idnits 2.17.00 (12 Aug 2021) /tmp/idnits35393/draft-ietf-httpbis-messaging-03.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The draft header indicates that this document obsoletes RFC7230, but the abstract doesn't seem to directly say this. It does mention RFC7230 though, so this could be OK. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == The document seems to contain a disclaimer for pre-RFC5378 work, but was first submitted on or after 10 November 2008. The disclaimer is usually necessary only for documents that revise or obsolete older RFCs, and that take significant amounts of text from those RFCs. If you can contact all authors of the source material and they are willing to grant the BCP78 rights to the IETF Trust, you can and should remove the disclaimer. Otherwise, the disclaimer is needed and you can ignore this comment. (See the Legal Provisions document at https://trustee.ietf.org/license-info for more information.) -- The document date (October 18, 2018) is 1310 days in the past. Is this intentional? Checking references for intended status: Proposed Standard ---------------------------------------------------------------------------- (See RFCs 3967 and 4897 for information about using normative references to lower-maturity documents in RFCs) == Unused Reference: 'RFC3986' is defined on line 2030, but no explicit reference was found in the text == Unused Reference: 'RFC7231' is defined on line 2109, but no explicit reference was found in the text == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-cache-03 -- Possible downref: Normative reference to a draft: ref.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reference to an Informational RFC: RFC 1951 ** Downref: Normative reference to an Informational RFC: RFC 1952 == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-semantics-03 -- Possible downref: Normative reference to a draft: ref.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 2068 (Obsoleted by RFC 2616) -- Duplicate reference: RFC7230, mentioned in 'RFC7230', was also mentioned in 'Err4667'. Summary: 3 errors (**), 0 flaws (~~), 6 warnings (==), 8 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: April 21, 2019 J. Reschke, Ed. 7 greenbytes 8 October 18, 2018 10 HTTP/1.1 Messaging 11 draft-ietf-httpbis-messaging-03 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.4. 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 April 21, 2019. 54 Copyright Notice 56 Copyright (c) 2018 IETF Trust and the persons identified as the 57 document authors. All rights reserved. 59 This document is subject to BCP 78 and the IETF Trust's Legal 60 Provisions Relating to IETF Documents 61 (https://trustee.ietf.org/license-info) in effect on the date of 62 publication of this document. Please review these documents 63 carefully, as they describe your rights and restrictions with respect 64 to this document. Code Components extracted from this document must 65 include Simplified BSD License text as described in Section 4.e of 66 the Trust Legal Provisions and are provided without warranty as 67 described in the Simplified BSD License. 69 This document may contain material from IETF Documents or IETF 70 Contributions published or made publicly available before November 71 10, 2008. The person(s) controlling the copyright in some of this 72 material may not have granted the IETF Trust the right to allow 73 modifications of such material outside the IETF Standards Process. 74 Without obtaining an adequate license from the person(s) controlling 75 the copyright in such materials, this document may not be modified 76 outside the IETF Standards Process, and derivative works of it may 77 not be created outside the IETF Standards Process, except to format 78 it for publication as an RFC or to translate it into languages other 79 than English. 81 Table of Contents 83 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 84 1.1. Requirements Notation . . . . . . . . . . . . . . . . . . 5 85 1.2. Syntax Notation . . . . . . . . . . . . . . . . . . . . . 5 86 2. Message . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 87 2.1. Message Format . . . . . . . . . . . . . . . . . . . . . 6 88 2.2. HTTP Version . . . . . . . . . . . . . . . . . . . . . . 6 89 2.3. Message Parsing . . . . . . . . . . . . . . . . . . . . . 7 90 3. Request Line . . . . . . . . . . . . . . . . . . . . . . . . 8 91 3.1. Method . . . . . . . . . . . . . . . . . . . . . . . . . 9 92 3.2. Request Target . . . . . . . . . . . . . . . . . . . . . 9 93 3.2.1. origin-form . . . . . . . . . . . . . . . . . . . . . 10 94 3.2.2. absolute-form . . . . . . . . . . . . . . . . . . . . 10 95 3.2.3. authority-form . . . . . . . . . . . . . . . . . . . 11 96 3.2.4. asterisk-form . . . . . . . . . . . . . . . . . . . . 11 98 3.3. Effective Request URI . . . . . . . . . . . . . . . . . . 12 99 4. Status Line . . . . . . . . . . . . . . . . . . . . . . . . . 13 100 5. Header Fields . . . . . . . . . . . . . . . . . . . . . . . . 14 101 5.1. Header Field Parsing . . . . . . . . . . . . . . . . . . 15 102 5.2. Obsolete Line Folding . . . . . . . . . . . . . . . . . . 15 103 6. Message Body . . . . . . . . . . . . . . . . . . . . . . . . 16 104 6.1. Transfer-Encoding . . . . . . . . . . . . . . . . . . . . 17 105 6.2. Content-Length . . . . . . . . . . . . . . . . . . . . . 18 106 6.3. Message Body Length . . . . . . . . . . . . . . . . . . . 19 107 7. Transfer Codings . . . . . . . . . . . . . . . . . . . . . . 21 108 7.1. Chunked Transfer Coding . . . . . . . . . . . . . . . . . 22 109 7.1.1. Chunk Extensions . . . . . . . . . . . . . . . . . . 23 110 7.1.2. Chunked Trailer Part . . . . . . . . . . . . . . . . 23 111 7.1.3. Decoding Chunked . . . . . . . . . . . . . . . . . . 24 112 7.2. Transfer Codings for Compression . . . . . . . . . . . . 25 113 7.3. Transfer Coding Registry . . . . . . . . . . . . . . . . 25 114 7.4. TE . . . . . . . . . . . . . . . . . . . . . . . . . . . 26 115 8. Handling Incomplete Messages . . . . . . . . . . . . . . . . 27 116 9. Connection Management . . . . . . . . . . . . . . . . . . . . 28 117 9.1. Connection . . . . . . . . . . . . . . . . . . . . . . . 28 118 9.2. Establishment . . . . . . . . . . . . . . . . . . . . . . 30 119 9.3. Associating a Response to a Request . . . . . . . . . . . 30 120 9.4. Persistence . . . . . . . . . . . . . . . . . . . . . . . 30 121 9.4.1. Retrying Requests . . . . . . . . . . . . . . . . . . 31 122 9.4.2. Pipelining . . . . . . . . . . . . . . . . . . . . . 32 123 9.5. Concurrency . . . . . . . . . . . . . . . . . . . . . . . 32 124 9.6. Failures and Timeouts . . . . . . . . . . . . . . . . . . 33 125 9.7. Tear-down . . . . . . . . . . . . . . . . . . . . . . . . 34 126 9.8. Upgrade . . . . . . . . . . . . . . . . . . . . . . . . . 35 127 9.8.1. Upgrade Protocol Names . . . . . . . . . . . . . . . 37 128 9.8.2. Upgrade Token Registry . . . . . . . . . . . . . . . 37 129 10. Enclosing Messages as Data . . . . . . . . . . . . . . . . . 38 130 10.1. Media Type message/http . . . . . . . . . . . . . . . . 38 131 10.2. Media Type application/http . . . . . . . . . . . . . . 39 132 11. Security Considerations . . . . . . . . . . . . . . . . . . . 40 133 11.1. Response Splitting . . . . . . . . . . . . . . . . . . . 40 134 11.2. Request Smuggling . . . . . . . . . . . . . . . . . . . 41 135 11.3. Message Integrity . . . . . . . . . . . . . . . . . . . 42 136 11.4. Message Confidentiality . . . . . . . . . . . . . . . . 42 137 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 42 138 12.1. Header Field Registration . . . . . . . . . . . . . . . 43 139 12.2. Media Type Registration . . . . . . . . . . . . . . . . 43 140 12.3. Transfer Coding Registration . . . . . . . . . . . . . . 43 141 12.4. Upgrade Token Registration . . . . . . . . . . . . . . . 43 142 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 43 143 13.1. Normative References . . . . . . . . . . . . . . . . . . 43 144 13.2. Informative References . . . . . . . . . . . . . . . . . 44 145 Appendix A. Collected ABNF . . . . . . . . . . . . . . . . . . . 47 146 Appendix B. Differences between HTTP and MIME . . . . . . . . . 48 147 B.1. MIME-Version . . . . . . . . . . . . . . . . . . . . . . 49 148 B.2. Conversion to Canonical Form . . . . . . . . . . . . . . 49 149 B.3. Conversion of Date Formats . . . . . . . . . . . . . . . 49 150 B.4. Conversion of Content-Encoding . . . . . . . . . . . . . 50 151 B.5. Conversion of Content-Transfer-Encoding . . . . . . . . . 50 152 B.6. MHTML and Line Length Limitations . . . . . . . . . . . . 50 153 Appendix C. HTTP Version History . . . . . . . . . . . . . . . . 50 154 C.1. Changes from HTTP/1.0 . . . . . . . . . . . . . . . . . . 51 155 C.1.1. Multihomed Web Servers . . . . . . . . . . . . . . . 51 156 C.1.2. Keep-Alive Connections . . . . . . . . . . . . . . . 52 157 C.1.3. Introduction of Transfer-Encoding . . . . . . . . . . 52 158 C.2. Changes from RFC 7230 . . . . . . . . . . . . . . . . . . 53 159 Appendix D. Change Log . . . . . . . . . . . . . . . . . . . . . 53 160 D.1. Between RFC7230 and draft 00 . . . . . . . . . . . . . . 53 161 D.2. Since draft-ietf-httpbis-messaging-00 . . . . . . . . . . 53 162 D.3. Since draft-ietf-httpbis-messaging-01 . . . . . . . . . . 54 163 D.4. Since draft-ietf-httpbis-messaging-02 . . . . . . . . . . 55 164 Index . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . 55 165 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 57 166 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 57 168 1. Introduction 170 The Hypertext Transfer Protocol (HTTP) is a stateless application- 171 level request/response protocol that uses extensible semantics and 172 self-descriptive messages for flexible interaction with network-based 173 hypertext information systems. HTTP is defined by a series of 174 documents that collectively form the HTTP/1.1 specification: 176 o "HTTP Semantics" [Semantics] 178 o "HTTP Caching" [Caching] 180 o "HTTP/1.1 Messaging" (this document) 182 This document defines HTTP/1.1 message syntax and framing 183 requirements and their associated connection management. Our goal is 184 to define all of the mechanisms necessary for HTTP/1.1 message 185 handling that are independent of message semantics, thereby defining 186 the complete set of requirements for message parsers and message- 187 forwarding intermediaries. 189 This document obsoletes the portions of RFC 7230 related to HTTP/1.1 190 messaging and connection management, with the changes being 191 summarized in Appendix C.2. The other parts of RFC 7230 are 192 obsoleted by "HTTP Semantics" [Semantics]. 194 1.1. Requirements Notation 196 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 197 "SHOULD", "SHOULD NOT", "RECOMMENDED", "MAY", and "OPTIONAL" in this 198 document are to be interpreted as described in [RFC2119]. 200 Conformance criteria and considerations regarding error handling are 201 defined in Section 3 of [Semantics]. 203 1.2. Syntax Notation 205 This specification uses the Augmented Backus-Naur Form (ABNF) 206 notation of [RFC5234] with a list extension, defined in Section 11 of 207 [Semantics], that allows for compact definition of comma-separated 208 lists using a '#' operator (similar to how the '*' operator indicates 209 repetition). Appendix A shows the collected grammar with all list 210 operators expanded to standard ABNF notation. 212 As a convention, ABNF rule names prefixed with "obs-" denote 213 "obsolete" grammar rules that appear for historical reasons. 215 The following core rules are included by reference, as defined in 216 [RFC5234], Appendix B.1: ALPHA (letters), CR (carriage return), CRLF 217 (CR LF), CTL (controls), DIGIT (decimal 0-9), DQUOTE (double quote), 218 HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line 219 feed), OCTET (any 8-bit sequence of data), SP (space), and VCHAR (any 220 visible [USASCII] character). 222 The rules below are defined in [Semantics]: 224 BWS = 225 OWS = 226 RWS = 227 absolute-URI = 228 absolute-path = 229 authority = 230 comment = 231 field-name = 232 field-value = 233 obs-text = 234 port = 235 query = 236 quoted-string = 237 token = 238 uri-host = 240 2. Message 242 2.1. Message Format 244 All HTTP/1.1 messages consist of a start-line followed by a sequence 245 of octets in a format similar to the Internet Message Format 246 [RFC5322]: zero or more header fields (collectively referred to as 247 the "headers" or the "header section"), an empty line indicating the 248 end of the header section, and an optional message body. 250 HTTP-message = start-line 251 *( header-field CRLF ) 252 CRLF 253 [ message-body ] 255 An HTTP message can be either a request from client to server or a 256 response from server to client. Syntactically, the two types of 257 message differ only in the start-line, which is either a request-line 258 (for requests) or a status-line (for responses), and in the algorithm 259 for determining the length of the message body (Section 6). 261 start-line = request-line / status-line 263 In theory, a client could receive requests and a server could receive 264 responses, distinguishing them by their different start-line formats. 265 In practice, servers are implemented to only expect a request (a 266 response is interpreted as an unknown or invalid request method) and 267 clients are implemented to only expect a response. 269 Although HTTP makes use of some protocol elements similar to the 270 Multipurpose Internet Mail Extensions (MIME) [RFC2045], see 271 Appendix B for the differences between HTTP and MIME messages. 273 2.2. HTTP Version 275 HTTP uses a "." numbering scheme to indicate versions 276 of the protocol. This specification defines version "1.1". 277 Section 3.5 of [Semantics] specifies the semantics of HTTP version 278 numbers. 280 The version of an HTTP/1.x message is indicated by an HTTP-version 281 field in the start-line. HTTP-version is case-sensitive. 283 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 284 HTTP-name = %x48.54.54.50 ; "HTTP", case-sensitive 286 When an HTTP/1.1 message is sent to an HTTP/1.0 recipient [RFC1945] 287 or a recipient whose version is unknown, the HTTP/1.1 message is 288 constructed such that it can be interpreted as a valid HTTP/1.0 289 message if all of the newer features are ignored. This specification 290 places recipient-version requirements on some new features so that a 291 conformant sender will only use compatible features until it has 292 determined, through configuration or the receipt of a message, that 293 the recipient supports HTTP/1.1. 