the SP separator 415 while ignoring preceding or trailing whitespace; such whitespace 416 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 417 (%x0C), or bare CR. However, lenient parsing can result in request 418 smuggling security vulnerabilities if there are multiple recipients 419 of the message and each has its own unique interpretation of 420 robustness (see Section 11.2). 422 HTTP does not place a predefined limit on the length of a request- 423 line, as described in Section 2 of [HTTP]. A server that receives a 424 method longer than any that it implements SHOULD respond with a 501 425 (Not Implemented) status code. A server that receives a request- 426 target longer than any URI it wishes to parse MUST respond with a 414 427 (URI Too Long) status code (see Section 15.5.15 of [HTTP]). 429 Various ad hoc limitations on request-line length are found in 430 practice. It is RECOMMENDED that all HTTP senders and recipients 431 support, at a minimum, request-line lengths of 8000 octets. 433 3.1. Method 435 The method token indicates the request method to be performed on the 436 target resource. The request method is case-sensitive. 438 method = token 440 The request methods defined by this specification can be found in 441 Section 9 of [HTTP], along with information regarding the HTTP method 442 registry and considerations for defining new methods. 444 3.2. Request Target 446 The request-target identifies the target resource upon which to apply 447 the request. The client derives a request-target from its desired 448 target URI. There are four distinct formats for the request-target, 449 depending on both the method being requested and whether the request 450 is to a proxy. 452 request-target = origin-form 453 / absolute-form 454 / authority-form 455 / asterisk-form 457 No whitespace is allowed in the request-target. Unfortunately, some 458 user agents fail to properly encode or exclude whitespace found in 459 hypertext references, resulting in those disallowed characters being 460 sent as the request-target in a malformed request-line. 462 Recipients of an invalid request-line SHOULD respond with either a 463 400 (Bad Request) error or a 301 (Moved Permanently) redirect with 464 the request-target properly encoded. A recipient SHOULD NOT attempt 465 to autocorrect and then process the request without a redirect, since 466 the invalid request-line might be deliberately crafted to bypass 467 security filters along the request chain. 469 A client MUST send a Host header field in all HTTP/1.1 request 470 messages. If the target URI includes an authority component, then a 471 client MUST send a field value for Host that is identical to that 472 authority component, excluding any userinfo subcomponent and its "@" 473 delimiter (Section 4.2.1 of [HTTP]). If the authority component is 474 missing or undefined for the target URI, then a client MUST send a 475 Host header field with an empty field value. 477 A server MUST respond with a 400 (Bad Request) status code to any 478 HTTP/1.1 request message that lacks a Host header field and to any 479 request message that contains more than one Host header field line or 480 a Host header field with an invalid field value. 482 3.2.1. origin-form 484 The most common form of request-target is the _origin-form_. 486 origin-form = absolute-path [ "?" query ] 488 When making a request directly to an origin server, other than a 489 CONNECT or server-wide OPTIONS request (as detailed below), a client 490 MUST send only the absolute path and query components of the target 491 URI as the request-target. If the target URI's path component is 492 empty, the client MUST send "/" as the path within the origin-form of 493 request-target. A Host header field is also sent, as defined in 494 Section 7.2 of [HTTP]. 496 For example, a client wishing to retrieve a representation of the 497 resource identified as 499 http://www.example.org/where?q=now 501 directly from the origin server would open (or reuse) a TCP 502 connection to port 80 of the host "www.example.org" and send the 503 lines: 505 GET /where?q=now HTTP/1.1 506 Host: www.example.org 508 followed by the remainder of the request message. 510 3.2.2. absolute-form 512 When making a request to a proxy, other than a CONNECT or server-wide 513 OPTIONS request (as detailed below), a client MUST send the target 514 URI in _absolute-form_ as the request-target. 516 absolute-form = absolute-URI 518 The proxy is requested to either service that request from a valid 519 cache, if possible, or make the same request on the client's behalf 520 to either the next inbound proxy server or directly to the origin 521 server indicated by the request-target. Requirements on such 522 "forwarding" of messages are defined in Section 7.6 of [HTTP]. 524 An example absolute-form of request-line would be: 526 GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1 528 A client MUST send a Host header field in an HTTP/1.1 request even if 529 the request-target is in the absolute-form, since this allows the 530 Host information to be forwarded through ancient HTTP/1.0 proxies 531 that might not have implemented Host. 533 When a proxy receives a request with an absolute-form of request- 534 target, the proxy MUST ignore the received Host header field (if any) 535 and instead replace it with the host information of the request- 536 target. A proxy that forwards such a request MUST generate a new 537 Host field value based on the received request-target rather than 538 forward the received Host field value. 540 When an origin server receives a request with an absolute-form of 541 request-target, the origin server MUST ignore the received Host 542 header field (if any) and instead use the host information of the 543 request-target. Note that if the request-target does not have an 544 authority component, an empty Host header field will be sent in this 545 case. 547 A server MUST accept the absolute-form in requests even though most 548 HTTP/1.1 clients will only send the absolute-form to a proxy. 550 3.2.3. authority-form 552 The _authority-form_ of request-target is only used for CONNECT 553 requests (Section 9.3.6 of [HTTP]). It consists of only the uri-host 554 and port number of the tunnel destination, separated by a colon 555 (":"). 557 authority-form = uri-host ":" port 559 When making a CONNECT request to establish a tunnel through one or 560 more proxies, a client MUST send only the host and port of the tunnel 561 destination as the request-target. The client obtains the host and 562 port from the target URI's authority component, except that it sends 563 the scheme's default port if the target URI elides the port. For 564 example, a CONNECT request to "http://www.example.com" looks like 566 CONNECT www.example.com:80 HTTP/1.1 567 Host: www.example.com 569 3.2.4. asterisk-form 571 The _asterisk-form_ of request-target is only used for a server-wide 572 OPTIONS request (Section 9.3.7 of [HTTP]). 574 asterisk-form = "*" 576 When a client wishes to request OPTIONS for the server as a whole, as 577 opposed to a specific named resource of that server, the client MUST 578 send only "*" (%x2A) as the request-target. For example, 580 OPTIONS * HTTP/1.1 582 If a proxy receives an OPTIONS request with an absolute-form of 583 request-target in which the URI has an empty path and no query 584 component, then the last proxy on the request chain MUST send a 585 request-target of "*" when it forwards the request to the indicated 586 origin server. 588 For example, the request 590 OPTIONS http://www.example.org:8001 HTTP/1.1 592 would be forwarded by the final proxy as 594 OPTIONS * HTTP/1.1 595 Host: www.example.org:8001 597 after connecting to port 8001 of host "www.example.org". 599 3.3. Reconstructing the Target URI 601 The target URI is the request-target when the request-target is in 602 absolute-form. In that case, a server will parse the URI into its 603 generic components for further evaluation. 605 Otherwise, the server reconstructs the target URI from the connection 606 context and various parts of the request message in order to identify 607 the target resource (Section 7.1 of [HTTP]): 609 * If the server's configuration provides for a fixed URI scheme, or 610 a scheme is provided by a trusted outbound gateway, that scheme is 611 used for the target URI. This is common in large-scale 612 deployments because a gateway server will receive the client's 613 connection context and replace that with their own connection to 614 the inbound server. Otherwise, if the request is received over a 615 secured connection, the target URI's scheme is "https"; if not, 616 the scheme is "http". 618 * If the request-target is in authority-form, the target URI's 619 authority component is the request-target. Otherwise, the target 620 URI's authority component is the field value of the Host header 621 field. If there is no Host header field or if its field value is 622 empty or invalid, the target URI's authority component is empty. 624 * If the request-target is in authority-form or asterisk-form, the 625 target URI's combined path and query component is empty. 626 Otherwise, the target URI's combined path and query component is 627 the request-target. 629 * The components of a reconstructed target URI, once determined as 630 above, can be recombined into absolute-URI form by concatenating 631 the scheme, "://", authority, and combined path and query 632 component. 634 Example 1: the following message received over a secure connection 636 GET /pub/WWW/TheProject.html HTTP/1.1 637 Host: www.example.org 639 has a target URI of 641 https://www.example.org/pub/WWW/TheProject.html 643 Example 2: the following message received over an insecure connection 645 OPTIONS * HTTP/1.1 646 Host: www.example.org:8080 648 has a target URI of 650 http://www.example.org:8080 652 If the target URI's authority component is empty and its URI scheme 653 requires a non-empty authority (as is the case for "http" and 654 "https"), the server can reject the request or determine whether a 655 configured default applies that is consistent with the incoming 656 connection's context. Context might include connection details like 657 address and port, what security has been applied, and locally-defined 658 information specific to that server's configuration. An empty 659 authority is replaced with the configured default before further 660 processing of the request. 662 Supplying a default name for authority within the context of a 663 secured connection is inherently unsafe if there is any chance that 664 the user agent's intended authority might differ from the default. A 665 server that can uniquely identify an authority from the request 666 context MAY use that identity as a default without this risk. 667 Alternatively, it might be better to redirect the request to a safe 668 resource that explains how to obtain a new client. 670 Note that reconstructing the client's target URI is only half of the 671 process for identifying a target resource. The other half is 672 determining whether that target URI identifies a resource for which 673 the server is willing and able to send a response, as defined in 674 Section 7.4 of [HTTP]. 676 4. Status Line 678 The first line of a response message is the status-line, consisting 679 of the protocol version, a space (SP), the status code, another 680 space, and ending with an OPTIONAL textual phrase describing the 681 status code. 683 status-line = HTTP-version SP status-code SP [reason-phrase] 685 Although the status-line grammar rule requires that each of the 686 component elements be separated by a single SP octet, recipients MAY 687 instead parse on whitespace-delimited word boundaries and, aside from 688 the line terminator, treat any form of whitespace as the SP separator 689 while ignoring preceding or trailing whitespace; such whitespace 690 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 691 (%x0C), or bare CR. However, lenient parsing can result in response 692 splitting security vulnerabilities if there are multiple recipients 693 of the message and each has its own unique interpretation of 694 robustness (see Section 11.1). 