idnits 2.17.1 draft-ietf-httpbis-messaging-15.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 (30 March 2021) is 1117 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'RFC7231' is defined on line 2032, but no explicit reference was found in the text == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-cache-15 -- Possible downref: Normative reference to a draft: ref. 'Caching' ** Downref: Normative reference to an Informational RFC: RFC 1950 ** Downref: Normative reference to an Informational RFC: RFC 1951 ** Downref: Normative reference to an Informational RFC: RFC 1952 == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-semantics-15 -- Possible downref: Normative reference to a draft: ref. 'Semantics' -- Possible downref: Non-RFC (?) normative reference: ref. 'USASCII' -- Possible downref: Non-RFC (?) normative reference: ref. 'Welch' -- Obsolete informational reference (is this intentional?): RFC 7230 (ref. 'Err4667') (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 2068 (Obsoleted by RFC 2616) -- Duplicate reference: RFC7230, mentioned in 'RFC7230', was also mentioned in 'Err4667'. -- Obsolete informational reference (is this intentional?): RFC 7230 (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 7231 (Obsoleted by RFC 9110) Summary: 3 errors (**), 0 flaws (~~), 6 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 HTTP Working Group R. Fielding, Ed. 3 Internet-Draft Adobe 4 Obsoletes: 7230 (if approved) M. Nottingham, Ed. 5 Intended status: Standards Track Fastly 6 Expires: 1 October 2021 J. Reschke, Ed. 7 greenbytes 8 30 March 2021 10 HTTP/1.1 11 draft-ietf-httpbis-messaging-15 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.16. 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 1 October 2021. 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 . . . . . . . . . . . . . . . . . . . . 13 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 . . . . . . . . . . . . . . . . . . . . . 19 104 6.3. Message Body Length . . . . . . . . . . . . . . . . . . . 20 105 7. Transfer Codings . . . . . . . . . . . . . . . . . . . . . . 22 106 7.1. Chunked Transfer Coding . . . . . . . . . . . . . . . . . 23 107 7.1.1. Chunk Extensions . . . . . . . . . . . . . . . . . . 23 108 7.1.2. Chunked Trailer Section . . . . . . . . . . . . . . . 24 109 7.1.3. Decoding Chunked . . . . . . . . . . . . . . . . . . 24 110 7.2. Transfer Codings for Compression . . . . . . . . . . . . 25 111 7.3. Transfer Coding Registry . . . . . . . . . . . . . . . . 25 112 7.4. Negotiating Transfer Codings . . . . . . . . . . . . . . 26 113 8. Handling Incomplete Messages . . . . . . . . . . . . . . . . 27 114 9. Connection Management . . . . . . . . . . . . . . . . . . . . 28 115 9.1. Establishment . . . . . . . . . . . . . . . . . . . . . . 28 116 9.2. Associating a Response to a Request . . . . . . . . . . . 28 117 9.3. Persistence . . . . . . . . . . . . . . . . . . . . . . . 29 118 9.3.1. Retrying Requests . . . . . . . . . . . . . . . . . . 30 119 9.3.2. Pipelining . . . . . . . . . . . . . . . . . . . . . 30 120 9.4. Concurrency . . . . . . . . . . . . . . . . . . . . . . . 31 121 9.5. Failures and Timeouts . . . . . . . . . . . . . . . . . . 31 122 9.6. Tear-down . . . . . . . . . . . . . . . . . . . . . . . . 32 123 9.7. TLS Connection Initiation . . . . . . . . . . . . . . . . 34 124 9.8. TLS Connection Closure . . . . . . . . . . . . . . . . . 34 125 10. Enclosing Messages as Data . . . . . . . . . . . . . . . . . 35 126 10.1. Media Type message/http . . . . . . . . . . . . . . . . 35 127 10.2. Media Type application/http . . . . . . . . . . . . . . 36 128 11. Security Considerations . . . . . . . . . . . . . . . . . . . 37 129 11.1. Response Splitting . . . . . . . . . . . . . . . . . . . 37 130 11.2. Request Smuggling . . . . . . . . . . . . . . . . . . . 38 131 11.3. Message Integrity . . . . . . . . . . . . . . . . . . . 39 132 11.4. Message Confidentiality . . . . . . . . . . . . . . . . 39 133 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 40 134 12.1. Field Name Registration . . . . . . . . . . . . . . . . 40 135 12.2. Media Type Registration . . . . . . . . . . . . . . . . 40 136 12.3. Transfer Coding Registration . . . . . . . . . . . . . . 40 137 12.4. ALPN Protocol ID Registration . . . . . . . . . . . . . 41 138 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 42 139 13.1. Normative References . . . . . . . . . . . . . . . . . . 42 140 13.2. Informative References . . . . . . . . . . . . . . . . . 43 141 Appendix A. Collected ABNF . . . . . . . . . . . . . . . . . . . 44 142 Appendix B. Differences between HTTP and MIME . . . . . . . . . 46 143 B.1. MIME-Version . . . . . . . . . . . . . . . . . . . . . . 46 144 B.2. Conversion to Canonical Form . . . . . . . . . . . . . . 46 145 B.3. Conversion of Date Formats . . . . . . . . . . . . . . . 47 146 B.4. Conversion of Content-Encoding . . . . . . . . . . . . . 47 147 B.5. Conversion of Content-Transfer-Encoding . . . . . . . . . 47 148 B.6. MHTML and Line Length Limitations . . . . . . . . . . . . 48 149 Appendix C. Changes from previous RFCs . . . . . . . . . . . . . 48 150 C.1. Changes from HTTP/0.9 . . . . . . . . . . . . . . . . . . 48 151 C.2. Changes from HTTP/1.0 . . . . . . . . . . . . . . . . . . 48 152 C.2.1. Multihomed Web Servers . . . . . . . . . . . . . . . 48 153 C.2.2. Keep-Alive Connections . . . . . . . . . . . . . . . 49 154 C.2.3. Introduction of Transfer-Encoding . . . . . . . . . . 49 155 C.3. Changes from RFC 7230 . . . . . . . . . . . . . . . . . . 49 156 Appendix D. Change Log . . . . . . . . . . . . . . . . . . . . . 50 157 D.1. Between RFC7230 and draft 00 . . . . . . . . . . . . . . 50 158 D.2. Since draft-ietf-httpbis-messaging-00 . . . . . . . . . . 51 159 D.3. Since draft-ietf-httpbis-messaging-01 . . . . . . . . . . 51 160 D.4. Since draft-ietf-httpbis-messaging-02 . . . . . . . . . . 52 161 D.5. Since draft-ietf-httpbis-messaging-03 . . . . . . . . . . 52 162 D.6. Since draft-ietf-httpbis-messaging-04 . . . . . . . . . . 52 163 D.7. Since draft-ietf-httpbis-messaging-05 . . . . . . . . . . 52 164 D.8. Since draft-ietf-httpbis-messaging-06 . . . . . . . . . . 53 165 D.9. Since draft-ietf-httpbis-messaging-07 . . . . . . . . . . 53 166 D.10. Since draft-ietf-httpbis-messaging-08 . . . . . . . . . . 54 167 D.11. Since draft-ietf-httpbis-messaging-09 . . . . . . . . . . 54 168 D.12. Since draft-ietf-httpbis-messaging-10 . . . . . . . . . . 54 169 D.13. Since draft-ietf-httpbis-messaging-11 . . . . . . . . . . 54 170 D.14. Since draft-ietf-httpbis-messaging-12 . . . . . . . . . . 55 171 D.15. Since draft-ietf-httpbis-messaging-13 . . . . . . . . . . 55 172 D.16. Since draft-ietf-httpbis-messaging-14 . . . . . . . . . . 55 173 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 56 174 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 56 176 1. Introduction 178 The Hypertext Transfer Protocol (HTTP) is a stateless application- 179 level request/response protocol that uses extensible semantics and 180 self-descriptive messages for flexible interaction with network-based 181 hypertext information systems. HTTP/1.1 is defined by: 183 * This document 185 * "HTTP Semantics" [Semantics] 187 * "HTTP Caching" [Caching] 188 This document specifies how HTTP semantics are conveyed using the 189 HTTP/1.1 message syntax, framing and connection management 190 mechanisms. Its goal is to define the complete set of requirements 191 for HTTP/1.1 message parsers and message-forwarding intermediaries. 193 This document obsoletes the portions of RFC 7230 related to HTTP/1.1 194 messaging and connection management, with the changes being 195 summarized in Appendix C.3. The other parts of RFC 7230 are 196 obsoleted by "HTTP Semantics" [Semantics]. 198 1.1. Requirements Notation 200 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 201 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 202 "OPTIONAL" in this document are to be interpreted as described in BCP 203 14 [RFC2119] [RFC8174] when, and only when, they appear in all 204 capitals, as shown here. 206 Conformance criteria and considerations regarding error handling are 207 defined in Section 2 of [Semantics]. 209 1.2. Syntax Notation 211 This specification uses the Augmented Backus-Naur Form (ABNF) 212 notation of [RFC5234], extended with the notation for case- 213 sensitivity in strings defined in [RFC7405]. 215 It also uses a list extension, defined in Section 5.6.1 of 216 [Semantics], that allows for compact definition of comma-separated 217 lists using a '#' operator (similar to how the '*' operator indicates 218 repetition). Appendix A shows the collected grammar with all list 219 operators expanded to standard ABNF notation. 221 As a convention, ABNF rule names prefixed with "obs-" denote 222 "obsolete" grammar rules that appear for historical reasons. 224 The following core rules are included by reference, as defined in 225 [RFC5234], Appendix B.1: ALPHA (letters), CR (carriage return), CRLF 226 (CR LF), CTL (controls), DIGIT (decimal 0-9), DQUOTE (double quote), 227 HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line 228 feed), OCTET (any 8-bit sequence of data), SP (space), and VCHAR (any 229 visible [USASCII] character). 231 The rules below are defined in [Semantics]: 233 BWS = 234 OWS = 235 RWS = 236 absolute-path = 237 field-name = 238 field-value = 239 obs-text = 240 quoted-string = 241 token = 242 transfer-coding = 243 245 The rules below are defined in [RFC3986]: 247 absolute-URI = 248 authority = 249 uri-host = 250 port = 251 query = 253 2. Message 255 2.1. Message Format 257 An HTTP/1.1 message consists of a start-line followed by a CRLF and a 258 sequence of octets in a format similar to the Internet Message Format 259 [RFC5322]: zero or more header field lines (collectively referred to 260 as the "headers" or the "header section"), an empty line indicating 261 the end of the header section, and an optional message body. 263 HTTP-message = start-line CRLF 264 *( field-line CRLF ) 265 CRLF 266 [ message-body ] 268 A message can be either a request from client to server or a response 269 from server to client. Syntactically, the two types of message 270 differ only in the start-line, which is either a request-line (for 271 requests) or a status-line (for responses), and in the algorithm for 272 determining the length of the message body (Section 6). 274 start-line = request-line / status-line 276 In theory, a client could receive requests and a server could receive 277 responses, distinguishing them by their different start-line formats. 278 In practice, servers are implemented to only expect a request (a 279 response is interpreted as an unknown or invalid request method) and 280 clients are implemented to only expect a response. 282 Although HTTP makes use of some protocol elements similar to the 283 Multipurpose Internet Mail Extensions (MIME) [RFC2045], see 284 Appendix B for the differences between HTTP and MIME messages. 286 2.2. Message Parsing 288 The normal procedure for parsing an HTTP message is to read the 289 start-line into a structure, read each header field line into a hash 290 table by field name until the empty line, and then use the parsed 291 data to determine if a message body is expected. If a message body 292 has been indicated, then it is read as a stream until an amount of 293 octets equal to the message body length is read or the connection is 294 closed. 296 A recipient MUST parse an HTTP message as a sequence of octets in an 297 encoding that is a superset of US-ASCII [USASCII]. Parsing an HTTP 298 message as a stream of Unicode characters, without regard for the 299 specific encoding, creates security vulnerabilities due to the 300 varying ways that string processing libraries handle invalid 301 multibyte character sequences that contain the octet LF (%x0A). 302 String-based parsers can only be safely used within protocol elements 303 after the element has been extracted from the message, such as within 304 a header field line value after message parsing has delineated the 305 individual field lines. 307 Although the line terminator for the start-line and fields is the 308 sequence CRLF, a recipient MAY recognize a single LF as a line 309 terminator and ignore any preceding CR. 311 A sender MUST NOT generate a bare CR (a CR character not immediately 312 followed by LF) within any protocol elements other than the content. 313 A recipient of such a bare CR MUST consider that element to be 314 invalid or replace each bare CR with SP before processing the element 315 or forwarding the message. 317 Older HTTP/1.0 user agent implementations might send an extra CRLF 318 after a POST request as a workaround for some early server 319 applications that failed to read message body content that was not 320 terminated by a line-ending. An HTTP/1.1 user agent MUST NOT preface 321 or follow a request with an extra CRLF. If terminating the request 322 message body with a line-ending is desired, then the user agent MUST 323 count the terminating CRLF octets as part of the message body length. 325 In the interest of robustness, a server that is expecting to receive 326 and parse a request-line SHOULD ignore at least one empty line (CRLF) 327 received prior to the request-line. 329 A sender MUST NOT send whitespace between the start-line and the 330 first header field. A recipient that receives whitespace between the 331 start-line and the first header field MUST either reject the message 332 as invalid or consume each whitespace-preceded line without further 333 processing of it (i.e., ignore the entire line, along with any 334 subsequent lines preceded by whitespace, until a properly formed 335 header field is received or the header section is terminated). 