idnits 2.17.1 draft-ietf-httpbis-messaging-11.txt: Checking boilerplate required by RFC 5378 and the IETF Trust (see https://trustee.ietf.org/license-info): ---------------------------------------------------------------------------- No issues found here. Checking nits according to https://www.ietf.org/id-info/1id-guidelines.txt: ---------------------------------------------------------------------------- == There is 1 instance of lines with non-ascii characters in the document. Checking nits according to https://www.ietf.org/id-info/checklist : ---------------------------------------------------------------------------- -- The draft header indicates that this document obsoletes RFC7230, but the abstract doesn't seem to directly say this. It does mention RFC7230 though, so this could be OK. Miscellaneous warnings: ---------------------------------------------------------------------------- == The copyright year in the IETF Trust and authors Copyright Line does not match the current year == Line 710 has weird spacing: '...ncoding stan...' == 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 (August 27, 2020) is 1336 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 1899, but no explicit reference was found in the text == Unused Reference: 'RFC7231' is defined on line 1988, but no explicit reference was found in the text == Outdated reference: A later version (-19) exists of draft-ietf-httpbis-cache-11 -- Possible downref: Normative reference to a draft: ref.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-11 -- Possible downref: Normative reference to a draft: ref. 'Semantics' -- Possible downref: Non-RFC (?) normative reference: ref. 'USASCII' -- Possible downref: Non-RFC (?) normative reference: ref. 'Welch' -- Obsolete informational reference (is this intentional?): RFC 7230 (ref. 'Err4667') (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 2068 (Obsoleted by RFC 2616) -- Duplicate reference: RFC7230, mentioned in 'RFC7230', was also mentioned in 'Err4667'. -- Obsolete informational reference (is this intentional?): RFC 7230 (Obsoleted by RFC 9110, RFC 9112) -- Obsolete informational reference (is this intentional?): RFC 7231 (Obsoleted by RFC 9110) Summary: 3 errors (**), 0 flaws (~~), 8 warnings (==), 11 comments (--). Run idnits with the --verbose option for more detailed information about the items above. -------------------------------------------------------------------------------- 2 HTTP Working Group R. Fielding, Ed. 3 Internet-Draft Adobe 4 Obsoletes: 7230 (if approved) M. Nottingham, Ed. 5 Intended status: Standards Track Fastly 6 Expires: February 28, 2021 J. F. Reschke, Ed. 7 greenbytes 8 August 27, 2020 10 HTTP/1.1 Messaging 11 draft-ietf-httpbis-messaging-11 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.12. 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 February 28, 2021. 54 Copyright Notice 56 Copyright (c) 2020 IETF Trust and the persons identified as the 57 document authors. All rights reserved. 59 This document is subject to BCP 78 and the IETF Trust's Legal 60 Provisions Relating to IETF Documents (https://trustee.ietf.org/ 61 license-info) in effect on the date of publication of this document. 62 Please review these documents carefully, as they describe your rights 63 and restrictions with respect to this document. Code Components 64 extracted from this document must include Simplified BSD License text 65 as described in Section 4.e of the Trust Legal Provisions and are 66 provided without warranty as described in the Simplified BSD License. 68 This document may contain material from IETF Documents or IETF 69 Contributions published or made publicly available before November 70 10, 2008. The person(s) controlling the copyright in some of this 71 material may not have granted the IETF Trust the right to allow 72 modifications of such material outside the IETF Standards Process. 73 Without obtaining an adequate license from the person(s) controlling 74 the copyright in such materials, this document may not be modified 75 outside the IETF Standards Process, and derivative works of it may 76 not be created outside the IETF Standards Process, except to format 77 it for publication as an RFC or to translate it into languages other 78 than English. 80 Table of Contents 82 1. Introduction . . . . . . . . . . . . . . . . . . . . . . . . 4 83 1.1. Requirements Notation . . . . . . . . . . . . . . . . . . 5 84 1.2. Syntax Notation . . . . . . . . . . . . . . . . . . . . . 5 85 2. Message . . . . . . . . . . . . . . . . . . . . . . . . . . . 6 86 2.1. Message Format . . . . . . . . . . . . . . . . . . . . . 6 87 2.2. Message Parsing . . . . . . . . . . . . . . . . . . . . . 7 88 2.3. HTTP Version . . . . . . . . . . . . . . . . . . . . . . 8 89 3. Request Line . . . . . . . . . . . . . . . . . . . . . . . . 9 90 3.1. Method . . . . . . . . . . . . . . . . . . . . . . . . . 10 91 3.2. Request Target . . . . . . . . . . . . . . . . . . . . . 10 92 3.2.1. origin-form . . . . . . . . . . . . . . . . . . . . . 11 93 3.2.2. absolute-form . . . . . . . . . . . . . . . . . . . . 11 94 3.2.3. authority-form . . . . . . . . . . . . . . . . . . . 12 95 3.2.4. asterisk-form . . . . . . . . . . . . . . . . . . . . 12 96 3.3. Reconstructing the Target URI . . . . . . . . . . . . . . 13 97 4. Status Line . . . . . . . . . . . . . . . . . . . . . . . . . 14 98 5. Field Syntax . . . . . . . . . . . . . . . . . . . . . . . . 15 99 5.1. Field Line Parsing . . . . . . . . . . . . . . . . . . . 16 100 5.2. Obsolete Line Folding . . . . . . . . . . . . . . . . . . 17 101 6. Message Body . . . . . . . . . . . . . . . . . . . . . . . . 17 102 6.1. Transfer-Encoding . . . . . . . . . . . . . . . . . . . . 18 103 6.2. Content-Length . . . . . . . . . . . . . . . . . . . . . 19 104 6.3. Message Body Length . . . . . . . . . . . . . . . . . . . 20 105 7. Transfer Codings . . . . . . . . . . . . . . . . . . . . . . 22 106 7.1. Chunked Transfer Coding . . . . . . . . . . . . . . . . . 23 107 7.1.1. Chunk Extensions . . . . . . . . . . . . . . . . . . 24 108 7.1.2. Chunked Trailer Section . . . . . . . . . . . . . . . 25 109 7.1.3. Decoding Chunked . . . . . . . . . . . . . . . . . . 25 110 7.2. Transfer Codings for Compression . . . . . . . . . . . . 26 111 7.3. Transfer Coding Registry . . . . . . . . . . . . . . . . 26 112 7.4. Negotiating Transfer Codings . . . . . . . . . . . . . . 27 113 8. Handling Incomplete Messages . . . . . . . . . . . . . . . . 27 114 9. Connection Management . . . . . . . . . . . . . . . . . . . . 28 115 9.1. Establishment . . . . . . . . . . . . . . . . . . . . . . 28 116 9.2. Associating a Response to a Request . . . . . . . . . . . 29 117 9.3. Persistence . . . . . . . . . . . . . . . . . . . . . . . 29 118 9.3.1. Retrying Requests . . . . . . . . . . . . . . . . . . 30 119 9.3.2. Pipelining . . . . . . . . . . . . . . . . . . . . . 30 120 9.4. Concurrency . . . . . . . . . . . . . . . . . . . . . . . 31 121 9.5. Failures and Timeouts . . . . . . . . . . . . . . . . . . 32 122 9.6. Tear-down . . . . . . . . . . . . . . . . . . . . . . . . 32 123 9.7. TLS Connection Initiation . . . . . . . . . . . . . . . . 34 124 9.8. TLS Connection Closure . . . . . . . . . . . . . . . . . 34 125 10. Enclosing Messages as Data . . . . . . . . . . . . . . . . . 35 126 10.1. Media Type message/http . . . . . . . . . . . . . . . . 35 127 10.2. Media Type application/http . . . . . . . . . . . . . . 36 128 11. Security Considerations . . . . . . . . . . . . . . . . . . . 37 129 11.1. Response Splitting . . . . . . . . . . . . . . . . . . . 37 130 11.2. Request Smuggling . . . . . . . . . . . . . . . . . . . 38 131 11.3. Message Integrity . . . . . . . . . . . . . . . . . . . 38 132 11.4. Message Confidentiality . . . . . . . . . . . . . . . . 39 133 12. IANA Considerations . . . . . . . . . . . . . . . . . . . . . 39 134 12.1. Field Name Registration . . . . . . . . . . . . . . . . 39 135 12.2. Media Type Registration . . . . . . . . . . . . . . . . 39 136 12.3. Transfer Coding Registration . . . . . . . . . . . . . . 40 137 12.4. Upgrade Token Registration . . . . . . . . . . . . . . . 40 138 12.5. ALPN Protocol ID Registration . . . . . . . . . . . . . 40 139 13. References . . . . . . . . . . . . . . . . . . . . . . . . . 40 140 13.1. Normative References . . . . . . . . . . . . . . . . . . 40 141 13.2. Informative References . . . . . . . . . . . . . . . . . 41 142 Appendix A. Collected ABNF . . . . . . . . . . . . . . . . . . . 43 143 Appendix B. Differences between HTTP and MIME . . . . . . . . . 44 144 B.1. MIME-Version . . . . . . . . . . . . . . . . . . . . . . 45 145 B.2. Conversion to Canonical Form . . . . . . . . . . . . . . 45 146 B.3. Conversion of Date Formats . . . . . . . . . . . . . . . 45 147 B.4. Conversion of Content-Encoding . . . . . . . . . . . . . 46 148 B.5. Conversion of Content-Transfer-Encoding . . . . . . . . . 46 149 B.6. MHTML and Line Length Limitations . . . . . . . . . . . . 46 150 Appendix C. HTTP Version History . . . . . . . . . . . . . . . . 46 151 C.1. Changes from HTTP/1.0 . . . . . . . . . . . . . . . . . . 47 152 C.1.1. Multihomed Web Servers . . . . . . . . . . . . . . . 47 153 C.1.2. Keep-Alive Connections . . . . . . . . . . . . . . . 48 154 C.1.3. Introduction of Transfer-Encoding . . . . . . . . . . 48 155 C.2. Changes from RFC 7230 . . . . . . . . . . . . . . . . . . 49 156 Appendix D. Change Log . . . . . . . . . . . . . . . . . . . . . 49 157 D.1. Between RFC7230 and draft 00 . . . . . . . . . . . . . . 49 158 D.2. Since draft-ietf-httpbis-messaging-00 . . . . . . . . . . 50 159 D.3. Since draft-ietf-httpbis-messaging-01 . . . . . . . . . . 50 160 D.4. Since draft-ietf-httpbis-messaging-02 . . . . . . . . . . 51 161 D.5. Since draft-ietf-httpbis-messaging-03 . . . . . . . . . . 51 162 D.6. Since draft-ietf-httpbis-messaging-04 . . . . . . . . . . 51 163 D.7. Since draft-ietf-httpbis-messaging-05 . . . . . . . . . . 52 164 D.8. Since draft-ietf-httpbis-messaging-06 . . . . . . . . . . 52 165 D.9. Since draft-ietf-httpbis-messaging-07 . . . . . . . . . . 52 166 D.10. Since draft-ietf-httpbis-messaging-08 . . . . . . . . . . 53 167 D.11. Since draft-ietf-httpbis-messaging-09 . . . . . . . . . . 53 168 D.12. Since draft-ietf-httpbis-messaging-10 . . . . . . . . . . 53 169 Acknowledgments . . . . . . . . . . . . . . . . . . . . . . . . . 53 170 Authors' Addresses . . . . . . . . . . . . . . . . . . . . . . . 54 172 1. Introduction 174 The Hypertext Transfer Protocol (HTTP) is a stateless application- 175 level request/response protocol that uses extensible semantics and 176 self-descriptive messages for flexible interaction with network-based 177 hypertext information systems. HTTP is defined by a series of 178 documents that collectively form the HTTP/1.1 specification: 180 o "HTTP Semantics" [Semantics] 182 o "HTTP Caching" [Caching] 184 o "HTTP/1.1 Messaging" (this document) 185 This document defines HTTP/1.1 message syntax and framing 186 requirements and their associated connection management. Our goal is 187 to define all of the mechanisms necessary for HTTP/1.1 message 188 handling that are independent of message semantics, thereby defining 189 the complete set of requirements for message parsers and message- 190 forwarding intermediaries. 