696 The status-code element is a 3-digit integer code describing the 697 result of the server's attempt to understand and satisfy the client's 698 corresponding request. A recipient parses and interprets the 699 remainder of the response message in light of the semantics defined 700 for that status code, if the status code is recognized by that 701 recipient, or in accordance with the class of that status code when 702 the specific code is unrecognized. 704 status-code = 3DIGIT 706 HTTP's core status codes are defined in Section 15 of [HTTP], along 707 with the classes of status codes, considerations for the definition 708 of new status codes, and the IANA registry for collecting such 709 definitions. 711 The reason-phrase element exists for the sole purpose of providing a 712 textual description associated with the numeric status code, mostly 713 out of deference to earlier Internet application protocols that were 714 more frequently used with interactive text clients. 716 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 718 A client SHOULD ignore the reason-phrase content because it is not a 719 reliable channel for information (it might be translated for a given 720 locale, overwritten by intermediaries, or discarded when the message 721 is forwarded via other versions of HTTP). A server MUST send the 722 space that separates status-code from the reason-phrase even when the 723 reason-phrase is absent (i.e., the status-line would end with the 724 three octets SP CR LF). 726 5. Field Syntax 728 Each field line consists of a case-insensitive field name followed by 729 a colon (":"), optional leading whitespace, the field line value, and 730 optional trailing whitespace. 732 field-line = field-name ":" OWS field-value OWS 734 Most HTTP field names and the rules for parsing within field values 735 are defined in Section 6.3 of [HTTP]. This section covers the 736 generic syntax for header field inclusion within, and extraction 737 from, HTTP/1.1 messages. 739 5.1. Field Line Parsing 741 Messages are parsed using a generic algorithm, independent of the 742 individual field names. The contents within a given field line value 743 are not parsed until a later stage of message interpretation (usually 744 after the message's entire field section has been processed). 746 No whitespace is allowed between the field name and colon. In the 747 past, differences in the handling of such whitespace have led to 748 security vulnerabilities in request routing and response handling. A 749 server MUST reject, with a response status code of 400 (Bad Request), 750 any received request message that contains whitespace between a 751 header field name and colon. A proxy MUST remove any such whitespace 752 from a response message before forwarding the message downstream. 754 A field line value might be preceded and/or followed by optional 755 whitespace (OWS); a single SP preceding the field line value is 756 preferred for consistent readability by humans. The field line value 757 does not include that leading or trailing whitespace: OWS occurring 758 before the first non-whitespace octet of the field line value, or 759 after the last non-whitespace octet of the field line value, is 760 excluded by parsers when extracting the field line value from a field 761 line. 763 5.2. Obsolete Line Folding 765 Historically, HTTP/1.x field values could be extended over multiple 766 lines by preceding each extra line with at least one space or 767 horizontal tab (obs-fold). This specification deprecates such line 768 folding except within the message/http media type (Section 10.1). 770 obs-fold = OWS CRLF RWS 771 ; obsolete line folding 773 A sender MUST NOT generate a message that includes line folding 774 (i.e., that has any field line value that contains a match to the 775 obs-fold rule) unless the message is intended for packaging within 776 the message/http media type. 778 A server that receives an obs-fold in a request message that is not 779 within a message/http container MUST either reject the message by 780 sending a 400 (Bad Request), preferably with a representation 781 explaining that obsolete line folding is unacceptable, or replace 782 each received obs-fold with one or more SP octets prior to 783 interpreting the field value or forwarding the message downstream. 785 A proxy or gateway that receives an obs-fold in a response message 786 that is not within a message/http container MUST either discard the 787 message and replace it with a 502 (Bad Gateway) response, preferably 788 with a representation explaining that unacceptable line folding was 789 received, or replace each received obs-fold with one or more SP 790 octets prior to interpreting the field value or forwarding the 791 message downstream. 793 A user agent that receives an obs-fold in a response message that is 794 not within a message/http container MUST replace each received 795 obs-fold with one or more SP octets prior to interpreting the field 796 value. 798 6. Message Body 800 The message body (if any) of an HTTP/1.1 message is used to carry 801 content (Section 6.4 of [HTTP]) for the request or response. The 802 message body is identical to the content unless a transfer coding has 803 been applied, as described in Section 6.1. 805 message-body = *OCTET 807 The rules for determining when a message body is present in an 808 HTTP/1.1 message differ for requests and responses. 810 The presence of a message body in a request is signaled by a 811 Content-Length or Transfer-Encoding header field. Request message 812 framing is independent of method semantics. 814 The presence of a message body in a response depends on both the 815 request method to which it is responding and the response status code 816 (Section 4), and corresponds to when content is allowed; see 817 Section 6.4 of [HTTP]. 819 6.1. Transfer-Encoding 821 The Transfer-Encoding header field lists the transfer coding names 822 corresponding to the sequence of transfer codings that have been (or 823 will be) applied to the content in order to form the message body. 824 Transfer codings are defined in Section 7. 826 Transfer-Encoding = #transfer-coding 827 ; defined in [HTTP], Section 10.1.4 829 Transfer-Encoding is analogous to the Content-Transfer-Encoding field 830 of MIME, which was designed to enable safe transport of binary data 831 over a 7-bit transport service ([RFC2045], Section 6). However, safe 832 transport has a different focus for an 8bit-clean transfer protocol. 833 In HTTP's case, Transfer-Encoding is primarily intended to accurately 834 delimit dynamically generated content. It also serves to distinguish 835 encodings that are only applied in transit from the encodings that 836 are a characteristic of the selected representation. 838 A recipient MUST be able to parse the chunked transfer coding 839 (Section 7.1) because it plays a crucial role in framing messages 840 when the content size is not known in advance. A sender MUST NOT 841 apply the chunked transfer coding more than once to a message body 842 (i.e., chunking an already chunked message is not allowed). If any 843 transfer coding other than chunked is applied to a request's content, 844 the sender MUST apply chunked as the final transfer coding to ensure 845 that the message is properly framed. If any transfer coding other 846 than chunked is applied to a response's content, the sender MUST 847 either apply chunked as the final transfer coding or terminate the 848 message by closing the connection. 850 For example, 852 Transfer-Encoding: gzip, chunked 853 indicates that the content has been compressed using the gzip coding 854 and then chunked using the chunked coding while forming the message 855 body. 857 Unlike Content-Encoding (Section 8.4.1 of [HTTP]), Transfer-Encoding 858 is a property of the message, not of the representation, and any 859 recipient along the request/response chain MAY decode the received 860 transfer coding(s) or apply additional transfer coding(s) to the 861 message body, assuming that corresponding changes are made to the 862 Transfer-Encoding field value. Additional information about the 863 encoding parameters can be provided by other header fields not 864 defined by this specification. 866 Transfer-Encoding MAY be sent in a response to a HEAD request or in a 867 304 (Not Modified) response (Section 15.4.5 of [HTTP]) to a GET 868 request, neither of which includes a message body, to indicate that 869 the origin server would have applied a transfer coding to the message 870 body if the request had been an unconditional GET. This indication 871 is not required, however, because any recipient on the response chain 872 (including the origin server) can remove transfer codings when they 873 are not needed. 875 A server MUST NOT send a Transfer-Encoding header field in any 876 response with a status code of 1xx (Informational) or 204 (No 877 Content). A server MUST NOT send a Transfer-Encoding header field in 878 any 2xx (Successful) response to a CONNECT request (Section 9.3.6 of 879 [HTTP]). 881 A server that receives a request message with a transfer coding it 882 does not understand SHOULD respond with 501 (Not Implemented). 884 Transfer-Encoding was added in HTTP/1.1. It is generally assumed 885 that implementations advertising only HTTP/1.0 support will not 886 understand how to process transfer-encoded content, and that an 887 HTTP/1.0 message received with a Transfer-Encoding is likely to have 888 been forwarded without proper handling of the chunked encoding in 889 transit. 891 A client MUST NOT send a request containing Transfer-Encoding unless 892 it knows the server will handle HTTP/1.1 requests (or later minor 893 revisions); such knowledge might be in the form of specific user 894 configuration or by remembering the version of a prior received 895 response. A server MUST NOT send a response containing Transfer- 896 Encoding unless the corresponding request indicates HTTP/1.1 (or 897 later minor revisions). 899 Early implementations of Transfer-Encoding would occasionally send 900 both a chunked encoding for message framing and an estimated Content- 901 Length header field for use by progress bars. This is why Transfer- 902 Encoding is defined as overriding Content-Length, as opposed to them 903 being mutually incompatible. Unfortunately, forwarding such a 904 message can lead to vulnerabilities regarding request smuggling 905 (Section 11.2) or response splitting (Section 11.1) attacks if any 906 downstream recipient fails to parse the message according to this 907 specification, particularly when a downstream recipient only 908 implements HTTP/1.0. 910 A server MAY reject a request that contains both Content-Length and 911 Transfer-Encoding or process such a request in accordance with the 912 Transfer-Encoding alone. Regardless, the server MUST close the 913 connection after responding to such a request to avoid the potential 914 attacks. 916 A server or client that receives an HTTP/1.0 message containing a 917 Transfer-Encoding header field MUST treat the message as if the 918 framing is faulty, even if a Content-Length is present, and close the 919 connection after processing the message. The message sender might 920 have retained a portion of the message, in buffer, that could be 921 misinterpreted by further use of the connection. 923 6.2. Content-Length 925 When a message does not have a Transfer-Encoding header field, a 926 Content-Length header field (Section 8.6 of [HTTP]) can provide the 927 anticipated size, as a decimal number of octets, for potential 928 content. For messages that do include content, the Content-Length 929 field value provides the framing information necessary for 930 determining where the data (and message) ends. For messages that do 931 not include content, the Content-Length indicates the size of the 932 selected representation (Section 8.6 of [HTTP]). 934 A sender MUST NOT send a Content-Length header field in any message 935 that contains a Transfer-Encoding header field. 937 | *Note:* HTTP's use of Content-Length for message framing 938 | differs significantly from the same field's use in MIME, where 939 | it is an optional field used only within the "message/external- 940 | body" media-type. 942 6.3. Message Body Length 944 The length of a message body is determined by one of the following 945 (in order of precedence): 947 1. Any response to a HEAD request and any response with a 1xx 948 (Informational), 204 (No Content), or 304 (Not Modified) status 949 code is always terminated by the first empty line after the 950 header fields, regardless of the header fields present in the 951 message, and thus cannot contain a message body or trailer 952 section. 