337 The presence of such whitespace in a request might be an attempt to 338 trick a server into ignoring that field line or processing the line 339 after it as a new request, either of which might result in a security 340 vulnerability if other implementations within the request chain 341 interpret the same message differently. Likewise, the presence of 342 such whitespace in a response might be ignored by some clients or 343 cause others to cease parsing. 345 When a server listening only for HTTP request messages, or processing 346 what appears from the start-line to be an HTTP request message, 347 receives a sequence of octets that does not match the HTTP-message 348 grammar aside from the robustness exceptions listed above, the server 349 SHOULD respond with a 400 (Bad Request) response and close the 350 connection. 352 2.3. HTTP Version 354 HTTP uses a "." numbering scheme to indicate versions 355 of the protocol. This specification defines version "1.1". 356 Section 2.5 of [Semantics] specifies the semantics of HTTP version 357 numbers. 359 The version of an HTTP/1.x message is indicated by an HTTP-version 360 field in the start-line. HTTP-version is case-sensitive. 362 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 363 HTTP-name = %s"HTTP" 365 When an HTTP/1.1 message is sent to an HTTP/1.0 recipient [RFC1945] 366 or a recipient whose version is unknown, the HTTP/1.1 message is 367 constructed such that it can be interpreted as a valid HTTP/1.0 368 message if all of the newer features are ignored. This specification 369 places recipient-version requirements on some new features so that a 370 conformant sender will only use compatible features until it has 371 determined, through configuration or the receipt of a message, that 372 the recipient supports HTTP/1.1. 374 Intermediaries that process HTTP messages (i.e., all intermediaries 375 other than those acting as tunnels) MUST send their own HTTP-version 376 in forwarded messages, unless it is purposefully downgraded as a 377 workaround for an upstream issue. In other words, an intermediary is 378 not allowed to blindly forward the start-line without ensuring that 379 the protocol version in that message matches a version to which that 380 intermediary is conformant for both the receiving and sending of 381 messages. Forwarding an HTTP message without rewriting the HTTP- 382 version might result in communication errors when downstream 383 recipients use the message sender's version to determine what 384 features are safe to use for later communication with that sender. 386 A server MAY send an HTTP/1.0 response to an HTTP/1.1 request if it 387 is known or suspected that the client incorrectly implements the HTTP 388 specification and is incapable of correctly processing later version 389 responses, such as when a client fails to parse the version number 390 correctly or when an intermediary is known to blindly forward the 391 HTTP-version even when it doesn't conform to the given minor version 392 of the protocol. Such protocol downgrades SHOULD NOT be performed 393 unless triggered by specific client attributes, such as when one or 394 more of the request header fields (e.g., User-Agent) uniquely match 395 the values sent by a client known to be in error. 397 3. Request Line 399 A request-line begins with a method token, followed by a single space 400 (SP), the request-target, another single space (SP), and ends with 401 the protocol version. 403 request-line = method SP request-target SP HTTP-version 405 Although the request-line grammar rule requires that each of the 406 component elements be separated by a single SP octet, recipients MAY 407 instead parse on whitespace-delimited word boundaries and, aside from 408 the CRLF terminator, treat any form of whitespace as the SP separator 409 while ignoring preceding or trailing whitespace; such whitespace 410 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 411 (%x0C), or bare CR. However, lenient parsing can result in request 412 smuggling security vulnerabilities if there are multiple recipients 413 of the message and each has its own unique interpretation of 414 robustness (see Section 11.2). 416 HTTP does not place a predefined limit on the length of a request- 417 line, as described in Section 2 of [Semantics]. A server that 418 receives a method longer than any that it implements SHOULD respond 419 with a 501 (Not Implemented) status code. A server that receives a 420 request-target longer than any URI it wishes to parse MUST respond 421 with a 414 (URI Too Long) status code (see Section 15.5.15 of 422 [Semantics]). 424 Various ad hoc limitations on request-line length are found in 425 practice. It is RECOMMENDED that all HTTP senders and recipients 426 support, at a minimum, request-line lengths of 8000 octets. 428 3.1. Method 430 The method token indicates the request method to be performed on the 431 target resource. The request method is case-sensitive. 433 method = token 435 The request methods defined by this specification can be found in 436 Section 9 of [Semantics], along with information regarding the HTTP 437 method registry and considerations for defining new methods. 439 3.2. Request Target 441 The request-target identifies the target resource upon which to apply 442 the request. The client derives a request-target from its desired 443 target URI. There are four distinct formats for the request-target, 444 depending on both the method being requested and whether the request 445 is to a proxy. 447 request-target = origin-form 448 / absolute-form 449 / authority-form 450 / asterisk-form 452 No whitespace is allowed in the request-target. Unfortunately, some 453 user agents fail to properly encode or exclude whitespace found in 454 hypertext references, resulting in those disallowed characters being 455 sent as the request-target in a malformed request-line. 457 Recipients of an invalid request-line SHOULD respond with either a 458 400 (Bad Request) error or a 301 (Moved Permanently) redirect with 459 the request-target properly encoded. A recipient SHOULD NOT attempt 460 to autocorrect and then process the request without a redirect, since 461 the invalid request-line might be deliberately crafted to bypass 462 security filters along the request chain. 464 A client MUST send a Host header field in all HTTP/1.1 request 465 messages. If the target URI includes an authority component, then a 466 client MUST send a field value for Host that is identical to that 467 authority component, excluding any userinfo subcomponent and its "@" 468 delimiter (Section 4.2.1 of [Semantics]). If the authority component 469 is missing or undefined for the target URI, then a client MUST send a 470 Host header field with an empty field value. 472 A server MUST respond with a 400 (Bad Request) status code to any 473 HTTP/1.1 request message that lacks a Host header field and to any 474 request message that contains more than one Host header field line or 475 a Host header field with an invalid field value. 477 3.2.1. origin-form 479 The most common form of request-target is the _origin-form_. 481 origin-form = absolute-path [ "?" query ] 483 When making a request directly to an origin server, other than a 484 CONNECT or server-wide OPTIONS request (as detailed below), a client 485 MUST send only the absolute path and query components of the target 486 URI as the request-target. If the target URI's path component is 487 empty, the client MUST send "/" as the path within the origin-form of 488 request-target. A Host header field is also sent, as defined in 489 Section 7.2 of [Semantics]. 491 For example, a client wishing to retrieve a representation of the 492 resource identified as 494 http://www.example.org/where?q=now 496 directly from the origin server would open (or reuse) a TCP 497 connection to port 80 of the host "www.example.org" and send the 498 lines: 500 GET /where?q=now HTTP/1.1 501 Host: www.example.org 503 followed by the remainder of the request message. 505 3.2.2. absolute-form 507 When making a request to a proxy, other than a CONNECT or server-wide 508 OPTIONS request (as detailed below), a client MUST send the target 509 URI in _absolute-form_ as the request-target. 511 absolute-form = absolute-URI 513 The proxy is requested to either service that request from a valid 514 cache, if possible, or make the same request on the client's behalf 515 to either the next inbound proxy server or directly to the origin 516 server indicated by the request-target. Requirements on such 517 "forwarding" of messages are defined in Section 7.6 of [Semantics]. 519 An example absolute-form of request-line would be: 521 GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1 523 A client MUST send a Host header field in an HTTP/1.1 request even if 524 the request-target is in the absolute-form, since this allows the 525 Host information to be forwarded through ancient HTTP/1.0 proxies 526 that might not have implemented Host. 528 When a proxy receives a request with an absolute-form of request- 529 target, the proxy MUST ignore the received Host header field (if any) 530 and instead replace it with the host information of the request- 531 target. A proxy that forwards such a request MUST generate a new 532 Host field value based on the received request-target rather than 533 forward the received Host field value. 535 When an origin server receives a request with an absolute-form of 536 request-target, the origin server MUST ignore the received Host 537 header field (if any) and instead use the host information of the 538 request-target. Note that if the request-target does not have an 539 authority component, an empty Host header field will be sent in this 540 case. 542 To allow for transition to the absolute-form for all requests in some 543 future version of HTTP, a server MUST accept the absolute-form in 544 requests, even though HTTP/1.1 clients will only send them in 545 requests to proxies. 547 3.2.3. authority-form 549 The _authority-form_ of request-target is only used for CONNECT 550 requests (Section 9.3.6 of [Semantics]). It consists of only the 551 uri-host and port number of the tunnel destination, separated by a 552 colon (":"). 554 authority-form = uri-host ":" port 556 When making a CONNECT request to establish a tunnel through one or 557 more proxies, a client MUST send only the host and port of the tunnel 558 destination as the request-target. The client obtains the host and 559 port from the target URI's authority component, except that it sends 560 the scheme's default port if the target URI elides the port. For 561 example, a CONNECT request to "http://www.example.com" looks like 563 CONNECT www.example.com:80 HTTP/1.1 564 Host: www.example.com 566 3.2.4. asterisk-form 568 The _asterisk-form_ of request-target is only used for a server-wide 569 OPTIONS request (Section 9.3.7 of [Semantics]). 571 asterisk-form = "*" 573 When a client wishes to request OPTIONS for the server as a whole, as 574 opposed to a specific named resource of that server, the client MUST 575 send only "*" (%x2A) as the request-target. For example, 577 OPTIONS * HTTP/1.1 579 If a proxy receives an OPTIONS request with an absolute-form of 580 request-target in which the URI has an empty path and no query 581 component, then the last proxy on the request chain MUST send a 582 request-target of "*" when it forwards the request to the indicated 583 origin server. 585 For example, the request 587 OPTIONS http://www.example.org:8001 HTTP/1.1 589 would be forwarded by the final proxy as 591 OPTIONS * HTTP/1.1 592 Host: www.example.org:8001 594 after connecting to port 8001 of host "www.example.org". 596 3.3. Reconstructing the Target URI 598 The target URI is the request-target when the request-target is in 599 absolute-form. In that case, a server will parse the URI into its 600 generic components for further evaluation. 602 Otherwise, the server reconstructs the target URI from the connection 603 context and various parts of the request message in order to identify 604 the target resource (Section 7.1 of [Semantics]): 606 * If the server's configuration provides for a fixed URI scheme, or 607 a scheme is provided by a trusted outbound gateway, that scheme is 608 used for the target URI. This is common in large-scale 609 deployments because a gateway server will receive the client's 610 connection context and replace that with their own connection to 611 the inbound server. Otherwise, if the request is received over a 612 secured connection, the target URI's scheme is "https"; if not, 613 the scheme is "http". 615 * If the request-target is in authority-form, the target URI's 616 authority component is the request-target. Otherwise, the target 617 URI's authority component is the field value of the Host header 618 field. If there is no Host header field or if its field value is 619 empty or invalid, the target URI's authority component is empty. 621 * If the request-target is in authority-form or asterisk-form, the 622 target URI's combined path and query component is empty. 623 Otherwise, the target URI's combined path and query component is 624 the request-target. 626 * The components of a reconstructed target URI, once determined as 627 above, can be recombined into absolute-URI form by concatenating 628 the scheme, "://", authority, and combined path and query 629 component. 631 Example 1: the following message received over a secure connection 633 GET /pub/WWW/TheProject.html HTTP/1.1 634 Host: www.example.org 636 has a target URI of 638 https://www.example.org/pub/WWW/TheProject.html 640 Example 2: the following message received over an insecure connection 642 OPTIONS * HTTP/1.1 643 Host: www.example.org:8080 645 has a target URI of 647 http://www.example.org:8080 649 If the target URI's authority component is empty and its URI scheme 650 requires a non-empty authority (as is the case for "http" and 651 "https"), the server can reject the request or determine whether a 652 configured default applies that is consistent with the incoming 653 connection's context. Context might include connection details like 654 address and port, what security has been applied, and locally-defined 655 information specific to that server's configuration. An empty 656 authority is replaced with the configured default before further 657 processing of the request. 