192 This document obsoletes the portions of RFC 7230 related to HTTP/1.1 193 messaging and connection management, with the changes being 194 summarized in Appendix C.2. The other parts of RFC 7230 are 195 obsoleted by "HTTP Semantics" [Semantics]. 197 1.1. Requirements Notation 199 The key words "MUST", "MUST NOT", "REQUIRED", "SHALL", "SHALL NOT", 200 "SHOULD", "SHOULD NOT", "RECOMMENDED", "NOT RECOMMENDED", "MAY", and 201 "OPTIONAL" in this document are to be interpreted as described in BCP 202 14 [RFC2119] [RFC8174] when, and only when, they appear in all 203 capitals, as shown here. 205 Conformance criteria and considerations regarding error handling are 206 defined in Section 3 of [Semantics]. 208 1.2. Syntax Notation 210 This specification uses the Augmented Backus-Naur Form (ABNF) 211 notation of [RFC5234], extended with the notation for case- 212 sensitivity in strings defined in [RFC7405]. 214 It also uses a list extension, defined in Section 5.5 of [Semantics], 215 that allows for compact definition of comma-separated lists using a 216 '#' operator (similar to how the '*' operator indicates repetition). 217 Appendix A shows the collected grammar with all list operators 218 expanded to standard ABNF notation. 220 As a convention, ABNF rule names prefixed with "obs-" denote 221 "obsolete" grammar rules that appear for historical reasons. 223 The following core rules are included by reference, as defined in 224 [RFC5234], Appendix B.1: ALPHA (letters), CR (carriage return), CRLF 225 (CR LF), CTL (controls), DIGIT (decimal 0-9), DQUOTE (double quote), 226 HEXDIG (hexadecimal 0-9/A-F/a-f), HTAB (horizontal tab), LF (line 227 feed), OCTET (any 8-bit sequence of data), SP (space), and VCHAR (any 228 visible [USASCII] character). 230 The rules below are defined in [Semantics]: 232 BWS = 233 OWS = 234 RWS = 235 absolute-URI = 236 absolute-path = 237 authority = 238 comment = 239 field-name = 240 field-value = 241 obs-text = 242 port = 243 query = 244 quoted-string = 245 token = 246 uri-host = 248 2. Message 250 2.1. Message Format 252 An HTTP/1.1 message consists of a start-line followed by a CRLF and a 253 sequence of octets in a format similar to the Internet Message Format 254 [RFC5322]: zero or more header field lines (collectively referred to 255 as the "headers" or the "header section"), an empty line indicating 256 the end of the header section, and an optional message body. 258 HTTP-message = start-line CRLF 259 *( field-line CRLF ) 260 CRLF 261 [ message-body ] 263 A message can be either a request from client to server or a response 264 from server to client. Syntactically, the two types of message 265 differ only in the start-line, which is either a request-line (for 266 requests) or a status-line (for responses), and in the algorithm for 267 determining the length of the message body (Section 6). 269 start-line = request-line / status-line 271 In theory, a client could receive requests and a server could receive 272 responses, distinguishing them by their different start-line formats. 273 In practice, servers are implemented to only expect a request (a 274 response is interpreted as an unknown or invalid request method) and 275 clients are implemented to only expect a response. 277 Although HTTP makes use of some protocol elements similar to the 278 Multipurpose Internet Mail Extensions (MIME) [RFC2045], see 279 Appendix B for the differences between HTTP and MIME messages. 281 2.2. Message Parsing 283 The normal procedure for parsing an HTTP message is to read the 284 start-line into a structure, read each header field line into a hash 285 table by field name until the empty line, and then use the parsed 286 data to determine if a message body is expected. If a message body 287 has been indicated, then it is read as a stream until an amount of 288 octets equal to the message body length is read or the connection is 289 closed. 291 A recipient MUST parse an HTTP message as a sequence of octets in an 292 encoding that is a superset of US-ASCII [USASCII]. Parsing an HTTP 293 message as a stream of Unicode characters, without regard for the 294 specific encoding, creates security vulnerabilities due to the 295 varying ways that string processing libraries handle invalid 296 multibyte character sequences that contain the octet LF (%x0A). 297 String-based parsers can only be safely used within protocol elements 298 after the element has been extracted from the message, such as within 299 a header field line value after message parsing has delineated the 300 individual field lines. 302 Although the line terminator for the start-line and header fields is 303 the sequence CRLF, a recipient MAY recognize a single LF as a line 304 terminator and ignore any preceding CR. 306 A sender MUST NOT generate a bare CR (a CR character not immediately 307 followed by LF) within any protocol elements other than the payload 308 body. A recipient of such a bare CR MUST consider that element to be 309 invalid or replace each bare CR with SP before processing the element 310 or forwarding the message. 312 Older HTTP/1.0 user agent implementations might send an extra CRLF 313 after a POST request as a workaround for some early server 314 applications that failed to read message body content that was not 315 terminated by a line-ending. An HTTP/1.1 user agent MUST NOT preface 316 or follow a request with an extra CRLF. If terminating the request 317 message body with a line-ending is desired, then the user agent MUST 318 count the terminating CRLF octets as part of the message body length. 320 In the interest of robustness, a server that is expecting to receive 321 and parse a request-line SHOULD ignore at least one empty line (CRLF) 322 received prior to the request-line. 324 A sender MUST NOT send whitespace between the start-line and the 325 first header field. A recipient that receives whitespace between the 326 start-line and the first header field MUST either reject the message 327 as invalid or consume each whitespace-preceded line without further 328 processing of it (i.e., ignore the entire line, along with any 329 subsequent lines preceded by whitespace, until a properly formed 330 header field is received or the header section is terminated). 332 The presence of such whitespace in a request might be an attempt to 333 trick a server into ignoring that field line or processing the line 334 after it as a new request, either of which might result in a security 335 vulnerability if other implementations within the request chain 336 interpret the same message differently. Likewise, the presence of 337 such whitespace in a response might be ignored by some clients or 338 cause others to cease parsing. 340 When a server listening only for HTTP request messages, or processing 341 what appears from the start-line to be an HTTP request message, 342 receives a sequence of octets that does not match the HTTP-message 343 grammar aside from the robustness exceptions listed above, the server 344 SHOULD respond with a 400 (Bad Request) response. 346 2.3. HTTP Version 348 HTTP uses a "." numbering scheme to indicate versions 349 of the protocol. This specification defines version "1.1". 350 Section 4.2 of [Semantics] specifies the semantics of HTTP version 351 numbers. 353 The version of an HTTP/1.x message is indicated by an HTTP-version 354 field in the start-line. HTTP-version is case-sensitive. 356 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 357 HTTP-name = %s"HTTP" 359 When an HTTP/1.1 message is sent to an HTTP/1.0 recipient [RFC1945] 360 or a recipient whose version is unknown, the HTTP/1.1 message is 361 constructed such that it can be interpreted as a valid HTTP/1.0 362 message if all of the newer features are ignored. This specification 363 places recipient-version requirements on some new features so that a 364 conformant sender will only use compatible features until it has 365 determined, through configuration or the receipt of a message, that 366 the recipient supports HTTP/1.1. 368 Intermediaries that process HTTP messages (i.e., all intermediaries 369 other than those acting as tunnels) MUST send their own HTTP-version 370 in forwarded messages. In other words, they are not allowed to 371 blindly forward the start-line without ensuring that the protocol 372 version in that message matches a version to which that intermediary 373 is conformant for both the receiving and sending of messages. 374 Forwarding an HTTP message without rewriting the HTTP-version might 375 result in communication errors when downstream recipients use the 376 message sender's version to determine what features are safe to use 377 for later communication with that sender. 379 A server MAY send an HTTP/1.0 response to an HTTP/1.1 request if it 380 is known or suspected that the client incorrectly implements the HTTP 381 specification and is incapable of correctly processing later version 382 responses, such as when a client fails to parse the version number 383 correctly or when an intermediary is known to blindly forward the 384 HTTP-version even when it doesn't conform to the given minor version 385 of the protocol. Such protocol downgrades SHOULD NOT be performed 386 unless triggered by specific client attributes, such as when one or 387 more of the request header fields (e.g., User-Agent) uniquely match 388 the values sent by a client known to be in error. 390 3. Request Line 392 A request-line begins with a method token, followed by a single space 393 (SP), the request-target, another single space (SP), and ends with 394 the protocol version. 396 request-line = method SP request-target SP HTTP-version 398 Although the request-line grammar rule requires that each of the 399 component elements be separated by a single SP octet, recipients MAY 400 instead parse on whitespace-delimited word boundaries and, aside from 401 the CRLF terminator, treat any form of whitespace as the SP separator 402 while ignoring preceding or trailing whitespace; such whitespace 403 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 404 (%x0C), or bare CR. However, lenient parsing can result in request 405 smuggling security vulnerabilities if there are multiple recipients 406 of the message and each has its own unique interpretation of 407 robustness (see Section 11.2). 409 HTTP does not place a predefined limit on the length of a request- 410 line, as described in Section 3 of [Semantics]. A server that 411 receives a method longer than any that it implements SHOULD respond 412 with a 501 (Not Implemented) status code. A server that receives a 413 request-target longer than any URI it wishes to parse MUST respond 414 with a 414 (URI Too Long) status code (see Section 10.5.15 of 415 [Semantics]). 417 Various ad hoc limitations on request-line length are found in 418 practice. It is RECOMMENDED that all HTTP senders and recipients 419 support, at a minimum, request-line lengths of 8000 octets. 421 3.1. Method 423 The method token indicates the request method to be performed on the 424 target resource. The request method is case-sensitive. 426 method = token 428 The request methods defined by this specification can be found in 429 Section 8 of [Semantics], along with information regarding the HTTP 430 method registry and considerations for defining new methods. 