954 2. Any 2xx (Successful) response to a CONNECT request implies that 955 the connection will become a tunnel immediately after the empty 956 line that concludes the header fields. A client MUST ignore any 957 Content-Length or Transfer-Encoding header fields received in 958 such a message. 960 3. If a message is received with both a Transfer-Encoding and a 961 Content-Length header field, the Transfer-Encoding overrides the 962 Content-Length. Such a message might indicate an attempt to 963 perform request smuggling (Section 11.2) or response splitting 964 (Section 11.1) and ought to be handled as an error. An 965 intermediary that chooses to forward the message MUST first 966 remove the received Content-Length field and process the 967 Transfer-Encoding (as described below) prior to forwarding the 968 message downstream. 970 4. If a Transfer-Encoding header field is present and the chunked 971 transfer coding (Section 7.1) is the final encoding, the message 972 body length is determined by reading and decoding the chunked 973 data until the transfer coding indicates the data is complete. 975 If a Transfer-Encoding header field is present in a response and 976 the chunked transfer coding is not the final encoding, the 977 message body length is determined by reading the connection until 978 it is closed by the server. 980 If a Transfer-Encoding header field is present in a request and 981 the chunked transfer coding is not the final encoding, the 982 message body length cannot be determined reliably; the server 983 MUST respond with the 400 (Bad Request) status code and then 984 close the connection. 986 5. If a message is received without Transfer-Encoding and with an 987 invalid Content-Length header field, then the message framing is 988 invalid and the recipient MUST treat it as an unrecoverable 989 error, unless the field value can be successfully parsed as a 990 comma-separated list (Section 5.6.1 of [HTTP]), all values in the 991 list are valid, and all values in the list are the same (in which 992 case the message is processed with that single value used as the 993 Content-Length field value). If the unrecoverable error is in a 994 request message, the server MUST respond with a 400 (Bad Request) 995 status code and then close the connection. If it is in a 996 response message received by a proxy, the proxy MUST close the 997 connection to the server, discard the received response, and send 998 a 502 (Bad Gateway) response to the client. If it is in a 999 response message received by a user agent, the user agent MUST 1000 close the connection to the server and discard the received 1001 response. 1003 6. If a valid Content-Length header field is present without 1004 Transfer-Encoding, its decimal value defines the expected message 1005 body length in octets. If the sender closes the connection or 1006 the recipient times out before the indicated number of octets are 1007 received, the recipient MUST consider the message to be 1008 incomplete and close the connection. 1010 7. If this is a request message and none of the above are true, then 1011 the message body length is zero (no message body is present). 1013 8. Otherwise, this is a response message without a declared message 1014 body length, so the message body length is determined by the 1015 number of octets received prior to the server closing the 1016 connection. 1018 Since there is no way to distinguish a successfully completed, close- 1019 delimited response message from a partially received message 1020 interrupted by network failure, a server SHOULD generate encoding or 1021 length-delimited messages whenever possible. The close-delimiting 1022 feature exists primarily for backwards compatibility with HTTP/1.0. 1024 | *Note:* Request messages are never close-delimited because they 1025 | are always explicitly framed by length or transfer coding, with 1026 | the absence of both implying the request ends immediately after 1027 | the header section. 1029 A server MAY reject a request that contains a message body but not a 1030 Content-Length by responding with 411 (Length Required). 1032 Unless a transfer coding other than chunked has been applied, a 1033 client that sends a request containing a message body SHOULD use a 1034 valid Content-Length header field if the message body length is known 1035 in advance, rather than the chunked transfer coding, since some 1036 existing services respond to chunked with a 411 (Length Required) 1037 status code even though they understand the chunked transfer coding. 1038 This is typically because such services are implemented via a gateway 1039 that requires a content-length in advance of being called and the 1040 server is unable or unwilling to buffer the entire request before 1041 processing. 1043 A user agent that sends a request that contains a message body MUST 1044 send either a valid Content-Length header field or use the chunked 1045 transfer coding. A client MUST NOT use the chunked transfer encoding 1046 unless it knows the server will handle HTTP/1.1 (or later) requests; 1047 such knowledge can be in the form of specific user configuration or 1048 by remembering the version of a prior received response. 1050 If the final response to the last request on a connection has been 1051 completely received and there remains additional data to read, a user 1052 agent MAY discard the remaining data or attempt to determine if that 1053 data belongs as part of the prior message body, which might be the 1054 case if the prior message's Content-Length value is incorrect. A 1055 client MUST NOT process, cache, or forward such extra data as a 1056 separate response, since such behavior would be vulnerable to cache 1057 poisoning. 1059 7. Transfer Codings 1061 Transfer coding names are used to indicate an encoding transformation 1062 that has been, can be, or might need to be applied to a message's 1063 content in order to ensure "safe transport" through the network. 1064 This differs from a content coding in that the transfer coding is a 1065 property of the message rather than a property of the representation 1066 that is being transferred. 1068 All transfer-coding names are case-insensitive and ought to be 1069 registered within the HTTP Transfer Coding registry, as defined in 1070 Section 7.3. They are used in the Transfer-Encoding (Section 6.1) 1071 and TE (Section 10.1.4 of [HTTP]) header fields (the latter also 1072 defining the "transfer-coding" grammar). 1074 7.1. Chunked Transfer Coding 1076 The chunked transfer coding wraps content in order to transfer it as 1077 a series of chunks, each with its own size indicator, followed by an 1078 OPTIONAL trailer section containing trailer fields. Chunked enables 1079 content streams of unknown size to be transferred as a sequence of 1080 length-delimited buffers, which enables the sender to retain 1081 connection persistence and the recipient to know when it has received 1082 the entire message. 1084 chunked-body = *chunk 1085 last-chunk 1086 trailer-section 1087 CRLF 1089 chunk = chunk-size [ chunk-ext ] CRLF 1090 chunk-data CRLF 1091 chunk-size = 1*HEXDIG 1092 last-chunk = 1*("0") [ chunk-ext ] CRLF 1094 chunk-data = 1*OCTET ; a sequence of chunk-size octets 1096 The chunk-size field is a string of hex digits indicating the size of 1097 the chunk-data in octets. The chunked transfer coding is complete 1098 when a chunk with a chunk-size of zero is received, possibly followed 1099 by a trailer section, and finally terminated by an empty line. 1101 A recipient MUST be able to parse and decode the chunked transfer 1102 coding. 1104 HTTP/1.1 does not define any means to limit the size of a chunked 1105 response such that an intermediary can be assured of buffering the 1106 entire response. Additionally, very large chunk sizes may cause 1107 overflows or loss of precision if their values are not represented 1108 accurately in a receiving implementation. Therefore, recipients MUST 1109 anticipate potentially large hexadecimal numerals and prevent parsing 1110 errors due to integer conversion overflows or precision loss due to 1111 integer representation. 1113 The chunked encoding does not define any parameters. Their presence 1114 SHOULD be treated as an error. 1116 7.1.1. Chunk Extensions 1118 The chunked encoding allows each chunk to include zero or more chunk 1119 extensions, immediately following the chunk-size, for the sake of 1120 supplying per-chunk metadata (such as a signature or hash), mid- 1121 message control information, or randomization of message body size. 1123 chunk-ext = *( BWS ";" BWS chunk-ext-name 1124 [ BWS "=" BWS chunk-ext-val ] ) 1126 chunk-ext-name = token 1127 chunk-ext-val = token / quoted-string 1129 The chunked encoding is specific to each connection and is likely to 1130 be removed or recoded by each recipient (including intermediaries) 1131 before any higher-level application would have a chance to inspect 1132 the extensions. Hence, use of chunk extensions is generally limited 1133 to specialized HTTP services such as "long polling" (where client and 1134 server can have shared expectations regarding the use of chunk 1135 extensions) or for padding within an end-to-end secured connection. 1137 A recipient MUST ignore unrecognized chunk extensions. A server 1138 ought to limit the total length of chunk extensions received in a 1139 request to an amount reasonable for the services provided, in the 1140 same way that it applies length limitations and timeouts for other 1141 parts of a message, and generate an appropriate 4xx (Client Error) 1142 response if that amount is exceeded. 1144 7.1.2. Chunked Trailer Section 1146 A trailer section allows the sender to include additional fields at 1147 the end of a chunked message in order to supply metadata that might 1148 be dynamically generated while the content is sent, such as a message 1149 integrity check, digital signature, or post-processing status. The 1150 proper use and limitations of trailer fields are defined in 1151 Section 6.5 of [HTTP]. 1153 trailer-section = *( field-line CRLF ) 1155 A recipient that decodes and removes the chunked encoding from a 1156 message (e.g., for storage or forwarding to a non-HTTP/1.1 peer) MUST 1157 discard any received trailer fields, store/forward them separately 1158 from the header fields, or selectively merge into the header section 1159 only those trailer fields corresponding to header field definitions 1160 that are understood by the recipient to explicitly permit and define 1161 how their corresponding trailer field value can be safely merged. 1163 7.1.3. Decoding Chunked 1165 A process for decoding the chunked transfer coding can be represented 1166 in pseudo-code as: 1168 length := 0 1169 read chunk-size, chunk-ext (if any), and CRLF 1170 while (chunk-size > 0) { 1171 read chunk-data and CRLF 1172 append chunk-data to content 1173 length := length + chunk-size 1174 read chunk-size, chunk-ext (if any), and CRLF 1175 } 1176 read trailer field 1177 while (trailer field is not empty) { 1178 if (trailer fields are stored/forwarded separately) { 1179 append trailer field to existing trailer fields 1180 } 1181 else if (trailer field is understood and defined as mergeable) { 1182 merge trailer field with existing header fields 1183 } 1184 else { 1185 discard trailer field 1186 } 1187 read trailer field 1188 } 1189 Content-Length := length 1190 Remove "chunked" from Transfer-Encoding 1192 7.2. Transfer Codings for Compression 1194 The following transfer coding names for compression are defined by 1195 the same algorithm as their corresponding content coding: 1197 compress (and x-compress) 1198 See Section 8.4.1.1 of [HTTP]. 1200 deflate 1201 See Section 8.4.1.2 of [HTTP]. 1203 gzip (and x-gzip) 1204 See Section 8.4.1.3 of [HTTP]. 1206 The compression codings do not define any parameters. The presence 1207 of parameters with any of these compression codings SHOULD be treated 1208 as an error. 1210 7.3. Transfer Coding Registry 1212 The "HTTP Transfer Coding Registry" defines the namespace for 1213 transfer coding names. It is maintained at 1214 . 