295 Intermediaries that process HTTP messages (i.e., all intermediaries 296 other than those acting as tunnels) MUST send their own HTTP-version 297 in forwarded messages. In other words, they are not allowed to 298 blindly forward the start-line without ensuring that the protocol 299 version in that message matches a version to which that intermediary 300 is conformant for both the receiving and sending of messages. 301 Forwarding an HTTP message without rewriting the HTTP-version might 302 result in communication errors when downstream recipients use the 303 message sender's version to determine what features are safe to use 304 for later communication with that sender. 306 A server MAY send an HTTP/1.0 response to an HTTP/1.1 request if it 307 is known or suspected that the client incorrectly implements the HTTP 308 specification and is incapable of correctly processing later version 309 responses, such as when a client fails to parse the version number 310 correctly or when an intermediary is known to blindly forward the 311 HTTP-version even when it doesn't conform to the given minor version 312 of the protocol. Such protocol downgrades SHOULD NOT be performed 313 unless triggered by specific client attributes, such as when one or 314 more of the request header fields (e.g., User-Agent) uniquely match 315 the values sent by a client known to be in error. 317 2.3. Message Parsing 319 The normal procedure for parsing an HTTP message is to read the 320 start-line into a structure, read each header field into a hash table 321 by field name until the empty line, and then use the parsed data to 322 determine if a message body is expected. If a message body has been 323 indicated, then it is read as a stream until an amount of octets 324 equal to the message body length is read or the connection is closed. 326 A recipient MUST parse an HTTP message as a sequence of octets in an 327 encoding that is a superset of US-ASCII [USASCII]. Parsing an HTTP 328 message as a stream of Unicode characters, without regard for the 329 specific encoding, creates security vulnerabilities due to the 330 varying ways that string processing libraries handle invalid 331 multibyte character sequences that contain the octet LF (%x0A). 332 String-based parsers can only be safely used within protocol elements 333 after the element has been extracted from the message, such as within 334 a header field-value after message parsing has delineated the 335 individual fields. 337 Although the line terminator for the start-line and header fields is 338 the sequence CRLF, a recipient MAY recognize a single LF as a line 339 terminator and ignore any preceding CR. 341 Older HTTP/1.0 user agent implementations might send an extra CRLF 342 after a POST request as a workaround for some early server 343 applications that failed to read message body content that was not 344 terminated by a line-ending. An HTTP/1.1 user agent MUST NOT preface 345 or follow a request with an extra CRLF. If terminating the request 346 message body with a line-ending is desired, then the user agent MUST 347 count the terminating CRLF octets as part of the message body length. 349 In the interest of robustness, a server that is expecting to receive 350 and parse a request-line SHOULD ignore at least one empty line (CRLF) 351 received prior to the request-line. 353 A sender MUST NOT send whitespace between the start-line and the 354 first header field. A recipient that receives whitespace between the 355 start-line and the first header field MUST either reject the message 356 as invalid or consume each whitespace-preceded line without further 357 processing of it (i.e., ignore the entire line, along with any 358 subsequent lines preceded by whitespace, until a properly formed 359 header field is received or the header section is terminated). 361 The presence of such whitespace in a request might be an attempt to 362 trick a server into ignoring that field or processing the line after 363 it as a new request, either of which might result in a security 364 vulnerability if other implementations within the request chain 365 interpret the same message differently. Likewise, the presence of 366 such whitespace in a response might be ignored by some clients or 367 cause others to cease parsing. 369 When a server listening only for HTTP request messages, or processing 370 what appears from the start-line to be an HTTP request message, 371 receives a sequence of octets that does not match the HTTP-message 372 grammar aside from the robustness exceptions listed above, the server 373 SHOULD respond with a 400 (Bad Request) response. 375 3. Request Line 377 A request-line begins with a method token, followed by a single space 378 (SP), the request-target, another single space (SP), the protocol 379 version, and ends with CRLF. 381 request-line = method SP request-target SP HTTP-version CRLF 383 Although the request-line grammar rule requires that each of the 384 component elements be separated by a single SP octet, recipients MAY 385 instead parse on whitespace-delimited word boundaries and, aside from 386 the CRLF terminator, treat any form of whitespace as the SP separator 387 while ignoring preceding or trailing whitespace; such whitespace 388 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 389 (%x0C), or bare CR. However, lenient parsing can result in request 390 smuggling security vulnerabilities if there are multiple recipients 391 of the message and each has its own unique interpretation of 392 robustness (see Section 11.2). 394 HTTP does not place a predefined limit on the length of a request- 395 line, as described in Section 3 of [Semantics]. A server that 396 receives a method longer than any that it implements SHOULD respond 397 with a 501 (Not Implemented) status code. A server that receives a 398 request-target longer than any URI it wishes to parse MUST respond 399 with a 414 (URI Too Long) status code (see Section 9.5.15 of 400 [Semantics]). 402 Various ad hoc limitations on request-line length are found in 403 practice. It is RECOMMENDED that all HTTP senders and recipients 404 support, at a minimum, request-line lengths of 8000 octets. 406 3.1. Method 408 The method token indicates the request method to be performed on the 409 target resource. The request method is case-sensitive. 411 method = token 413 The request methods defined by this specification can be found in 414 Section 7 of [Semantics], along with information regarding the HTTP 415 method registry and considerations for defining new methods. 417 3.2. Request Target 419 The request-target identifies the target resource upon which to apply 420 the request. The client derives a request-target from its desired 421 target URI. There are four distinct formats for the request-target, 422 depending on both the method being requested and whether the request 423 is to a proxy. 425 request-target = origin-form 426 / absolute-form 427 / authority-form 428 / asterisk-form 430 No whitespace is allowed in the request-target. Unfortunately, some 431 user agents fail to properly encode or exclude whitespace found in 432 hypertext references, resulting in those disallowed characters being 433 sent as the request-target in a malformed request-line. 435 Recipients of an invalid request-line SHOULD respond with either a 436 400 (Bad Request) error or a 301 (Moved Permanently) redirect with 437 the request-target properly encoded. A recipient SHOULD NOT attempt 438 to autocorrect and then process the request without a redirect, since 439 the invalid request-line might be deliberately crafted to bypass 440 security filters along the request chain. 442 3.2.1. origin-form 444 The most common form of request-target is the origin-form. 446 origin-form = absolute-path [ "?" query ] 448 When making a request directly to an origin server, other than a 449 CONNECT or server-wide OPTIONS request (as detailed below), a client 450 MUST send only the absolute path and query components of the target 451 URI as the request-target. If the target URI's path component is 452 empty, the client MUST send "/" as the path within the origin-form of 453 request-target. A Host header field is also sent, as defined in 454 Section 5.4 of [Semantics]. 456 For example, a client wishing to retrieve a representation of the 457 resource identified as 459 http://www.example.org/where?q=now 461 directly from the origin server would open (or reuse) a TCP 462 connection to port 80 of the host "www.example.org" and send the 463 lines: 465 GET /where?q=now HTTP/1.1 466 Host: www.example.org 468 followed by the remainder of the request message. 470 3.2.2. absolute-form 472 When making a request to a proxy, other than a CONNECT or server-wide 473 OPTIONS request (as detailed below), a client MUST send the target 474 URI in absolute-form as the request-target. 476 absolute-form = absolute-URI 478 The proxy is requested to either service that request from a valid 479 cache, if possible, or make the same request on the client's behalf 480 to either the next inbound proxy server or directly to the origin 481 server indicated by the request-target. Requirements on such 482 "forwarding" of messages are defined in Section 5.5 of [Semantics]. 484 An example absolute-form of request-line would be: 486 GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1 488 To allow for transition to the absolute-form for all requests in some 489 future version of HTTP, a server MUST accept the absolute-form in 490 requests, even though HTTP/1.1 clients will only send them in 491 requests to proxies. 493 3.2.3. authority-form 495 The authority-form of request-target is only used for CONNECT 496 requests (Section 7.3.6 of [Semantics]). 498 authority-form = authority 500 When making a CONNECT request to establish a tunnel through one or 501 more proxies, a client MUST send only the target URI's authority 502 component (excluding any userinfo and its "@" delimiter) as the 503 request-target. For example, 505 CONNECT www.example.com:80 HTTP/1.1 507 3.2.4. asterisk-form 509 The asterisk-form of request-target is only used for a server-wide 510 OPTIONS request (Section 7.3.7 of [Semantics]). 512 asterisk-form = "*" 514 When a client wishes to request OPTIONS for the server as a whole, as 515 opposed to a specific named resource of that server, the client MUST 516 send only "*" (%x2A) as the request-target. For example, 518 OPTIONS * HTTP/1.1 520 If a proxy receives an OPTIONS request with an absolute-form of 521 request-target in which the URI has an empty path and no query 522 component, then the last proxy on the request chain MUST send a 523 request-target of "*" when it forwards the request to the indicated 524 origin server. 526 For example, the request 528 OPTIONS http://www.example.org:8001 HTTP/1.1 530 would be forwarded by the final proxy as 532 OPTIONS * HTTP/1.1 533 Host: www.example.org:8001 535 after connecting to port 8001 of host "www.example.org". 537 3.3. Effective Request URI 539 Since the request-target often contains only part of the user agent's 540 target URI, a server reconstructs the intended target as an effective 541 request URI to properly service the request (Section 5.3 of 542 [Semantics]). 544 If the request-target is in absolute-form, the effective request URI 545 is the same as the request-target. Otherwise, the effective request 546 URI is constructed as follows: 548 If the server's configuration (or outbound gateway) provides a 549 fixed URI scheme, that scheme is used for the effective request 550 URI. Otherwise, if the request is received over a TLS-secured TCP 551 connection, the effective request URI's scheme is "https"; if not, 552 the scheme is "http". 554 If the server's configuration (or outbound gateway) provides a 555 fixed URI authority component, that authority is used for the 556 effective request URI. If not, then if the request-target is in 557 authority-form, the effective request URI's authority component is 558 the same as the request-target. If not, then if a Host header 559 field is supplied with a non-empty field-value, the authority 560 component is the same as the Host field-value. Otherwise, the 561 authority component is assigned the default name configured for 562 the server and, if the connection's incoming TCP port number 563 differs from the default port for the effective request URI's 564 scheme, then a colon (":") and the incoming port number (in 565 decimal form) are appended to the authority component. 567 If the request-target is in authority-form or asterisk-form, the 568 effective request URI's combined path and query component is 569 empty. Otherwise, the combined path and query component is the 570 same as the request-target. 572 The components of the effective request URI, once determined as 573 above, can be combined into absolute-URI form by concatenating the 574 scheme, "://", authority, and combined path and query component. 576 Example 1: the following message received over an insecure TCP 577 connection 579 GET /pub/WWW/TheProject.html HTTP/1.1 580 Host: www.example.org:8080 582 has an effective request URI of 584 http://www.example.org:8080/pub/WWW/TheProject.html 586 Example 2: the following message received over a TLS-secured TCP 587 connection 589 OPTIONS * HTTP/1.1 590 Host: www.example.org 592 has an effective request URI of 594 https://www.example.org 596 Recipients of an HTTP/1.0 request that lacks a Host header field 597 might need to use heuristics (e.g., examination of the URI path for 598 something unique to a particular host) in order to guess the 599 effective request URI's authority component. 601 4. Status Line 603 The first line of a response message is the status-line, consisting 604 of the protocol version, a space (SP), the status code, another 605 space, a possibly empty textual phrase describing the status code, 606 and ending with CRLF. 608 status-line = HTTP-version SP status-code SP reason-phrase CRLF 610 Although the status-line grammar rule requires that each of the 611 component elements be separated by a single SP octet, recipients MAY 612 instead parse on whitespace-delimited word boundaries and, aside from 613 the line terminator, treat any form of whitespace as the SP separator 614 while ignoring preceding or trailing whitespace; such whitespace 615 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 616 (%x0C), or bare CR. However, lenient parsing can result in response 617 splitting security vulnerabilities if there are multiple recipients 618 of the message and each has its own unique interpretation of 619 robustness (see Section 11.1). 