659 Supplying a default name for authority within the context of a 660 secured connection is inherently unsafe if there is any chance that 661 the user agent's intended authority might differ from the selected 662 default. A server that can uniquely identify an authority from the 663 request context MAY use that identity as a default without this risk. 664 Alternatively, it might be better to redirect the request to a safe 665 resource that explains how to obtain a new client. 667 Note that reconstructing the client's target URI is only half of the 668 process for identifying a target resource. The other half is 669 determining whether that target URI identifies a resource for which 670 the server is willing and able to send a response, as defined in 671 Section 7.4 of [Semantics]. 673 4. Status Line 675 The first line of a response message is the status-line, consisting 676 of the protocol version, a space (SP), the status code, another 677 space, and ending with an OPTIONAL textual phrase describing the 678 status code. 680 status-line = HTTP-version SP status-code SP [reason-phrase] 682 Although the status-line grammar rule requires that each of the 683 component elements be separated by a single SP octet, recipients MAY 684 instead parse on whitespace-delimited word boundaries and, aside from 685 the line terminator, treat any form of whitespace as the SP separator 686 while ignoring preceding or trailing whitespace; such whitespace 687 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 688 (%x0C), or bare CR. However, lenient parsing can result in response 689 splitting security vulnerabilities if there are multiple recipients 690 of the message and each has its own unique interpretation of 691 robustness (see Section 11.1). 693 The status-code element is a 3-digit integer code describing the 694 result of the server's attempt to understand and satisfy the client's 695 corresponding request. The rest of the response message is to be 696 interpreted in light of the semantics defined for that status code. 697 See Section 15 of [Semantics] for information about the semantics of 698 status codes, including the classes of status code (indicated by the 699 first digit), the status codes defined by this specification, 700 considerations for the definition of new status codes, and the IANA 701 registry. 703 status-code = 3DIGIT 705 The reason-phrase element exists for the sole purpose of providing a 706 textual description associated with the numeric status code, mostly 707 out of deference to earlier Internet application protocols that were 708 more frequently used with interactive text clients. 710 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 712 A client SHOULD ignore the reason-phrase content because it is not a 713 reliable channel for information (it might be translated for a given 714 locale, overwritten by intermediaries, or discarded when the message 715 is forwarded via other versions of HTTP). A server MUST send the 716 space that separates status-code from the reason-phrase even when the 717 reason-phrase is absent (i.e., the status-line would end with the 718 three octets SP CR LF). 720 5. Field Syntax 722 Each field line consists of a case-insensitive field name followed by 723 a colon (":"), optional leading whitespace, the field line value, and 724 optional trailing whitespace. 726 field-line = field-name ":" OWS field-value OWS 728 Most HTTP field names and the rules for parsing within field values 729 are defined in Section 6.3 of [Semantics]. This section covers the 730 generic syntax for header field inclusion within, and extraction 731 from, HTTP/1.1 messages. 733 5.1. Field Line Parsing 735 Messages are parsed using a generic algorithm, independent of the 736 individual field names. The contents within a given field line value 737 are not parsed until a later stage of message interpretation (usually 738 after the message's entire field section has been processed). 740 No whitespace is allowed between the field name and colon. In the 741 past, differences in the handling of such whitespace have led to 742 security vulnerabilities in request routing and response handling. A 743 server MUST reject any received request message that contains 744 whitespace between a header field name and colon with a response 745 status code of 400 (Bad Request). A proxy MUST remove any such 746 whitespace from a response message before forwarding the message 747 downstream. 749 A field line value might be preceded and/or followed by optional 750 whitespace (OWS); a single SP preceding the field line value is 751 preferred for consistent readability by humans. The field line value 752 does not include any leading or trailing whitespace: OWS occurring 753 before the first non-whitespace octet of the field line value or 754 after the last non-whitespace octet of the field line value ought to 755 be excluded by parsers when extracting the field line value from a 756 field line. 758 5.2. Obsolete Line Folding 760 Historically, HTTP field line values could be extended over multiple 761 lines by preceding each extra line with at least one space or 762 horizontal tab (obs-fold). This specification deprecates such line 763 folding except within the message/http media type (Section 10.1). 765 obs-fold = OWS CRLF RWS 766 ; obsolete line folding 768 A sender MUST NOT generate a message that includes line folding 769 (i.e., that has any field line value that contains a match to the 770 obs-fold rule) unless the message is intended for packaging within 771 the message/http media type. 773 A server that receives an obs-fold in a request message that is not 774 within a message/http container MUST either reject the message by 775 sending a 400 (Bad Request), preferably with a representation 776 explaining that obsolete line folding is unacceptable, or replace 777 each received obs-fold with one or more SP octets prior to 778 interpreting the field value or forwarding the message downstream. 780 A proxy or gateway that receives an obs-fold in a response message 781 that is not within a message/http container MUST either discard the 782 message and replace it with a 502 (Bad Gateway) response, preferably 783 with a representation explaining that unacceptable line folding was 784 received, or replace each received obs-fold with one or more SP 785 octets prior to interpreting the field value or forwarding the 786 message downstream. 788 A user agent that receives an obs-fold in a response message that is 789 not within a message/http container MUST replace each received 790 obs-fold with one or more SP octets prior to interpreting the field 791 value. 793 6. Message Body 795 The message body (if any) of an HTTP/1.1 message is used to carry 796 content (Section 6.4 of [Semantics]) for the request or response. 797 The message body is identical to the content unless a transfer coding 798 has been applied, as described in Section 6.1. 800 message-body = *OCTET 802 The rules for determining when a message body is present in an 803 HTTP/1.1 message differ for requests and responses. 805 The presence of a message body in a request is signaled by a 806 Content-Length or Transfer-Encoding header field. Request message 807 framing is independent of method semantics. 809 The presence of a message body in a response depends on both the 810 request method to which it is responding and the response status code 811 (Section 4), and corresponds to when content is allowed; see 812 Section 6.4 of [Semantics]. 814 6.1. Transfer-Encoding 816 The Transfer-Encoding header field lists the transfer coding names 817 corresponding to the sequence of transfer codings that have been (or 818 will be) applied to the content in order to form the message body. 819 Transfer codings are defined in Section 7. 821 Transfer-Encoding = #transfer-coding 822 ; defined in [Semantics], Section 10.1.4 824 Transfer-Encoding is analogous to the Content-Transfer-Encoding field 825 of MIME, which was designed to enable safe transport of binary data 826 over a 7-bit transport service ([RFC2045], Section 6). However, safe 827 transport has a different focus for an 8bit-clean transfer protocol. 828 In HTTP's case, Transfer-Encoding is primarily intended to accurately 829 delimit dynamically generated content and to distinguish encodings 830 that are only applied for transport efficiency or security from those 831 that are characteristics of the selected resource. 833 A recipient MUST be able to parse the chunked transfer coding 834 (Section 7.1) because it plays a crucial role in framing messages 835 when the content size is not known in advance. A sender MUST NOT 836 apply chunked more than once to a message body (i.e., chunking an 837 already chunked message is not allowed). If any transfer coding 838 other than chunked is applied to a request's content, the sender MUST 839 apply chunked as the final transfer coding to ensure that the message 840 is properly framed. If any transfer coding other than chunked is 841 applied to a response's content, the sender MUST either apply chunked 842 as the final transfer coding or terminate the message by closing the 843 connection. 845 For example, 847 Transfer-Encoding: gzip, chunked 849 indicates that the content has been compressed using the gzip coding 850 and then chunked using the chunked coding while forming the message 851 body. 853 Unlike Content-Encoding (Section 8.4.1 of [Semantics]), Transfer- 854 Encoding is a property of the message, not of the representation, and 855 any recipient along the request/response chain MAY decode the 856 received transfer coding(s) or apply additional transfer coding(s) to 857 the message body, assuming that corresponding changes are made to the 858 Transfer-Encoding field value. Additional information about the 859 encoding parameters can be provided by other header fields not 860 defined by this specification. 862 Transfer-Encoding MAY be sent in a response to a HEAD request or in a 863 304 (Not Modified) response (Section 15.4.5 of [Semantics]) to a GET 864 request, neither of which includes a message body, to indicate that 865 the origin server would have applied a transfer coding to the message 866 body if the request had been an unconditional GET. This indication 867 is not required, however, because any recipient on the response chain 868 (including the origin server) can remove transfer codings when they 869 are not needed. 871 A server MUST NOT send a Transfer-Encoding header field in any 872 response with a status code of 1xx (Informational) or 204 (No 873 Content). A server MUST NOT send a Transfer-Encoding header field in 874 any 2xx (Successful) response to a CONNECT request (Section 9.3.6 of 875 [Semantics]). 877 Transfer-Encoding was added in HTTP/1.1. It is generally assumed 878 that implementations advertising only HTTP/1.0 support will not 879 understand how to process transfer-encoded content. A client MUST 880 NOT send a request containing Transfer-Encoding unless it knows the 881 server will handle HTTP/1.1 requests (or later minor revisions); such 882 knowledge might be in the form of specific user configuration or by 883 remembering the version of a prior received response. A server MUST 884 NOT send a response containing Transfer-Encoding unless the 885 corresponding request indicates HTTP/1.1 (or later minor revisions). 887 A server that receives a request message with a transfer coding it 888 does not understand SHOULD respond with 501 (Not Implemented). 890 6.2. Content-Length 892 When a message does not have a Transfer-Encoding header field, a 893 Content-Length header field (Section 8.6 of [Semantics]) can provide 894 the anticipated size, as a decimal number of octets, for potential 895 content. For messages that do include content, the Content-Length 896 field value provides the framing information necessary for 897 determining where the data (and message) ends. For messages that do 898 not include content, the Content-Length indicates the size of the 899 selected representation (Section 8.6 of [Semantics]). 901 A sender MUST NOT send a Content-Length header field in any message 902 that contains a Transfer-Encoding header field. 904 | *Note:* HTTP's use of Content-Length for message framing 905 | differs significantly from the same field's use in MIME, where 906 | it is an optional field used only within the "message/external- 907 | body" media-type. 909 6.3. Message Body Length 911 The length of a message body is determined by one of the following 912 (in order of precedence): 914 1. Any response to a HEAD request and any response with a 1xx 915 (Informational), 204 (No Content), or 304 (Not Modified) status 916 code is always terminated by the first empty line after the 917 header fields, regardless of the header fields present in the 918 message, and thus cannot contain a message body or trailer 919 section. 921 2. Any 2xx (Successful) response to a CONNECT request implies that 922 the connection will become a tunnel immediately after the empty 923 line that concludes the header fields. A client MUST ignore any 924 Content-Length or Transfer-Encoding header fields received in 925 such a message. 927 3. If a message is received with both a Transfer-Encoding and a 928 Content-Length header field, the Transfer-Encoding overrides the 929 Content-Length. Such a message might indicate an attempt to 930 perform request smuggling (Section 11.2) or response splitting 931 (Section 11.1) and ought to be handled as an error. An 932 intermediary that chooses to forward the message MUST first 933 remove the received Content-Length field and process the 934 Transfer-Encoding (as described below) prior to forwarding the 935 message downstream. 