1216 Registrations MUST include the following fields: 1218 * Name 1220 * Description 1222 * Pointer to specification text 1224 Names of transfer codings MUST NOT overlap with names of content 1225 codings (Section 8.4.1 of [HTTP]) unless the encoding transformation 1226 is identical, as is the case for the compression codings defined in 1227 Section 7.2. 1229 The TE header field (Section 10.1.4 of [HTTP]) uses a pseudo 1230 parameter named "q" as rank value when multiple transfer codings are 1231 acceptable. Future registrations of transfer codings SHOULD NOT 1232 define parameters called "q" (case-insensitively) in order to avoid 1233 ambiguities. 1235 Values to be added to this namespace require IETF Review (see 1236 Section 4.8 of [RFC8126]), and MUST conform to the purpose of 1237 transfer coding defined in this specification. 1239 Use of program names for the identification of encoding formats is 1240 not desirable and is discouraged for future encodings. 1242 7.4. Negotiating Transfer Codings 1244 The TE field (Section 10.1.4 of [HTTP]) is used in HTTP/1.1 to 1245 indicate what transfer-codings, besides chunked, the client is 1246 willing to accept in the response, and whether the client is willing 1247 to preserve trailer fields in a chunked transfer coding. 1249 A client MUST NOT send the chunked transfer coding name in TE; 1250 chunked is always acceptable for HTTP/1.1 recipients. 1252 Three examples of TE use are below. 1254 TE: deflate 1255 TE: 1256 TE: trailers, deflate;q=0.5 1258 When multiple transfer codings are acceptable, the client MAY rank 1259 the codings by preference using a case-insensitive "q" parameter 1260 (similar to the qvalues used in content negotiation fields, 1261 Section 12.4.2 of [HTTP]). The rank value is a real number in the 1262 range 0 through 1, where 0.001 is the least preferred and 1 is the 1263 most preferred; a value of 0 means "not acceptable". 1265 If the TE field value is empty or if no TE field is present, the only 1266 acceptable transfer coding is chunked. A message with no transfer 1267 coding is always acceptable. 1269 The keyword "trailers" indicates that the sender will not discard 1270 trailer fields, as described in Section 6.5 of [HTTP]. 1272 Since the TE header field only applies to the immediate connection, a 1273 sender of TE MUST also send a "TE" connection option within the 1274 Connection header field (Section 7.6.1 of [HTTP]) in order to prevent 1275 the TE header field from being forwarded by intermediaries that do 1276 not support its semantics. 1278 8. Handling Incomplete Messages 1280 A server that receives an incomplete request message, usually due to 1281 a canceled request or a triggered timeout exception, MAY send an 1282 error response prior to closing the connection. 1284 A client that receives an incomplete response message, which can 1285 occur when a connection is closed prematurely or when decoding a 1286 supposedly chunked transfer coding fails, MUST record the message as 1287 incomplete. Cache requirements for incomplete responses are defined 1288 in Section 3 of [CACHING]. 1290 If a response terminates in the middle of the header section (before 1291 the empty line is received) and the status code might rely on header 1292 fields to convey the full meaning of the response, then the client 1293 cannot assume that meaning has been conveyed; the client might need 1294 to repeat the request in order to determine what action to take next. 1296 A message body that uses the chunked transfer coding is incomplete if 1297 the zero-sized chunk that terminates the encoding has not been 1298 received. A message that uses a valid Content-Length is incomplete 1299 if the size of the message body received (in octets) is less than the 1300 value given by Content-Length. A response that has neither chunked 1301 transfer coding nor Content-Length is terminated by closure of the 1302 connection and, if the header section was received intact, is 1303 considered complete unless an error was indicated by the underlying 1304 connection (e.g., an "incomplete close" in TLS would leave the 1305 response incomplete, as described in Section 9.8). 1307 9. Connection Management 1309 HTTP messaging is independent of the underlying transport- or 1310 session-layer connection protocol(s). HTTP only presumes a reliable 1311 transport with in-order delivery of requests and the corresponding 1312 in-order delivery of responses. The mapping of HTTP request and 1313 response structures onto the data units of an underlying transport 1314 protocol is outside the scope of this specification. 1316 As described in Section 7.3 of [HTTP], the specific connection 1317 protocols to be used for an HTTP interaction are determined by client 1318 configuration and the target URI. For example, the "http" URI scheme 1319 (Section 4.2.1 of [HTTP]) indicates a default connection of TCP over 1320 IP, with a default TCP port of 80, but the client might be configured 1321 to use a proxy via some other connection, port, or protocol. 1323 HTTP implementations are expected to engage in connection management, 1324 which includes maintaining the state of current connections, 1325 establishing a new connection or reusing an existing connection, 1326 processing messages received on a connection, detecting connection 1327 failures, and closing each connection. Most clients maintain 1328 multiple connections in parallel, including more than one connection 1329 per server endpoint. Most servers are designed to maintain thousands 1330 of concurrent connections, while controlling request queues to enable 1331 fair use and detect denial-of-service attacks. 1333 9.1. Establishment 1335 It is beyond the scope of this specification to describe how 1336 connections are established via various transport- or session-layer 1337 protocols. Each HTTP connection maps to one underlying transport 1338 connection. 1340 9.2. Associating a Response to a Request 1342 HTTP/1.1 does not include a request identifier for associating a 1343 given request message with its corresponding one or more response 1344 messages. Hence, it relies on the order of response arrival to 1345 correspond exactly to the order in which requests are made on the 1346 same connection. More than one response message per request only 1347 occurs when one or more informational responses (1xx, see 1348 Section 15.2 of [HTTP]) precede a final response to the same request. 1350 A client that has more than one outstanding request on a connection 1351 MUST maintain a list of outstanding requests in the order sent and 1352 MUST associate each received response message on that connection to 1353 the first outstanding request that has not yet received a final (non- 1354 1xx) response. 1356 If a client receives data on a connection that doesn't have 1357 outstanding requests, the client MUST NOT consider that data to be a 1358 valid response; the client SHOULD close the connection, since message 1359 delimitation is now ambiguous, unless the data consists only of one 1360 or more CRLF (which can be discarded, as per Section 2.2). 1362 9.3. Persistence 1364 HTTP/1.1 defaults to the use of _persistent connections_, allowing 1365 multiple requests and responses to be carried over a single 1366 connection. HTTP implementations SHOULD support persistent 1367 connections. 1369 A recipient determines whether a connection is persistent or not 1370 based on the protocol version and Connection header field 1371 (Section 7.6.1 of [HTTP]) in the most recently received message, if 1372 any: 1374 * If the close connection option is present (Section 9.6), the 1375 connection will not persist after the current response; else, 1377 * If the received protocol is HTTP/1.1 (or later), the connection 1378 will persist after the current response; else, 1380 * If the received protocol is HTTP/1.0, the "keep-alive" connection 1381 option is present, either the recipient is not a proxy or the 1382 message is a response, and the recipient wishes to honor the 1383 HTTP/1.0 "keep-alive" mechanism, the connection will persist after 1384 the current response; otherwise, 1386 * The connection will close after the current response. 1388 A client that does not support persistent connections MUST send the 1389 close connection option in every request message. 1391 A server that does not support persistent connections MUST send the 1392 close connection option in every response message that does not have 1393 a 1xx (Informational) status code. 1395 A client MAY send additional requests on a persistent connection 1396 until it sends or receives a close connection option or receives an 1397 HTTP/1.0 response without a "keep-alive" connection option. 1399 In order to remain persistent, all messages on a connection need to 1400 have a self-defined message length (i.e., one not defined by closure 1401 of the connection), as described in Section 6. A server MUST read 1402 the entire request message body or close the connection after sending 1403 its response, since otherwise the remaining data on a persistent 1404 connection would be misinterpreted as the next request. Likewise, a 1405 client MUST read the entire response message body if it intends to 1406 reuse the same connection for a subsequent request. 1408 A proxy server MUST NOT maintain a persistent connection with an 1409 HTTP/1.0 client (see Section 19.7.1 of [RFC2068] for information and 1410 discussion of the problems with the Keep-Alive header field 1411 implemented by many HTTP/1.0 clients). 1413 See Appendix C.2.2 for more information on backwards compatibility 1414 with HTTP/1.0 clients. 1416 9.3.1. Retrying Requests 1418 Connections can be closed at any time, with or without intention. 1419 Implementations ought to anticipate the need to recover from 1420 asynchronous close events. The conditions under which a client can 1421 automatically retry a sequence of outstanding requests are defined in 1422 Section 9.2.2 of [HTTP]. 1424 9.3.2. Pipelining 1426 A client that supports persistent connections MAY _pipeline_ its 1427 requests (i.e., send multiple requests without waiting for each 1428 response). A server MAY process a sequence of pipelined requests in 1429 parallel if they all have safe methods (Section 9.2.1 of [HTTP]), but 1430 it MUST send the corresponding responses in the same order that the 1431 requests were received. 1433 A client that pipelines requests SHOULD retry unanswered requests if 1434 the connection closes before it receives all of the corresponding 1435 responses. When retrying pipelined requests after a failed 1436 connection (a connection not explicitly closed by the server in its 1437 last complete response), a client MUST NOT pipeline immediately after 1438 connection establishment, since the first remaining request in the 1439 prior pipeline might have caused an error response that can be lost 1440 again if multiple requests are sent on a prematurely closed 1441 connection (see the TCP reset problem described in Section 9.6). 1443 Idempotent methods (Section 9.2.2 of [HTTP]) are significant to 1444 pipelining because they can be automatically retried after a 1445 connection failure. A user agent SHOULD NOT pipeline requests after 1446 a non-idempotent method, until the final response status code for 1447 that method has been received, unless the user agent has a means to 1448 detect and recover from partial failure conditions involving the 1449 pipelined sequence. 1451 An intermediary that receives pipelined requests MAY pipeline those 1452 requests when forwarding them inbound, since it can rely on the 1453 outbound user agent(s) to determine what requests can be safely 1454 pipelined. If the inbound connection fails before receiving a 1455 response, the pipelining intermediary MAY attempt to retry a sequence 1456 of requests that have yet to receive a response if the requests all 1457 have idempotent methods; otherwise, the pipelining intermediary 1458 SHOULD forward any received responses and then close the 1459 corresponding outbound connection(s) so that the outbound user 1460 agent(s) can recover accordingly. 1462 9.4. Concurrency 1464 A client ought to limit the number of simultaneous open connections 1465 that it maintains to a given server. 1467 Previous revisions of HTTP gave a specific number of connections as a 1468 ceiling, but this was found to be impractical for many applications. 1469 As a result, this specification does not mandate a particular maximum 1470 number of connections but, instead, encourages clients to be 1471 conservative when opening multiple connections. 