621 The status-code element is a 3-digit integer code describing the 622 result of the server's attempt to understand and satisfy the client's 623 corresponding request. The rest of the response message is to be 624 interpreted in light of the semantics defined for that status code. 625 See Section 9 of [Semantics] for information about the semantics of 626 status codes, including the classes of status code (indicated by the 627 first digit), the status codes defined by this specification, 628 considerations for the definition of new status codes, and the IANA 629 registry. 631 status-code = 3DIGIT 633 The reason-phrase element exists for the sole purpose of providing a 634 textual description associated with the numeric status code, mostly 635 out of deference to earlier Internet application protocols that were 636 more frequently used with interactive text clients. 638 A client SHOULD ignore the reason-phrase content because it is not a 639 reliable channel for information (it might be discarded or 640 overwritten by intermediaries, and it is not transmitted in other 641 versions of HTTP). 643 reason-phrase = *( HTAB / SP / VCHAR / obs-text ) 645 5. Header Fields 647 Each header field consists of a case-insensitive field name followed 648 by a colon (":"), optional leading whitespace, the field value, and 649 optional trailing whitespace. 651 header-field = field-name ":" OWS field-value OWS 653 Most HTTP field names and the rules for parsing within field values 654 are defined in Section 4 of [Semantics]. This section covers the 655 generic syntax for header field inclusion within, and extraction 656 from, HTTP/1.1 messages. In addition, the following header fields 657 are defined by this document because they are specific to HTTP/1.1 658 message processing: 660 +-------------------+----------+----------+---------------+ 661 | Header Field Name | Protocol | Status | Reference | 662 +-------------------+----------+----------+---------------+ 663 | Connection | http | standard | Section 9.1 | 664 | MIME-Version | http | standard | Appendix B.1 | 665 | TE | http | standard | Section 7.4 | 666 | Transfer-Encoding | http | standard | Section 6.1 | 667 | Upgrade | http | standard | Section 9.8 | 668 +-------------------+----------+----------+---------------+ 669 Furthermore, the field name "Close" is reserved, since using that 670 name as an HTTP header field might conflict with the "close" 671 connection option of the Connection header field (Section 9.1). 673 +-------------------+----------+----------+------------+ 674 | Header Field Name | Protocol | Status | Reference | 675 +-------------------+----------+----------+------------+ 676 | Close | http | reserved | Section 5 | 677 +-------------------+----------+----------+------------+ 679 5.1. Header Field Parsing 681 Messages are parsed using a generic algorithm, independent of the 682 individual header field names. The contents within a given field 683 value are not parsed until a later stage of message interpretation 684 (usually after the message's entire header section has been 685 processed). 687 No whitespace is allowed between the header field-name and colon. In 688 the past, differences in the handling of such whitespace have led to 689 security vulnerabilities in request routing and response handling. A 690 server MUST reject any received request message that contains 691 whitespace between a header field-name and colon with a response 692 status code of 400 (Bad Request). A proxy MUST remove any such 693 whitespace from a response message before forwarding the message 694 downstream. 696 A field value might be preceded and/or followed by optional 697 whitespace (OWS); a single SP preceding the field-value is preferred 698 for consistent readability by humans. The field value does not 699 include any leading or trailing whitespace: OWS occurring before the 700 first non-whitespace octet of the field value or after the last non- 701 whitespace octet of the field value ought to be excluded by parsers 702 when extracting the field value from a header field. 704 5.2. Obsolete Line Folding 706 Historically, HTTP header field values could be extended over 707 multiple lines by preceding each extra line with at least one space 708 or horizontal tab (obs-fold). This specification deprecates such 709 line folding except within the message/http media type 710 (Section 10.1). 712 obs-fold = CRLF 1*( SP / HTAB ) 713 ; obsolete line folding 715 A sender MUST NOT generate a message that includes line folding 716 (i.e., that has any field-value that contains a match to the obs-fold 717 rule) unless the message is intended for packaging within the 718 message/http media type. 720 A server that receives an obs-fold in a request message that is not 721 within a message/http container MUST either reject the message by 722 sending a 400 (Bad Request), preferably with a representation 723 explaining that obsolete line folding is unacceptable, or replace 724 each received obs-fold with one or more SP octets prior to 725 interpreting the field value or forwarding the message downstream. 727 A proxy or gateway that receives an obs-fold in a response message 728 that is not within a message/http container MUST either discard the 729 message and replace it with a 502 (Bad Gateway) response, preferably 730 with a representation explaining that unacceptable line folding was 731 received, or replace each received obs-fold with one or more SP 732 octets prior to interpreting the field value or forwarding the 733 message downstream. 735 A user agent that receives an obs-fold in a response message that is 736 not within a message/http container MUST replace each received obs- 737 fold with one or more SP octets prior to interpreting the field 738 value. 740 6. Message Body 742 The message body (if any) of an HTTP message is used to carry the 743 payload body of that request or response. The message body is 744 identical to the payload body unless a transfer coding has been 745 applied, as described in Section 6.1. 747 message-body = *OCTET 749 The rules for when a message body is allowed in a message differ for 750 requests and responses. 752 The presence of a message body in a request is signaled by a Content- 753 Length or Transfer-Encoding header field. Request message framing is 754 independent of method semantics, even if the method does not define 755 any use for a message body. 757 The presence of a message body in a response depends on both the 758 request method to which it is responding and the response status code 759 (Section 4). Responses to the HEAD request method (Section 7.3.2 of 760 [Semantics]) never include a message body because the associated 761 response header fields (e.g., Transfer-Encoding, Content-Length, 762 etc.), if present, indicate only what their values would have been if 763 the request method had been GET (Section 7.3.1 of [Semantics]). 2xx 764 (Successful) responses to a CONNECT request method (Section 7.3.6 of 766 [Semantics]) switch to tunnel mode instead of having a message body. 767 All 1xx (Informational), 204 (No Content), and 304 (Not Modified) 768 responses do not include a message body. All other responses do 769 include a message body, although the body might be of zero length. 771 6.1. Transfer-Encoding 773 The Transfer-Encoding header field lists the transfer coding names 774 corresponding to the sequence of transfer codings that have been (or 775 will be) applied to the payload body in order to form the message 776 body. Transfer codings are defined in Section 7. 778 Transfer-Encoding = 1#transfer-coding 780 Transfer-Encoding is analogous to the Content-Transfer-Encoding field 781 of MIME, which was designed to enable safe transport of binary data 782 over a 7-bit transport service ([RFC2045], Section 6). However, safe 783 transport has a different focus for an 8bit-clean transfer protocol. 784 In HTTP's case, Transfer-Encoding is primarily intended to accurately 785 delimit a dynamically generated payload and to distinguish payload 786 encodings that are only applied for transport efficiency or security 787 from those that are characteristics of the selected resource. 789 A recipient MUST be able to parse the chunked transfer coding 790 (Section 7.1) because it plays a crucial role in framing messages 791 when the payload body size is not known in advance. A sender MUST 792 NOT apply chunked more than once to a message body (i.e., chunking an 793 already chunked message is not allowed). If any transfer coding 794 other than chunked is applied to a request payload body, the sender 795 MUST apply chunked as the final transfer coding to ensure that the 796 message is properly framed. If any transfer coding other than 797 chunked is applied to a response payload body, the sender MUST either 798 apply chunked as the final transfer coding or terminate the message 799 by closing the connection. 801 For example, 803 Transfer-Encoding: gzip, chunked 805 indicates that the payload body has been compressed using the gzip 806 coding and then chunked using the chunked coding while forming the 807 message body. 809 Unlike Content-Encoding (Section 6.1.2 of [Semantics]), Transfer- 810 Encoding is a property of the message, not of the representation, and 811 any recipient along the request/response chain MAY decode the 812 received transfer coding(s) or apply additional transfer coding(s) to 813 the message body, assuming that corresponding changes are made to the 814 Transfer-Encoding field-value. Additional information about the 815 encoding parameters can be provided by other header fields not 816 defined by this specification. 818 Transfer-Encoding MAY be sent in a response to a HEAD request or in a 819 304 (Not Modified) response (Section 9.4.5 of [Semantics]) to a GET 820 request, neither of which includes a message body, to indicate that 821 the origin server would have applied a transfer coding to the message 822 body if the request had been an unconditional GET. This indication 823 is not required, however, because any recipient on the response chain 824 (including the origin server) can remove transfer codings when they 825 are not needed. 827 A server MUST NOT send a Transfer-Encoding header field in any 828 response with a status code of 1xx (Informational) or 204 (No 829 Content). A server MUST NOT send a Transfer-Encoding header field in 830 any 2xx (Successful) response to a CONNECT request (Section 7.3.6 of 831 [Semantics]). 833 Transfer-Encoding was added in HTTP/1.1. It is generally assumed 834 that implementations advertising only HTTP/1.0 support will not 835 understand how to process a transfer-encoded payload. A client MUST 836 NOT send a request containing Transfer-Encoding unless it knows the 837 server will handle HTTP/1.1 (or later) requests; such knowledge might 838 be in the form of specific user configuration or by remembering the 839 version of a prior received response. A server MUST NOT send a 840 response containing Transfer-Encoding unless the corresponding 841 request indicates HTTP/1.1 (or later). 843 A server that receives a request message with a transfer coding it 844 does not understand SHOULD respond with 501 (Not Implemented). 846 6.2. Content-Length 848 When a message does not have a Transfer-Encoding header field, a 849 Content-Length header field can provide the anticipated size, as a 850 decimal number of octets, for a potential payload body. For messages 851 that do include a payload body, the Content-Length field-value 852 provides the framing information necessary for determining where the 853 body (and message) ends. For messages that do not include a payload 854 body, the Content-Length indicates the size of the selected 855 representation (Section 6.2.4 of [Semantics]). 857 Note: HTTP's use of Content-Length for message framing differs 858 significantly from the same field's use in MIME, where it is an 859 optional field used only within the "message/external-body" media- 860 type. 862 6.3. Message Body Length 864 The length of a message body is determined by one of the following 865 (in order of precedence): 867 1. Any response to a HEAD request and any response with a 1xx 868 (Informational), 204 (No Content), or 304 (Not Modified) status 869 code is always terminated by the first empty line after the 870 header fields, regardless of the header fields present in the 871 message, and thus cannot contain a message body. 873 2. Any 2xx (Successful) response to a CONNECT request implies that 874 the connection will become a tunnel immediately after the empty 875 line that concludes the header fields. A client MUST ignore any 876 Content-Length or Transfer-Encoding header fields received in 877 such a message. 879 3. If a Transfer-Encoding header field is present and the chunked 880 transfer coding (Section 7.1) is the final encoding, the message 881 body length is determined by reading and decoding the chunked 882 data until the transfer coding indicates the data is complete. 884 If a Transfer-Encoding header field is present in a response and 885 the chunked transfer coding is not the final encoding, the 886 message body length is determined by reading the connection until 887 it is closed by the server. If a Transfer-Encoding header field 888 is present in a request and the chunked transfer coding is not 889 the final encoding, the message body length cannot be determined 890 reliably; the server MUST respond with the 400 (Bad Request) 891 status code and then close the connection. 893 If a message is received with both a Transfer-Encoding and a 894 Content-Length header field, the Transfer-Encoding overrides the 895 Content-Length. Such a message might indicate an attempt to 896 perform request smuggling (Section 11.2) or response splitting 897 (Section 11.1) and ought to be handled as an error. A sender 898 MUST remove the received Content-Length field prior to forwarding 899 such a message downstream. 