937 4. If a Transfer-Encoding header field is present and the chunked 938 transfer coding (Section 7.1) is the final encoding, the message 939 body length is determined by reading and decoding the chunked 940 data until the transfer coding indicates the data is complete. 942 If a Transfer-Encoding header field is present in a response and 943 the chunked transfer coding is not the final encoding, the 944 message body length is determined by reading the connection until 945 it is closed by the server. 947 If a Transfer-Encoding header field is present in a request and 948 the chunked transfer coding is not the final encoding, the 949 message body length cannot be determined reliably; the server 950 MUST respond with the 400 (Bad Request) status code and then 951 close the connection. 953 5. If a message is received without Transfer-Encoding and with an 954 invalid Content-Length header field, then the message framing is 955 invalid and the recipient MUST treat it as an unrecoverable 956 error, unless the field value can be successfully parsed as a 957 comma-separated list (Section 5.6.1 of [Semantics]), all values 958 in the list are valid, and all values in the list are the same. 959 If this is a request message, the server MUST respond with a 400 960 (Bad Request) status code and then close the connection. If this 961 is a response message received by a proxy, the proxy MUST close 962 the connection to the server, discard the received response, and 963 send a 502 (Bad Gateway) response to the client. If this is a 964 response message received by a user agent, the user agent MUST 965 close the connection to the server and discard the received 966 response. 968 6. If a valid Content-Length header field is present without 969 Transfer-Encoding, its decimal value defines the expected message 970 body length in octets. If the sender closes the connection or 971 the recipient times out before the indicated number of octets are 972 received, the recipient MUST consider the message to be 973 incomplete and close the connection. 975 7. If this is a request message and none of the above are true, then 976 the message body length is zero (no message body is present). 978 8. Otherwise, this is a response message without a declared message 979 body length, so the message body length is determined by the 980 number of octets received prior to the server closing the 981 connection. 983 Since there is no way to distinguish a successfully completed, close- 984 delimited response message from a partially received message 985 interrupted by network failure, a server SHOULD generate encoding or 986 length-delimited messages whenever possible. The close-delimiting 987 feature exists primarily for backwards compatibility with HTTP/1.0. 989 | *Note:* Request messages are never close-delimited because they 990 | are always explicitly framed by length or transfer coding, with 991 | the absence of both implying the request ends immediately after 992 | the header section. 994 A server MAY reject a request that contains a message body but not a 995 Content-Length by responding with 411 (Length Required). 997 Unless a transfer coding other than chunked has been applied, a 998 client that sends a request containing a message body SHOULD use a 999 valid Content-Length header field if the message body length is known 1000 in advance, rather than the chunked transfer coding, since some 1001 existing services respond to chunked with a 411 (Length Required) 1002 status code even though they understand the chunked transfer coding. 1003 This is typically because such services are implemented via a gateway 1004 that requires a content-length in advance of being called and the 1005 server is unable or unwilling to buffer the entire request before 1006 processing. 1008 A user agent that sends a request that contains a message body MUST 1009 send either a valid Content-Length header field or use the chunked 1010 transfer coding. A client MUST NOT use the chunked transfer encoding 1011 unless it knows the server will handle HTTP/1.1 (or later) requests; 1012 such knowledge can be in the form of specific user configuration or 1013 by remembering the version of a prior received response. 1015 If the final response to the last request on a connection has been 1016 completely received and there remains additional data to read, a user 1017 agent MAY discard the remaining data or attempt to determine if that 1018 data belongs as part of the prior message body, which might be the 1019 case if the prior message's Content-Length value is incorrect. A 1020 client MUST NOT process, cache, or forward such extra data as a 1021 separate response, since such behavior would be vulnerable to cache 1022 poisoning. 1024 7. Transfer Codings 1026 Transfer coding names are used to indicate an encoding transformation 1027 that has been, can be, or might need to be applied to a message's 1028 content in order to ensure "safe transport" through the network. 1029 This differs from a content coding in that the transfer coding is a 1030 property of the message rather than a property of the representation 1031 that is being transferred. 1033 All transfer-coding names are case-insensitive and ought to be 1034 registered within the HTTP Transfer Coding registry, as defined in 1035 Section 7.3. They are used in the Transfer-Encoding (Section 6.1) 1036 and TE (Section 10.1.4 of [Semantics]) header fields (the latter also 1037 defining the "transfer-coding" grammar). 1039 7.1. Chunked Transfer Coding 1041 The chunked transfer coding wraps content in order to transfer it as 1042 a series of chunks, each with its own size indicator, followed by an 1043 OPTIONAL trailer section containing trailer fields. Chunked enables 1044 content streams of unknown size to be transferred as a sequence of 1045 length-delimited buffers, which enables the sender to retain 1046 connection persistence and the recipient to know when it has received 1047 the entire message. 1049 chunked-body = *chunk 1050 last-chunk 1051 trailer-section 1052 CRLF 1054 chunk = chunk-size [ chunk-ext ] CRLF 1055 chunk-data CRLF 1056 chunk-size = 1*HEXDIG 1057 last-chunk = 1*("0") [ chunk-ext ] CRLF 1059 chunk-data = 1*OCTET ; a sequence of chunk-size octets 1061 The chunk-size field is a string of hex digits indicating the size of 1062 the chunk-data in octets. The chunked transfer coding is complete 1063 when a chunk with a chunk-size of zero is received, possibly followed 1064 by a trailer section, and finally terminated by an empty line. 1066 A recipient MUST be able to parse and decode the chunked transfer 1067 coding. 1069 HTTP/1.1 does not define any means to limit the size of a chunked 1070 response such that an intermediary can be assured of buffering the 1071 entire response. Additionally, very large chunk sizes may cause 1072 overflows or loss of precision if their values are not represented 1073 accurately in a receiving implementation. Therefore, recipients MUST 1074 anticipate potentially large decimal numerals and prevent parsing 1075 errors due to integer conversion overflows or precision loss due to 1076 integer representation. 1078 The chunked encoding does not define any parameters. Their presence 1079 SHOULD be treated as an error. 1081 7.1.1. Chunk Extensions 1083 The chunked encoding allows each chunk to include zero or more chunk 1084 extensions, immediately following the chunk-size, for the sake of 1085 supplying per-chunk metadata (such as a signature or hash), mid- 1086 message control information, or randomization of message body size. 1088 chunk-ext = *( BWS ";" BWS chunk-ext-name 1089 [ BWS "=" BWS chunk-ext-val ] ) 1091 chunk-ext-name = token 1092 chunk-ext-val = token / quoted-string 1094 The chunked encoding is specific to each connection and is likely to 1095 be removed or recoded by each recipient (including intermediaries) 1096 before any higher-level application would have a chance to inspect 1097 the extensions. Hence, use of chunk extensions is generally limited 1098 to specialized HTTP services such as "long polling" (where client and 1099 server can have shared expectations regarding the use of chunk 1100 extensions) or for padding within an end-to-end secured connection. 1102 A recipient MUST ignore unrecognized chunk extensions. A server 1103 ought to limit the total length of chunk extensions received in a 1104 request to an amount reasonable for the services provided, in the 1105 same way that it applies length limitations and timeouts for other 1106 parts of a message, and generate an appropriate 4xx (Client Error) 1107 response if that amount is exceeded. 1109 7.1.2. Chunked Trailer Section 1111 A trailer section allows the sender to include additional fields at 1112 the end of a chunked message in order to supply metadata that might 1113 be dynamically generated while the content is sent, such as a message 1114 integrity check, digital signature, or post-processing status. The 1115 proper use and limitations of trailer fields are defined in 1116 Section 6.5 of [Semantics]. 1118 trailer-section = *( field-line CRLF ) 1120 A recipient that decodes and removes the chunked encoding from a 1121 message (e.g., for storage or forwarding to a non-HTTP/1.1 peer) MUST 1122 discard any received trailer fields, store/forward them separately 1123 from the header fields, or selectively merge into the header section 1124 only those trailer fields corresponding to header field definitions 1125 that are understood by the recipient to explicitly permit and define 1126 how their corresponding trailer field value can be safely merged. 1128 7.1.3. Decoding Chunked 1130 A process for decoding the chunked transfer coding can be represented 1131 in pseudo-code as: 1133 length := 0 1134 read chunk-size, chunk-ext (if any), and CRLF 1135 while (chunk-size > 0) { 1136 read chunk-data and CRLF 1137 append chunk-data to content 1138 length := length + chunk-size 1139 read chunk-size, chunk-ext (if any), and CRLF 1140 } 1141 read trailer field 1142 while (trailer field is not empty) { 1143 if (trailer fields are stored/forwarded separately) { 1144 append trailer field to existing trailer fields 1145 } 1146 else if (trailer field is understood and defined as mergeable) { 1147 merge trailer field with existing header fields 1148 } 1149 else { 1150 discard trailer field 1151 } 1152 read trailer field 1153 } 1154 Content-Length := length 1155 Remove "chunked" from Transfer-Encoding 1157 7.2. Transfer Codings for Compression 1159 The following transfer coding names for compression are defined by 1160 the same algorithm as their corresponding content coding: 1162 compress (and x-compress) 1163 See Section 8.4.1.1 of [Semantics]. 1165 deflate 1166 See Section 8.4.1.2 of [Semantics]. 1168 gzip (and x-gzip) 1169 See Section 8.4.1.3 of [Semantics]. 1171 The compression codings do not define any parameters. Their presence 1172 SHOULD be treated as an error. 1174 7.3. Transfer Coding Registry 1176 The "HTTP Transfer Coding Registry" defines the namespace for 1177 transfer coding names. It is maintained at 1178 . 1180 Registrations MUST include the following fields: 1182 * Name 1184 * Description 1186 * Pointer to specification text 1188 Names of transfer codings MUST NOT overlap with names of content 1189 codings (Section 8.4.1 of [Semantics]) unless the encoding 1190 transformation is identical, as is the case for the compression 1191 codings defined in Section 7.2. 1193 The TE header field (Section 10.1.4 of [Semantics]) uses a pseudo 1194 parameter named "q" as rank value when multiple transfer codings are 1195 acceptable. Future registrations of transfer codings SHOULD NOT 1196 define parameters called "q" (case-insensitively) in order to avoid 1197 ambiguities. 1199 Values to be added to this namespace require IETF Review (see 1200 Section 4.8 of [RFC8126]), and MUST conform to the purpose of 1201 transfer coding defined in this specification. 1203 Use of program names for the identification of encoding formats is 1204 not desirable and is discouraged for future encodings. 1206 7.4. Negotiating Transfer Codings 1208 The TE field (Section 10.1.4 of [Semantics]) is used in HTTP/1.1 to 1209 indicate what transfer-codings, besides chunked, the client is 1210 willing to accept in the response, and whether or not the client is 1211 willing to preserve trailer fields in a chunked transfer coding. 1213 A client MUST NOT send the chunked transfer coding name in TE; 1214 chunked is always acceptable for HTTP/1.1 recipients. 1216 Three examples of TE use are below. 1218 TE: deflate 1219 TE: 1220 TE: trailers, deflate;q=0.5 1222 When multiple transfer codings are acceptable, the client MAY rank 1223 the codings by preference using a case-insensitive "q" parameter 1224 (similar to the qvalues used in content negotiation fields, 1225 Section 12.4.2 of [Semantics]). The rank value is a real number in 1226 the range 0 through 1, where 0.001 is the least preferred and 1 is 1227 the most preferred; a value of 0 means "not acceptable". 1229 If the TE field value is empty or if no TE field is present, the only 1230 acceptable transfer coding is chunked. A message with no transfer 1231 coding is always acceptable. 1233 The keyword "trailers" indicates that the sender will not discard 1234 trailer fields, as described in Section 6.5 of [Semantics]. 1236 Since the TE header field only applies to the immediate connection, a 1237 sender of TE MUST also send a "TE" connection option within the 1238 Connection header field (Section 7.6.1 of [Semantics]) in order to 1239 prevent the TE header field from being forwarded by intermediaries 1240 that do not support its semantics. 