1473 Multiple connections are typically used to avoid the "head-of-line 1474 blocking" problem, wherein a request that takes significant server- 1475 side processing and/or transfers very large content would block 1476 subsequent requests on the same connection. However, each connection 1477 consumes server resources. 1479 Furthermore, using multiple connections can cause undesirable side 1480 effects in congested networks. Using larger numbers of connections 1481 can also cause side effects in otherwise uncongested networks, 1482 because their aggregate and initially synchronized sending behavior 1483 can cause congestion that would not have been present if fewer 1484 parallel connections had been used. 1486 Note that a server might reject traffic that it deems abusive or 1487 characteristic of a denial-of-service attack, such as an excessive 1488 number of open connections from a single client. 1490 9.5. Failures and Timeouts 1492 Servers will usually have some timeout value beyond which they will 1493 no longer maintain an inactive connection. Proxy servers might make 1494 this a higher value since it is likely that the client will be making 1495 more connections through the same proxy server. The use of 1496 persistent connections places no requirements on the length (or 1497 existence) of this timeout for either the client or the server. 1499 A client or server that wishes to time out SHOULD issue a graceful 1500 close on the connection. Implementations SHOULD constantly monitor 1501 open connections for a received closure signal and respond to it as 1502 appropriate, since prompt closure of both sides of a connection 1503 enables allocated system resources to be reclaimed. 1505 A client, server, or proxy MAY close the transport connection at any 1506 time. For example, a client might have started to send a new request 1507 at the same time that the server has decided to close the "idle" 1508 connection. From the server's point of view, the connection is being 1509 closed while it was idle, but from the client's point of view, a 1510 request is in progress. 1512 A server SHOULD sustain persistent connections, when possible, and 1513 allow the underlying transport's flow-control mechanisms to resolve 1514 temporary overloads, rather than terminate connections with the 1515 expectation that clients will retry. The latter technique can 1516 exacerbate network congestion or server load. 1518 A client sending a message body SHOULD monitor the network connection 1519 for an error response while it is transmitting the request. If the 1520 client sees a response that indicates the server does not wish to 1521 receive the message body and is closing the connection, the client 1522 SHOULD immediately cease transmitting the body and close its side of 1523 the connection. 1525 9.6. Tear-down 1527 The "close" connection option is defined as a signal that the sender 1528 will close this connection after completion of the response. A 1529 sender SHOULD send a Connection header field (Section 7.6.1 of 1530 [HTTP]) containing the close connection option when it intends to 1531 close a connection. For example, 1533 Connection: close 1535 as a request header field indicates that this is the last request 1536 that the client will send on this connection, while in a response the 1537 same field indicates that the server is going to close this 1538 connection after the response message is complete. 1540 Note that the field name "Close" is reserved, since using that name 1541 as a header field might conflict with the close connection option. 1543 A client that sends a close connection option MUST NOT send further 1544 requests on that connection (after the one containing the close) and 1545 MUST close the connection after reading the final response message 1546 corresponding to this request. 1548 A server that receives a close connection option MUST initiate 1549 closure of the connection (see below) after it sends the final 1550 response to the request that contained the close connection option. 1551 The server SHOULD send a close connection option in its final 1552 response on that connection. The server MUST NOT process any further 1553 requests received on that connection. 1555 A server that sends a close connection option MUST initiate closure 1556 of the connection (see below) after it sends the response containing 1557 the close connection option. The server MUST NOT process any further 1558 requests received on that connection. 1560 A client that receives a close connection option MUST cease sending 1561 requests on that connection and close the connection after reading 1562 the response message containing the close connection option; if 1563 additional pipelined requests had been sent on the connection, the 1564 client SHOULD NOT assume that they will be processed by the server. 1566 If a server performs an immediate close of a TCP connection, there is 1567 a significant risk that the client will not be able to read the last 1568 HTTP response. If the server receives additional data from the 1569 client on a fully closed connection, such as another request sent by 1570 the client before receiving the server's response, the server's TCP 1571 stack will send a reset packet to the client; unfortunately, the 1572 reset packet might erase the client's unacknowledged input buffers 1573 before they can be read and interpreted by the client's HTTP parser. 1575 To avoid the TCP reset problem, servers typically close a connection 1576 in stages. First, the server performs a half-close by closing only 1577 the write side of the read/write connection. The server then 1578 continues to read from the connection until it receives a 1579 corresponding close by the client, or until the server is reasonably 1580 certain that its own TCP stack has received the client's 1581 acknowledgement of the packet(s) containing the server's last 1582 response. Finally, the server fully closes the connection. 1584 It is unknown whether the reset problem is exclusive to TCP or might 1585 also be found in other transport connection protocols. 1587 Note that a TCP connection that is half-closed by the client does not 1588 delimit a request message, nor does it imply that the client is no 1589 longer interested in a response. In general, transport signals 1590 cannot be relied upon to signal edge cases, since HTTP/1.1 is 1591 independent of transport. 1593 9.7. TLS Connection Initiation 1595 Conceptually, HTTP/TLS is simply sending HTTP messages over a 1596 connection secured via TLS [TLS13]. 1598 The HTTP client also acts as the TLS client. It initiates a 1599 connection to the server on the appropriate port and sends the TLS 1600 ClientHello to begin the TLS handshake. When the TLS handshake has 1601 finished, the client may then initiate the first HTTP request. All 1602 HTTP data MUST be sent as TLS "application data", but is otherwise 1603 treated like a normal connection for HTTP (including potential reuse 1604 as a persistent connection). 1606 9.8. TLS Connection Closure 1608 TLS provides a facility for secure connection closure through an 1609 exchange of closure alerts prior to closing a connection [TLS13]. 1610 When a valid closure alert is received, an implementation can be 1611 assured that no further data will be received on that connection. 1613 When an implementation knows that it has sent or received all the 1614 message data that it cares about, typically by detecting HTTP message 1615 boundaries, it might generate an "incomplete close" by sending a 1616 closure alert and then closing the connection without waiting to 1617 receive the corresponding closure alert from its peer. 1619 An incomplete close does not call into question the security of the 1620 data already received, but it could indicate that subsequent data 1621 might have been truncated. As TLS is not directly aware of HTTP 1622 message framing, it is necessary to examine the HTTP data itself to 1623 determine whether messages were complete. Handling of incomplete 1624 messages is defined in Section 8. 1626 When encountering an incomplete close, a client SHOULD treat as 1627 completed all requests for which it has received as much data as 1628 specified in the Content-Length header or, when a Transfer-Encoding 1629 of chunked is used, for which the terminal zero-length chunk has been 1630 received. A response that has neither chunked transfer coding nor 1631 Content-Length is complete only if a valid closure alert has been 1632 received. Treating an incomplete message as complete could expose 1633 implementations to attack. 1635 A client detecting an incomplete close SHOULD recover gracefully. 1637 Clients MUST send a closure alert before closing the connection. 1638 Clients that do not expect to receive any more data MAY choose not to 1639 wait for the server's closure alert and simply close the connection, 1640 thus generating an incomplete close on the server side. 1642 Servers SHOULD be prepared to receive an incomplete close from the 1643 client, since the client can often determine when the end of server 1644 data is. 1646 Servers MUST attempt to initiate an exchange of closure alerts with 1647 the client before closing the connection. Servers MAY close the 1648 connection after sending the closure alert, thus generating an 1649 incomplete close on the client side. 1651 10. Enclosing Messages as Data 1653 10.1. Media Type message/http 1655 The message/http media type can be used to enclose a single HTTP 1656 request or response message, provided that it obeys the MIME 1657 restrictions for all "message" types regarding line length and 1658 encodings. Because of the line length limitations, field values 1659 within message/http are allowed to use line folding (obs-fold), as 1660 described in Section 5.2, to convey the field value over multiple 1661 lines. A recipient of message/http data MUST replace any obsolete 1662 line folding with one or more SP characters when the message is 1663 consumed. 1665 Type name: message 1667 Subtype name: http 1669 Required parameters: N/A 1671 Optional parameters: version, msgtype 1673 version: The HTTP-version number of the enclosed message (e.g., 1674 "1.1"). If not present, the version can be determined from the 1675 first line of the body. 1677 msgtype: The message type - "request" or "response". If not 1678 present, the type can be determined from the first line of the 1679 body. 1681 Encoding considerations: only "7bit", "8bit", or "binary" are 1682 permitted 1684 Security considerations: see Section 11 1686 Interoperability considerations: N/A 1688 Published specification: This specification (see Section 10.1). 1690 Applications that use this media type: N/A 1692 Fragment identifier considerations: N/A 1694 Additional information: Magic number(s): N/A 1696 Deprecated alias names for this type: N/A 1698 File extension(s): N/A 1700 Macintosh file type code(s): N/A 1702 Person and email address to contact for further information: See Aut 1703 hors' Addresses section. 1705 Intended usage: COMMON 1707 Restrictions on usage: N/A 1709 Author: See Authors' Addresses section. 1711 Change controller: IESG 1713 10.2. Media Type application/http 1715 The application/http media type can be used to enclose a pipeline of 1716 one or more HTTP request or response messages (not intermixed). 1718 Type name: application 1720 Subtype name: http 1722 Required parameters: N/A 1724 Optional parameters: version, msgtype 1726 version: The HTTP-version number of the enclosed messages (e.g., 1727 "1.1"). If not present, the version can be determined from the 1728 first line of the body. 1730 msgtype: The message type - "request" or "response". If not 1731 present, the type can be determined from the first line of the 1732 body. 1734 Encoding considerations: HTTP messages enclosed by this type are in 1735 "binary" format; use of an appropriate Content-Transfer-Encoding 1736 is required when transmitted via email. 1738 Security considerations: see Section 11 1740 Interoperability considerations: N/A 1742 Published specification: This specification (see Section 10.2). 1744 Applications that use this media type: N/A 1746 Fragment identifier considerations: N/A 1748 Additional information: Deprecated alias names for this type: N/A 1750 Magic number(s): N/A 1752 File extension(s): N/A 1754 Macintosh file type code(s): N/A 1756 Person and email address to contact for further information: See Aut 1757 hors' Addresses section. 