901 4. If a message is received without Transfer-Encoding and with 902 either multiple Content-Length header fields having differing 903 field-values or a single Content-Length header field having an 904 invalid value, then the message framing is invalid and the 905 recipient MUST treat it as an unrecoverable error. If this is a 906 request message, the server MUST respond with a 400 (Bad Request) 907 status code and then close the connection. If this is a response 908 message received by a proxy, the proxy MUST close the connection 909 to the server, discard the received response, and send a 502 (Bad 910 Gateway) response to the client. If this is a response message 911 received by a user agent, the user agent MUST close the 912 connection to the server and discard the received response. 914 5. If a valid Content-Length header field is present without 915 Transfer-Encoding, its decimal value defines the expected message 916 body length in octets. If the sender closes the connection or 917 the recipient times out before the indicated number of octets are 918 received, the recipient MUST consider the message to be 919 incomplete and close the connection. 921 6. If this is a request message and none of the above are true, then 922 the message body length is zero (no message body is present). 924 7. Otherwise, this is a response message without a declared message 925 body length, so the message body length is determined by the 926 number of octets received prior to the server closing the 927 connection. 929 Since there is no way to distinguish a successfully completed, close- 930 delimited message from a partially received message interrupted by 931 network failure, a server SHOULD generate encoding or length- 932 delimited messages whenever possible. The close-delimiting feature 933 exists primarily for backwards compatibility with HTTP/1.0. 935 A server MAY reject a request that contains a message body but not a 936 Content-Length by responding with 411 (Length Required). 938 Unless a transfer coding other than chunked has been applied, a 939 client that sends a request containing a message body SHOULD use a 940 valid Content-Length header field if the message body length is known 941 in advance, rather than the chunked transfer coding, since some 942 existing services respond to chunked with a 411 (Length Required) 943 status code even though they understand the chunked transfer coding. 944 This is typically because such services are implemented via a gateway 945 that requires a content-length in advance of being called and the 946 server is unable or unwilling to buffer the entire request before 947 processing. 949 A user agent that sends a request containing a message body MUST send 950 a valid Content-Length header field if it does not know the server 951 will handle HTTP/1.1 (or later) requests; such knowledge can be in 952 the form of specific user configuration or by remembering the version 953 of a prior received response. 955 If the final response to the last request on a connection has been 956 completely received and there remains additional data to read, a user 957 agent MAY discard the remaining data or attempt to determine if that 958 data belongs as part of the prior response body, which might be the 959 case if the prior message's Content-Length value is incorrect. A 960 client MUST NOT process, cache, or forward such extra data as a 961 separate response, since such behavior would be vulnerable to cache 962 poisoning. 964 7. Transfer Codings 966 Transfer coding names are used to indicate an encoding transformation 967 that has been, can be, or might need to be applied to a payload body 968 in order to ensure "safe transport" through the network. This 969 differs from a content coding in that the transfer coding is a 970 property of the message rather than a property of the representation 971 that is being transferred. 973 transfer-coding = "chunked" ; Section 7.1 974 / "compress" ; [Semantics], Section 6.1.2.1 975 / "deflate" ; [Semantics], Section 6.1.2.2 976 / "gzip" ; [Semantics], Section 6.1.2.3 977 / transfer-extension 978 transfer-extension = token *( OWS ";" OWS transfer-parameter ) 980 Parameters are in the form of a name=value pair. 982 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 984 All transfer-coding names are case-insensitive and ought to be 985 registered within the HTTP Transfer Coding registry, as defined in 986 Section 7.3. They are used in the TE (Section 7.4) and Transfer- 987 Encoding (Section 6.1) header fields. 989 +------------+------------------------------------------+-----------+ 990 | Name | Description | Reference | 991 +------------+------------------------------------------+-----------+ 992 | chunked | Transfer in a series of chunks | Section 7 | 993 | | | .1 | 994 | compress | UNIX "compress" data format [Welch] | Section 7 | 995 | | | .2 | 996 | deflate | "deflate" compressed data ([RFC1951]) | Section 7 | 997 | | inside the "zlib" data format | .2 | 998 | | ([RFC1950]) | | 999 | gzip | GZIP file format [RFC1952] | Section 7 | 1000 | | | .2 | 1001 | trailers | (reserved) | Section 7 | 1002 | x-compress | Deprecated (alias for compress) | Section 7 | 1003 | | | .2 | 1004 | x-gzip | Deprecated (alias for gzip) | Section 7 | 1005 | | | .2 | 1006 +------------+------------------------------------------+-----------+ 1008 Note: the coding name "trailers" is reserved because it would 1009 clash with the use of the keyword "trailers" in the TE header 1010 field (Section 7.4). 1012 7.1. Chunked Transfer Coding 1014 The chunked transfer coding wraps the payload body in order to 1015 transfer it as a series of chunks, each with its own size indicator, 1016 followed by an OPTIONAL trailer containing header fields. Chunked 1017 enables content streams of unknown size to be transferred as a 1018 sequence of length-delimited buffers, which enables the sender to 1019 retain connection persistence and the recipient to know when it has 1020 received the entire message. 1022 chunked-body = *chunk 1023 last-chunk 1024 trailer-part 1025 CRLF 1027 chunk = chunk-size [ chunk-ext ] CRLF 1028 chunk-data CRLF 1029 chunk-size = 1*HEXDIG 1030 last-chunk = 1*("0") [ chunk-ext ] CRLF 1032 chunk-data = 1*OCTET ; a sequence of chunk-size octets 1034 The chunk-size field is a string of hex digits indicating the size of 1035 the chunk-data in octets. The chunked transfer coding is complete 1036 when a chunk with a chunk-size of zero is received, possibly followed 1037 by a trailer, and finally terminated by an empty line. 1039 A recipient MUST be able to parse and decode the chunked transfer 1040 coding. 1042 7.1.1. Chunk Extensions 1044 The chunked encoding allows each chunk to include zero or more chunk 1045 extensions, immediately following the chunk-size, for the sake of 1046 supplying per-chunk metadata (such as a signature or hash), mid- 1047 message control information, or randomization of message body size. 1049 chunk-ext = *( BWS ";" BWS chunk-ext-name 1050 [ BWS "=" BWS chunk-ext-val ] ) 1052 chunk-ext-name = token 1053 chunk-ext-val = token / quoted-string 1055 The chunked encoding is specific to each connection and is likely to 1056 be removed or recoded by each recipient (including intermediaries) 1057 before any higher-level application would have a chance to inspect 1058 the extensions. Hence, use of chunk extensions is generally limited 1059 to specialized HTTP services such as "long polling" (where client and 1060 server can have shared expectations regarding the use of chunk 1061 extensions) or for padding within an end-to-end secured connection. 1063 A recipient MUST ignore unrecognized chunk extensions. A server 1064 ought to limit the total length of chunk extensions received in a 1065 request to an amount reasonable for the services provided, in the 1066 same way that it applies length limitations and timeouts for other 1067 parts of a message, and generate an appropriate 4xx (Client Error) 1068 response if that amount is exceeded. 1070 7.1.2. Chunked Trailer Part 1072 A trailer allows the sender to include additional fields at the end 1073 of a chunked message in order to supply metadata that might be 1074 dynamically generated while the message body is sent, such as a 1075 message integrity check, digital signature, or post-processing 1076 status. The trailer fields are identical to header fields, except 1077 they are sent in a chunked trailer instead of the message's header 1078 section. 1080 trailer-part = *( header-field CRLF ) 1082 A sender MUST NOT generate a trailer that contains a field necessary 1083 for message framing (e.g., Transfer-Encoding and Content-Length), 1084 routing (e.g., Host), request modifiers (e.g., controls and 1085 conditionals in Section 8 of [Semantics]), authentication (e.g., see 1086 Section 8.5 of [Semantics] and [RFC6265]), response control data 1087 (e.g., see Section 10.1 of [Semantics]), or determining how to 1088 process the payload (e.g., Content-Encoding, Content-Type, Content- 1089 Range, and Trailer). 1091 When a chunked message containing a non-empty trailer is received, 1092 the recipient MAY process the fields (aside from those forbidden 1093 above) as if they were appended to the message's header section. A 1094 recipient MUST ignore (or consider as an error) any fields that are 1095 forbidden to be sent in a trailer, since processing them as if they 1096 were present in the header section might bypass external security 1097 filters. 1099 Unless the request includes a TE header field indicating "trailers" 1100 is acceptable, as described in Section 7.4, a server SHOULD NOT 1101 generate trailer fields that it believes are necessary for the user 1102 agent to receive. Without a TE containing "trailers", the server 1103 ought to assume that the trailer fields might be silently discarded 1104 along the path to the user agent. This requirement allows 1105 intermediaries to forward a de-chunked message to an HTTP/1.0 1106 recipient without buffering the entire response. 1108 When a message includes a message body encoded with the chunked 1109 transfer coding and the sender desires to send metadata in the form 1110 of trailer fields at the end of the message, the sender SHOULD 1111 generate a Trailer header field before the message body to indicate 1112 which fields will be present in the trailers. This allows the 1113 recipient to prepare for receipt of that metadata before it starts 1114 processing the body, which is useful if the message is being streamed 1115 and the recipient wishes to confirm an integrity check on the fly. 1117 7.1.3. Decoding Chunked 1119 A process for decoding the chunked transfer coding can be represented 1120 in pseudo-code as: 1122 length := 0 1123 read chunk-size, chunk-ext (if any), and CRLF 1124 while (chunk-size > 0) { 1125 read chunk-data and CRLF 1126 append chunk-data to decoded-body 1127 length := length + chunk-size 1128 read chunk-size, chunk-ext (if any), and CRLF 1129 } 1130 read trailer field 1131 while (trailer field is not empty) { 1132 if (trailer field is allowed to be sent in a trailer) { 1133 append trailer field to existing header fields 1134 } 1135 read trailer-field 1136 } 1137 Content-Length := length 1138 Remove "chunked" from Transfer-Encoding 1139 Remove Trailer from existing header fields 1141 7.2. Transfer Codings for Compression 1143 The following transfer coding names for compression are defined by 1144 the same algorithm as their corresponding content coding: 1146 compress (and x-compress) 1147 See Section 6.1.2.1 of [Semantics]. 1149 deflate 1150 See Section 6.1.2.2 of [Semantics]. 1152 gzip (and x-gzip) 1153 See Section 6.1.2.3 of [Semantics]. 1155 7.3. Transfer Coding Registry 1157 The "HTTP Transfer Coding Registry" defines the namespace for 1158 transfer coding names. It is maintained at 1159 . 1161 Registrations MUST include the following fields: 1163 o Name 1165 o Description 1167 o Pointer to specification text 1168 Names of transfer codings MUST NOT overlap with names of content 1169 codings (Section 6.1.2 of [Semantics]) unless the encoding 1170 transformation is identical, as is the case for the compression 1171 codings defined in Section 7.2. 1173 The TE header field (Section 7.4) uses a pseudo parameter named "q" 1174 as rank value when multiple transfer codings are acceptable. Future 1175 registrations of transfer codings SHOULD NOT define parameters called 1176 "q" (case-insensitively) in order to avoid ambiguities. 1178 Values to be added to this namespace require IETF Review (see 1179 Section 4.8 of [RFC8126]), and MUST conform to the purpose of 1180 transfer coding defined in this specification. 1182 Use of program names for the identification of encoding formats is 1183 not desirable and is discouraged for future encodings. 1185 7.4. TE 1187 The "TE" header field in a request indicates what transfer codings, 1188 besides chunked, the client is willing to accept in response, and 1189 whether or not the client is willing to accept trailer fields in a 1190 chunked transfer coding. 1192 The TE field-value consists of a comma-separated list of transfer 1193 coding names, each allowing for optional parameters (as described in 1194 Section 7), and/or the keyword "trailers". A client MUST NOT send 1195 the chunked transfer coding name in TE; chunked is always acceptable 1196 for HTTP/1.1 recipients. 1198 TE = #t-codings 1199 t-codings = "trailers" / ( transfer-coding [ t-ranking ] ) 1200 t-ranking = OWS ";" OWS "q=" rank 1201 rank = ( "0" [ "." 0*3DIGIT ] ) 1202 / ( "1" [ "." 0*3("0") ] ) 1204 Three examples of TE use are below. 1206 TE: deflate 1207 TE: 1208 TE: trailers, deflate;q=0.5 1210 The presence of the keyword "trailers" indicates that the client is 1211 willing to accept trailer fields in a chunked transfer coding, as 1212 defined in Section 7.1.2, on behalf of itself and any downstream 1213 clients. For requests from an intermediary, this implies that 1214 either: (a) all downstream clients are willing to accept trailer 1215 fields in the forwarded response; or, (b) the intermediary will 1216 attempt to buffer the response on behalf of downstream recipients. 1217 Note that HTTP/1.1 does not define any means to limit the size of a 1218 chunked response such that an intermediary can be assured of 1219 buffering the entire response. 1221 When multiple transfer codings are acceptable, the client MAY rank 1222 the codings by preference using a case-insensitive "q" parameter 1223 (similar to the qvalues used in content negotiation fields, 1224 Section 8.4.1 of [Semantics]). The rank value is a real number in 1225 the range 0 through 1, where 0.001 is the least preferred and 1 is 1226 the most preferred; a value of 0 means "not acceptable". 1228 If the TE field-value is empty or if no TE field is present, the only 1229 acceptable transfer coding is chunked. A message with no transfer 1230 coding is always acceptable. 