432 3.2. Request Target 434 The request-target identifies the target resource upon which to apply 435 the request. The client derives a request-target from its desired 436 target URI. There are four distinct formats for the request-target, 437 depending on both the method being requested and whether the request 438 is to a proxy. 440 request-target = origin-form 441 / absolute-form 442 / authority-form 443 / asterisk-form 445 No whitespace is allowed in the request-target. Unfortunately, some 446 user agents fail to properly encode or exclude whitespace found in 447 hypertext references, resulting in those disallowed characters being 448 sent as the request-target in a malformed request-line. 450 Recipients of an invalid request-line SHOULD respond with either a 451 400 (Bad Request) error or a 301 (Moved Permanently) redirect with 452 the request-target properly encoded. A recipient SHOULD NOT attempt 453 to autocorrect and then process the request without a redirect, since 454 the invalid request-line might be deliberately crafted to bypass 455 security filters along the request chain. 457 A client MUST send a Host header field in all HTTP/1.1 request 458 messages. If the target URI includes an authority component, then a 459 client MUST send a field value for Host that is identical to that 460 authority component, excluding any userinfo subcomponent and its "@" 461 delimiter (Section 2.5.1 of [Semantics]). If the authority component 462 is missing or undefined for the target URI, then a client MUST send a 463 Host header field with an empty field value. 465 A server MUST respond with a 400 (Bad Request) status code to any 466 HTTP/1.1 request message that lacks a Host header field and to any 467 request message that contains more than one Host header field or a 468 Host header field with an invalid field value. 470 3.2.1. origin-form 472 The most common form of request-target is the origin-form. 474 origin-form = absolute-path [ "?" query ] 476 When making a request directly to an origin server, other than a 477 CONNECT or server-wide OPTIONS request (as detailed below), a client 478 MUST send only the absolute path and query components of the target 479 URI as the request-target. If the target URI's path component is 480 empty, the client MUST send "/" as the path within the origin-form of 481 request-target. A Host header field is also sent, as defined in 482 Section 6.5 of [Semantics]. 484 For example, a client wishing to retrieve a representation of the 485 resource identified as 487 http://www.example.org/where?q=now 489 directly from the origin server would open (or reuse) a TCP 490 connection to port 80 of the host "www.example.org" and send the 491 lines: 493 GET /where?q=now HTTP/1.1 494 Host: www.example.org 496 followed by the remainder of the request message. 498 3.2.2. absolute-form 500 When making a request to a proxy, other than a CONNECT or server-wide 501 OPTIONS request (as detailed below), a client MUST send the target 502 URI in absolute-form as the request-target. 504 absolute-form = absolute-URI 506 The proxy is requested to either service that request from a valid 507 cache, if possible, or make the same request on the client's behalf 508 to either the next inbound proxy server or directly to the origin 509 server indicated by the request-target. Requirements on such 510 "forwarding" of messages are defined in Section 6.6 of [Semantics]. 512 An example absolute-form of request-line would be: 514 GET http://www.example.org/pub/WWW/TheProject.html HTTP/1.1 516 A client MUST send a Host header field in an HTTP/1.1 request even if 517 the request-target is in the absolute-form, since this allows the 518 Host information to be forwarded through ancient HTTP/1.0 proxies 519 that might not have implemented Host. 521 When a proxy receives a request with an absolute-form of request- 522 target, the proxy MUST ignore the received Host header field (if any) 523 and instead replace it with the host information of the request- 524 target. A proxy that forwards such a request MUST generate a new 525 Host field value based on the received request-target rather than 526 forward the received Host field value. 528 When an origin server receives a request with an absolute-form of 529 request-target, the origin server MUST ignore the received Host 530 header field (if any) and instead use the host information of the 531 request-target. Note that if the request-target does not have an 532 authority component, an empty Host header field will be sent in this 533 case. 535 To allow for transition to the absolute-form for all requests in some 536 future version of HTTP, a server MUST accept the absolute-form in 537 requests, even though HTTP/1.1 clients will only send them in 538 requests to proxies. 540 3.2.3. authority-form 542 The authority-form of request-target is only used for CONNECT 543 requests (Section 8.3.6 of [Semantics]). 545 authority-form = authority 547 When making a CONNECT request to establish a tunnel through one or 548 more proxies, a client MUST send only the target URI's authority 549 component (excluding any userinfo and its "@" delimiter) as the 550 request-target. For example, 552 CONNECT www.example.com:80 HTTP/1.1 554 3.2.4. asterisk-form 556 The asterisk-form of request-target is only used for a server-wide 557 OPTIONS request (Section 8.3.7 of [Semantics]). 559 asterisk-form = "*" 561 When a client wishes to request OPTIONS for the server as a whole, as 562 opposed to a specific named resource of that server, the client MUST 563 send only "*" (%x2A) as the request-target. For example, 564 OPTIONS * HTTP/1.1 566 If a proxy receives an OPTIONS request with an absolute-form of 567 request-target in which the URI has an empty path and no query 568 component, then the last proxy on the request chain MUST send a 569 request-target of "*" when it forwards the request to the indicated 570 origin server. 572 For example, the request 574 OPTIONS http://www.example.org:8001 HTTP/1.1 576 would be forwarded by the final proxy as 578 OPTIONS * HTTP/1.1 579 Host: www.example.org:8001 581 after connecting to port 8001 of host "www.example.org". 583 3.3. Reconstructing the Target URI 585 Since the request-target often contains only part of the user agent's 586 target URI, a server constructs its value to properly service the 587 request (Section 6.1 of [Semantics]). 589 If the request-target is in absolute-form, the target URI is the same 590 as the request-target. Otherwise, the target URI is constructed as 591 follows: 593 o If the server's configuration (or outbound gateway) provides a 594 fixed URI scheme, that scheme is used for the target URI. 595 Otherwise, if the request is received over a secured connection, 596 the target URI's scheme is "https"; if not, the scheme is "http". 598 o If the server's configuration (or outbound gateway) provides a 599 fixed URI authority component, that authority is used for the 600 target URI. If not, then if the request-target is in 601 authority-form, the target URI's authority component is the same 602 as the request-target. If not, then if a Host header field is 603 supplied with a non-empty field-value, the authority component is 604 the same as the Host field-value. Otherwise, the authority 605 component is assigned the default name configured for the server 606 and, if the connection's incoming TCP port number differs from the 607 default port for the target URI's scheme, then a colon (":") and 608 the incoming port number (in decimal form) are appended to the 609 authority component. 611 o If the request-target is in authority-form or asterisk-form, the 612 target URI's combined path and query component is empty. 613 Otherwise, the combined path and query component is the same as 614 the request-target. 616 o The components of the target URI, once determined as above, can be 617 combined into absolute-URI form by concatenating the scheme, 618 "://", authority, and combined path and query component. 620 Example 1: the following message received over an insecure TCP 621 connection 623 GET /pub/WWW/TheProject.html HTTP/1.1 624 Host: www.example.org:8080 626 has a target URI of 628 http://www.example.org:8080/pub/WWW/TheProject.html 630 Example 2: the following message received over a secured connection 632 OPTIONS * HTTP/1.1 633 Host: www.example.org 635 has a target URI of 637 https://www.example.org 639 Recipients of an HTTP/1.0 request that lacks a Host header field 640 might need to use heuristics (e.g., examination of the URI path for 641 something unique to a particular host) in order to guess the target 642 URI's authority component. 644 4. Status Line 646 The first line of a response message is the status-line, consisting 647 of the protocol version, a space (SP), the status code, another 648 space, and ending with an OPTIONAL textual phrase describing the 649 status code. 651 status-line = HTTP-version SP status-code SP [reason-phrase] 653 Although the status-line grammar rule requires that each of the 654 component elements be separated by a single SP octet, recipients MAY 655 instead parse on whitespace-delimited word boundaries and, aside from 656 the line terminator, treat any form of whitespace as the SP separator 657 while ignoring preceding or trailing whitespace; such whitespace 658 includes one or more of the following octets: SP, HTAB, VT (%x0B), FF 659 (%x0C), or bare CR. However, lenient parsing can result in response 660 splitting security vulnerabilities if there are multiple recipients 661 of the message and each has its own unique interpretation of 662 robustness (see Section 11.1). 664 The status-code element is a 3-digit integer code describing the 665 result of the server's attempt to understand and satisfy the client's 666 corresponding request. The rest of the response message is to be 667 interpreted in light of the semantics defined for that status code. 668 See Section 10 of [Semantics] for information about the semantics of 669 status codes, including the classes of status code (indicated by the 670 first digit), the status codes defined by this specification, 671 considerations for the definition of new status codes, and the IANA 672 registry. 674 status-code = 3DIGIT 676 The reason-phrase element exists for the sole purpose of providing a 677 textual description associated with the numeric status code, mostly 678 out of deference to earlier Internet application protocols that were 679 more frequently used with interactive text clients. 681 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 683 A client SHOULD ignore the reason-phrase content because it is not a 684 reliable channel for information (it might be translated for a given 685 locale, overwritten by intermediaries, or discarded when the message 686 is forwarded via other versions of HTTP). A server MUST send the 687 space that separates status-code from the reason-phrase even when the 688 reason-phrase is absent (i.e., the status-line would end with the 689 three octets SP CR LF). 691 5. Field Syntax 693 Each field line consists of a case-insensitive field name followed by 694 a colon (":"), optional leading whitespace, the field line value, and 695 optional trailing whitespace. 