1242 8. Handling Incomplete Messages 1244 A server that receives an incomplete request message, usually due to 1245 a canceled request or a triggered timeout exception, MAY send an 1246 error response prior to closing the connection. 1248 A client that receives an incomplete response message, which can 1249 occur when a connection is closed prematurely or when decoding a 1250 supposedly chunked transfer coding fails, MUST record the message as 1251 incomplete. Cache requirements for incomplete responses are defined 1252 in Section 3 of [Caching]. 1254 If a response terminates in the middle of the header section (before 1255 the empty line is received) and the status code might rely on header 1256 fields to convey the full meaning of the response, then the client 1257 cannot assume that meaning has been conveyed; the client might need 1258 to repeat the request in order to determine what action to take next. 1260 A message body that uses the chunked transfer coding is incomplete if 1261 the zero-sized chunk that terminates the encoding has not been 1262 received. A message that uses a valid Content-Length is incomplete 1263 if the size of the message body received (in octets) is less than the 1264 value given by Content-Length. A response that has neither chunked 1265 transfer coding nor Content-Length is terminated by closure of the 1266 connection and, if the header section was received intact, is 1267 considered complete unless an error was indicated by the underlying 1268 connection (e.g., an "incomplete close" in TLS would leave the 1269 response incomplete, as described in Section 9.8). 1271 9. Connection Management 1273 HTTP messaging is independent of the underlying transport- or 1274 session-layer connection protocol(s). HTTP only presumes a reliable 1275 transport with in-order delivery of requests and the corresponding 1276 in-order delivery of responses. The mapping of HTTP request and 1277 response structures onto the data units of an underlying transport 1278 protocol is outside the scope of this specification. 1280 As described in Section 7.3 of [Semantics], the specific connection 1281 protocols to be used for an HTTP interaction are determined by client 1282 configuration and the target URI. For example, the "http" URI scheme 1283 (Section 4.2.1 of [Semantics]) indicates a default connection of TCP 1284 over IP, with a default TCP port of 80, but the client might be 1285 configured to use a proxy via some other connection, port, or 1286 protocol. 1288 HTTP implementations are expected to engage in connection management, 1289 which includes maintaining the state of current connections, 1290 establishing a new connection or reusing an existing connection, 1291 processing messages received on a connection, detecting connection 1292 failures, and closing each connection. Most clients maintain 1293 multiple connections in parallel, including more than one connection 1294 per server endpoint. Most servers are designed to maintain thousands 1295 of concurrent connections, while controlling request queues to enable 1296 fair use and detect denial-of-service attacks. 1298 9.1. Establishment 1300 It is beyond the scope of this specification to describe how 1301 connections are established via various transport- or session-layer 1302 protocols. Each connection applies to only one transport link. 1304 9.2. Associating a Response to a Request 1306 HTTP/1.1 does not include a request identifier for associating a 1307 given request message with its corresponding one or more response 1308 messages. Hence, it relies on the order of response arrival to 1309 correspond exactly to the order in which requests are made on the 1310 same connection. More than one response message per request only 1311 occurs when one or more informational responses (1xx, see 1312 Section 15.2 of [Semantics]) precede a final response to the same 1313 request. 1315 A client that has more than one outstanding request on a connection 1316 MUST maintain a list of outstanding requests in the order sent and 1317 MUST associate each received response message on that connection to 1318 the highest ordered request that has not yet received a final (non- 1319 1xx) response. 1321 If an HTTP/1.1 client receives data on a connection that doesn't have 1322 any outstanding requests, it MUST NOT consider them to be a response 1323 to a not-yet-issued request; it SHOULD close the connection, since 1324 message delimitation is now ambiguous, unless the data consists only 1325 of one or more CRLF (which can be discarded, as per Section 2.2). 1327 9.3. Persistence 1329 HTTP/1.1 defaults to the use of _persistent connections_, allowing 1330 multiple requests and responses to be carried over a single 1331 connection. HTTP implementations SHOULD support persistent 1332 connections. 1334 A recipient determines whether a connection is persistent or not 1335 based on the most recently received message's protocol version and 1336 Connection header field (Section 7.6.1 of [Semantics]), if any: 1338 * If the close connection option is present (Section 9.6), the 1339 connection will not persist after the current response; else, 1341 * If the received protocol is HTTP/1.1 (or later), the connection 1342 will persist after the current response; else, 1344 * If the received protocol is HTTP/1.0, the "keep-alive" connection 1345 option is present, either the recipient is not a proxy or the 1346 message is a response, and the recipient wishes to honor the 1347 HTTP/1.0 "keep-alive" mechanism, the connection will persist after 1348 the current response; otherwise, 1350 * The connection will close after the current response. 1352 A client that does not support persistent connections MUST send the 1353 close connection option in every request message. 1355 A server that does not support persistent connections MUST send the 1356 close connection option in every response message that does not have 1357 a 1xx (Informational) status code. 1359 A client MAY send additional requests on a persistent connection 1360 until it sends or receives a close connection option or receives an 1361 HTTP/1.0 response without a "keep-alive" connection option. 1363 In order to remain persistent, all messages on a connection need to 1364 have a self-defined message length (i.e., one not defined by closure 1365 of the connection), as described in Section 6. A server MUST read 1366 the entire request message body or close the connection after sending 1367 its response, since otherwise the remaining data on a persistent 1368 connection would be misinterpreted as the next request. Likewise, a 1369 client MUST read the entire response message body if it intends to 1370 reuse the same connection for a subsequent request. 1372 A proxy server MUST NOT maintain a persistent connection with an 1373 HTTP/1.0 client (see Section 19.7.1 of [RFC2068] for information and 1374 discussion of the problems with the Keep-Alive header field 1375 implemented by many HTTP/1.0 clients). 1377 See Appendix C.2.2 for more information on backwards compatibility 1378 with HTTP/1.0 clients. 1380 9.3.1. Retrying Requests 1382 Connections can be closed at any time, with or without intention. 1383 Implementations ought to anticipate the need to recover from 1384 asynchronous close events. The conditions under which a client can 1385 automatically retry a sequence of outstanding requests are defined in 1386 Section 9.2.2 of [Semantics]. 1388 9.3.2. Pipelining 1390 A client that supports persistent connections MAY _pipeline_ its 1391 requests (i.e., send multiple requests without waiting for each 1392 response). A server MAY process a sequence of pipelined requests in 1393 parallel if they all have safe methods (Section 9.2.1 of 1394 [Semantics]), but it MUST send the corresponding responses in the 1395 same order that the requests were received. 1397 A client that pipelines requests SHOULD retry unanswered requests if 1398 the connection closes before it receives all of the corresponding 1399 responses. When retrying pipelined requests after a failed 1400 connection (a connection not explicitly closed by the server in its 1401 last complete response), a client MUST NOT pipeline immediately after 1402 connection establishment, since the first remaining request in the 1403 prior pipeline might have caused an error response that can be lost 1404 again if multiple requests are sent on a prematurely closed 1405 connection (see the TCP reset problem described in Section 9.6). 1407 Idempotent methods (Section 9.2.2 of [Semantics]) are significant to 1408 pipelining because they can be automatically retried after a 1409 connection failure. A user agent SHOULD NOT pipeline requests after 1410 a non-idempotent method, until the final response status code for 1411 that method has been received, unless the user agent has a means to 1412 detect and recover from partial failure conditions involving the 1413 pipelined sequence. 1415 An intermediary that receives pipelined requests MAY pipeline those 1416 requests when forwarding them inbound, since it can rely on the 1417 outbound user agent(s) to determine what requests can be safely 1418 pipelined. If the inbound connection fails before receiving a 1419 response, the pipelining intermediary MAY attempt to retry a sequence 1420 of requests that have yet to receive a response if the requests all 1421 have idempotent methods; otherwise, the pipelining intermediary 1422 SHOULD forward any received responses and then close the 1423 corresponding outbound connection(s) so that the outbound user 1424 agent(s) can recover accordingly. 1426 9.4. Concurrency 1428 A client ought to limit the number of simultaneous open connections 1429 that it maintains to a given server. 1431 Previous revisions of HTTP gave a specific number of connections as a 1432 ceiling, but this was found to be impractical for many applications. 1433 As a result, this specification does not mandate a particular maximum 1434 number of connections but, instead, encourages clients to be 1435 conservative when opening multiple connections. 1437 Multiple connections are typically used to avoid the "head-of-line 1438 blocking" problem, wherein a request that takes significant server- 1439 side processing and/or transfers very large content would block 1440 subsequent requests on the same connection. However, each connection 1441 consumes server resources. Furthermore, using multiple connections 1442 can cause undesirable side effects in congested networks. 1444 Note that a server might reject traffic that it deems abusive or 1445 characteristic of a denial-of-service attack, such as an excessive 1446 number of open connections from a single client. 1448 9.5. Failures and Timeouts 1450 Servers will usually have some timeout value beyond which they will 1451 no longer maintain an inactive connection. Proxy servers might make 1452 this a higher value since it is likely that the client will be making 1453 more connections through the same proxy server. The use of 1454 persistent connections places no requirements on the length (or 1455 existence) of this timeout for either the client or the server. 1457 A client or server that wishes to time out SHOULD issue a graceful 1458 close on the connection. Implementations SHOULD constantly monitor 1459 open connections for a received closure signal and respond to it as 1460 appropriate, since prompt closure of both sides of a connection 1461 enables allocated system resources to be reclaimed. 1463 A client, server, or proxy MAY close the transport connection at any 1464 time. For example, a client might have started to send a new request 1465 at the same time that the server has decided to close the "idle" 1466 connection. From the server's point of view, the connection is being 1467 closed while it was idle, but from the client's point of view, a 1468 request is in progress. 1470 A server SHOULD sustain persistent connections, when possible, and 1471 allow the underlying transport's flow-control mechanisms to resolve 1472 temporary overloads, rather than terminate connections with the 1473 expectation that clients will retry. The latter technique can 1474 exacerbate network congestion or server load. 1476 A client sending a message body SHOULD monitor the network connection 1477 for an error response while it is transmitting the request. If the 1478 client sees a response that indicates the server does not wish to 1479 receive the message body and is closing the connection, the client 1480 SHOULD immediately cease transmitting the body and close its side of 1481 the connection. 1483 9.6. Tear-down 1485 The "close" connection option is defined as a signal that the sender 1486 will close this connection after completion of the response. A 1487 sender SHOULD send a Connection header field (Section 7.6.1 of 1488 [Semantics]) containing the close connection option when it intends 1489 to close a connection. For example, 1491 Connection: close 1493 as a request header field indicates that this is the last request 1494 that the client will send on this connection, while in a response the 1495 same field indicates that the server is going to close this 1496 connection after the response message is complete. 1498 Note that the field name "Close" is reserved, since using that name 1499 as a header field might conflict with the close connection option. 1501 A client that sends a close connection option MUST NOT send further 1502 requests on that connection (after the one containing the close) and 1503 MUST close the connection after reading the final response message 1504 corresponding to this request. 1506 A server that receives a close connection option MUST initiate 1507 closure of the connection (see below) after it sends the final 1508 response to the request that contained the close connection option. 