1232 Since the TE header field only applies to the immediate connection, a 1233 sender of TE MUST also send a "TE" connection option within the 1234 Connection header field (Section 9.1) in order to prevent the TE 1235 field from being forwarded by intermediaries that do not support its 1236 semantics. 1238 8. Handling Incomplete Messages 1240 A server that receives an incomplete request message, usually due to 1241 a canceled request or a triggered timeout exception, MAY send an 1242 error response prior to closing the connection. 1244 A client that receives an incomplete response message, which can 1245 occur when a connection is closed prematurely or when decoding a 1246 supposedly chunked transfer coding fails, MUST record the message as 1247 incomplete. Cache requirements for incomplete responses are defined 1248 in Section 3 of [Caching]. 1250 If a response terminates in the middle of the header section (before 1251 the empty line is received) and the status code might rely on header 1252 fields to convey the full meaning of the response, then the client 1253 cannot assume that meaning has been conveyed; the client might need 1254 to repeat the request in order to determine what action to take next. 1256 A message body that uses the chunked transfer coding is incomplete if 1257 the zero-sized chunk that terminates the encoding has not been 1258 received. A message that uses a valid Content-Length is incomplete 1259 if the size of the message body received (in octets) is less than the 1260 value given by Content-Length. A response that has neither chunked 1261 transfer coding nor Content-Length is terminated by closure of the 1262 connection and, thus, is considered complete regardless of the number 1263 of message body octets received, provided that the header section was 1264 received intact. 1266 9. Connection Management 1268 HTTP messaging is independent of the underlying transport- or 1269 session-layer connection protocol(s). HTTP only presumes a reliable 1270 transport with in-order delivery of requests and the corresponding 1271 in-order delivery of responses. The mapping of HTTP request and 1272 response structures onto the data units of an underlying transport 1273 protocol is outside the scope of this specification. 1275 As described in Section 5.2 of [Semantics], the specific connection 1276 protocols to be used for an HTTP interaction are determined by client 1277 configuration and the target URI. For example, the "http" URI scheme 1278 (Section 2.5.1 of [Semantics]) indicates a default connection of TCP 1279 over IP, with a default TCP port of 80, but the client might be 1280 configured to use a proxy via some other connection, port, or 1281 protocol. 1283 HTTP implementations are expected to engage in connection management, 1284 which includes maintaining the state of current connections, 1285 establishing a new connection or reusing an existing connection, 1286 processing messages received on a connection, detecting connection 1287 failures, and closing each connection. Most clients maintain 1288 multiple connections in parallel, including more than one connection 1289 per server endpoint. Most servers are designed to maintain thousands 1290 of concurrent connections, while controlling request queues to enable 1291 fair use and detect denial-of-service attacks. 1293 9.1. Connection 1295 The "Connection" header field allows the sender to indicate desired 1296 control options for the current connection. In order to avoid 1297 confusing downstream recipients, a proxy or gateway MUST remove or 1298 replace any received connection options before forwarding the 1299 message. 1301 When a header field aside from Connection is used to supply control 1302 information for or about the current connection, the sender MUST list 1303 the corresponding field-name within the Connection header field. A 1304 proxy or gateway MUST parse a received Connection header field before 1305 a message is forwarded and, for each connection-option in this field, 1306 remove any header field(s) from the message with the same name as the 1307 connection-option, and then remove the Connection header field itself 1308 (or replace it with the intermediary's own connection options for the 1309 forwarded message). 1311 Hence, the Connection header field provides a declarative way of 1312 distinguishing header fields that are only intended for the immediate 1313 recipient ("hop-by-hop") from those fields that are intended for all 1314 recipients on the chain ("end-to-end"), enabling the message to be 1315 self-descriptive and allowing future connection-specific extensions 1316 to be deployed without fear that they will be blindly forwarded by 1317 older intermediaries. 1319 The Connection header field's value has the following grammar: 1321 Connection = 1#connection-option 1322 connection-option = token 1324 Connection options are case-insensitive. 1326 A sender MUST NOT send a connection option corresponding to a header 1327 field that is intended for all recipients of the payload. For 1328 example, Cache-Control is never appropriate as a connection option 1329 (Section 5.2 of [Caching]). 1331 The connection options do not always correspond to a header field 1332 present in the message, since a connection-specific header field 1333 might not be needed if there are no parameters associated with a 1334 connection option. In contrast, a connection-specific header field 1335 that is received without a corresponding connection option usually 1336 indicates that the field has been improperly forwarded by an 1337 intermediary and ought to be ignored by the recipient. 1339 When defining new connection options, specification authors ought to 1340 survey existing header field names and ensure that the new connection 1341 option does not share the same name as an already deployed header 1342 field. Defining a new connection option essentially reserves that 1343 potential field-name for carrying additional information related to 1344 the connection option, since it would be unwise for senders to use 1345 that field-name for anything else. 1347 The "close" connection option is defined for a sender to signal that 1348 this connection will be closed after completion of the response. For 1349 example, 1351 Connection: close 1353 in either the request or the response header fields indicates that 1354 the sender is going to close the connection after the current 1355 request/response is complete (Section 9.7). 1357 A client that does not support persistent connections MUST send the 1358 "close" connection option in every request message. 1360 A server that does not support persistent connections MUST send the 1361 "close" connection option in every response message that does not 1362 have a 1xx (Informational) status code. 1364 9.2. Establishment 1366 It is beyond the scope of this specification to describe how 1367 connections are established via various transport- or session-layer 1368 protocols. Each connection applies to only one transport link. 1370 9.3. Associating a Response to a Request 1372 HTTP/1.1 does not include a request identifier for associating a 1373 given request message with its corresponding one or more response 1374 messages. Hence, it relies on the order of response arrival to 1375 correspond exactly to the order in which requests are made on the 1376 same connection. More than one response message per request only 1377 occurs when one or more informational responses (1xx, see Section 9.2 1378 of [Semantics]) precede a final response to the same request. 1380 A client that has more than one outstanding request on a connection 1381 MUST maintain a list of outstanding requests in the order sent and 1382 MUST associate each received response message on that connection to 1383 the highest ordered request that has not yet received a final (non- 1384 1xx) response. 1386 9.4. Persistence 1388 HTTP/1.1 defaults to the use of "persistent connections", allowing 1389 multiple requests and responses to be carried over a single 1390 connection. The "close" connection option is used to signal that a 1391 connection will not persist after the current request/response. HTTP 1392 implementations SHOULD support persistent connections. 1394 A recipient determines whether a connection is persistent or not 1395 based on the most recently received message's protocol version and 1396 Connection header field (if any): 1398 o If the "close" connection option is present, the connection will 1399 not persist after the current response; else, 1401 o If the received protocol is HTTP/1.1 (or later), the connection 1402 will persist after the current response; else, 1404 o If the received protocol is HTTP/1.0, the "keep-alive" connection 1405 option is present, either the recipient is not a proxy or the 1406 message is a response, and the recipient wishes to honor the 1407 HTTP/1.0 "keep-alive" mechanism, the connection will persist after 1408 the current response; otherwise, 1410 o The connection will close after the current response. 1412 A client MAY send additional requests on a persistent connection 1413 until it sends or receives a "close" connection option or receives an 1414 HTTP/1.0 response without a "keep-alive" connection option. 1416 In order to remain persistent, all messages on a connection need to 1417 have a self-defined message length (i.e., one not defined by closure 1418 of the connection), as described in Section 6. A server MUST read 1419 the entire request message body or close the connection after sending 1420 its response, since otherwise the remaining data on a persistent 1421 connection would be misinterpreted as the next request. Likewise, a 1422 client MUST read the entire response message body if it intends to 1423 reuse the same connection for a subsequent request. 1425 A proxy server MUST NOT maintain a persistent connection with an 1426 HTTP/1.0 client (see Section 19.7.1 of [RFC2068] for information and 1427 discussion of the problems with the Keep-Alive header field 1428 implemented by many HTTP/1.0 clients). 1430 See Appendix C.1.2 for more information on backwards compatibility 1431 with HTTP/1.0 clients. 1433 9.4.1. Retrying Requests 1435 Connections can be closed at any time, with or without intention. 1436 Implementations ought to anticipate the need to recover from 1437 asynchronous close events. 1439 When an inbound connection is closed prematurely, a client MAY open a 1440 new connection and automatically retransmit an aborted sequence of 1441 requests if all of those requests have idempotent methods 1442 (Section 7.2.2 of [Semantics]). A proxy MUST NOT automatically retry 1443 non-idempotent requests. 1445 A user agent MUST NOT automatically retry a request with a non- 1446 idempotent method unless it has some means to know that the request 1447 semantics are actually idempotent, regardless of the method, or some 1448 means to detect that the original request was never applied. For 1449 example, a user agent that knows (through design or configuration) 1450 that a POST request to a given resource is safe can repeat that 1451 request automatically. Likewise, a user agent designed specifically 1452 to operate on a version control repository might be able to recover 1453 from partial failure conditions by checking the target resource 1454 revision(s) after a failed connection, reverting or fixing any 1455 changes that were partially applied, and then automatically retrying 1456 the requests that failed. 1458 A client SHOULD NOT automatically retry a failed automatic retry. 1460 9.4.2. Pipelining 1462 A client that supports persistent connections MAY "pipeline" its 1463 requests (i.e., send multiple requests without waiting for each 1464 response). A server MAY process a sequence of pipelined requests in 1465 parallel if they all have safe methods (Section 7.2.1 of 1466 [Semantics]), but it MUST send the corresponding responses in the 1467 same order that the requests were received. 1469 A client that pipelines requests SHOULD retry unanswered requests if 1470 the connection closes before it receives all of the corresponding 1471 responses. When retrying pipelined requests after a failed 1472 connection (a connection not explicitly closed by the server in its 1473 last complete response), a client MUST NOT pipeline immediately after 1474 connection establishment, since the first remaining request in the 1475 prior pipeline might have caused an error response that can be lost 1476 again if multiple requests are sent on a prematurely closed 1477 connection (see the TCP reset problem described in Section 9.7). 1479 Idempotent methods (Section 7.2.2 of [Semantics]) are significant to 1480 pipelining because they can be automatically retried after a 1481 connection failure. A user agent SHOULD NOT pipeline requests after 1482 a non-idempotent method, until the final response status code for 1483 that method has been received, unless the user agent has a means to 1484 detect and recover from partial failure conditions involving the 1485 pipelined sequence. 1487 An intermediary that receives pipelined requests MAY pipeline those 1488 requests when forwarding them inbound, since it can rely on the 1489 outbound user agent(s) to determine what requests can be safely 1490 pipelined. If the inbound connection fails before receiving a 1491 response, the pipelining intermediary MAY attempt to retry a sequence 1492 of requests that have yet to receive a response if the requests all 1493 have idempotent methods; otherwise, the pipelining intermediary 1494 SHOULD forward any received responses and then close the 1495 corresponding outbound connection(s) so that the outbound user 1496 agent(s) can recover accordingly. 1498 9.5. Concurrency 1500 A client ought to limit the number of simultaneous open connections 1501 that it maintains to a given server. 1503 Previous revisions of HTTP gave a specific number of connections as a 1504 ceiling, but this was found to be impractical for many applications. 1505 As a result, this specification does not mandate a particular maximum 1506 number of connections but, instead, encourages clients to be 1507 conservative when opening multiple connections. 1509 Multiple connections are typically used to avoid the "head-of-line 1510 blocking" problem, wherein a request that takes significant server- 1511 side processing and/or has a large payload blocks subsequent requests 1512 on the same connection. However, each connection consumes server 1513 resources. Furthermore, using multiple connections can cause 1514 undesirable side effects in congested networks. 