697 field-line = field-name ":" OWS field-value OWS 699 Most HTTP field names and the rules for parsing within field values 700 are defined in Section 5 of [Semantics]. This section covers the 701 generic syntax for header field inclusion within, and extraction 702 from, HTTP/1.1 messages. In addition, the following header fields 703 are defined by this document because they are specific to HTTP/1.1 704 message processing: 706 ------------------- ---------- ------ 707 Field Name Status Ref. 708 ------------------- ---------- ------ 709 MIME-Version standard B.1 710 Transfer-Encoding standard 6.1 711 ------------------- ---------- ------ 713 Table 1 715 Furthermore, the field name "Close" is reserved, since using that 716 name as an HTTP header field might conflict with the "close" 717 connection option of the Connection header field (Section 6.8 of 718 [Semantics]). 720 ------------ ---------- ----------- ------------ 721 Field Name Status Reference Comments 722 ------------ ---------- ----------- ------------ 723 Close standard Section 5 (reserved) 724 ------------ ---------- ----------- ------------ 726 Table 2 728 5.1. Field Line Parsing 730 Messages are parsed using a generic algorithm, independent of the 731 individual field names. The contents within a given field line value 732 are not parsed until a later stage of message interpretation (usually 733 after the message's entire header section has been processed). 735 No whitespace is allowed between the field name and colon. In the 736 past, differences in the handling of such whitespace have led to 737 security vulnerabilities in request routing and response handling. A 738 server MUST reject any received request message that contains 739 whitespace between a header field name and colon with a response 740 status code of 400 (Bad Request). A proxy MUST remove any such 741 whitespace from a response message before forwarding the message 742 downstream. 744 A field line value might be preceded and/or followed by optional 745 whitespace (OWS); a single SP preceding the field line value is 746 preferred for consistent readability by humans. The field line value 747 does not include any leading or trailing whitespace: OWS occurring 748 before the first non-whitespace octet of the field line value or 749 after the last non-whitespace octet of the field line value ought to 750 be excluded by parsers when extracting the field line value from a 751 header field line. 753 5.2. Obsolete Line Folding 755 Historically, HTTP header field line values could be extended over 756 multiple lines by preceding each extra line with at least one space 757 or horizontal tab (obs-fold). This specification deprecates such 758 line folding except within the message/http media type 759 (Section 10.1). 761 obs-fold = OWS CRLF RWS 762 ; obsolete line folding 764 A sender MUST NOT generate a message that includes line folding 765 (i.e., that has any field line value that contains a match to the 766 obs-fold rule) unless the message is intended for packaging within 767 the message/http media type. 769 A server that receives an obs-fold in a request message that is not 770 within a message/http container MUST either reject the message by 771 sending a 400 (Bad Request), preferably with a representation 772 explaining that obsolete line folding is unacceptable, or replace 773 each received obs-fold with one or more SP octets prior to 774 interpreting the field value or forwarding the message downstream. 776 A proxy or gateway that receives an obs-fold in a response message 777 that is not within a message/http container MUST either discard the 778 message and replace it with a 502 (Bad Gateway) response, preferably 779 with a representation explaining that unacceptable line folding was 780 received, or replace each received obs-fold with one or more SP 781 octets prior to interpreting the field value or forwarding the 782 message downstream. 784 A user agent that receives an obs-fold in a response message that is 785 not within a message/http container MUST replace each received 786 obs-fold with one or more SP octets prior to interpreting the field 787 value. 789 6. Message Body 791 The message body (if any) of an HTTP message is used to carry the 792 payload body (Section 7.3.3 of [Semantics]) of that request or 793 response. The message body is identical to the payload body unless a 794 transfer coding has been applied, as described in Section 6.1. 796 message-body = *OCTET 798 The rules for determining when a message body is present in an 799 HTTP/1.1 message differ for requests and responses. 801 The presence of a message body in a request is signaled by a 802 Content-Length or Transfer-Encoding header field. Request message 803 framing is independent of method semantics, even if the method does 804 not define any use for a message body. 806 The presence of a message body in a response depends on both the 807 request method to which it is responding and the response status code 808 (Section 4), and corresponds to when a payload body is allowed; see 809 Section 7.3.3 of [Semantics]. 811 6.1. Transfer-Encoding 813 The Transfer-Encoding header field lists the transfer coding names 814 corresponding to the sequence of transfer codings that have been (or 815 will be) applied to the payload body in order to form the message 816 body. Transfer codings are defined in Section 7. 818 Transfer-Encoding = #transfer-coding 820 Transfer-Encoding is analogous to the Content-Transfer-Encoding field 821 of MIME, which was designed to enable safe transport of binary data 822 over a 7-bit transport service ([RFC2045], Section 6). However, safe 823 transport has a different focus for an 8bit-clean transfer protocol. 824 In HTTP's case, Transfer-Encoding is primarily intended to accurately 825 delimit a dynamically generated payload and to distinguish payload 826 encodings that are only applied for transport efficiency or security 827 from those that are characteristics of the selected resource. 829 A recipient MUST be able to parse the chunked transfer coding 830 (Section 7.1) because it plays a crucial role in framing messages 831 when the payload body size is not known in advance. A sender MUST 832 NOT apply chunked more than once to a message body (i.e., chunking an 833 already chunked message is not allowed). If any transfer coding 834 other than chunked is applied to a request payload body, the sender 835 MUST apply chunked as the final transfer coding to ensure that the 836 message is properly framed. If any transfer coding other than 837 chunked is applied to a response payload body, the sender MUST either 838 apply chunked as the final transfer coding or terminate the message 839 by closing the connection. 841 For example, 843 Transfer-Encoding: gzip, chunked 845 indicates that the payload body has been compressed using the gzip 846 coding and then chunked using the chunked coding while forming the 847 message body. 849 Unlike Content-Encoding (Section 7.1.2 of [Semantics]), Transfer- 850 Encoding is a property of the message, not of the representation, and 851 any recipient along the request/response chain MAY decode the 852 received transfer coding(s) or apply additional transfer coding(s) to 853 the message body, assuming that corresponding changes are made to the 854 Transfer-Encoding field value. Additional information about the 855 encoding parameters can be provided by other header fields not 856 defined by this specification. 858 Transfer-Encoding MAY be sent in a response to a HEAD request or in a 859 304 (Not Modified) response (Section 10.4.5 of [Semantics]) to a GET 860 request, neither of which includes a message body, to indicate that 861 the origin server would have applied a transfer coding to the message 862 body if the request had been an unconditional GET. This indication 863 is not required, however, because any recipient on the response chain 864 (including the origin server) can remove transfer codings when they 865 are not needed. 867 A server MUST NOT send a Transfer-Encoding header field in any 868 response with a status code of 1xx (Informational) or 204 (No 869 Content). A server MUST NOT send a Transfer-Encoding header field in 870 any 2xx (Successful) response to a CONNECT request (Section 8.3.6 of 871 [Semantics]). 873 Transfer-Encoding was added in HTTP/1.1. It is generally assumed 874 that implementations advertising only HTTP/1.0 support will not 875 understand how to process a transfer-encoded payload. A client MUST 876 NOT send a request containing Transfer-Encoding unless it knows the 877 server will handle HTTP/1.1 requests (or later minor revisions); such 878 knowledge might be in the form of specific user configuration or by 879 remembering the version of a prior received response. A server MUST 880 NOT send a response containing Transfer-Encoding unless the 881 corresponding request indicates HTTP/1.1 (or later minor revisions). 883 A server that receives a request message with a transfer coding it 884 does not understand SHOULD respond with 501 (Not Implemented). 886 6.2. Content-Length 888 When a message does not have a Transfer-Encoding header field, a 889 Content-Length header field can provide the anticipated size, as a 890 decimal number of octets, for a potential payload body. For messages 891 that do include a payload body, the Content-Length field value 892 provides the framing information necessary for determining where the 893 body (and message) ends. For messages that do not include a payload 894 body, the Content-Length indicates the size of the selected 895 representation (Section 7.2.4 of [Semantics]). 897 | *Note:* HTTP's use of Content-Length for message framing 898 | differs significantly from the same field's use in MIME, where 899 | it is an optional field used only within the "message/external- 900 | body" media-type. 902 6.3. Message Body Length 904 The length of a message body is determined by one of the following 905 (in order of precedence): 907 1. Any response to a HEAD request and any response with a 1xx 908 (Informational), 204 (No Content), or 304 (Not Modified) status 909 code is always terminated by the first empty line after the 910 header fields, regardless of the header fields present in the 911 message, and thus cannot contain a message body. 913 2. Any 2xx (Successful) response to a CONNECT request implies that 914 the connection will become a tunnel immediately after the empty 915 line that concludes the header fields. A client MUST ignore any 916 Content-Length or Transfer-Encoding header fields received in 917 such a message. 919 3. If a Transfer-Encoding header field is present and the chunked 920 transfer coding (Section 7.1) is the final encoding, the message 921 body length is determined by reading and decoding the chunked 922 data until the transfer coding indicates the data is complete. 924 If a Transfer-Encoding header field is present in a response and 925 the chunked transfer coding is not the final encoding, the 926 message body length is determined by reading the connection until 927 it is closed by the server. If a Transfer-Encoding header field 928 is present in a request and the chunked transfer coding is not 929 the final encoding, the message body length cannot be determined 930 reliably; the server MUST respond with the 400 (Bad Request) 931 status code and then close the connection. 