1509 The server SHOULD send a close connection option in its final 1510 response on that connection. The server MUST NOT process any further 1511 requests received on that connection. 1513 A server that sends a close connection option MUST initiate closure 1514 of the connection (see below) after it sends the response containing 1515 the close connection option. The server MUST NOT process any further 1516 requests received on that connection. 1518 A client that receives a close connection option MUST cease sending 1519 requests on that connection and close the connection after reading 1520 the response message containing the close connection option; if 1521 additional pipelined requests had been sent on the connection, the 1522 client SHOULD NOT assume that they will be processed by the server. 1524 If a server performs an immediate close of a TCP connection, there is 1525 a significant risk that the client will not be able to read the last 1526 HTTP response. If the server receives additional data from the 1527 client on a fully closed connection, such as another request sent by 1528 the client before receiving the server's response, the server's TCP 1529 stack will send a reset packet to the client; unfortunately, the 1530 reset packet might erase the client's unacknowledged input buffers 1531 before they can be read and interpreted by the client's HTTP parser. 1533 To avoid the TCP reset problem, servers typically close a connection 1534 in stages. First, the server performs a half-close by closing only 1535 the write side of the read/write connection. The server then 1536 continues to read from the connection until it receives a 1537 corresponding close by the client, or until the server is reasonably 1538 certain that its own TCP stack has received the client's 1539 acknowledgement of the packet(s) containing the server's last 1540 response. Finally, the server fully closes the connection. 1542 It is unknown whether the reset problem is exclusive to TCP or might 1543 also be found in other transport connection protocols. 1545 Note that a TCP connection that is half-closed by the client does not 1546 delimit a request message, nor does it imply that the client is no 1547 longer interested in a response. In general, transport signals 1548 cannot be relied upon to signal edge cases, since HTTP/1.1 is 1549 independent of transport. 1551 9.7. TLS Connection Initiation 1553 Conceptually, HTTP/TLS is simply sending HTTP messages over a 1554 connection secured via TLS [RFC8446]. 1556 The HTTP client also acts as the TLS client. It initiates a 1557 connection to the server on the appropriate port and sends the TLS 1558 ClientHello to begin the TLS handshake. When the TLS handshake has 1559 finished, the client may then initiate the first HTTP request. All 1560 HTTP data MUST be sent as TLS "application data", but is otherwise 1561 treated like a normal connection for HTTP (including potential reuse 1562 as a persistent connection). 1564 9.8. TLS Connection Closure 1566 TLS provides a facility for secure connection closure. When a valid 1567 closure alert is received, an implementation can be assured that no 1568 further data will be received on that connection. TLS 1569 implementations MUST initiate an exchange of closure alerts before 1570 closing a connection. A TLS implementation MAY, after sending a 1571 closure alert, close the connection without waiting for the peer to 1572 send its closure alert, generating an "incomplete close". This 1573 SHOULD only be done when the application knows (typically through 1574 detecting HTTP message boundaries) that it has sent or received all 1575 the message data that it cares about. 1577 An incomplete close does not call into question the security of the 1578 data already received, but it could indicate that subsequent data 1579 might have been truncated. As TLS is not directly aware of HTTP 1580 message framing, it is necessary to examine the HTTP data itself to 1581 determine whether messages were complete. Handing of incomplete 1582 messages is defined in Section 8. 1584 When encountering an incomplete close, a client SHOULD treat as 1585 completed all requests for which it has received as much data as 1586 specified in the Content-Length header or, when a Transfer-Encoding 1587 of chunked is used, for which the terminal zero-length chunk has been 1588 received. A response that has neither chunked transfer coding nor 1589 Content-Length is complete only if a valid closure alert has been 1590 received. Treating an incomplete message as complete could expose 1591 implementations to attack. 1593 A client detecting an incomplete close SHOULD recover gracefully. 1595 Clients MUST send a closure alert before closing the connection. 1596 Clients that do not expect to receive any more data MAY choose not to 1597 wait for the server's closure alert and simply close the connection, 1598 thus generating an incomplete close on the server side. 1600 Servers SHOULD be prepared to receive an incomplete close from the 1601 client, since the client can often determine when the end of server 1602 data is. 1604 Servers MUST attempt to initiate an exchange of closure alerts with 1605 the client before closing the connection. Servers MAY close the 1606 connection after sending the closure alert, thus generating an 1607 incomplete close on the client side. 1609 10. Enclosing Messages as Data 1611 10.1. Media Type message/http 1613 The message/http media type can be used to enclose a single HTTP 1614 request or response message, provided that it obeys the MIME 1615 restrictions for all "message" types regarding line length and 1616 encodings. 1618 Type name: message 1620 Subtype name: http 1622 Required parameters: N/A 1624 Optional parameters: version, msgtype 1626 version: The HTTP-version number of the enclosed message (e.g., 1627 "1.1"). If not present, the version can be determined from the 1628 first line of the body. 1630 msgtype: The message type - "request" or "response". If not 1631 present, the type can be determined from the first line of the 1632 body. 1634 Encoding considerations: only "7bit", "8bit", or "binary" are 1635 permitted 1637 Security considerations: see Section 11 1639 Interoperability considerations: N/A 1641 Published specification: This specification (see Section 10.1). 1643 Applications that use this media type: N/A 1645 Fragment identifier considerations: N/A 1647 Additional information: Magic number(s): N/A 1648 Deprecated alias names for this type: N/A 1650 File extension(s): N/A 1652 Macintosh file type code(s): N/A 1654 Person and email address to contact for further information: See Aut 1655 hors' Addresses section. 1657 Intended usage: COMMON 1659 Restrictions on usage: N/A 1661 Author: See Authors' Addresses section. 1663 Change controller: IESG 1665 10.2. Media Type application/http 1667 The application/http media type can be used to enclose a pipeline of 1668 one or more HTTP request or response messages (not intermixed). 1670 Type name: application 1672 Subtype name: http 1674 Required parameters: N/A 1676 Optional parameters: version, msgtype 1678 version: The HTTP-version number of the enclosed messages (e.g., 1679 "1.1"). If not present, the version can be determined from the 1680 first line of the body. 1682 msgtype: The message type - "request" or "response". If not 1683 present, the type can be determined from the first line of the 1684 body. 1686 Encoding considerations: HTTP messages enclosed by this type are in 1687 "binary" format; use of an appropriate Content-Transfer-Encoding 1688 is required when transmitted via email. 1690 Security considerations: see Section 11 1692 Interoperability considerations: N/A 1694 Published specification: This specification (see Section 10.2). 1696 Applications that use this media type: N/A 1698 Fragment identifier considerations: N/A 1700 Additional information: Deprecated alias names for this type: N/A 1702 Magic number(s): N/A 1704 File extension(s): N/A 1706 Macintosh file type code(s): N/A 1708 Person and email address to contact for further information: See Aut 1709 hors' Addresses section. 1711 Intended usage: COMMON 1713 Restrictions on usage: N/A 1715 Author: See Authors' Addresses section. 1717 Change controller: IESG 1719 11. Security Considerations 1721 This section is meant to inform developers, information providers, 1722 and users of known security considerations relevant to HTTP message 1723 syntax and parsing. Security considerations about HTTP semantics, 1724 content, and routing are addressed in [Semantics]. 1726 11.1. Response Splitting 1728 Response splitting (a.k.a, CRLF injection) is a common technique, 1729 used in various attacks on Web usage, that exploits the line-based 1730 nature of HTTP message framing and the ordered association of 1731 requests to responses on persistent connections [Klein]. This 1732 technique can be particularly damaging when the requests pass through 1733 a shared cache. 1735 Response splitting exploits a vulnerability in servers (usually 1736 within an application server) where an attacker can send encoded data 1737 within some parameter of the request that is later decoded and echoed 1738 within any of the response header fields of the response. If the 1739 decoded data is crafted to look like the response has ended and a 1740 subsequent response has begun, the response has been split and the 1741 content within the apparent second response is controlled by the 1742 attacker. The attacker can then make any other request on the same 1743 persistent connection and trick the recipients (including 1744 intermediaries) into believing that the second half of the split is 1745 an authoritative answer to the second request. 1747 For example, a parameter within the request-target might be read by 1748 an application server and reused within a redirect, resulting in the 1749 same parameter being echoed in the Location header field of the 1750 response. If the parameter is decoded by the application and not 1751 properly encoded when placed in the response field, the attacker can 1752 send encoded CRLF octets and other content that will make the 1753 application's single response look like two or more responses. 1755 A common defense against response splitting is to filter requests for 1756 data that looks like encoded CR and LF (e.g., "%0D" and "%0A"). 1757 However, that assumes the application server is only performing URI 1758 decoding, rather than more obscure data transformations like charset 1759 transcoding, XML entity translation, base64 decoding, sprintf 1760 reformatting, etc. A more effective mitigation is to prevent 1761 anything other than the server's core protocol libraries from sending 1762 a CR or LF within the header section, which means restricting the 1763 output of header fields to APIs that filter for bad octets and not 1764 allowing application servers to write directly to the protocol 1765 stream. 1767 11.2. Request Smuggling 1769 Request smuggling ([Linhart]) is a technique that exploits 1770 differences in protocol parsing among various recipients to hide 1771 additional requests (which might otherwise be blocked or disabled by 1772 policy) within an apparently harmless request. Like response 1773 splitting, request smuggling can lead to a variety of attacks on HTTP 1774 usage. 1776 This specification has introduced new requirements on request 1777 parsing, particularly with regard to message framing in Section 6.3, 1778 to reduce the effectiveness of request smuggling. 1780 11.3. Message Integrity 1782 HTTP does not define a specific mechanism for ensuring message 1783 integrity, instead relying on the error-detection ability of 1784 underlying transport protocols and the use of length or chunk- 1785 delimited framing to detect completeness. Historically, the lack of 1786 a single integrity mechanism has been justified by the informal 1787 nature of most HTTP communication. However, the prevalence of HTTP 1788 as an information access mechanism has resulted in its increasing use 1789 within environments where verification of message integrity is 1790 crucial. 1792 The mechanisms provided with the "https" scheme, such as 1793 authenticated encryption, provide protection against modification of 1794 messages. Care is needed however to ensure that connection closure 1795 cannot be used to truncate messages (see Section 9.8). User agents 1796 might refuse to accept incomplete messages or treat them specially. 1797 For example, a browser being used to view medical history or drug 1798 interaction information needs to indicate to the user when such 1799 information is detected by the protocol to be incomplete, expired, or 1800 corrupted during transfer. Such mechanisms might be selectively 1801 enabled via user agent extensions or the presence of message 1802 integrity metadata in a response. 1804 The "http" scheme provides no protection against accidental or 1805 malicious modification of messages. 1807 Extensions to the protocol might be used to mitigate the risk of 1808 unwanted modification of messages by intermediaries, even when the 1809 "https" scheme is used. Integrity might be assured by using hash 1810 functions or digital signatures that are selectively added to 1811 messages via extensible metadata fields. 1813 11.4. Message Confidentiality 1815 HTTP relies on underlying transport protocols to provide message 1816 confidentiality when that is desired. HTTP has been specifically 1817 designed to be independent of the transport protocol, such that it 1818 can be used over many different forms of encrypted connection, with 1819 the selection of such transports being identified by the choice of 1820 URI scheme or within user agent configuration. 1822 The "https" scheme can be used to identify resources that require a 1823 confidential connection, as described in Section 4.2.2 of 1824 [Semantics]. 