1516 Note that a server might reject traffic that it deems abusive or 1517 characteristic of a denial-of-service attack, such as an excessive 1518 number of open connections from a single client. 1520 9.6. Failures and Timeouts 1522 Servers will usually have some timeout value beyond which they will 1523 no longer maintain an inactive connection. Proxy servers might make 1524 this a higher value since it is likely that the client will be making 1525 more connections through the same proxy server. The use of 1526 persistent connections places no requirements on the length (or 1527 existence) of this timeout for either the client or the server. 1529 A client or server that wishes to time out SHOULD issue a graceful 1530 close on the connection. Implementations SHOULD constantly monitor 1531 open connections for a received closure signal and respond to it as 1532 appropriate, since prompt closure of both sides of a connection 1533 enables allocated system resources to be reclaimed. 1535 A client, server, or proxy MAY close the transport connection at any 1536 time. For example, a client might have started to send a new request 1537 at the same time that the server has decided to close the "idle" 1538 connection. From the server's point of view, the connection is being 1539 closed while it was idle, but from the client's point of view, a 1540 request is in progress. 1542 A server SHOULD sustain persistent connections, when possible, and 1543 allow the underlying transport's flow-control mechanisms to resolve 1544 temporary overloads, rather than terminate connections with the 1545 expectation that clients will retry. The latter technique can 1546 exacerbate network congestion. 1548 A client sending a message body SHOULD monitor the network connection 1549 for an error response while it is transmitting the request. If the 1550 client sees a response that indicates the server does not wish to 1551 receive the message body and is closing the connection, the client 1552 SHOULD immediately cease transmitting the body and close its side of 1553 the connection. 1555 9.7. Tear-down 1557 The Connection header field (Section 9.1) provides a "close" 1558 connection option that a sender SHOULD send when it wishes to close 1559 the connection after the current request/response pair. 1561 A client that sends a "close" connection option MUST NOT send further 1562 requests on that connection (after the one containing "close") and 1563 MUST close the connection after reading the final response message 1564 corresponding to this request. 1566 A server that receives a "close" connection option MUST initiate a 1567 close of the connection (see below) after it sends the final response 1568 to the request that contained "close". The server SHOULD send a 1569 "close" connection option in its final response on that connection. 1570 The server MUST NOT process any further requests received on that 1571 connection. 1573 A server that sends a "close" connection option MUST initiate a close 1574 of the connection (see below) after it sends the response containing 1575 "close". The server MUST NOT process any further requests received 1576 on that connection. 1578 A client that receives a "close" connection option MUST cease sending 1579 requests on that connection and close the connection after reading 1580 the response message containing the "close"; if additional pipelined 1581 requests had been sent on the connection, the client SHOULD NOT 1582 assume that they will be processed by the server. 1584 If a server performs an immediate close of a TCP connection, there is 1585 a significant risk that the client will not be able to read the last 1586 HTTP response. If the server receives additional data from the 1587 client on a fully closed connection, such as another request that was 1588 sent by the client before receiving the server's response, the 1589 server's TCP stack will send a reset packet to the client; 1590 unfortunately, the reset packet might erase the client's 1591 unacknowledged input buffers before they can be read and interpreted 1592 by the client's HTTP parser. 1594 To avoid the TCP reset problem, servers typically close a connection 1595 in stages. First, the server performs a half-close by closing only 1596 the write side of the read/write connection. The server then 1597 continues to read from the connection until it receives a 1598 corresponding close by the client, or until the server is reasonably 1599 certain that its own TCP stack has received the client's 1600 acknowledgement of the packet(s) containing the server's last 1601 response. Finally, the server fully closes the connection. 1603 It is unknown whether the reset problem is exclusive to TCP or might 1604 also be found in other transport connection protocols. 1606 9.8. Upgrade 1608 The "Upgrade" header field is intended to provide a simple mechanism 1609 for transitioning from HTTP/1.1 to some other protocol on the same 1610 connection. A client MAY send a list of protocols in the Upgrade 1611 header field of a request to invite the server to switch to one or 1612 more of those protocols, in order of descending preference, before 1613 sending the final response. A server MAY ignore a received Upgrade 1614 header field if it wishes to continue using the current protocol on 1615 that connection. Upgrade cannot be used to insist on a protocol 1616 change. 1618 Upgrade = 1#protocol 1620 protocol = protocol-name ["/" protocol-version] 1621 protocol-name = token 1622 protocol-version = token 1624 A server that sends a 101 (Switching Protocols) response MUST send an 1625 Upgrade header field to indicate the new protocol(s) to which the 1626 connection is being switched; if multiple protocol layers are being 1627 switched, the sender MUST list the protocols in layer-ascending 1628 order. A server MUST NOT switch to a protocol that was not indicated 1629 by the client in the corresponding request's Upgrade header field. A 1630 server MAY choose to ignore the order of preference indicated by the 1631 client and select the new protocol(s) based on other factors, such as 1632 the nature of the request or the current load on the server. 1634 A server that sends a 426 (Upgrade Required) response MUST send an 1635 Upgrade header field to indicate the acceptable protocols, in order 1636 of descending preference. 1638 A server MAY send an Upgrade header field in any other response to 1639 advertise that it implements support for upgrading to the listed 1640 protocols, in order of descending preference, when appropriate for a 1641 future request. 1643 The following is a hypothetical example sent by a client: 1645 GET /hello.txt HTTP/1.1 1646 Host: www.example.com 1647 Connection: upgrade 1648 Upgrade: HTTP/2.0, SHTTP/1.3, IRC/6.9, RTA/x11 1650 The capabilities and nature of the application-level communication 1651 after the protocol change is entirely dependent upon the new 1652 protocol(s) chosen. However, immediately after sending the 101 1653 (Switching Protocols) response, the server is expected to continue 1654 responding to the original request as if it had received its 1655 equivalent within the new protocol (i.e., the server still has an 1656 outstanding request to satisfy after the protocol has been changed, 1657 and is expected to do so without requiring the request to be 1658 repeated). 1660 For example, if the Upgrade header field is received in a GET request 1661 and the server decides to switch protocols, it first responds with a 1662 101 (Switching Protocols) message in HTTP/1.1 and then immediately 1663 follows that with the new protocol's equivalent of a response to a 1664 GET on the target resource. This allows a connection to be upgraded 1665 to protocols with the same semantics as HTTP without the latency cost 1666 of an additional round trip. A server MUST NOT switch protocols 1667 unless the received message semantics can be honored by the new 1668 protocol; an OPTIONS request can be honored by any protocol. 1670 The following is an example response to the above hypothetical 1671 request: 1673 HTTP/1.1 101 Switching Protocols 1674 Connection: upgrade 1675 Upgrade: HTTP/2.0 1677 [... data stream switches to HTTP/2.0 with an appropriate response 1678 (as defined by new protocol) to the "GET /hello.txt" request ...] 1680 When Upgrade is sent, the sender MUST also send a Connection header 1681 field (Section 9.1) that contains an "upgrade" connection option, in 1682 order to prevent Upgrade from being accidentally forwarded by 1683 intermediaries that might not implement the listed protocols. A 1684 server MUST ignore an Upgrade header field that is received in an 1685 HTTP/1.0 request. 1687 A client cannot begin using an upgraded protocol on the connection 1688 until it has completely sent the request message (i.e., the client 1689 can't change the protocol it is sending in the middle of a message). 1691 If a server receives both an Upgrade and an Expect header field with 1692 the "100-continue" expectation (Section 8.1.1 of [Semantics]), the 1693 server MUST send a 100 (Continue) response before sending a 101 1694 (Switching Protocols) response. 1696 The Upgrade header field only applies to switching protocols on top 1697 of the existing connection; it cannot be used to switch the 1698 underlying connection (transport) protocol, nor to switch the 1699 existing communication to a different connection. For those 1700 purposes, it is more appropriate to use a 3xx (Redirection) response 1701 (Section 9.4 of [Semantics]). 1703 9.8.1. Upgrade Protocol Names 1705 This specification only defines the protocol name "HTTP" for use by 1706 the family of Hypertext Transfer Protocols, as defined by the HTTP 1707 version rules of Section 3.5 of [Semantics] and future updates to 1708 this specification. Additional protocol names ought to be registered 1709 using the registration procedure defined in Section 9.8.2. 1711 +------+-------------------+--------------------+-------------------+ 1712 | Name | Description | Expected Version | Reference | 1713 | | | Tokens | | 1714 +------+-------------------+--------------------+-------------------+ 1715 | HTTP | Hypertext | any DIGIT.DIGIT | Section 3.5 of | 1716 | | Transfer Protocol | (e.g, "2.0") | [Semantics] | 1717 +------+-------------------+--------------------+-------------------+ 1719 9.8.2. Upgrade Token Registry 1721 The "Hypertext Transfer Protocol (HTTP) Upgrade Token Registry" 1722 defines the namespace for protocol-name tokens used to identify 1723 protocols in the Upgrade header field. The registry is maintained at 1724 . 1726 Each registered protocol name is associated with contact information 1727 and an optional set of specifications that details how the connection 1728 will be processed after it has been upgraded. 1730 Registrations happen on a "First Come First Served" basis (see 1731 Section 4.4 of [RFC8126]) and are subject to the following rules: 1733 1. A protocol-name token, once registered, stays registered forever. 1735 2. The registration MUST name a responsible party for the 1736 registration. 1738 3. The registration MUST name a point of contact. 1740 4. The registration MAY name a set of specifications associated with 1741 that token. Such specifications need not be publicly available. 1743 5. The registration SHOULD name a set of expected "protocol-version" 1744 tokens associated with that token at the time of registration. 1746 6. The responsible party MAY change the registration at any time. 1747 The IANA will keep a record of all such changes, and make them 1748 available upon request. 1750 7. The IESG MAY reassign responsibility for a protocol token. This 1751 will normally only be used in the case when a responsible party 1752 cannot be contacted. 1754 10. Enclosing Messages as Data 1756 10.1. Media Type message/http 1758 The message/http media type can be used to enclose a single HTTP 1759 request or response message, provided that it obeys the MIME 1760 restrictions for all "message" types regarding line length and 1761 encodings. 1763 Type name: message 1765 Subtype name: http 1767 Required parameters: N/A 1769 Optional parameters: version, msgtype 1771 version: The HTTP-version number of the enclosed message (e.g., 1772 "1.1"). If not present, the version can be determined from the 1773 first line of the body. 1775 msgtype: The message type -- "request" or "response". If not 1776 present, the type can be determined from the first line of the 1777 body. 1779 Encoding considerations: only "7bit", "8bit", or "binary" are 1780 permitted 1782 Security considerations: see Section 11 1784 Interoperability considerations: N/A 1786 Published specification: This specification (see Section 10.1). 1788 Applications that use this media type: N/A 1790 Fragment identifier considerations: N/A 1792 Additional information: 1794 Magic number(s): N/A 1796 Deprecated alias names for this type: N/A 1798 File extension(s): N/A 1800 Macintosh file type code(s): N/A 1802 Person and email address to contact for further information: 1803 See Authors' Addresses section. 1805 Intended usage: COMMON 1807 Restrictions on usage: N/A 1809 Author: See Authors' Addresses section. 1811 Change controller: IESG 1813 10.2. Media Type application/http 1815 The application/http media type can be used to enclose a pipeline of 1816 one or more HTTP request or response messages (not intermixed). 1818 Type name: application 1820 Subtype name: http 1822 Required parameters: N/A 1824 Optional parameters: version, msgtype 1826 version: The HTTP-version number of the enclosed messages (e.g., 1827 "1.1"). If not present, the version can be determined from the 1828 first line of the body. 1830 msgtype: The message type -- "request" or "response". If not 1831 present, the type can be determined from the first line of the 1832 body. 1834 Encoding considerations: HTTP messages enclosed by this type are in 1835 "binary" format; use of an appropriate Content-Transfer-Encoding 1836 is required when transmitted via email. 1838 Security considerations: see Section 11 1840 Interoperability considerations: N/A 1842 Published specification: This specification (see Section 10.2). 