933 If a message is received with both a Transfer-Encoding and a 934 Content-Length header field, the Transfer-Encoding overrides the 935 Content-Length. Such a message might indicate an attempt to 936 perform request smuggling (Section 11.2) or response splitting 937 (Section 11.1) and ought to be handled as an error. A sender 938 MUST remove the received Content-Length field prior to forwarding 939 such a message downstream. 941 4. If a message is received without Transfer-Encoding and with an 942 invalid Content-Length header field, then the message framing is 943 invalid and the recipient MUST treat it as an unrecoverable 944 error, unless the field value can be successfully parsed as a 945 comma-separated list (Section 5.5 of [Semantics]), all values in 946 the list are valid, and all values in the list are the same. If 947 this is a request message, the server MUST respond with a 400 948 (Bad Request) status code and then close the connection. If this 949 is a response message received by a proxy, the proxy MUST close 950 the connection to the server, discard the received response, and 951 send a 502 (Bad Gateway) response to the client. If this is a 952 response message received by a user agent, the user agent MUST 953 close the connection to the server and discard the received 954 response. 956 5. If a valid Content-Length header field is present without 957 Transfer-Encoding, its decimal value defines the expected message 958 body length in octets. If the sender closes the connection or 959 the recipient times out before the indicated number of octets are 960 received, the recipient MUST consider the message to be 961 incomplete and close the connection. 963 6. If this is a request message and none of the above are true, then 964 the message body length is zero (no message body is present). 966 7. Otherwise, this is a response message without a declared message 967 body length, so the message body length is determined by the 968 number of octets received prior to the server closing the 969 connection. 971 Since there is no way to distinguish a successfully completed, close- 972 delimited response message from a partially received message 973 interrupted by network failure, a server SHOULD generate encoding or 974 length-delimited messages whenever possible. The close-delimiting 975 feature exists primarily for backwards compatibility with HTTP/1.0. 977 | *Note:* Request messages are never close-delimited because they 978 | are always explicitly framed by length or transfer coding, with 979 | the absence of both implying the request ends immediately after 980 | the header section. 982 A server MAY reject a request that contains a message body but not a 983 Content-Length by responding with 411 (Length Required). 985 Unless a transfer coding other than chunked has been applied, a 986 client that sends a request containing a message body SHOULD use a 987 valid Content-Length header field if the message body length is known 988 in advance, rather than the chunked transfer coding, since some 989 existing services respond to chunked with a 411 (Length Required) 990 status code even though they understand the chunked transfer coding. 991 This is typically because such services are implemented via a gateway 992 that requires a content-length in advance of being called and the 993 server is unable or unwilling to buffer the entire request before 994 processing. 996 A user agent that sends a request containing a message body MUST send 997 a valid Content-Length header field if it does not know the server 998 will handle HTTP/1.1 (or later) requests; such knowledge can be in 999 the form of specific user configuration or by remembering the version 1000 of a prior received response. 1002 If the final response to the last request on a connection has been 1003 completely received and there remains additional data to read, a user 1004 agent MAY discard the remaining data or attempt to determine if that 1005 data belongs as part of the prior response body, which might be the 1006 case if the prior message's Content-Length value is incorrect. A 1007 client MUST NOT process, cache, or forward such extra data as a 1008 separate response, since such behavior would be vulnerable to cache 1009 poisoning. 1011 7. Transfer Codings 1013 Transfer coding names are used to indicate an encoding transformation 1014 that has been, can be, or might need to be applied to a payload body 1015 in order to ensure "safe transport" through the network. This 1016 differs from a content coding in that the transfer coding is a 1017 property of the message rather than a property of the representation 1018 that is being transferred. 1020 transfer-coding = token *( OWS ";" OWS transfer-parameter ) 1022 Parameters are in the form of a name=value pair. 1024 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 1026 All transfer-coding names are case-insensitive and ought to be 1027 registered within the HTTP Transfer Coding registry, as defined in 1028 Section 7.3. They are used in the TE (Section 5.6.5 of [Semantics]) 1029 and Transfer-Encoding (Section 6.1) header fields. 1031 ------------ ------------------------------- ----------- 1032 Name Description Reference 1033 ------------ ------------------------------- ----------- 1034 chunked Transfer in a series of Section 1035 chunks 7.1 1036 compress UNIX "compress" data format Section 1037 [Welch] 7.2 1038 deflate "deflate" compressed data Section 1039 ([RFC1951]) inside the "zlib" 7.2 1040 data format ([RFC1950]) 1041 gzip GZIP file format [RFC1952] Section 1042 7.2 1043 trailers (reserved) Section 7 1044 x-compress Deprecated (alias for Section 1045 compress) 7.2 1046 x-gzip Deprecated (alias for gzip) Section 1047 7.2 1048 ------------ ------------------------------- ----------- 1050 Table 3 1052 | *Note:* the coding name "trailers" is reserved because its use 1053 | would conflict with the keyword "trailers" in the TE header 1054 | field (Section 5.6.5 of [Semantics]). 1056 7.1. Chunked Transfer Coding 1058 The chunked transfer coding wraps the payload body in order to 1059 transfer it as a series of chunks, each with its own size indicator, 1060 followed by an OPTIONAL trailer section containing trailer fields. 1061 Chunked enables content streams of unknown size to be transferred as 1062 a sequence of length-delimited buffers, which enables the sender to 1063 retain connection persistence and the recipient to know when it has 1064 received the entire message. 1066 chunked-body = *chunk 1067 last-chunk 1068 trailer-section 1069 CRLF 1071 chunk = chunk-size [ chunk-ext ] CRLF 1072 chunk-data CRLF 1073 chunk-size = 1*HEXDIG 1074 last-chunk = 1*("0") [ chunk-ext ] CRLF 1076 chunk-data = 1*OCTET ; a sequence of chunk-size octets 1078 The chunk-size field is a string of hex digits indicating the size of 1079 the chunk-data in octets. The chunked transfer coding is complete 1080 when a chunk with a chunk-size of zero is received, possibly followed 1081 by a trailer section, and finally terminated by an empty line. 1083 A recipient MUST be able to parse and decode the chunked transfer 1084 coding. 1086 Note that HTTP/1.1 does not define any means to limit the size of a 1087 chunked response such that an intermediary can be assured of 1088 buffering the entire response. 1090 The chunked encoding does not define any parameters. Their presence 1091 SHOULD be treated as an error. 1093 7.1.1. Chunk Extensions 1095 The chunked encoding allows each chunk to include zero or more chunk 1096 extensions, immediately following the chunk-size, for the sake of 1097 supplying per-chunk metadata (such as a signature or hash), mid- 1098 message control information, or randomization of message body size. 1100 chunk-ext = *( BWS ";" BWS chunk-ext-name 1101 [ BWS "=" BWS chunk-ext-val ] ) 1103 chunk-ext-name = token 1104 chunk-ext-val = token / quoted-string 1106 The chunked encoding is specific to each connection and is likely to 1107 be removed or recoded by each recipient (including intermediaries) 1108 before any higher-level application would have a chance to inspect 1109 the extensions. Hence, use of chunk extensions is generally limited 1110 to specialized HTTP services such as "long polling" (where client and 1111 server can have shared expectations regarding the use of chunk 1112 extensions) or for padding within an end-to-end secured connection. 1114 A recipient MUST ignore unrecognized chunk extensions. A server 1115 ought to limit the total length of chunk extensions received in a 1116 request to an amount reasonable for the services provided, in the 1117 same way that it applies length limitations and timeouts for other 1118 parts of a message, and generate an appropriate 4xx (Client Error) 1119 response if that amount is exceeded. 1121 7.1.2. Chunked Trailer Section 1123 A trailer section allows the sender to include additional fields at 1124 the end of a chunked message in order to supply metadata that might 1125 be dynamically generated while the message body is sent, such as a 1126 message integrity check, digital signature, or post-processing 1127 status. The proper use and limitations of trailer fields are defined 1128 in Section 5.6 of [Semantics]. 1130 trailer-section = *( field-line CRLF ) 1132 A recipient that decodes and removes the chunked encoding from a 1133 message (e.g., for storage or forwarding to a non-HTTP/1.1 peer) MUST 1134 discard any received trailer fields, store/forward them separately 1135 from the header fields, or selectively merge into the header section 1136 only those trailer fields corresponding to header field definitions 1137 that are understood by the recipient to explicitly permit and define 1138 how their corresponding trailer field value can be safely merged. 1140 7.1.3. Decoding Chunked 1142 A process for decoding the chunked transfer coding can be represented 1143 in pseudo-code as: 1145 length := 0 1146 read chunk-size, chunk-ext (if any), and CRLF 1147 while (chunk-size > 0) { 1148 read chunk-data and CRLF 1149 append chunk-data to decoded-body 1150 length := length + chunk-size 1151 read chunk-size, chunk-ext (if any), and CRLF 1152 } 1153 read trailer field 1154 while (trailer field is not empty) { 1155 if (trailer fields are stored/forwarded separately) { 1156 append trailer field to existing trailer fields 1157 } 1158 else if (trailer field is understood and defined as mergeable) { 1159 merge trailer field with existing header fields 1160 } 1161 else { 1162 discard trailer field 1163 } 1164 read trailer field 1165 } 1166 Content-Length := length 1167 Remove "chunked" from Transfer-Encoding 1168 Remove Trailer from existing header fields 1170 7.2. Transfer Codings for Compression 1172 The following transfer coding names for compression are defined by 1173 the same algorithm as their corresponding content coding: 1175 compress (and x-compress) 1176 See Section 7.1.2.1 of [Semantics]. 1178 deflate 1179 See Section 7.1.2.2 of [Semantics]. 1181 gzip (and x-gzip) 1182 See Section 7.1.2.3 of [Semantics]. 1184 The compression codings do not define any parameters. Their presence 1185 SHOULD be treated as an error. 1187 7.3. Transfer Coding Registry 1189 The "HTTP Transfer Coding Registry" defines the namespace for 1190 transfer coding names. It is maintained at 1191 . 