1759 Intended usage: COMMON 1761 Restrictions on usage: N/A 1763 Author: See Authors' Addresses section. 1765 Change controller: IESG 1767 11. Security Considerations 1769 This section is meant to inform developers, information providers, 1770 and users about known security considerations relevant to HTTP 1771 message syntax and parsing. Security considerations about HTTP 1772 semantics, content, and routing are addressed in [HTTP]. 1774 11.1. Response Splitting 1776 Response splitting (a.k.a., CRLF injection) is a common technique, 1777 used in various attacks on Web usage, that exploits the line-based 1778 nature of HTTP message framing and the ordered association of 1779 requests to responses on persistent connections [Klein]. This 1780 technique can be particularly damaging when the requests pass through 1781 a shared cache. 1783 Response splitting exploits a vulnerability in servers (usually 1784 within an application server) where an attacker can send encoded data 1785 within some parameter of the request that is later decoded and echoed 1786 within any of the response header fields of the response. If the 1787 decoded data is crafted to look like the response has ended and a 1788 subsequent response has begun, the response has been split and the 1789 content within the apparent second response is controlled by the 1790 attacker. The attacker can then make any other request on the same 1791 persistent connection and trick the recipients (including 1792 intermediaries) into believing that the second half of the split is 1793 an authoritative answer to the second request. 1795 For example, a parameter within the request-target might be read by 1796 an application server and reused within a redirect, resulting in the 1797 same parameter being echoed in the Location header field of the 1798 response. If the parameter is decoded by the application and not 1799 properly encoded when placed in the response field, the attacker can 1800 send encoded CRLF octets and other content that will make the 1801 application's single response look like two or more responses. 1803 A common defense against response splitting is to filter requests for 1804 data that looks like encoded CR and LF (e.g., "%0D" and "%0A"). 1805 However, that assumes the application server is only performing URI 1806 decoding, rather than more obscure data transformations like charset 1807 transcoding, XML entity translation, base64 decoding, sprintf 1808 reformatting, etc. A more effective mitigation is to prevent 1809 anything other than the server's core protocol libraries from sending 1810 a CR or LF within the header section, which means restricting the 1811 output of header fields to APIs that filter for bad octets and not 1812 allowing application servers to write directly to the protocol 1813 stream. 1815 11.2. Request Smuggling 1817 Request smuggling ([Linhart]) is a technique that exploits 1818 differences in protocol parsing among various recipients to hide 1819 additional requests (which might otherwise be blocked or disabled by 1820 policy) within an apparently harmless request. Like response 1821 splitting, request smuggling can lead to a variety of attacks on HTTP 1822 usage. 1824 This specification has introduced new requirements on request 1825 parsing, particularly with regard to message framing in Section 6.3, 1826 to reduce the effectiveness of request smuggling. 1828 11.3. Message Integrity 1830 HTTP does not define a specific mechanism for ensuring message 1831 integrity, instead relying on the error-detection ability of 1832 underlying transport protocols and the use of length or chunk- 1833 delimited framing to detect completeness. Historically, the lack of 1834 a single integrity mechanism has been justified by the informal 1835 nature of most HTTP communication. However, the prevalence of HTTP 1836 as an information access mechanism has resulted in its increasing use 1837 within environments where verification of message integrity is 1838 crucial. 1840 The mechanisms provided with the "https" scheme, such as 1841 authenticated encryption, provide protection against modification of 1842 messages. Care is needed however to ensure that connection closure 1843 cannot be used to truncate messages (see Section 9.8). User agents 1844 might refuse to accept incomplete messages or treat them specially. 1845 For example, a browser being used to view medical history or drug 1846 interaction information needs to indicate to the user when such 1847 information is detected by the protocol to be incomplete, expired, or 1848 corrupted during transfer. Such mechanisms might be selectively 1849 enabled via user agent extensions or the presence of message 1850 integrity metadata in a response. 1852 The "http" scheme provides no protection against accidental or 1853 malicious modification of messages. 1855 Extensions to the protocol might be used to mitigate the risk of 1856 unwanted modification of messages by intermediaries, even when the 1857 "https" scheme is used. Integrity might be assured by using message 1858 authentication codes or digital signatures that are selectively added 1859 to messages via extensible metadata fields. 1861 11.4. Message Confidentiality 1863 HTTP relies on underlying transport protocols to provide message 1864 confidentiality when that is desired. HTTP has been specifically 1865 designed to be independent of the transport protocol, such that it 1866 can be used over many forms of encrypted connection, with the 1867 selection of such transports being identified by the choice of URI 1868 scheme or within user agent configuration. 1870 The "https" scheme can be used to identify resources that require a 1871 confidential connection, as described in Section 4.2.2 of [HTTP]. 1873 12. IANA Considerations 1875 The change controller for the following registrations is: "IETF 1876 (iesg@ietf.org) - Internet Engineering Task Force". 1878 12.1. Field Name Registration 1880 First, introduce the new "Hypertext Transfer Protocol (HTTP) Field 1881 Name Registry" at as 1882 described in Section 18.4 of [HTTP]. 1884 Then, please update the registry with the field names listed in the 1885 table below: 1887 +===================+==========+======+============+ 1888 | Field Name | Status | Ref. | Comments | 1889 +===================+==========+======+============+ 1890 | Close | standard | 9.6 | (reserved) | 1891 +-------------------+----------+------+------------+ 1892 | MIME-Version | standard | B.1 | | 1893 +-------------------+----------+------+------------+ 1894 | Transfer-Encoding | standard | 6.1 | | 1895 +-------------------+----------+------+------------+ 1897 Table 1 1899 12.2. Media Type Registration 1901 Please update the "Media Types" registry at 1902 with the registration 1903 information in Section 10.1 and Section 10.2 for the media types 1904 "message/http" and "application/http", respectively. 1906 12.3. Transfer Coding Registration 1908 Please update the "HTTP Transfer Coding Registry" at 1909 with the 1910 registration procedure of Section 7.3 and the content coding names 1911 summarized in the table below. 1913 +============+===============================+===========+ 1914 | Name | Description | Reference | 1915 +============+===============================+===========+ 1916 | chunked | Transfer in a series of | Section | 1917 | | chunks | 7.1 | 1918 +------------+-------------------------------+-----------+ 1919 | compress | UNIX "compress" data format | Section | 1920 | | [Welch] | 7.2 | 1921 +------------+-------------------------------+-----------+ 1922 | deflate | "deflate" compressed data | Section | 1923 | | ([RFC1951]) inside the "zlib" | 7.2 | 1924 | | data format ([RFC1950]) | | 1925 +------------+-------------------------------+-----------+ 1926 | gzip | GZIP file format [RFC1952] | Section | 1927 | | | 7.2 | 1928 +------------+-------------------------------+-----------+ 1929 | trailers | (reserved) | Section | 1930 | | | 12.3 | 1931 +------------+-------------------------------+-----------+ 1932 | x-compress | Deprecated (alias for | Section | 1933 | | compress) | 7.2 | 1934 +------------+-------------------------------+-----------+ 1935 | x-gzip | Deprecated (alias for gzip) | Section | 1936 | | | 7.2 | 1937 +------------+-------------------------------+-----------+ 1939 Table 2 1941 | *Note:* the coding name "trailers" is reserved because its use 1942 | would conflict with the keyword "trailers" in the TE header 1943 | field (Section 10.1.4 of [HTTP]). 1945 12.4. ALPN Protocol ID Registration 1947 Please update the "TLS Application-Layer Protocol Negotiation (ALPN) 1948 Protocol IDs" registry at with the 1950 registration below: 1952 +==========+=============================+================+ 1953 | Protocol | Identification Sequence | Reference | 1954 +==========+=============================+================+ 1955 | HTTP/1.1 | 0x68 0x74 0x74 0x70 0x2f | (this | 1956 | | 0x31 0x2e 0x31 ("http/1.1") | specification) | 1957 +----------+-----------------------------+----------------+ 1959 Table 3 1961 13. References 1963 13.1. Normative References 1965 [CACHING] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 1966 Ed., "HTTP Caching", Work in Progress, Internet-Draft, 1967 draft-ietf-httpbis-cache-18, 18 August 2021, 1968 . 1971 [HTTP] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 1972 Ed., "HTTP Semantics", Work in Progress, Internet-Draft, 1973 draft-ietf-httpbis-semantics-18, 18 August 2021, 1974 . 1977 [RFC1950] Deutsch, L.P. and J-L. Gailly, "ZLIB Compressed Data 1978 Format Specification version 3.3", RFC 1950, 1979 DOI 10.17487/RFC1950, May 1996, 1980 . 1982 [RFC1951] Deutsch, P., "DEFLATE Compressed Data Format Specification 1983 version 1.3", RFC 1951, DOI 10.17487/RFC1951, May 1996, 1984 . 1986 [RFC1952] Deutsch, P., Gailly, J-L., Adler, M., Deutsch, L.P., and 1987 G. Randers-Pehrson, "GZIP file format specification 1988 version 4.3", RFC 1952, DOI 10.17487/RFC1952, May 1996, 1989 . 1991 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1992 Requirement Levels", BCP 14, RFC 2119, 1993 DOI 10.17487/RFC2119, March 1997, 1994 . 1996 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 1997 Specifications: ABNF", STD 68, RFC 5234, 1998 DOI 10.17487/RFC5234, January 2008, 1999 . 2001 [RFC7405] Kyzivat, P., "Case-Sensitive String Support in ABNF", 2002 RFC 7405, DOI 10.17487/RFC7405, December 2014, 2003 . 2005 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 2006 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 2007 May 2017, . 2009 [TLS13] Rescorla, E., "The Transport Layer Security (TLS) Protocol 2010 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 2011 . 2013 [URI] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 2014 Resource Identifier (URI): Generic Syntax", STD 66, 2015 RFC 3986, DOI 10.17487/RFC3986, January 2005, 2016 . 2018 [USASCII] American National Standards Institute, "Coded Character 2019 Set -- 7-bit American Standard Code for Information 2020 Interchange", ANSI X3.4, 1986. 2022 [Welch] Welch, T. A., "A Technique for High-Performance Data 2023 Compression", IEEE Computer 17(6), June 1984. 2025 13.2. Informative References 2027 [Err4667] RFC Errata, Erratum ID 4667, RFC 7230, 2028 . 2030 [HTTP/1.0] Berners-Lee, T., Fielding, R.T., and H.F. Nielsen, 2031 "Hypertext Transfer Protocol -- HTTP/1.0", RFC 1945, 2032 DOI 10.17487/RFC1945, May 1996, 2033 . 2035 [Klein] Klein, A., "Divide and Conquer - HTTP Response Splitting, 2036 Web Cache Poisoning Attacks, and Related Topics", March 2037 2004, . 2040 [Linhart] Linhart, C., Klein, A., Heled, R., and S. Orrin, "HTTP 2041 Request Smuggling", June 2005, 2042 . 2045 [RFC2045] Freed, N. and N.S. Borenstein, "Multipurpose Internet Mail 2046 Extensions (MIME) Part One: Format of Internet Message 2047 Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996, 2048 . 2050 [RFC2046] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2051 Extensions (MIME) Part Two: Media Types", RFC 2046, 2052 DOI 10.17487/RFC2046, November 1996, 2053 . 2055 [RFC2049] Freed, N. and N.S. Borenstein, "Multipurpose Internet Mail 2056 Extensions (MIME) Part Five: Conformance Criteria and 2057 Examples", RFC 2049, DOI 10.17487/RFC2049, November 1996, 2058 . 