1844 Applications that use this media type: N/A 1846 Fragment identifier considerations: N/A 1848 Additional information: 1850 Deprecated alias names for this type: N/A 1852 Magic number(s): N/A 1854 File extension(s): N/A 1856 Macintosh file type code(s): N/A 1858 Person and email address to contact for further information: 1859 See Authors' Addresses section. 1861 Intended usage: COMMON 1863 Restrictions on usage: N/A 1865 Author: See Authors' Addresses section. 1867 Change controller: IESG 1869 11. Security Considerations 1871 This section is meant to inform developers, information providers, 1872 and users of known security considerations relevant to HTTP message 1873 syntax, parsing, and routing. Security considerations about HTTP 1874 semantics and payloads are addressed in [Semantics]. 1876 11.1. Response Splitting 1878 Response splitting (a.k.a, CRLF injection) is a common technique, 1879 used in various attacks on Web usage, that exploits the line-based 1880 nature of HTTP message framing and the ordered association of 1881 requests to responses on persistent connections [Klein]. This 1882 technique can be particularly damaging when the requests pass through 1883 a shared cache. 1885 Response splitting exploits a vulnerability in servers (usually 1886 within an application server) where an attacker can send encoded data 1887 within some parameter of the request that is later decoded and echoed 1888 within any of the response header fields of the response. If the 1889 decoded data is crafted to look like the response has ended and a 1890 subsequent response has begun, the response has been split and the 1891 content within the apparent second response is controlled by the 1892 attacker. The attacker can then make any other request on the same 1893 persistent connection and trick the recipients (including 1894 intermediaries) into believing that the second half of the split is 1895 an authoritative answer to the second request. 1897 For example, a parameter within the request-target might be read by 1898 an application server and reused within a redirect, resulting in the 1899 same parameter being echoed in the Location header field of the 1900 response. If the parameter is decoded by the application and not 1901 properly encoded when placed in the response field, the attacker can 1902 send encoded CRLF octets and other content that will make the 1903 application's single response look like two or more responses. 1905 A common defense against response splitting is to filter requests for 1906 data that looks like encoded CR and LF (e.g., "%0D" and "%0A"). 1907 However, that assumes the application server is only performing URI 1908 decoding, rather than more obscure data transformations like charset 1909 transcoding, XML entity translation, base64 decoding, sprintf 1910 reformatting, etc. A more effective mitigation is to prevent 1911 anything other than the server's core protocol libraries from sending 1912 a CR or LF within the header section, which means restricting the 1913 output of header fields to APIs that filter for bad octets and not 1914 allowing application servers to write directly to the protocol 1915 stream. 1917 11.2. Request Smuggling 1919 Request smuggling ([Linhart]) is a technique that exploits 1920 differences in protocol parsing among various recipients to hide 1921 additional requests (which might otherwise be blocked or disabled by 1922 policy) within an apparently harmless request. Like response 1923 splitting, request smuggling can lead to a variety of attacks on HTTP 1924 usage. 1926 This specification has introduced new requirements on request 1927 parsing, particularly with regard to message framing in Section 6.3, 1928 to reduce the effectiveness of request smuggling. 1930 11.3. Message Integrity 1932 HTTP does not define a specific mechanism for ensuring message 1933 integrity, instead relying on the error-detection ability of 1934 underlying transport protocols and the use of length or chunk- 1935 delimited framing to detect completeness. Additional integrity 1936 mechanisms, such as hash functions or digital signatures applied to 1937 the content, can be selectively added to messages via extensible 1938 metadata header fields. Historically, the lack of a single integrity 1939 mechanism has been justified by the informal nature of most HTTP 1940 communication. However, the prevalence of HTTP as an information 1941 access mechanism has resulted in its increasing use within 1942 environments where verification of message integrity is crucial. 1944 User agents are encouraged to implement configurable means for 1945 detecting and reporting failures of message integrity such that those 1946 means can be enabled within environments for which integrity is 1947 necessary. For example, a browser being used to view medical history 1948 or drug interaction information needs to indicate to the user when 1949 such information is detected by the protocol to be incomplete, 1950 expired, or corrupted during transfer. Such mechanisms might be 1951 selectively enabled via user agent extensions or the presence of 1952 message integrity metadata in a response. At a minimum, user agents 1953 ought to provide some indication that allows a user to distinguish 1954 between a complete and incomplete response message (Section 8) when 1955 such verification is desired. 1957 11.4. Message Confidentiality 1959 HTTP relies on underlying transport protocols to provide message 1960 confidentiality when that is desired. HTTP has been specifically 1961 designed to be independent of the transport protocol, such that it 1962 can be used over many different forms of encrypted connection, with 1963 the selection of such transports being identified by the choice of 1964 URI scheme or within user agent configuration. 1966 The "https" scheme can be used to identify resources that require a 1967 confidential connection, as described in Section 2.5.2 of 1968 [Semantics]. 1970 12. IANA Considerations 1972 The change controller for the following registrations is: "IETF 1973 (iesg@ietf.org) - Internet Engineering Task Force". 1975 12.1. Header Field Registration 1977 Please update the "Message Headers" registry of "Permanent Message 1978 Header Field Names" at with the header field names listed in the two tables of 1980 Section 5. 1982 12.2. Media Type Registration 1984 Please update the "Media Types" registry at 1985 with the registration 1986 information in Section 10.1 and Section 10.2 for the media types 1987 "message/http" and "application/http", respectively. 1989 12.3. Transfer Coding Registration 1991 Please update the "HTTP Transfer Coding Registry" at 1992 with the 1993 registration procedure of Section 7.3 and the content coding names 1994 summarized in the table of Section 7. 1996 12.4. Upgrade Token Registration 1998 Please update the "Hypertext Transfer Protocol (HTTP) Upgrade Token 1999 Registry" at 2000 with the registration procedure of Section 9.8.2 and the upgrade 2001 token names summarized in the table of Section 9.8.1. 2003 13. References 2005 13.1. Normative References 2007 [Caching] Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 2008 Ed., "HTTP Caching", draft-ietf-httpbis-cache-03 (work in 2009 progress), October 2018. 2011 [RFC1950] Deutsch, L. and J-L. Gailly, "ZLIB Compressed Data Format 2012 Specification version 3.3", RFC 1950, 2013 DOI 10.17487/RFC1950, May 1996, 2014 . 2016 [RFC1951] Deutsch, P., "DEFLATE Compressed Data Format Specification 2017 version 1.3", RFC 1951, DOI 10.17487/RFC1951, May 1996, 2018 . 2020 [RFC1952] Deutsch, P., Gailly, J-L., Adler, M., Deutsch, L., and G. 2021 Randers-Pehrson, "GZIP file format specification version 2022 4.3", RFC 1952, DOI 10.17487/RFC1952, May 1996, 2023 . 2025 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 2026 Requirement Levels", BCP 14, RFC 2119, 2027 DOI 10.17487/RFC2119, March 1997, 2028 . 2030 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 2031 Resource Identifier (URI): Generic Syntax", STD 66, 2032 RFC 3986, DOI 10.17487/RFC3986, January 2005, 2033 . 2035 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 2036 Specifications: ABNF", STD 68, RFC 5234, 2037 DOI 10.17487/RFC5234, January 2008, 2038 . 2040 [Semantics] 2041 Fielding, R., Ed., Nottingham, M., Ed., and J. Reschke, 2042 Ed., "HTTP Semantics", draft-ietf-httpbis-semantics-03 2043 (work in progress), October 2018. 2045 [USASCII] American National Standards Institute, "Coded Character 2046 Set -- 7-bit American Standard Code for Information 2047 Interchange", ANSI X3.4, 1986. 2049 [Welch] Welch, T., "A Technique for High-Performance Data 2050 Compression", IEEE Computer 17(6), June 1984. 2052 13.2. Informative References 2054 [Err4667] RFC Errata, Erratum ID 4667, RFC 7230, 2055 . 2057 [Klein] Klein, A., "Divide and Conquer - HTTP Response Splitting, 2058 Web Cache Poisoning Attacks, and Related Topics", March 2059 2004, . 2062 [Linhart] Linhart, C., Klein, A., Heled, R., and S. Orrin, "HTTP 2063 Request Smuggling", June 2005, 2064 . 2066 [RFC1945] Berners-Lee, T., Fielding, R., and H. Nielsen, "Hypertext 2067 Transfer Protocol -- HTTP/1.0", RFC 1945, 2068 DOI 10.17487/RFC1945, May 1996, 2069 . 2071 [RFC2045] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2072 Extensions (MIME) Part One: Format of Internet Message 2073 Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996, 2074 . 2076 [RFC2046] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2077 Extensions (MIME) Part Two: Media Types", RFC 2046, 2078 DOI 10.17487/RFC2046, November 1996, 2079 . 2081 [RFC2049] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 2082 Extensions (MIME) Part Five: Conformance Criteria and 2083 Examples", RFC 2049, DOI 10.17487/RFC2049, November 1996, 2084 . 2086 [RFC2068] Fielding, R., Gettys, J., Mogul, J., Nielsen, H., and T. 2087 Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", 2088 RFC 2068, DOI 10.17487/RFC2068, January 1997, 2089 . 2091 [RFC2557] Palme, F., Hopmann, A., Shelness, N., and E. Stefferud, 2092 "MIME Encapsulation of Aggregate Documents, such as HTML 2093 (MHTML)", RFC 2557, DOI 10.17487/RFC2557, March 1999, 2094 . 2096 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 2097 DOI 10.17487/RFC5322, October 2008, 2098 . 2100 [RFC6265] Barth, A., "HTTP State Management Mechanism", RFC 6265, 2101 DOI 10.17487/RFC6265, April 2011, 2102 . 2104 [RFC7230] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2105 Protocol (HTTP/1.1): Message Syntax and Routing", 2106 RFC 7230, DOI 10.17487/RFC7230, June 2014, 2107 . 2109 [RFC7231] Fielding, R., Ed. and J. Reschke, Ed., "Hypertext Transfer 2110 Protocol (HTTP/1.1): Semantics and Content", RFC 7231, 2111 DOI 10.17487/RFC7231, June 2014, 2112 . 2114 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 2115 Writing an IANA Considerations Section in RFCs", BCP 26, 2116 RFC 8126, DOI 10.17487/RFC8126, June 2017, 2117 . 2119 Appendix A. Collected ABNF 2121 In the collected ABNF below, list rules are expanded as per 2122 Section 11 of [Semantics]. 2124 BWS = 2126 Connection = *( "," OWS ) connection-option *( OWS "," [ OWS 2127 connection-option ] ) 2129 HTTP-message = start-line *( header-field CRLF ) CRLF [ message-body 2130 ] 2131 HTTP-name = %x48.54.54.50 ; HTTP 2132 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 2134 OWS = 2136 RWS = 2138 TE = [ ( "," / t-codings ) *( OWS "," [ OWS t-codings ] ) ] 2139 Transfer-Encoding = *( "," OWS ) transfer-coding *( OWS "," [ OWS 2140 transfer-coding ] ) 2142 Upgrade = *( "," OWS ) protocol *( OWS "," [ OWS protocol ] ) 2144 absolute-URI = 2145 absolute-form = absolute-URI 2146 absolute-path = 2147 asterisk-form = "*" 2148 authority = 2149 authority-form = authority 2151 chunk = chunk-size [ chunk-ext ] CRLF chunk-data CRLF 2152 chunk-data = 1*OCTET 2153 chunk-ext = *( BWS ";" BWS chunk-ext-name [ BWS "=" BWS chunk-ext-val 2154 ] ) 2155 chunk-ext-name = token 2156 chunk-ext-val = token / quoted-string 2157 chunk-size = 1*HEXDIG 2158 chunked-body = *chunk last-chunk trailer-part CRLF 2159 comment = 2160 connection-option = token 2162 field-name = 2163 field-value = 2165 header-field = field-name ":" OWS field-value OWS 2166 last-chunk = 1*"0" [ chunk-ext ] CRLF 2168 message-body = *OCTET 2169 method = token 2171 obs-fold = CRLF 1*( SP / HTAB ) 2172 obs-text = 2173 origin-form = absolute-path [ "?" query ] 2175 port = 2176 protocol = protocol-name [ "/" protocol-version ] 2177 protocol-name = token 2178 protocol-version = token 2180 query = 2181 quoted-string = 2183 rank = ( "0" [ "." *3DIGIT ] ) / ( "1" [ "." *3"0" ] ) 2184 reason-phrase = *( HTAB / SP / VCHAR / obs-text ) 2185 request-line = method SP request-target SP HTTP-version CRLF 2186 request-target = origin-form / absolute-form / authority-form / 2187 asterisk-form 2189 start-line = request-line / status-line 2190 status-code = 3DIGIT 2191 status-line = HTTP-version SP status-code SP reason-phrase CRLF 2193 t-codings = "trailers" / ( transfer-coding [ t-ranking ] ) 2194 t-ranking = OWS ";" OWS "q=" rank 2195 token = 2196 trailer-part = *( header-field CRLF ) 2197 transfer-coding = "chunked" / "compress" / "deflate" / "gzip" / 2198 transfer-extension 2199 transfer-extension = token *( OWS ";" OWS transfer-parameter ) 2200 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 2202 uri-host = 2204 Appendix B. Differences between HTTP and MIME 2206 HTTP/1.1 uses many of the constructs defined for the Internet Message 2207 Format [RFC5322] and the Multipurpose Internet Mail Extensions (MIME) 2208 [RFC2045] to allow a message body to be transmitted in an open 2209 variety of representations and with extensible header fields. 2210 However, RFC 2045 is focused only on email; applications of HTTP have 2211 many characteristics that differ from email; hence, HTTP has features 2212 that differ from MIME. These differences were carefully chosen to 2213 optimize performance over binary connections, to allow greater 2214 freedom in the use of new media types, to make date comparisons 2215 easier, and to acknowledge the practice of some early HTTP servers 2216 and clients. 2218 This appendix describes specific areas where HTTP differs from MIME. 2219 Proxies and gateways to and from strict MIME environments need to be 2220 aware of these differences and provide the appropriate conversions 2221 where necessary. 2223 B.1. MIME-Version 2225 HTTP is not a MIME-compliant protocol. However, messages can include 2226 a single MIME-Version header field to indicate what version of the 2227 MIME protocol was used to construct the message. Use of the MIME- 2228 Version header field indicates that the message is in full 2229 conformance with the MIME protocol (as defined in [RFC2045]). 