1826 12. IANA Considerations 1828 The change controller for the following registrations is: "IETF 1829 (iesg@ietf.org) - Internet Engineering Task Force". 1831 12.1. Field Name Registration 1833 First, introduce the new "Hypertext Transfer Protocol (HTTP) Field 1834 Name Registry" at as 1835 described in Section 18.4 of [Semantics]. 1837 Then, please update the registry with the field names listed in the 1838 table below: 1840 +===================+==========+======+============+ 1841 | Field Name | Status | Ref. | Comments | 1842 +===================+==========+======+============+ 1843 | Close | standard | 9.6 | (reserved) | 1844 +-------------------+----------+------+------------+ 1845 | MIME-Version | standard | B.1 | | 1846 +-------------------+----------+------+------------+ 1847 | Transfer-Encoding | standard | 6.1 | | 1848 +-------------------+----------+------+------------+ 1850 Table 1 1852 12.2. Media Type Registration 1854 Please update the "Media Types" registry at 1855 with the registration 1856 information in Section 10.1 and Section 10.2 for the media types 1857 "message/http" and "application/http", respectively. 1859 12.3. Transfer Coding Registration 1861 Please update the "HTTP Transfer Coding Registry" at 1862 with the 1863 registration procedure of Section 7.3 and the content coding names 1864 summarized in the table below. 1866 +============+===============================+===========+ 1867 | Name | Description | Reference | 1868 +============+===============================+===========+ 1869 | chunked | Transfer in a series of | Section | 1870 | | chunks | 7.1 | 1871 +------------+-------------------------------+-----------+ 1872 | compress | UNIX "compress" data format | Section | 1873 | | [Welch] | 7.2 | 1874 +------------+-------------------------------+-----------+ 1875 | deflate | "deflate" compressed data | Section | 1876 | | ([RFC1951]) inside the "zlib" | 7.2 | 1877 | | data format ([RFC1950]) | | 1878 +------------+-------------------------------+-----------+ 1879 | gzip | GZIP file format [RFC1952] | Section | 1880 | | | 7.2 | 1881 +------------+-------------------------------+-----------+ 1882 | trailers | (reserved) | Section | 1883 | | | 12.3 | 1884 +------------+-------------------------------+-----------+ 1885 | x-compress | Deprecated (alias for | Section | 1886 | | compress) | 7.2 | 1887 +------------+-------------------------------+-----------+ 1888 | x-gzip | Deprecated (alias for gzip) | Section | 1889 | | | 7.2 | 1890 +------------+-------------------------------+-----------+ 1892 Table 2 1894 | *Note:* the coding name "trailers" is reserved because its use 1895 | would conflict with the keyword "trailers" in the TE header 1896 | field (Section 10.1.4 of [Semantics]). 1898 12.4. ALPN Protocol ID Registration 1900 Please update the "TLS Application-Layer Protocol Negotiation (ALPN) 1901 Protocol IDs" registry at with the 1903 registration below: 1905 +==========+=============================+================+ 1906 | Protocol | Identification Sequence | Reference | 1907 +==========+=============================+================+ 1908 | HTTP/1.1 | 0x68 0x74 0x74 0x70 0x2f | (this | 1909 | | 0x31 0x2e 0x31 ("http/1.1") | specification) | 1910 +----------+-----------------------------+----------------+ 1912 Table 3 1914 13. References 1916 13.1. Normative References 1918 [Caching] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 1919 Ed., "HTTP Caching", Work in Progress, Internet-Draft, 1920 draft-ietf-httpbis-cache-15, 30 March 2021, 1921 . 1923 [RFC1950] Deutsch, L.P. and J-L. Gailly, "ZLIB Compressed Data 1924 Format Specification version 3.3", RFC 1950, 1925 DOI 10.17487/RFC1950, May 1996, 1926 . 1928 [RFC1951] Deutsch, P., "DEFLATE Compressed Data Format Specification 1929 version 1.3", RFC 1951, DOI 10.17487/RFC1951, May 1996, 1930 . 1932 [RFC1952] Deutsch, P., Gailly, J-L., Adler, M., Deutsch, L.P., and 1933 G. Randers-Pehrson, "GZIP file format specification 1934 version 4.3", RFC 1952, DOI 10.17487/RFC1952, May 1996, 1935 . 1937 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1938 Requirement Levels", BCP 14, RFC 2119, 1939 DOI 10.17487/RFC2119, March 1997, 1940 . 1942 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 1943 Resource Identifier (URI): Generic Syntax", STD 66, 1944 RFC 3986, DOI 10.17487/RFC3986, January 2005, 1945 . 1947 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 1948 Specifications: ABNF", STD 68, RFC 5234, 1949 DOI 10.17487/RFC5234, January 2008, 1950 . 1952 [RFC7405] Kyzivat, P., "Case-Sensitive String Support in ABNF", 1953 RFC 7405, DOI 10.17487/RFC7405, December 2014, 1954 . 1956 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1957 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1958 May 2017, . 1960 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 1961 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 1962 . 1964 [Semantics] 1965 Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 1966 Ed., "HTTP Semantics", Work in Progress, Internet-Draft, 1967 draft-ietf-httpbis-semantics-15, 30 March 2021, 1968 . 1971 [USASCII] American National Standards Institute, "Coded Character 1972 Set -- 7-bit American Standard Code for Information 1973 Interchange", ANSI X3.4, 1986. 1975 [Welch] Welch, T. A., "A Technique for High-Performance Data 1976 Compression", IEEE Computer 17(6), June 1984. 1978 13.2. Informative References 1980 [Err4667] RFC Errata, Erratum ID 4667, RFC 7230, 1981 . 1983 [Klein] Klein, A., "Divide and Conquer - HTTP Response Splitting, 1984 Web Cache Poisoning Attacks, and Related Topics", March 1985 2004, . 1988 [Linhart] Linhart, C., Klein, A., Heled, R., and S. Orrin, "HTTP 1989 Request Smuggling", June 2005, 1990 . 1993 [RFC1945] Berners-Lee, T., Fielding, R.T., and H.F. Nielsen, 1994 "Hypertext Transfer Protocol -- HTTP/1.0", RFC 1945, 1995 DOI 10.17487/RFC1945, May 1996, 1996 . 1998 [RFC2045] Freed, N. and N.S. Borenstein, "Multipurpose Internet Mail 1999 Extensions (MIME) Part One: Format of Internet Message 2000 Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996, 2001 . 2003 [RFC2046] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2004 Extensions (MIME) Part Two: Media Types", RFC 2046, 2005 DOI 10.17487/RFC2046, November 1996, 2006 . 2008 [RFC2049] Freed, N. and N.S. Borenstein, "Multipurpose Internet Mail 2009 Extensions (MIME) Part Five: Conformance Criteria and 2010 Examples", RFC 2049, DOI 10.17487/RFC2049, November 1996, 2011 . 2013 [RFC2068] Fielding, R., Gettys, J., Mogul, J., Nielsen, H., and T. 2014 Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", 2015 RFC 2068, DOI 10.17487/RFC2068, January 1997, 2016 . 2018 [RFC2557] Palme, F., Hopmann, A., Shelness, N., and E. Stefferud, 2019 "MIME Encapsulation of Aggregate Documents, such as HTML 2020 (MHTML)", RFC 2557, DOI 10.17487/RFC2557, March 1999, 2021 . 2023 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 2024 DOI 10.17487/RFC5322, October 2008, 2025 . 2027 [RFC7230] Fielding, R., Ed. and J. F. Reschke, Ed., "Hypertext 2028 Transfer Protocol (HTTP/1.1): Message Syntax and Routing", 2029 RFC 7230, DOI 10.17487/RFC7230, June 2014, 2030 . 2032 [RFC7231] Fielding, R., Ed. and J. F. Reschke, Ed., "Hypertext 2033 Transfer Protocol (HTTP/1.1): Semantics and Content", 2034 RFC 7231, DOI 10.17487/RFC7231, June 2014, 2035 . 2037 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2038 Writing an IANA Considerations Section in RFCs", BCP 26, 2039 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2040 . 2042 Appendix A. Collected ABNF 2044 In the collected ABNF below, list rules are expanded as per 2045 Section 5.6.1.1 of [Semantics]. 2047 BWS = 2049 HTTP-message = start-line CRLF *( field-line CRLF ) CRLF [ 2050 message-body ] 2051 HTTP-name = %x48.54.54.50 ; HTTP 2052 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 2054 OWS = 2055 RWS = 2057 Transfer-Encoding = [ transfer-coding *( OWS "," OWS transfer-coding 2058 ) ] 2060 absolute-URI = 2061 absolute-form = absolute-URI 2062 absolute-path = 2063 asterisk-form = "*" 2064 authority = 2065 authority-form = uri-host ":" port 2067 chunk = chunk-size [ chunk-ext ] CRLF chunk-data CRLF 2068 chunk-data = 1*OCTET 2069 chunk-ext = *( BWS ";" BWS chunk-ext-name [ BWS "=" BWS chunk-ext-val 2070 ] ) 2071 chunk-ext-name = token 2072 chunk-ext-val = token / quoted-string 2073 chunk-size = 1*HEXDIG 2074 chunked-body = *chunk last-chunk trailer-section CRLF 2076 field-line = field-name ":" OWS field-value OWS 2077 field-name = 2078 field-value = 2080 last-chunk = 1*"0" [ chunk-ext ] CRLF 2082 message-body = *OCTET 2083 method = token 2085 obs-fold = OWS CRLF RWS 2086 obs-text = 2087 origin-form = absolute-path [ "?" query ] 2089 port = 2091 query = 2092 quoted-string = 2094 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 2095 request-line = method SP request-target SP HTTP-version 2096 request-target = origin-form / absolute-form / authority-form / 2097 asterisk-form 2099 start-line = request-line / status-line 2100 status-code = 3DIGIT 2101 status-line = HTTP-version SP status-code SP [ reason-phrase ] 2102 token = 2103 trailer-section = *( field-line CRLF ) 2104 transfer-coding = 2106 uri-host = 2108 Appendix B. Differences between HTTP and MIME 2110 HTTP/1.1 uses many of the constructs defined for the Internet Message 2111 Format [RFC5322] and the Multipurpose Internet Mail Extensions (MIME) 2112 [RFC2045] to allow a message body to be transmitted in an open 2113 variety of representations and with extensible fields. However, RFC 2114 2045 is focused only on email; applications of HTTP have many 2115 characteristics that differ from email; hence, HTTP has features that 2116 differ from MIME. These differences were carefully chosen to 2117 optimize performance over binary connections, to allow greater 2118 freedom in the use of new media types, to make date comparisons 2119 easier, and to acknowledge the practice of some early HTTP servers 2120 and clients. 2122 This appendix describes specific areas where HTTP differs from MIME. 2123 Proxies and gateways to and from strict MIME environments need to be 2124 aware of these differences and provide the appropriate conversions 2125 where necessary. 2127 B.1. MIME-Version 2129 HTTP is not a MIME-compliant protocol. However, messages can include 2130 a single MIME-Version header field to indicate what version of the 2131 MIME protocol was used to construct the message. Use of the MIME- 2132 Version header field indicates that the message is in full 2133 conformance with the MIME protocol (as defined in [RFC2045]). 2134 Senders are responsible for ensuring full conformance (where 2135 possible) when exporting HTTP messages to strict MIME environments. 2137 B.2. Conversion to Canonical Form 2139 MIME requires that an Internet mail body part be converted to 2140 canonical form prior to being transferred, as described in Section 4 2141 of [RFC2049], and that content with a type of "text" represent line 2142 breaks as CRLF, forbidding the use of CR or LF outside of line break 2143 sequences [RFC2046]. In contrast, HTTP does not care whether CRLF, 2144 bare CR, or bare LF are used to indicate a line break within content. 2146 A proxy or gateway from HTTP to a strict MIME environment ought to 2147 translate all line breaks within text media types to the RFC 2049 2148 canonical form of CRLF. Note, however, this might be complicated by 2149 the presence of a Content-Encoding and by the fact that HTTP allows 2150 the use of some charsets that do not use octets 13 and 10 to 2151 represent CR and LF, respectively. 2153 Conversion will break any cryptographic checksums applied to the 2154 original content unless the original content is already in canonical 2155 form. Therefore, the canonical form is recommended for any content 2156 that uses such checksums in HTTP. 2158 B.3. Conversion of Date Formats 2160 HTTP/1.1 uses a restricted set of date formats (Section 5.6.7 of 2161 [Semantics]) to simplify the process of date comparison. Proxies and 2162 gateways from other protocols ought to ensure that any Date header 2163 field present in a message conforms to one of the HTTP/1.1 formats 2164 and rewrite the date if necessary. 2166 B.4. Conversion of Content-Encoding 2168 MIME does not include any concept equivalent to HTTP/1.1's Content- 2169 Encoding header field. Since this acts as a modifier on the media 2170 type, proxies and gateways from HTTP to MIME-compliant protocols 2171 ought to either change the value of the Content-Type header field or 2172 decode the representation before forwarding the message. (Some 2173 experimental applications of Content-Type for Internet mail have used 2174 a media-type parameter of ";conversions=" to perform 2175 a function equivalent to Content-Encoding. However, this parameter 2176 is not part of the MIME standards). 2178 B.5. Conversion of Content-Transfer-Encoding 2180 HTTP does not use the Content-Transfer-Encoding field of MIME. 2181 Proxies and gateways from MIME-compliant protocols to HTTP need to 2182 remove any Content-Transfer-Encoding prior to delivering the response 2183 message to an HTTP client. 2185 Proxies and gateways from HTTP to MIME-compliant protocols are 2186 responsible for ensuring that the message is in the correct format 2187 and encoding for safe transport on that protocol, where "safe 2188 transport" is defined by the limitations of the protocol being used. 2189 Such a proxy or gateway ought to transform and label the data with an 2190 appropriate Content-Transfer-Encoding if doing so will improve the 2191 likelihood of safe transport over the destination protocol. 2193 B.6. MHTML and Line Length Limitations 2195 HTTP implementations that share code with MHTML [RFC2557] 2196 implementations need to be aware of MIME line length limitations. 