2060 [RFC2068] Fielding, R., Gettys, J., Mogul, J., Nielsen, H., and T. 2061 Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", 2062 RFC 2068, DOI 10.17487/RFC2068, January 1997, 2063 . 2065 [RFC2557] Palme, F., Hopmann, A., Shelness, N., and E. Stefferud, 2066 "MIME Encapsulation of Aggregate Documents, such as HTML 2067 (MHTML)", RFC 2557, DOI 10.17487/RFC2557, March 1999, 2068 . 2070 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 2071 DOI 10.17487/RFC5322, October 2008, 2072 . 2074 [RFC7230] Fielding, R., Ed. and J. F. Reschke, Ed., "Hypertext 2075 Transfer Protocol (HTTP/1.1): Message Syntax and Routing", 2076 RFC 7230, DOI 10.17487/RFC7230, June 2014, 2077 . 2079 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2080 Writing an IANA Considerations Section in RFCs", BCP 26, 2081 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2082 . 2084 Appendix A. Collected ABNF 2086 In the collected ABNF below, list rules are expanded as per 2087 Section 5.6.1.1 of [HTTP]. 2089 BWS = 2091 HTTP-message = start-line CRLF *( field-line CRLF ) CRLF [ 2092 message-body ] 2093 HTTP-name = %x48.54.54.50 ; HTTP 2094 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 2096 OWS = 2098 RWS = 2100 Transfer-Encoding = [ transfer-coding *( OWS "," OWS transfer-coding 2101 ) ] 2103 absolute-URI = 2104 absolute-form = absolute-URI 2105 absolute-path = 2106 asterisk-form = "*" 2107 authority = 2108 authority-form = uri-host ":" port 2110 chunk = chunk-size [ chunk-ext ] CRLF chunk-data CRLF 2111 chunk-data = 1*OCTET 2112 chunk-ext = *( BWS ";" BWS chunk-ext-name [ BWS "=" BWS chunk-ext-val 2113 ] ) 2114 chunk-ext-name = token 2115 chunk-ext-val = token / quoted-string 2116 chunk-size = 1*HEXDIG 2117 chunked-body = *chunk last-chunk trailer-section CRLF 2119 field-line = field-name ":" OWS field-value OWS 2120 field-name = 2121 field-value = 2123 last-chunk = 1*"0" [ chunk-ext ] CRLF 2125 message-body = *OCTET 2126 method = token 2128 obs-fold = OWS CRLF RWS 2129 obs-text = 2130 origin-form = absolute-path [ "?" query ] 2132 port = 2134 query = 2135 quoted-string = 2137 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 2138 request-line = method SP request-target SP HTTP-version 2139 request-target = origin-form / absolute-form / authority-form / 2140 asterisk-form 2142 start-line = request-line / status-line 2143 status-code = 3DIGIT 2144 status-line = HTTP-version SP status-code SP [ reason-phrase ] 2146 token = 2147 trailer-section = *( field-line CRLF ) 2148 transfer-coding = 2150 uri-host = 2152 Appendix B. Differences between HTTP and MIME 2154 HTTP/1.1 uses many of the constructs defined for the Internet Message 2155 Format [RFC5322] and the Multipurpose Internet Mail Extensions (MIME) 2156 [RFC2045] to allow a message body to be transmitted in an open 2157 variety of representations and with extensible fields. However, RFC 2158 2045 is focused only on email; applications of HTTP have many 2159 characteristics that differ from email; hence, HTTP has features that 2160 differ from MIME. These differences were carefully chosen to 2161 optimize performance over binary connections, to allow greater 2162 freedom in the use of new media types, to make date comparisons 2163 easier, and to acknowledge the practice of some early HTTP servers 2164 and clients. 2166 This appendix describes specific areas where HTTP differs from MIME. 2167 Proxies and gateways to and from strict MIME environments need to be 2168 aware of these differences and provide the appropriate conversions 2169 where necessary. 2171 B.1. MIME-Version 2173 HTTP is not a MIME-compliant protocol. However, messages can include 2174 a single MIME-Version header field to indicate what version of the 2175 MIME protocol was used to construct the message. Use of the MIME- 2176 Version header field indicates that the message is in full 2177 conformance with the MIME protocol (as defined in [RFC2045]). 2178 Senders are responsible for ensuring full conformance (where 2179 possible) when exporting HTTP messages to strict MIME environments. 2181 B.2. Conversion to Canonical Form 2183 MIME requires that an Internet mail body part be converted to 2184 canonical form prior to being transferred, as described in Section 4 2185 of [RFC2049], and that content with a type of "text" represent line 2186 breaks as CRLF, forbidding the use of CR or LF outside of line break 2187 sequences [RFC2046]. In contrast, HTTP does not care whether CRLF, 2188 bare CR, or bare LF are used to indicate a line break within content. 2190 A proxy or gateway from HTTP to a strict MIME environment ought to 2191 translate all line breaks within text media types to the RFC 2049 2192 canonical form of CRLF. Note, however, this might be complicated by 2193 the presence of a Content-Encoding and by the fact that HTTP allows 2194 the use of some charsets that do not use octets 13 and 10 to 2195 represent CR and LF, respectively. 2197 Conversion will break any cryptographic checksums applied to the 2198 original content unless the original content is already in canonical 2199 form. Therefore, the canonical form is recommended for any content 2200 that uses such checksums in HTTP. 2202 B.3. Conversion of Date Formats 2204 HTTP/1.1 uses a restricted set of date formats (Section 5.6.7 of 2205 [HTTP]) to simplify the process of date comparison. Proxies and 2206 gateways from other protocols ought to ensure that any Date header 2207 field present in a message conforms to one of the HTTP/1.1 formats 2208 and rewrite the date if necessary. 2210 B.4. Conversion of Content-Encoding 2212 MIME does not include any concept equivalent to HTTP/1.1's Content- 2213 Encoding header field. Since this acts as a modifier on the media 2214 type, proxies and gateways from HTTP to MIME-compliant protocols 2215 ought to either change the value of the Content-Type header field or 2216 decode the representation before forwarding the message. (Some 2217 experimental applications of Content-Type for Internet mail have used 2218 a media-type parameter of ";conversions=" to perform 2219 a function equivalent to Content-Encoding. However, this parameter 2220 is not part of the MIME standards). 2222 B.5. Conversion of Content-Transfer-Encoding 2224 HTTP does not use the Content-Transfer-Encoding field of MIME. 2225 Proxies and gateways from MIME-compliant protocols to HTTP need to 2226 remove any Content-Transfer-Encoding prior to delivering the response 2227 message to an HTTP client. 2229 Proxies and gateways from HTTP to MIME-compliant protocols are 2230 responsible for ensuring that the message is in the correct format 2231 and encoding for safe transport on that protocol, where "safe 2232 transport" is defined by the limitations of the protocol being used. 2233 Such a proxy or gateway ought to transform and label the data with an 2234 appropriate Content-Transfer-Encoding if doing so will improve the 2235 likelihood of safe transport over the destination protocol. 2237 B.6. MHTML and Line Length Limitations 2239 HTTP implementations that share code with MHTML [RFC2557] 2240 implementations need to be aware of MIME line length limitations. 2241 Since HTTP does not have this limitation, HTTP does not fold long 2242 lines. MHTML messages being transported by HTTP follow all 2243 conventions of MHTML, including line length limitations and folding, 2244 canonicalization, etc., since HTTP transfers message-bodies without 2245 modification and, aside from the "multipart/byteranges" type 2246 (Section 14.6 of [HTTP]), does not interpret the content or any MIME 2247 header lines that might be contained therein. 2249 Appendix C. Changes from previous RFCs 2251 C.1. Changes from HTTP/0.9 2253 Since HTTP/0.9 did not support header fields in a request, there is 2254 no mechanism for it to support name-based virtual hosts (selection of 2255 resource by inspection of the Host header field). Any server that 2256 implements name-based virtual hosts ought to disable support for 2257 HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in fact, 2258 badly constructed HTTP/1.x requests caused by a client failing to 2259 properly encode the request-target. 2261 C.2. Changes from HTTP/1.0 2263 C.2.1. Multihomed Web Servers 2265 The requirements that clients and servers support the Host header 2266 field (Section 7.2 of [HTTP]), report an error if it is missing from 2267 an HTTP/1.1 request, and accept absolute URIs (Section 3.2) are among 2268 the most important changes defined by HTTP/1.1. 2270 Older HTTP/1.0 clients assumed a one-to-one relationship of IP 2271 addresses and servers; there was no other established mechanism for 2272 distinguishing the intended server of a request than the IP address 2273 to which that request was directed. The Host header field was 2274 introduced during the development of HTTP/1.1 and, though it was 2275 quickly implemented by most HTTP/1.0 browsers, additional 2276 requirements were placed on all HTTP/1.1 requests in order to ensure 2277 complete adoption. At the time of this writing, most HTTP-based 2278 services are dependent upon the Host header field for targeting 2279 requests. 2281 C.2.2. Keep-Alive Connections 2283 In HTTP/1.0, each connection is established by the client prior to 2284 the request and closed by the server after sending the response. 2285 However, some implementations implement the explicitly negotiated 2286 ("Keep-Alive") version of persistent connections described in 2287 Section 19.7.1 of [RFC2068]. 2289 Some clients and servers might wish to be compatible with these 2290 previous approaches to persistent connections, by explicitly 2291 negotiating for them with a "Connection: keep-alive" request header 2292 field. However, some experimental implementations of HTTP/1.0 2293 persistent connections are faulty; for example, if an HTTP/1.0 proxy 2294 server doesn't understand Connection, it will erroneously forward 2295 that header field to the next inbound server, which would result in a 2296 hung connection. 2298 One attempted solution was the introduction of a Proxy-Connection 2299 header field, targeted specifically at proxies. In practice, this 2300 was also unworkable, because proxies are often deployed in multiple 2301 layers, bringing about the same problem discussed above. 2303 As a result, clients are encouraged not to send the Proxy-Connection 2304 header field in any requests. 2306 Clients are also encouraged to consider the use of Connection: keep- 2307 alive in requests carefully; while they can enable persistent 2308 connections with HTTP/1.0 servers, clients using them will need to 2309 monitor the connection for "hung" requests (which indicate that the 2310 client ought to stop sending the header field), and this mechanism 2311 ought not be used by clients at all when a proxy is being used. 2313 C.2.3. Introduction of Transfer-Encoding 2315 HTTP/1.1 introduces the Transfer-Encoding header field (Section 6.1). 2316 Transfer codings need to be decoded prior to forwarding an HTTP 2317 message over a MIME-compliant protocol. 2319 C.3. Changes from RFC 7230 2321 Most of the sections introducing HTTP's design goals, history, 2322 architecture, conformance criteria, protocol versioning, URIs, 2323 message routing, and header fields have been moved to [HTTP]. This 2324 document has been reduced to just the messaging syntax and connection 2325 management requirements specific to HTTP/1.1. 2327 Bare CRs have been prohibited outside of content. (Section 2.2) 2329 The ABNF definition of authority-form has changed from the more 2330 general authority component of a URI (in which port is optional) to 2331 the specific host:port format that is required by CONNECT. 2332 (Section 3.2.3) 2334 Required recipients to avoid smuggling/splitting attacks when 2335 processing an ambiguous message framing. (Section 6.1) 2337 In the ABNF for chunked extensions, re-introduced (bad) whitespace 2338 around ";" and "=". Whitespace was removed in [RFC7230], but that 2339 change was found to break existing implementations (see [Err4667]). 