1193 Registrations MUST include the following fields: 1195 o Name 1197 o Description 1199 o Pointer to specification text 1201 Names of transfer codings MUST NOT overlap with names of content 1202 codings (Section 7.1.2 of [Semantics]) unless the encoding 1203 transformation is identical, as is the case for the compression 1204 codings defined in Section 7.2. 1206 The TE header field (Section 5.6.5 of [Semantics]) uses a pseudo 1207 parameter named "q" as rank value when multiple transfer codings are 1208 acceptable. Future registrations of transfer codings SHOULD NOT 1209 define parameters called "q" (case-insensitively) in order to avoid 1210 ambiguities. 1212 Values to be added to this namespace require IETF Review (see 1213 Section 4.8 of [RFC8126]), and MUST conform to the purpose of 1214 transfer coding defined in this specification. 1216 Use of program names for the identification of encoding formats is 1217 not desirable and is discouraged for future encodings. 1219 7.4. Negotiating Transfer Codings 1221 The TE field (Section 5.6.5 of [Semantics]) is used in HTTP/1.1 to 1222 indicate what transfer-codings, besides chunked, the client is 1223 willing to accept in the response, and whether or not the client is 1224 willing to accept trailer fields in a chunked transfer coding. 1226 A client MUST NOT send the chunked transfer coding name in TE; 1227 chunked is always acceptable for HTTP/1.1 recipients. 1229 Three examples of TE use are below. 1231 TE: deflate 1232 TE: 1233 TE: trailers, deflate;q=0.5 1235 When multiple transfer codings are acceptable, the client MAY rank 1236 the codings by preference using a case-insensitive "q" parameter 1237 (similar to the qvalues used in content negotiation fields, 1238 Section 7.4.4 of [Semantics]). The rank value is a real number in 1239 the range 0 through 1, where 0.001 is the least preferred and 1 is 1240 the most preferred; a value of 0 means "not acceptable". 1242 If the TE field value is empty or if no TE field is present, the only 1243 acceptable transfer coding is chunked. A message with no transfer 1244 coding is always acceptable. 1246 The keyword "trailers" indicates that the sender will not discard 1247 trailer fields, as described in Section 5.6 of [Semantics]. 1249 Since the TE header field only applies to the immediate connection, a 1250 sender of TE MUST also send a "TE" connection option within the 1251 Connection header field (Section 6.8 of [Semantics]) in order to 1252 prevent the TE field from being forwarded by intermediaries that do 1253 not support its semantics. 1255 8. Handling Incomplete Messages 1257 A server that receives an incomplete request message, usually due to 1258 a canceled request or a triggered timeout exception, MAY send an 1259 error response prior to closing the connection. 1261 A client that receives an incomplete response message, which can 1262 occur when a connection is closed prematurely or when decoding a 1263 supposedly chunked transfer coding fails, MUST record the message as 1264 incomplete. Cache requirements for incomplete responses are defined 1265 in Section 3 of [Caching]. 1267 If a response terminates in the middle of the header section (before 1268 the empty line is received) and the status code might rely on header 1269 fields to convey the full meaning of the response, then the client 1270 cannot assume that meaning has been conveyed; the client might need 1271 to repeat the request in order to determine what action to take next. 1273 A message body that uses the chunked transfer coding is incomplete if 1274 the zero-sized chunk that terminates the encoding has not been 1275 received. A message that uses a valid Content-Length is incomplete 1276 if the size of the message body received (in octets) is less than the 1277 value given by Content-Length. A response that has neither chunked 1278 transfer coding nor Content-Length is terminated by closure of the 1279 connection and, thus, is considered complete regardless of the number 1280 of message body octets received, provided that the header section was 1281 received intact. 1283 9. Connection Management 1285 HTTP messaging is independent of the underlying transport- or 1286 session-layer connection protocol(s). HTTP only presumes a reliable 1287 transport with in-order delivery of requests and the corresponding 1288 in-order delivery of responses. The mapping of HTTP request and 1289 response structures onto the data units of an underlying transport 1290 protocol is outside the scope of this specification. 1292 As described in Section 6.3 of [Semantics], the specific connection 1293 protocols to be used for an HTTP interaction are determined by client 1294 configuration and the target URI. For example, the "http" URI scheme 1295 (Section 2.5.1 of [Semantics]) indicates a default connection of TCP 1296 over IP, with a default TCP port of 80, but the client might be 1297 configured to use a proxy via some other connection, port, or 1298 protocol. 1300 HTTP implementations are expected to engage in connection management, 1301 which includes maintaining the state of current connections, 1302 establishing a new connection or reusing an existing connection, 1303 processing messages received on a connection, detecting connection 1304 failures, and closing each connection. Most clients maintain 1305 multiple connections in parallel, including more than one connection 1306 per server endpoint. Most servers are designed to maintain thousands 1307 of concurrent connections, while controlling request queues to enable 1308 fair use and detect denial-of-service attacks. 1310 9.1. Establishment 1312 It is beyond the scope of this specification to describe how 1313 connections are established via various transport- or session-layer 1314 protocols. Each connection applies to only one transport link. 1316 9.2. Associating a Response to a Request 1318 HTTP/1.1 does not include a request identifier for associating a 1319 given request message with its corresponding one or more response 1320 messages. Hence, it relies on the order of response arrival to 1321 correspond exactly to the order in which requests are made on the 1322 same connection. More than one response message per request only 1323 occurs when one or more informational responses (1xx, see 1324 Section 10.2 of [Semantics]) precede a final response to the same 1325 request. 1327 A client that has more than one outstanding request on a connection 1328 MUST maintain a list of outstanding requests in the order sent and 1329 MUST associate each received response message on that connection to 1330 the highest ordered request that has not yet received a final (non- 1331 1xx) response. 1333 If an HTTP/1.1 client receives data on a connection that doesn't have 1334 any outstanding requests, it MUST NOT consider them to be a response 1335 to a not-yet-issued request; it SHOULD close the connection, since 1336 message delimitation is now ambiguous, unless the data consists only 1337 of one or more CRLF (which can be discarded, as per Section 2.2). 1339 9.3. Persistence 1341 HTTP/1.1 defaults to the use of "persistent connections", allowing 1342 multiple requests and responses to be carried over a single 1343 connection. The "close" connection option is used to signal that a 1344 connection will not persist after the current request/response. HTTP 1345 implementations SHOULD support persistent connections. 1347 A recipient determines whether a connection is persistent or not 1348 based on the most recently received message's protocol version and 1349 Connection header field (Section 6.8 of [Semantics]), if any: 1351 o If the "close" connection option is present, the connection will 1352 not persist after the current response; else, 1354 o If the received protocol is HTTP/1.1 (or later), the connection 1355 will persist after the current response; else, 1357 o If the received protocol is HTTP/1.0, the "keep-alive" connection 1358 option is present, either the recipient is not a proxy or the 1359 message is a response, and the recipient wishes to honor the 1360 HTTP/1.0 "keep-alive" mechanism, the connection will persist after 1361 the current response; otherwise, 1363 o The connection will close after the current response. 1365 A client that does not support persistent connections MUST send the 1366 "close" connection option in every request message. 1368 A server that does not support persistent connections MUST send the 1369 "close" connection option in every response message that does not 1370 have a 1xx (Informational) status code. 1372 A client MAY send additional requests on a persistent connection 1373 until it sends or receives a "close" connection option or receives an 1374 HTTP/1.0 response without a "keep-alive" connection option. 1376 In order to remain persistent, all messages on a connection need to 1377 have a self-defined message length (i.e., one not defined by closure 1378 of the connection), as described in Section 6. A server MUST read 1379 the entire request message body or close the connection after sending 1380 its response, since otherwise the remaining data on a persistent 1381 connection would be misinterpreted as the next request. Likewise, a 1382 client MUST read the entire response message body if it intends to 1383 reuse the same connection for a subsequent request. 1385 A proxy server MUST NOT maintain a persistent connection with an 1386 HTTP/1.0 client (see Section 19.7.1 of [RFC2068] for information and 1387 discussion of the problems with the Keep-Alive header field 1388 implemented by many HTTP/1.0 clients). 1390 See Appendix C.1.2 for more information on backwards compatibility 1391 with HTTP/1.0 clients. 1393 9.3.1. Retrying Requests 1395 Connections can be closed at any time, with or without intention. 1396 Implementations ought to anticipate the need to recover from 1397 asynchronous close events. The conditions under which a client can 1398 automatically retry a sequence of outstanding requests are defined in 1399 Section 8.2.2 of [Semantics]. 1401 9.3.2. Pipelining 1403 A client that supports persistent connections MAY "pipeline" its 1404 requests (i.e., send multiple requests without waiting for each 1405 response). A server MAY process a sequence of pipelined requests in 1406 parallel if they all have safe methods (Section 8.2.1 of 1407 [Semantics]), but it MUST send the corresponding responses in the 1408 same order that the requests were received. 1410 A client that pipelines requests SHOULD retry unanswered requests if 1411 the connection closes before it receives all of the corresponding 1412 responses. When retrying pipelined requests after a failed 1413 connection (a connection not explicitly closed by the server in its 1414 last complete response), a client MUST NOT pipeline immediately after 1415 connection establishment, since the first remaining request in the 1416 prior pipeline might have caused an error response that can be lost 1417 again if multiple requests are sent on a prematurely closed 1418 connection (see the TCP reset problem described in Section 9.6). 