2197 Since HTTP does not have this limitation, HTTP does not fold long 2198 lines. MHTML messages being transported by HTTP follow all 2199 conventions of MHTML, including line length limitations and folding, 2200 canonicalization, etc., since HTTP transfers message-bodies without 2201 modification and, aside from the "multipart/byteranges" type 2202 (Section 14.6 of [Semantics]), does not interpret the content or any 2203 MIME header lines that might be contained therein. 2205 Appendix C. Changes from previous RFCs 2207 C.1. Changes from HTTP/0.9 2209 Since HTTP/0.9 did not support header fields in a request, there is 2210 no mechanism for it to support name-based virtual hosts (selection of 2211 resource by inspection of the Host header field). Any server that 2212 implements name-based virtual hosts ought to disable support for 2213 HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in fact, 2214 badly constructed HTTP/1.x requests caused by a client failing to 2215 properly encode the request-target. 2217 C.2. Changes from HTTP/1.0 2219 C.2.1. Multihomed Web Servers 2221 The requirements that clients and servers support the Host header 2222 field (Section 7.2 of [Semantics]), report an error if it is missing 2223 from an HTTP/1.1 request, and accept absolute URIs (Section 3.2) are 2224 among the most important changes defined by HTTP/1.1. 2226 Older HTTP/1.0 clients assumed a one-to-one relationship of IP 2227 addresses and servers; there was no other established mechanism for 2228 distinguishing the intended server of a request than the IP address 2229 to which that request was directed. The Host header field was 2230 introduced during the development of HTTP/1.1 and, though it was 2231 quickly implemented by most HTTP/1.0 browsers, additional 2232 requirements were placed on all HTTP/1.1 requests in order to ensure 2233 complete adoption. At the time of this writing, most HTTP-based 2234 services are dependent upon the Host header field for targeting 2235 requests. 2237 C.2.2. Keep-Alive Connections 2239 In HTTP/1.0, each connection is established by the client prior to 2240 the request and closed by the server after sending the response. 2241 However, some implementations implement the explicitly negotiated 2242 ("Keep-Alive") version of persistent connections described in 2243 Section 19.7.1 of [RFC2068]. 2245 Some clients and servers might wish to be compatible with these 2246 previous approaches to persistent connections, by explicitly 2247 negotiating for them with a "Connection: keep-alive" request header 2248 field. However, some experimental implementations of HTTP/1.0 2249 persistent connections are faulty; for example, if an HTTP/1.0 proxy 2250 server doesn't understand Connection, it will erroneously forward 2251 that header field to the next inbound server, which would result in a 2252 hung connection. 2254 One attempted solution was the introduction of a Proxy-Connection 2255 header field, targeted specifically at proxies. In practice, this 2256 was also unworkable, because proxies are often deployed in multiple 2257 layers, bringing about the same problem discussed above. 2259 As a result, clients are encouraged not to send the Proxy-Connection 2260 header field in any requests. 2262 Clients are also encouraged to consider the use of Connection: keep- 2263 alive in requests carefully; while they can enable persistent 2264 connections with HTTP/1.0 servers, clients using them will need to 2265 monitor the connection for "hung" requests (which indicate that the 2266 client ought stop sending the header field), and this mechanism ought 2267 not be used by clients at all when a proxy is being used. 2269 C.2.3. Introduction of Transfer-Encoding 2271 HTTP/1.1 introduces the Transfer-Encoding header field (Section 6.1). 2272 Transfer codings need to be decoded prior to forwarding an HTTP 2273 message over a MIME-compliant protocol. 2275 C.3. Changes from RFC 7230 2277 Most of the sections introducing HTTP's design goals, history, 2278 architecture, conformance criteria, protocol versioning, URIs, 2279 message routing, and header fields have been moved to [Semantics]. 2280 This document has been reduced to just the messaging syntax and 2281 connection management requirements specific to HTTP/1.1. 2283 Prohibited generation of bare CRs outside of content. (Section 2.2) 2284 The ABNF definition of authority-form has changed from the more 2285 general authority component of a URI (in which port is optional) to 2286 the specific host:port format that is required by CONNECT. 2287 (Section 3.2.3) 2289 In the ABNF for chunked extensions, re-introduced (bad) whitespace 2290 around ";" and "=". Whitespace was removed in [RFC7230], but that 2291 change was found to break existing implementations (see [Err4667]). 2292 (Section 7.1.1) 2294 Trailer field semantics now transcend the specifics of chunked 2295 encoding. The decoding algorithm for chunked (Section 7.1.3) has 2296 been updated to encourage storage/forwarding of trailer fields 2297 separately from the header section, to only allow merging into the 2298 header section if the recipient knows the corresponding field 2299 definition permits and defines how to merge, and otherwise to discard 2300 the trailer fields instead of merging. The trailer part is now 2301 called the trailer section to be more consistent with the header 2302 section and more distinct from a body part. (Section 7.1.2) 2304 Disallowed transfer coding parameters called "q" in order to avoid 2305 conflicts with the use of ranks in the TE header field. 2306 (Section 7.3) 2308 Appendix D. Change Log 2310 This section is to be removed before publishing as an RFC. 2312 D.1. Between RFC7230 and draft 00 2314 The changes were purely editorial: 2316 * Change boilerplate and abstract to indicate the "draft" status, 2317 and update references to ancestor specifications. 2319 * Adjust historical notes. 2321 * Update links to sibling specifications. 2323 * Replace sections listing changes from RFC 2616 by new empty 2324 sections referring to RFC 723x. 2326 * Remove acknowledgements specific to RFC 723x. 2328 * Move "Acknowledgements" to the very end and make them unnumbered. 2330 D.2. Since draft-ietf-httpbis-messaging-00 2332 The changes in this draft are editorial, with respect to HTTP as a 2333 whole, to move all core HTTP semantics into [Semantics]: 2335 * Moved introduction, architecture, conformance, and ABNF extensions 2336 from RFC 7230 (Messaging) to semantics [Semantics]. 2338 * Moved discussion of MIME differences from RFC 7231 (Semantics) to 2339 Appendix B since they mostly cover transforming 1.1 messages. 2341 * Moved all extensibility tips, registration procedures, and 2342 registry tables from the IANA considerations to normative 2343 sections, reducing the IANA considerations to just instructions 2344 that will be removed prior to publication as an RFC. 2346 D.3. Since draft-ietf-httpbis-messaging-01 2348 * Cite RFC 8126 instead of RFC 5226 () 2351 * Resolved erratum 4779, no change needed here 2352 (, 2353 ) 2355 * In Section 7, fixed prose claiming transfer parameters allow bare 2356 names (, 2357 ) 2359 * Resolved erratum 4225, no change needed here 2360 (, 2361 ) 2363 * Replace "response code" with "response status code" 2364 (, 2365 ) 2367 * In Section 9.3, clarify statement about HTTP/1.0 keep-alive 2368 (, 2369 ) 2371 * In Section 7.1.1, re-introduce (bad) whitespace around ";" and "=" 2372 (, 2373 , ) 2376 * In Section 7.3, state that transfer codings should not use 2377 parameters named "q" (, ) 2380 * In Section 7, mark coding name "trailers" as reserved in the IANA 2381 registry () 2383 D.4. Since draft-ietf-httpbis-messaging-02 2385 * In Section 4, explain why the reason phrase should be ignored by 2386 clients (). 2388 * Add Section 9.2 to explain how request/response correlation is 2389 performed () 2391 D.5. Since draft-ietf-httpbis-messaging-03 2393 * In Section 9.2, caution against treating data on a connection as 2394 part of a not-yet-issued request () 2397 * In Section 7, remove the predefined codings from the ABNF and make 2398 it generic instead () 2401 * Use RFC 7405 ABNF notation for case-sensitive string constants 2402 () 2404 D.6. Since draft-ietf-httpbis-messaging-04 2406 * In Section 7.8 of [Semantics], clarify that protocol-name is to be 2407 matched case-insensitively () 2410 * In Section 5.2, add leading optional whitespace to obs-fold ABNF 2411 (, 2412 ) 2414 * In Section 4, add clarifications about empty reason phrases 2415 () 2417 * Move discussion of retries from Section 9.3.1 into [Semantics] 2418 () 2420 D.7. Since draft-ietf-httpbis-messaging-05 2421 * In Section 7.1.2, the trailer part has been renamed the trailer 2422 section (for consistency with the header section) and trailers are 2423 no longer merged as header fields by default, but rather can be 2424 discarded, kept separate from header fields, or merged with header 2425 fields only if understood and defined as being mergeable 2426 () 2428 * In Section 2.1 and related Sections, move the trailing CRLF from 2429 the line grammars into the message format 2430 () 2432 * Moved Section 2.3 down () 2435 * In Section 7.8 of [Semantics], use 'websocket' instead of 2436 'HTTP/2.0' in examples () 2439 * Move version non-specific text from Section 6 into semantics as 2440 "payload" () 2442 * In Section 9.8, add text from RFC 2818 2443 () 2445 D.8. Since draft-ietf-httpbis-messaging-06 2447 * In Section 12.4, update the APLN protocol id for HTTP/1.1 2448 () 2450 * In Section 5, align with updates to field terminology in semantics 2451 () 2453 * In Section 7.6.1 of [Semantics], clarify that new connection 2454 options indeed need to be registered () 2457 * In Section 1.1, reference RFC 8174 as well 2458 () 2460 D.9. Since draft-ietf-httpbis-messaging-07 2462 * Move TE: trailers into [Semantics] () 2465 * In Section 6.3, adjust requirements for handling multiple content- 2466 length values () 2468 * Throughout, replace "effective request URI" with "target URI" 2469 () 2471 * In Section 6.1, don't claim Transfer-Encoding is supported by 2472 HTTP/2 or later () 2474 D.10. Since draft-ietf-httpbis-messaging-08 2476 * In Section 2.2, disallow bare CRs () 2479 * Appendix A now uses the sender variant of the "#" list expansion 2480 () 2482 * In Section 5, adjust IANA "Close" entry for new registry format 2483 () 2485 D.11. Since draft-ietf-httpbis-messaging-09 2487 * Switch to xml2rfc v3 mode for draft generation 2488 () 2490 D.12. Since draft-ietf-httpbis-messaging-10 2492 * In Section 6.3, note that TCP half-close does not delimit a 2493 request; talk about corresponding server-side behaviour in 2494 Section 9.6 () 2496 * Moved requirements specific to HTTP/1.1 from [Semantics] into 2497 Section 3.2 () 2499 * In Section 6.1 (Transfer-Encoding), adjust ABNF to allow empty 2500 lists () 2502 * In Section 9.7, add text from RFC 2818 2503 () 2505 * Moved definitions of "TE" and "Upgrade" into [Semantics] 2506 () 2508 * Moved definition of "Connection" into [Semantics] 2509 () 2511 D.13. Since draft-ietf-httpbis-messaging-11 2513 * Move IANA Upgrade Token Registry instructions to [Semantics] 2514 () 2516 D.14. Since draft-ietf-httpbis-messaging-12 2518 * Moved content of history appendix to Semantics 2519 () 2521 * Moved note about "close" being reserved as field name to 2522 Section 9.3 () 2524 * Moved table of transfer codings into Section 12.3 2525 () 2527 * In Section 13.2, updated the URI for the [Linhart] paper 2528 () 2530 * Changed document title to just "HTTP/1.1" 2531 () 2533 * In Section 7, moved transfer-coding ABNF to Section 10.1.4 of 2534 [Semantics] () 2536 * Changed to using "payload data" when defining requirements about 2537 the data being conveyed within a message, instead of the terms 2538 "payload body" or "response body" or "representation body", since 2539 they often get confused with the HTTP/1.1 message body (which 2540 includes transfer coding) () 2543 D.15. Since draft-ietf-httpbis-messaging-13 2545 * In Section 6.3, clarify that a message needs to be checked for 2546 both Content-Length and Transfer-Encoding, before processing 2547 Transfer-Encoding, and that ought to be treated as an error, but 2548 an intermediary can choose to forward the message downstream after 2549 removing the Content-Length and processing the Transfer-Encoding 2550 () 2552 * Changed to using "content" instead of "payload" or "payload data" 2553 to avoid confusion with the payload of version-specific messaging 2554 frames () 2556 D.16. Since draft-ietf-httpbis-messaging-14 2558 * In Section 9.6, define the close connection option, since its 2559 definition was removed from the Connection header field for being 2560 specific to 1.1 () 2562 * In Section 3.3, clarify how the target URI is reconstructed when 2563 the request-target is not in absolute-form and highlight risk in 2564 selecting a default host () 2567 * In Section 7.1, clarify large chunk handling issues 2568 () 2570 * In Section 2.2, explicitly close the connection after sending a 2571 400 () 2573 * In Section 2.3, refine version requirements for intermediaries 2574 () 2576 * In Section 7.1.3, don't remove the Trailer header field 2577 () 2579 * In Section 3.2.3, changed the ABNF definition of authority-form 2580 from the authority component (in which port is optional) to the 2581 host:port format that has always been required by CONNECT 2582 () 2584 Acknowledgments 2586 See Appendix "Acknowledgments" of [Semantics]. 2588 Authors' Addresses 2590 Roy T. Fielding (editor) 2591 Adobe 2592 345 Park Ave 2593 San Jose, CA 95110 2594 United States of America 2596 Email: fielding@gbiv.com 2597 URI: https://roy.gbiv.com/ 2599 Mark Nottingham (editor) 2600 Fastly 2601 Prahran VIC 2602 Australia 2604 Email: mnot@mnot.net 2605 URI: https://www.mnot.net/ 2606 Julian Reschke (editor) 2607 greenbytes GmbH 2608 Hafenweg 16 2609 48155 Münster 2610 Germany 2612 Email: julian.reschke@greenbytes.de 2613 URI: https://greenbytes.de/tech/webdav/