2230 Senders are responsible for ensuring full conformance (where 2231 possible) when exporting HTTP messages to strict MIME environments. 2233 B.2. Conversion to Canonical Form 2235 MIME requires that an Internet mail body part be converted to 2236 canonical form prior to being transferred, as described in Section 4 2237 of [RFC2049]. Section 6.1.1.2 of [Semantics] describes the forms 2238 allowed for subtypes of the "text" media type when transmitted over 2239 HTTP. [RFC2046] requires that content with a type of "text" 2240 represent line breaks as CRLF and forbids the use of CR or LF outside 2241 of line break sequences. HTTP allows CRLF, bare CR, and bare LF to 2242 indicate a line break within text content. 2244 A proxy or gateway from HTTP to a strict MIME environment ought to 2245 translate all line breaks within text media types to the RFC 2049 2246 canonical form of CRLF. Note, however, this might be complicated by 2247 the presence of a Content-Encoding and by the fact that HTTP allows 2248 the use of some charsets that do not use octets 13 and 10 to 2249 represent CR and LF, respectively. 2251 Conversion will break any cryptographic checksums applied to the 2252 original content unless the original content is already in canonical 2253 form. Therefore, the canonical form is recommended for any content 2254 that uses such checksums in HTTP. 2256 B.3. Conversion of Date Formats 2258 HTTP/1.1 uses a restricted set of date formats (Section 10.1.1.1 of 2259 [Semantics]) to simplify the process of date comparison. Proxies and 2260 gateways from other protocols ought to ensure that any Date header 2261 field present in a message conforms to one of the HTTP/1.1 formats 2262 and rewrite the date if necessary. 2264 B.4. Conversion of Content-Encoding 2266 MIME does not include any concept equivalent to HTTP/1.1's Content- 2267 Encoding header field. Since this acts as a modifier on the media 2268 type, proxies and gateways from HTTP to MIME-compliant protocols 2269 ought to either change the value of the Content-Type header field or 2270 decode the representation before forwarding the message. (Some 2271 experimental applications of Content-Type for Internet mail have used 2272 a media-type parameter of ";conversions=" to perform 2273 a function equivalent to Content-Encoding. However, this parameter 2274 is not part of the MIME standards). 2276 B.5. Conversion of Content-Transfer-Encoding 2278 HTTP does not use the Content-Transfer-Encoding field of MIME. 2279 Proxies and gateways from MIME-compliant protocols to HTTP need to 2280 remove any Content-Transfer-Encoding prior to delivering the response 2281 message to an HTTP client. 2283 Proxies and gateways from HTTP to MIME-compliant protocols are 2284 responsible for ensuring that the message is in the correct format 2285 and encoding for safe transport on that protocol, where "safe 2286 transport" is defined by the limitations of the protocol being used. 2287 Such a proxy or gateway ought to transform and label the data with an 2288 appropriate Content-Transfer-Encoding if doing so will improve the 2289 likelihood of safe transport over the destination protocol. 2291 B.6. MHTML and Line Length Limitations 2293 HTTP implementations that share code with MHTML [RFC2557] 2294 implementations need to be aware of MIME line length limitations. 2295 Since HTTP does not have this limitation, HTTP does not fold long 2296 lines. MHTML messages being transported by HTTP follow all 2297 conventions of MHTML, including line length limitations and folding, 2298 canonicalization, etc., since HTTP transfers message-bodies as 2299 payload and, aside from the "multipart/byteranges" type 2300 (Section 6.3.4 of [Semantics]), does not interpret the content or any 2301 MIME header lines that might be contained therein. 2303 Appendix C. HTTP Version History 2305 HTTP has been in use since 1990. The first version, later referred 2306 to as HTTP/0.9, was a simple protocol for hypertext data transfer 2307 across the Internet, using only a single request method (GET) and no 2308 metadata. HTTP/1.0, as defined by [RFC1945], added a range of 2309 request methods and MIME-like messaging, allowing for metadata to be 2310 transferred and modifiers placed on the request/response semantics. 2311 However, HTTP/1.0 did not sufficiently take into consideration the 2312 effects of hierarchical proxies, caching, the need for persistent 2313 connections, or name-based virtual hosts. The proliferation of 2314 incompletely implemented applications calling themselves "HTTP/1.0" 2315 further necessitated a protocol version change in order for two 2316 communicating applications to determine each other's true 2317 capabilities. 2319 HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent 2320 requirements that enable reliable implementations, adding only those 2321 features that can either be safely ignored by an HTTP/1.0 recipient 2322 or only be sent when communicating with a party advertising 2323 conformance with HTTP/1.1. 2325 HTTP/1.1 has been designed to make supporting previous versions easy. 2326 A general-purpose HTTP/1.1 server ought to be able to understand any 2327 valid request in the format of HTTP/1.0, responding appropriately 2328 with an HTTP/1.1 message that only uses features understood (or 2329 safely ignored) by HTTP/1.0 clients. Likewise, an HTTP/1.1 client 2330 can be expected to understand any valid HTTP/1.0 response. 2332 Since HTTP/0.9 did not support header fields in a request, there is 2333 no mechanism for it to support name-based virtual hosts (selection of 2334 resource by inspection of the Host header field). Any server that 2335 implements name-based virtual hosts ought to disable support for 2336 HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in fact, 2337 badly constructed HTTP/1.x requests caused by a client failing to 2338 properly encode the request-target. 2340 C.1. Changes from HTTP/1.0 2342 This section summarizes major differences between versions HTTP/1.0 2343 and HTTP/1.1. 2345 C.1.1. Multihomed Web Servers 2347 The requirements that clients and servers support the Host header 2348 field (Section 5.4 of [Semantics]), report an error if it is missing 2349 from an HTTP/1.1 request, and accept absolute URIs (Section 3.2) are 2350 among the most important changes defined by HTTP/1.1. 2352 Older HTTP/1.0 clients assumed a one-to-one relationship of IP 2353 addresses and servers; there was no other established mechanism for 2354 distinguishing the intended server of a request than the IP address 2355 to which that request was directed. The Host header field was 2356 introduced during the development of HTTP/1.1 and, though it was 2357 quickly implemented by most HTTP/1.0 browsers, additional 2358 requirements were placed on all HTTP/1.1 requests in order to ensure 2359 complete adoption. At the time of this writing, most HTTP-based 2360 services are dependent upon the Host header field for targeting 2361 requests. 2363 C.1.2. Keep-Alive Connections 2365 In HTTP/1.0, each connection is established by the client prior to 2366 the request and closed by the server after sending the response. 2367 However, some implementations implement the explicitly negotiated 2368 ("Keep-Alive") version of persistent connections described in 2369 Section 19.7.1 of [RFC2068]. 2371 Some clients and servers might wish to be compatible with these 2372 previous approaches to persistent connections, by explicitly 2373 negotiating for them with a "Connection: keep-alive" request header 2374 field. However, some experimental implementations of HTTP/1.0 2375 persistent connections are faulty; for example, if an HTTP/1.0 proxy 2376 server doesn't understand Connection, it will erroneously forward 2377 that header field to the next inbound server, which would result in a 2378 hung connection. 2380 One attempted solution was the introduction of a Proxy-Connection 2381 header field, targeted specifically at proxies. In practice, this 2382 was also unworkable, because proxies are often deployed in multiple 2383 layers, bringing about the same problem discussed above. 2385 As a result, clients are encouraged not to send the Proxy-Connection 2386 header field in any requests. 2388 Clients are also encouraged to consider the use of Connection: keep- 2389 alive in requests carefully; while they can enable persistent 2390 connections with HTTP/1.0 servers, clients using them will need to 2391 monitor the connection for "hung" requests (which indicate that the 2392 client ought stop sending the header field), and this mechanism ought 2393 not be used by clients at all when a proxy is being used. 2395 C.1.3. Introduction of Transfer-Encoding 2397 HTTP/1.1 introduces the Transfer-Encoding header field (Section 6.1). 2398 Transfer codings need to be decoded prior to forwarding an HTTP 2399 message over a MIME-compliant protocol. 2401 C.2. Changes from RFC 7230 2403 Most of the sections introducing HTTP's design goals, history, 2404 architecture, conformance criteria, protocol versioning, URIs, 2405 message routing, and header field values have been moved to 2406 [Semantics]. This document has been reduced to just the messaging 2407 syntax and connection management requirements specific to HTTP/1.1. 2409 Furthermore: 2411 In the ABNF for chunked extensions, re-introduce (bad) whitespace 2412 around ";" and "=". Whitespace was removed in [RFC7230], but later 2413 this change was found to break existing implementations (see 2414 [Err4667]). (Section 7.1.1) 2416 Disallow transfer coding parameters called "q" in order to avoid 2417 conflicts with the use of ranks in the TE header field. 2418 (Section 7.3) 2420 Appendix D. Change Log 2422 This section is to be removed before publishing as an RFC. 2424 D.1. Between RFC7230 and draft 00 2426 The changes were purely editorial: 2428 o Change boilerplate and abstract to indicate the "draft" status, 2429 and update references to ancestor specifications. 2431 o Adjust historical notes. 2433 o Update links to sibling specifications. 2435 o Replace sections listing changes from RFC 2616 by new empty 2436 sections referring to RFC 723x. 2438 o Remove acknowledgements specific to RFC 723x. 2440 o Move "Acknowledgements" to the very end and make them unnumbered. 2442 D.2. Since draft-ietf-httpbis-messaging-00 2444 The changes in this draft are editorial, with respect to HTTP as a 2445 whole, to move all core HTTP semantics into [Semantics]: 2447 o Moved introduction, architecture, conformance, and ABNF extensions 2448 from RFC 7230 (Messaging) to semantics [Semantics]. 2450 o Moved discussion of MIME differences from RFC 7231 (Semantics) to 2451 Appendix B since they mostly cover transforming 1.1 messages. 2453 o Moved all extensibility tips, registration procedures, and 2454 registry tables from the IANA considerations to normative 2455 sections, reducing the IANA considerations to just instructions 2456 that will be removed prior to publication as an RFC. 2458 D.3. Since draft-ietf-httpbis-messaging-01 2460 o Cite RFC 8126 instead of RFC 5226 () 2463 o Resolved erratum 4779, no change needed here 2464 (, 2465 ) 2467 o In Section 7, fixed prose claiming transfer parameters allow bare 2468 names (, 2469 ) 2471 o Resolved erratum 4225, no change needed here 2472 (, 2473 ) 2475 o Replace "response code" with "response status code" 2476 (, 2477 ) 2479 o In Section 9.4, clarify statement about HTTP/1.0 keep-alive 2480 (, 2481 ) 2483 o In Section 7.1.1, re-introduce (bad) whitespace around ";" and "=" 2484 (, 2485 , ) 2488 o In Section 7.3, state that transfer codings should not use 2489 parameters named "q" (, ) 2492 o In Section 7, mark coding name "trailers" as reserved in the IANA 2493 registry () 2495 D.4. Since draft-ietf-httpbis-messaging-02 2497 o In Section 4, explain why the reason phrase should be ignored by 2498 clients (). 2500 o Add Section 9.3 to explain how request/response correlation is 2501 performed () 2503 Index 2505 A 2506 absolute-form (of request-target) 10 2507 application/http Media Type 39 2508 asterisk-form (of request-target) 11 2509 authority-form (of request-target) 11 2511 C 2512 Connection header field 28, 34 2513 Content-Length header field 18 2514 Content-Transfer-Encoding header field 50 2515 chunked (Coding Format) 17, 19 2516 chunked (transfer coding) 22 2517 close 28, 34 2518 compress (transfer coding) 25 2520 D 2521 deflate (transfer coding) 25 2523 E 2524 effective request URI 12 2526 G 2527 Grammar 2528 absolute-form 9-10 2529 ALPHA 5 2530 asterisk-form 9, 11 2531 authority-form 9, 11 2532 chunk 22 2533 chunk-data 22 2534 chunk-ext 22-23 2535 chunk-ext-name 23 2536 chunk-ext-val 23 2537 chunk-size 22 2538 chunked-body 22-23 2539 Connection 29 2540 connection-option 29 2541 CR 5 2542 CRLF 5 2543 CTL 5 2544 DIGIT 5 2545 DQUOTE 5 2546 field-name 14 2547 field-value 14 2548 header-field 14, 23 2549 HEXDIG 5 2550 HTAB 5 2551 HTTP-message 6 2552 HTTP-name 6 2553 HTTP-version 6 2554 last-chunk 22 2555 LF 5 2556 message-body 16 2557 method 9 2558 obs-fold 15 2559 OCTET 5 2560 origin-form 9-10 2561 rank 26 2562 reason-phrase 14 2563 request-line 8 2564 request-target 9 2565 SP 5 2566 start-line 6 2567 status-code 14 2568 status-line 13 2569 t-codings 26 2570 t-ranking 26 2571 TE 26 2572 trailer-part 22-23 2573 transfer-coding 21 2574 Transfer-Encoding 17 2575 transfer-extension 21 2576 transfer-parameter 21 2577 Upgrade 35 2578 VCHAR 5 2579 gzip (transfer coding) 25 2581 H 2582 header field 6 2583 header section 6 2584 headers 6 2586 M 2587 MIME-Version header field 49 2588 Media Type 2589 application/http 39 2590 message/http 38 2592 message/http Media Type 38 2593 method 9 2595 O 2596 origin-form (of request-target) 10 2598 R 2599 request-target 9 2601 T 2602 TE header field 26 2603 Transfer-Encoding header field 17 2605 U 2606 Upgrade header field 35 2608 X 2609 x-compress (transfer coding) 25 2610 x-gzip (transfer coding) 25 2612 Acknowledgments 2614 See Appendix "Acknowledgments" of [Semantics]. 2616 Authors' Addresses 2618 Roy T. Fielding (editor) 2619 Adobe 2620 345 Park Ave 2621 San Jose, CA 95110 2622 USA 2624 EMail: fielding@gbiv.com 2625 URI: https://roy.gbiv.com/ 2627 Mark Nottingham (editor) 2628 Fastly 2630 EMail: mnot@mnot.net 2631 URI: https://www.mnot.net/ 2632 Julian F. Reschke (editor) 2633 greenbytes GmbH 2634 Hafenweg 16 2635 Muenster, NW 48155 2636 Germany 2638 EMail: julian.reschke@greenbytes.de 2639 URI: https://greenbytes.de/tech/webdav/