1420 Idempotent methods (Section 8.2.2 of [Semantics]) are significant to 1421 pipelining because they can be automatically retried after a 1422 connection failure. A user agent SHOULD NOT pipeline requests after 1423 a non-idempotent method, until the final response status code for 1424 that method has been received, unless the user agent has a means to 1425 detect and recover from partial failure conditions involving the 1426 pipelined sequence. 1428 An intermediary that receives pipelined requests MAY pipeline those 1429 requests when forwarding them inbound, since it can rely on the 1430 outbound user agent(s) to determine what requests can be safely 1431 pipelined. If the inbound connection fails before receiving a 1432 response, the pipelining intermediary MAY attempt to retry a sequence 1433 of requests that have yet to receive a response if the requests all 1434 have idempotent methods; otherwise, the pipelining intermediary 1435 SHOULD forward any received responses and then close the 1436 corresponding outbound connection(s) so that the outbound user 1437 agent(s) can recover accordingly. 1439 9.4. Concurrency 1441 A client ought to limit the number of simultaneous open connections 1442 that it maintains to a given server. 1444 Previous revisions of HTTP gave a specific number of connections as a 1445 ceiling, but this was found to be impractical for many applications. 1446 As a result, this specification does not mandate a particular maximum 1447 number of connections but, instead, encourages clients to be 1448 conservative when opening multiple connections. 1450 Multiple connections are typically used to avoid the "head-of-line 1451 blocking" problem, wherein a request that takes significant server- 1452 side processing and/or has a large payload blocks subsequent requests 1453 on the same connection. However, each connection consumes server 1454 resources. Furthermore, using multiple connections can cause 1455 undesirable side effects in congested networks. 1457 Note that a server might reject traffic that it deems abusive or 1458 characteristic of a denial-of-service attack, such as an excessive 1459 number of open connections from a single client. 1461 9.5. Failures and Timeouts 1463 Servers will usually have some timeout value beyond which they will 1464 no longer maintain an inactive connection. Proxy servers might make 1465 this a higher value since it is likely that the client will be making 1466 more connections through the same proxy server. The use of 1467 persistent connections places no requirements on the length (or 1468 existence) of this timeout for either the client or the server. 1470 A client or server that wishes to time out SHOULD issue a graceful 1471 close on the connection. Implementations SHOULD constantly monitor 1472 open connections for a received closure signal and respond to it as 1473 appropriate, since prompt closure of both sides of a connection 1474 enables allocated system resources to be reclaimed. 1476 A client, server, or proxy MAY close the transport connection at any 1477 time. For example, a client might have started to send a new request 1478 at the same time that the server has decided to close the "idle" 1479 connection. From the server's point of view, the connection is being 1480 closed while it was idle, but from the client's point of view, a 1481 request is in progress. 1483 A server SHOULD sustain persistent connections, when possible, and 1484 allow the underlying transport's flow-control mechanisms to resolve 1485 temporary overloads, rather than terminate connections with the 1486 expectation that clients will retry. The latter technique can 1487 exacerbate network congestion. 1489 A client sending a message body SHOULD monitor the network connection 1490 for an error response while it is transmitting the request. If the 1491 client sees a response that indicates the server does not wish to 1492 receive the message body and is closing the connection, the client 1493 SHOULD immediately cease transmitting the body and close its side of 1494 the connection. 1496 9.6. Tear-down 1498 The Connection header field (Section 6.8 of [Semantics]) provides a 1499 "close" connection option that a sender SHOULD send when it wishes to 1500 close the connection after the current request/response pair. 1502 A client that sends a "close" connection option MUST NOT send further 1503 requests on that connection (after the one containing "close") and 1504 MUST close the connection after reading the final response message 1505 corresponding to this request. 1507 A server that receives a "close" connection option MUST initiate a 1508 close of the connection (see below) after it sends the final response 1509 to the request that contained "close". The server SHOULD send a 1510 "close" connection option in its final response on that connection. 1511 The server MUST NOT process any further requests received on that 1512 connection. 1514 A server that sends a "close" connection option MUST initiate a close 1515 of the connection (see below) after it sends the response containing 1516 "close". The server MUST NOT process any further requests received 1517 on that connection. 1519 A client that receives a "close" connection option MUST cease sending 1520 requests on that connection and close the connection after reading 1521 the response message containing the "close"; if additional pipelined 1522 requests had been sent on the connection, the client SHOULD NOT 1523 assume that they will be processed by the server. 1525 If a server performs an immediate close of a TCP connection, there is 1526 a significant risk that the client will not be able to read the last 1527 HTTP response. If the server receives additional data from the 1528 client on a fully closed connection, such as another request that was 1529 sent by the client before receiving the server's response, the 1530 server's TCP stack will send a reset packet to the client; 1531 unfortunately, the reset packet might erase the client's 1532 unacknowledged input buffers before they can be read and interpreted 1533 by the client's HTTP parser. 1535 To avoid the TCP reset problem, servers typically close a connection 1536 in stages. First, the server performs a half-close by closing only 1537 the write side of the read/write connection. The server then 1538 continues to read from the connection until it receives a 1539 corresponding close by the client, or until the server is reasonably 1540 certain that its own TCP stack has received the client's 1541 acknowledgement of the packet(s) containing the server's last 1542 response. Finally, the server fully closes the connection. 1544 It is unknown whether the reset problem is exclusive to TCP or might 1545 also be found in other transport connection protocols. 1547 Note that a TCP connection that is half-closed by the client does not 1548 delimit a request message, nor does it imply that the client is no 1549 longer interested in a response. In general, transport signals 1550 cannot be relied upon to signal edge cases, since HTTP/1.1 is 1551 independent of transport. 1553 9.7. TLS Connection Initiation 1555 Conceptually, HTTP/TLS is simply sending HTTP messages over a 1556 connection secured via TLS [RFC8446]. 1558 The HTTP client also acts as the TLS client. It initiates a 1559 connection to the server on the appropriate port and sends the TLS 1560 ClientHello to begin the TLS handshake. When the TLS handshake has 1561 finished, the client may then initiate the first HTTP request. All 1562 HTTP data MUST be sent as TLS "application data", but is otherwise 1563 treated like a normal connection for HTTP (including potential reuse 1564 as a persistent connection). 1566 9.8. TLS Connection Closure 1568 TLS provides a facility for secure connection closure. When a valid 1569 closure alert is received, an implementation can be assured that no 1570 further data will be received on that connection. TLS 1571 implementations MUST initiate an exchange of closure alerts before 1572 closing a connection. A TLS implementation MAY, after sending a 1573 closure alert, close the connection without waiting for the peer to 1574 send its closure alert, generating an "incomplete close". Note that 1575 an implementation which does this MAY choose to reuse the session. 1576 This SHOULD only be done when the application knows (typically 1577 through detecting HTTP message boundaries) that it has received all 1578 the message data that it cares about. 1580 As specified in [RFC8446], any implementation which receives a 1581 connection close without first receiving a valid closure alert (a 1582 "premature close") MUST NOT reuse that session. Note that a 1583 premature close does not call into question the security of the data 1584 already received, but simply indicates that subsequent data might 1585 have been truncated. Because TLS is oblivious to HTTP request/ 1586 response boundaries, it is necessary to examine the HTTP data itself 1587 (specifically the Content-Length header) to determine whether the 1588 truncation occurred inside a message or between messages. 1590 When encountering a premature close, a client SHOULD treat as 1591 completed all requests for which it has received as much data as 1592 specified in the Content-Length header. 1594 A client detecting an incomplete close SHOULD recover gracefully. It 1595 MAY resume a TLS session closed in this fashion. 1597 Clients MUST send a closure alert before closing the connection. 1598 Clients which are unprepared to receive any more data MAY choose not 1599 to wait for the server's closure alert and simply close the 1600 connection, thus generating an incomplete close on the server side. 1602 Servers SHOULD be prepared to receive an incomplete close from the 1603 client, since the client can often determine when the end of server 1604 data is. Servers SHOULD be willing to resume TLS sessions closed in 1605 this fashion. 1607 Servers MUST attempt to initiate an exchange of closure alerts with 1608 the client before closing the connection. Servers MAY close the 1609 connection after sending the closure alert, thus generating an 1610 incomplete close on the client side. 1612 10. Enclosing Messages as Data 1614 10.1. Media Type message/http 1616 The message/http media type can be used to enclose a single HTTP 1617 request or response message, provided that it obeys the MIME 1618 restrictions for all "message" types regarding line length and 1619 encodings. 1621 Type name: message 1623 Subtype name: http 1625 Required parameters: N/A 1627 Optional parameters: version, msgtype 1629 version: The HTTP-version number of the enclosed message (e.g., 1630 "1.1"). If not present, the version can be determined from the 1631 first line of the body. 1633 msgtype: The message type - "request" or "response". If not 1634 present, the type can be determined from the first line of the 1635 body. 1637 Encoding considerations: only "7bit", "8bit", or "binary" are 1638 permitted 1640 Security considerations: see Section 11 1642 Interoperability considerations: N/A 1644 Published specification: This specification (see Section 10.1). 1646 Applications that use this media type: N/A 1648 Fragment identifier considerations: N/A 1649 Additional information: Magic number(s): N/A 1651 Deprecated alias names for this type: N/A 1653 File extension(s): N/A 1655 Macintosh file type code(s): N/A 1657 Person and email address to contact for further information: See Aut 1658 hors' Addresses section. 