2340 (Section 7.1.1) 2341 Trailer field semantics now transcend the specifics of chunked 2342 encoding. The decoding algorithm for chunked (Section 7.1.3) has 2343 been updated to encourage storage/forwarding of trailer fields 2344 separately from the header section, to only allow merging into the 2345 header section if the recipient knows the corresponding field 2346 definition permits and defines how to merge, and otherwise to discard 2347 the trailer fields instead of merging. The trailer part is now 2348 called the trailer section to be more consistent with the header 2349 section and more distinct from a body part. (Section 7.1.2) 2351 Disallowed transfer coding parameters called "q" in order to avoid 2352 conflicts with the use of ranks in the TE header field. 2353 (Section 7.3) 2355 Appendix D. Change Log 2357 This section is to be removed before publishing as an RFC. 2359 D.1. Between RFC7230 and draft 00 2361 The changes were purely editorial: 2363 * Change boilerplate and abstract to indicate the "draft" status, 2364 and update references to ancestor specifications. 2366 * Adjust historical notes. 2368 * Update links to sibling specifications. 2370 * Replace sections listing changes from RFC 2616 by new empty 2371 sections referring to RFC 723x. 2373 * Remove acknowledgements specific to RFC 723x. 2375 * Move "Acknowledgements" to the very end and make them unnumbered. 2377 D.2. Since draft-ietf-httpbis-messaging-00 2379 The changes in this draft are editorial, with respect to HTTP as a 2380 whole, to move all core HTTP semantics into [HTTP]: 2382 * Moved introduction, architecture, conformance, and ABNF extensions 2383 from RFC 7230 (Messaging) to semantics [HTTP]. 2385 * Moved discussion of MIME differences from RFC 7231 (Semantics) to 2386 Appendix B since they mostly cover transforming 1.1 messages. 2388 * Moved all extensibility tips, registration procedures, and 2389 registry tables from the IANA considerations to normative 2390 sections, reducing the IANA considerations to just instructions 2391 that will be removed prior to publication as an RFC. 2393 D.3. Since draft-ietf-httpbis-messaging-01 2395 * Cite RFC 8126 instead of RFC 5226 () 2398 * Resolved erratum 4779, no change needed here 2399 (, 2400 ) 2402 * In Section 7, fixed prose claiming transfer parameters allow bare 2403 names (, 2404 ) 2406 * Resolved erratum 4225, no change needed here 2407 (, 2408 ) 2410 * Replace "response code" with "response status code" 2411 (, 2412 ) 2414 * In Section 9.3, clarify statement about HTTP/1.0 keep-alive 2415 (, 2416 ) 2418 * In Section 7.1.1, re-introduce (bad) whitespace around ";" and "=" 2419 (, 2420 , ) 2423 * In Section 7.3, state that transfer codings should not use 2424 parameters named "q" (, ) 2427 * In Section 7, mark coding name "trailers" as reserved in the IANA 2428 registry () 2430 D.4. Since draft-ietf-httpbis-messaging-02 2432 * In Section 4, explain why the reason phrase should be ignored by 2433 clients (). 2435 * Add Section 9.2 to explain how request/response correlation is 2436 performed () 2438 D.5. Since draft-ietf-httpbis-messaging-03 2440 * In Section 9.2, caution against treating data on a connection as 2441 part of a not-yet-issued request () 2444 * In Section 7, remove the predefined codings from the ABNF and make 2445 it generic instead () 2448 * Use RFC 7405 ABNF notation for case-sensitive string constants 2449 () 2451 D.6. Since draft-ietf-httpbis-messaging-04 2453 * In Section 7.8 of [HTTP], clarify that protocol-name is to be 2454 matched case-insensitively () 2457 * In Section 5.2, add leading optional whitespace to obs-fold ABNF 2458 (, 2459 ) 2461 * In Section 4, add clarifications about empty reason phrases 2462 () 2464 * Move discussion of retries from Section 9.3.1 into [HTTP] 2465 () 2467 D.7. Since draft-ietf-httpbis-messaging-05 2469 * In Section 7.1.2, the trailer part has been renamed the trailer 2470 section (for consistency with the header section) and trailers are 2471 no longer merged as header fields by default, but rather can be 2472 discarded, kept separate from header fields, or merged with header 2473 fields only if understood and defined as being mergeable 2474 () 2476 * In Section 2.1 and related Sections, move the trailing CRLF from 2477 the line grammars into the message format 2478 () 2480 * Moved Section 2.3 down () 2483 * In Section 7.8 of [HTTP], use 'websocket' instead of 'HTTP/2.0' in 2484 examples () 2486 * Move version non-specific text from Section 6 into semantics as 2487 "payload" () 2489 * In Section 9.8, add text from RFC 2818 2490 () 2492 D.8. Since draft-ietf-httpbis-messaging-06 2494 * In Section 12.4, update the ALPN protocol ID for HTTP/1.1 2495 () 2497 * In Section 5, align with updates to field terminology in semantics 2498 () 2500 * In Section 7.6.1 of [HTTP], clarify that new connection options 2501 indeed need to be registered () 2504 * In Section 1.1, reference RFC 8174 as well 2505 () 2507 D.9. Since draft-ietf-httpbis-messaging-07 2509 * Move TE: trailers into [HTTP] () 2512 * In Section 6.3, adjust requirements for handling multiple content- 2513 length values () 2515 * Throughout, replace "effective request URI" with "target URI" 2516 () 2518 * In Section 6.1, don't claim Transfer-Encoding is supported by 2519 HTTP/2 or later () 2521 D.10. Since draft-ietf-httpbis-messaging-08 2523 * In Section 2.2, disallow bare CRs () 2526 * Appendix A now uses the sender variant of the "#" list expansion 2527 () 2529 * In Section 5, adjust IANA "Close" entry for new registry format 2530 () 2532 D.11. Since draft-ietf-httpbis-messaging-09 2534 * Switch to xml2rfc v3 mode for draft generation 2535 () 2537 D.12. Since draft-ietf-httpbis-messaging-10 2539 * In Section 6.3, note that TCP half-close does not delimit a 2540 request; talk about corresponding server-side behaviour in 2541 Section 9.6 () 2543 * Moved requirements specific to HTTP/1.1 from [HTTP] into 2544 Section 3.2 () 2546 * In Section 6.1 (Transfer-Encoding), adjust ABNF to allow empty 2547 lists () 2549 * In Section 9.7, add text from RFC 2818 2550 () 2552 * Moved definitions of "TE" and "Upgrade" into [HTTP] 2553 () 2555 * Moved definition of "Connection" into [HTTP] 2556 () 2558 D.13. Since draft-ietf-httpbis-messaging-11 2560 * Move IANA Upgrade Token Registry instructions to [HTTP] 2561 () 2563 D.14. Since draft-ietf-httpbis-messaging-12 2565 * Moved content of history appendix to Semantics 2566 () 2568 * Moved note about "close" being reserved as field name to 2569 Section 9.3 () 2571 * Moved table of transfer codings into Section 12.3 2572 () 2574 * In Section 13.2, updated the URI for the [Linhart] paper 2575 () 2577 * Changed document title to just "HTTP/1.1" 2578 () 2580 * In Section 7, moved transfer-coding ABNF to Section 10.1.4 of 2581 [HTTP] () 2583 * Changed to using "payload data" when defining requirements about 2584 the data being conveyed within a message, instead of the terms 2585 "payload body" or "response body" or "representation body", since 2586 they often get confused with the HTTP/1.1 message body (which 2587 includes transfer coding) () 2590 D.15. Since draft-ietf-httpbis-messaging-13 2592 * In Section 6.3, clarify that a message needs to be checked for 2593 both Content-Length and Transfer-Encoding, before processing 2594 Transfer-Encoding, and that ought to be treated as an error, but 2595 an intermediary can choose to forward the message downstream after 2596 removing the Content-Length and processing the Transfer-Encoding 2597 () 2599 * Changed to using "content" instead of "payload" or "payload data" 2600 to avoid confusion with the payload of version-specific messaging 2601 frames () 2603 D.16. Since draft-ietf-httpbis-messaging-14 2605 * In Section 9.6, define the close connection option, since its 2606 definition was removed from the Connection header field for being 2607 specific to 1.1 () 2609 * In Section 3.3, clarify how the target URI is reconstructed when 2610 the request-target is not in absolute-form and highlight risk in 2611 selecting a default host () 2614 * In Section 7.1, clarify large chunk handling issues 2615 () 2617 * In Section 2.2, explicitly close the connection after sending a 2618 400 () 2620 * In Section 2.3, refine version requirements for intermediaries 2621 () 2623 * In Section 7.1.3, don't remove the Trailer header field 2624 () 2626 * In Section 3.2.3, changed the ABNF definition of authority-form 2627 from the authority component (in which port is optional) to the 2628 host:port format that has always been required by CONNECT 2629 () 2631 D.17. Since draft-ietf-httpbis-messaging-15 2633 * None. 2635 D.18. Since draft-ietf-httpbis-messaging-16 2637 This draft addresses mostly editorial issues raised during or past 2638 IETF Last Call; see for a summary. 2641 Furthermore: 2643 * In Section 6.1, require recipients to avoid smuggling/splitting 2644 attacks when processing an ambiguous message framing 2645 () 2647 D.19. Since draft-ietf-httpbis-messaging-17 2649 * In Section 4, rephrase text about status code definitions in 2650 [HTTP] () 2652 * In Section 9.2, clarify how to match responses to requests 2653 () 2655 * Made reference to [RFC5322] normative, as it is referenced from 2656 the ABNF (for "From" header field) () 2659 * In Section 5.2, include text about message/http that previously 2660 was in [HTTP] () 2662 * Throughout, disambiguate "selected representation" and "selected 2663 response" (now "chosen response") () 2666 Acknowledgements 2668 See Appendix "Acknowledgements" of [HTTP]. 2670 Index 2672 A C D F G H M O R T X 2673 A 2675 absolute-form (of request-target) Section 3.2.2 2676 application/http Media Type Section 10.2 2677 asterisk-form (of request-target) Section 3.2.4 2678 authority-form (of request-target) Section 3.2.3 2680 C 2682 Connection header field Section 9.6 2683 Content-Length header field Section 6.2 2684 Content-Transfer-Encoding header field Appendix B.5 2685 chunked (Coding Format) Section 6.1; Section 6.3 2686 chunked (transfer coding) Section 7.1 2687 close Section 9.3; Section 9.6 2688 compress (transfer coding) Section 7.2 2690 D 2692 deflate (transfer coding) Section 7.2 2694 F 2696 Fields 2697 Close Section 9.6, Paragraph 4 2698 MIME-Version Appendix B.1 2699 Transfer-Encoding Section 6.1 2701 G 2703 Grammar 2704 ALPHA Section 1.2 2705 CR Section 1.2 2706 CRLF Section 1.2 2707 CTL Section 1.2 2708 DIGIT Section 1.2 2709 DQUOTE Section 1.2 2710 HEXDIG Section 1.2 2711 HTAB Section 1.2 2712 HTTP-message Section 2.1 2713 HTTP-name Section 2.3 2714 HTTP-version Section 2.3 2715 LF Section 1.2 2716 OCTET Section 1.2 2717 SP Section 1.2 2718 Transfer-Encoding Section 6.1 2719 VCHAR Section 1.2 2720 absolute-form Section 3.2; Section 3.2.2 2721 asterisk-form Section 3.2; Section 3.2.4 2722 authority-form Section 3.2; Section 3.2.3 2723 chunk Section 7.1 2724 chunk-data Section 7.1 2725 chunk-ext Section 7.1; Section 7.1.1 2726 chunk-ext-name Section 7.1.1 2727 chunk-ext-val Section 7.1.1 2728 chunk-size Section 7.1 2729 chunked-body Section 7.1 2730 field-line Section 5; Section 7.1.2 2731 field-name Section 5 2732 field-value Section 5 2733 last-chunk Section 7.1 2734 message-body Section 6 2735 method Section 3.1 2736 obs-fold Section 5.2 2737 origin-form Section 3.2; Section 3.2.1 2738 reason-phrase Section 4 2739 request-line Section 3 2740 request-target Section 3.2 2741 start-line Section 2.1 2742 status-code Section 4 2743 status-line Section 4 2744 trailer-section Section 7.1; Section 7.1.2 2745 gzip (transfer coding) Section 7.2 2747 H 2749 Header Fields 2750 MIME-Version Appendix B.1 2751 Transfer-Encoding Section 6.1 2752 header line Section 2.1 2753 header section Section 2.1 2754 headers Section 2.1 2756 M 2758 MIME-Version header field Appendix B.1 2759 Media Type 2760 application/http Section 10.2 2761 message/http Section 10.1 2762 message/http Media Type Section 10.1 2763 method Section 3.1 2765 O 2767 origin-form (of request-target) Section 3.2.1 2769 R 2771 request-target Section 3.2 2773 T 2775 Transfer-Encoding header field Section 6.1 2777 X 2779 x-compress (transfer coding) Section 7.2 2780 x-gzip (transfer coding) Section 7.2 2782 Authors' Addresses 2784 Roy T. Fielding (editor) 2785 Adobe 2786 345 Park Ave 2787 San Jose, CA 95110 2788 United States of America 2790 Email: fielding@gbiv.com 2791 URI: https://roy.gbiv.com/ 2793 Mark Nottingham (editor) 2794 Fastly 2795 Prahran VIC 2796 Australia 2798 Email: mnot@mnot.net 2799 URI: https://www.mnot.net/ 2801 Julian Reschke (editor) 2802 greenbytes GmbH 2803 Hafenweg 16 2804 48155 Münster 2805 Germany 2807 Email: julian.reschke@greenbytes.de 2808 URI: https://greenbytes.de/tech/webdav/