1660 Intended usage: COMMON 1662 Restrictions on usage: N/A 1664 Author: See Authors' Addresses section. 1666 Change controller: IESG 1668 10.2. Media Type application/http 1670 The application/http media type can be used to enclose a pipeline of 1671 one or more HTTP request or response messages (not intermixed). 1673 Type name: application 1675 Subtype name: http 1677 Required parameters: N/A 1679 Optional parameters: version, msgtype 1681 version: The HTTP-version number of the enclosed messages (e.g., 1682 "1.1"). If not present, the version can be determined from the 1683 first line of the body. 1685 msgtype: The message type - "request" or "response". If not 1686 present, the type can be determined from the first line of the 1687 body. 1689 Encoding considerations: HTTP messages enclosed by this type are in 1690 "binary" format; use of an appropriate Content-Transfer-Encoding 1691 is required when transmitted via email. 1693 Security considerations: see Section 11 1695 Interoperability considerations: N/A 1696 Published specification: This specification (see Section 10.2). 1698 Applications that use this media type: N/A 1700 Fragment identifier considerations: N/A 1702 Additional information: Deprecated alias names for this type: N/A 1704 Magic number(s): N/A 1706 File extension(s): N/A 1708 Macintosh file type code(s): N/A 1710 Person and email address to contact for further information: See Aut 1711 hors' Addresses section. 1713 Intended usage: COMMON 1715 Restrictions on usage: N/A 1717 Author: See Authors' Addresses section. 1719 Change controller: IESG 1721 11. Security Considerations 1723 This section is meant to inform developers, information providers, 1724 and users of known security considerations relevant to HTTP message 1725 syntax, parsing, and routing. Security considerations about HTTP 1726 semantics and payloads are addressed in [Semantics]. 1728 11.1. Response Splitting 1730 Response splitting (a.k.a, CRLF injection) is a common technique, 1731 used in various attacks on Web usage, that exploits the line-based 1732 nature of HTTP message framing and the ordered association of 1733 requests to responses on persistent connections [Klein]. This 1734 technique can be particularly damaging when the requests pass through 1735 a shared cache. 1737 Response splitting exploits a vulnerability in servers (usually 1738 within an application server) where an attacker can send encoded data 1739 within some parameter of the request that is later decoded and echoed 1740 within any of the response header fields of the response. If the 1741 decoded data is crafted to look like the response has ended and a 1742 subsequent response has begun, the response has been split and the 1743 content within the apparent second response is controlled by the 1744 attacker. The attacker can then make any other request on the same 1745 persistent connection and trick the recipients (including 1746 intermediaries) into believing that the second half of the split is 1747 an authoritative answer to the second request. 1749 For example, a parameter within the request-target might be read by 1750 an application server and reused within a redirect, resulting in the 1751 same parameter being echoed in the Location header field of the 1752 response. If the parameter is decoded by the application and not 1753 properly encoded when placed in the response field, the attacker can 1754 send encoded CRLF octets and other content that will make the 1755 application's single response look like two or more responses. 1757 A common defense against response splitting is to filter requests for 1758 data that looks like encoded CR and LF (e.g., "%0D" and "%0A"). 1759 However, that assumes the application server is only performing URI 1760 decoding, rather than more obscure data transformations like charset 1761 transcoding, XML entity translation, base64 decoding, sprintf 1762 reformatting, etc. A more effective mitigation is to prevent 1763 anything other than the server's core protocol libraries from sending 1764 a CR or LF within the header section, which means restricting the 1765 output of header fields to APIs that filter for bad octets and not 1766 allowing application servers to write directly to the protocol 1767 stream. 1769 11.2. Request Smuggling 1771 Request smuggling ([Linhart]) is a technique that exploits 1772 differences in protocol parsing among various recipients to hide 1773 additional requests (which might otherwise be blocked or disabled by 1774 policy) within an apparently harmless request. Like response 1775 splitting, request smuggling can lead to a variety of attacks on HTTP 1776 usage. 1778 This specification has introduced new requirements on request 1779 parsing, particularly with regard to message framing in Section 6.3, 1780 to reduce the effectiveness of request smuggling. 1782 11.3. Message Integrity 1784 HTTP does not define a specific mechanism for ensuring message 1785 integrity, instead relying on the error-detection ability of 1786 underlying transport protocols and the use of length or chunk- 1787 delimited framing to detect completeness. Additional integrity 1788 mechanisms, such as hash functions or digital signatures applied to 1789 the content, can be selectively added to messages via extensible 1790 metadata fields. Historically, the lack of a single integrity 1791 mechanism has been justified by the informal nature of most HTTP 1792 communication. However, the prevalence of HTTP as an information 1793 access mechanism has resulted in its increasing use within 1794 environments where verification of message integrity is crucial. 1796 User agents are encouraged to implement configurable means for 1797 detecting and reporting failures of message integrity such that those 1798 means can be enabled within environments for which integrity is 1799 necessary. For example, a browser being used to view medical history 1800 or drug interaction information needs to indicate to the user when 1801 such information is detected by the protocol to be incomplete, 1802 expired, or corrupted during transfer. Such mechanisms might be 1803 selectively enabled via user agent extensions or the presence of 1804 message integrity metadata in a response. At a minimum, user agents 1805 ought to provide some indication that allows a user to distinguish 1806 between a complete and incomplete response message (Section 8) when 1807 such verification is desired. 1809 11.4. Message Confidentiality 1811 HTTP relies on underlying transport protocols to provide message 1812 confidentiality when that is desired. HTTP has been specifically 1813 designed to be independent of the transport protocol, such that it 1814 can be used over many different forms of encrypted connection, with 1815 the selection of such transports being identified by the choice of 1816 URI scheme or within user agent configuration. 1818 The "https" scheme can be used to identify resources that require a 1819 confidential connection, as described in Section 2.5.2 of 1820 [Semantics]. 1822 12. IANA Considerations 1824 The change controller for the following registrations is: "IETF 1825 (iesg@ietf.org) - Internet Engineering Task Force". 1827 12.1. Field Name Registration 1829 Please update the "Hypertext Transfer Protocol (HTTP) Field Name 1830 Registry" at with the 1831 field names listed in the two tables of Section 5. 1833 12.2. Media Type Registration 1835 Please update the "Media Types" registry at 1836 with the registration 1837 information in Section 10.1 and Section 10.2 for the media types 1838 "message/http" and "application/http", respectively. 1840 12.3. Transfer Coding Registration 1842 Please update the "HTTP Transfer Coding Registry" at 1843 with the 1844 registration procedure of Section 7.3 and the content coding names 1845 summarized in the table of Section 7. 1847 12.4. Upgrade Token Registration 1849 Please update the "Hypertext Transfer Protocol (HTTP) Upgrade Token 1850 Registry" at 1851 with the registration procedure of Section 6.7.2 of [Semantics] and 1852 the upgrade token names summarized in the table of Section 6.7.1 of 1853 [Semantics]. 1855 12.5. ALPN Protocol ID Registration 1857 Please update the "TLS Application-Layer Protocol Negotiation (ALPN) 1858 Protocol IDs" registry at with the 1860 registration below: 1862 ---------- ----------------------------- ---------------- 1863 Protocol Identification Sequence Reference 1864 ---------- ----------------------------- ---------------- 1865 HTTP/1.1 0x68 0x74 0x74 0x70 0x2f (this 1866 0x31 0x2e 0x31 ("http/1.1") specification) 1867 ---------- ----------------------------- ---------------- 1869 Table 4 1871 13. References 1873 13.1. Normative References 1875 [Caching] Fielding, R., Ed., Nottingham, M., Ed., and J. F. Reschke, 1876 Ed., "HTTP Caching", Work in Progress, Internet-Draft, 1877 draft-ietf-httpbis-cache-11, August 27, 2020, 1878 . 1880 [RFC1950] Deutsch, L.P. and J-L. Gailly, "ZLIB Compressed Data 1881 Format Specification version 3.3", RFC 1950, 1882 DOI 10.17487/RFC1950, May 1996, 1883 . 1885 [RFC1951] Deutsch, P., "DEFLATE Compressed Data Format Specification 1886 version 1.3", RFC 1951, DOI 10.17487/RFC1951, May 1996, 1887 . 1889 [RFC1952] Deutsch, P., Gailly, J-L., Adler, M., Deutsch, L.P., and 1890 G. Randers-Pehrson, "GZIP file format specification 1891 version 4.3", RFC 1952, DOI 10.17487/RFC1952, May 1996, 1892 . 1894 [RFC2119] Bradner, S., "Key words for use in RFCs to Indicate 1895 Requirement Levels", BCP 14, RFC 2119, 1896 DOI 10.17487/RFC2119, March 1997, 1897 . 1899 [RFC3986] Berners-Lee, T., Fielding, R., and L. Masinter, "Uniform 1900 Resource Identifier (URI): Generic Syntax", STD 66, 1901 RFC 3986, DOI 10.17487/RFC3986, January 2005, 1902 . 1904 [RFC5234] Crocker, D., Ed. and P. Overell, "Augmented BNF for Syntax 1905 Specifications: ABNF", STD 68, RFC 5234, 1906 DOI 10.17487/RFC5234, January 2008, 1907 . 1909 [RFC7405] Kyzivat, P., "Case-Sensitive String Support in ABNF", 1910 RFC 7405, DOI 10.17487/RFC7405, December 2014, 1911 . 1913 [RFC8174] Leiba, B., "Ambiguity of Uppercase vs Lowercase in RFC 1914 2119 Key Words", BCP 14, RFC 8174, DOI 10.17487/RFC8174, 1915 May 2017, . 1917 [RFC8446] Rescorla, E., "The Transport Layer Security (TLS) Protocol 1918 Version 1.3", RFC 8446, DOI 10.17487/RFC8446, August 2018, 1919 . 1921 [Semantics] 1922 Fielding, R., Ed., Nottingham, M., Ed., and J. F. Reschke, 1923 Ed., "HTTP Semantics", Work in Progress, Internet-Draft, 1924 draft-ietf-httpbis-semantics-11, August 27, 2020, 1925 . 1928 [USASCII] American National Standards Institute, "Coded Character 1929 Set -- 7-bit American Standard Code for Information 1930 Interchange", ANSI X3.4, 1986. 1932 [Welch] Welch, T. A., "A Technique for High-Performance Data 1933 Compression", IEEE Computer 17(6), June 1984. 1935 13.2. Informative References 1937 [Err4667] RFC Errata, Erratum ID 4667, RFC 7230, 1938 . 1940 [Klein] Klein, A., "Divide and Conquer - HTTP Response Splitting, 1941 Web Cache Poisoning Attacks, and Related Topics", March 1942 2004, . 1945 [Linhart] Linhart, C., Klein, A., Heled, R., and S. Orrin, "HTTP 1946 Request Smuggling", June 2005, 1947 . 1949 [RFC1945] Berners-Lee, T., Fielding, R.T., and H.F. Nielsen, 1950 "Hypertext Transfer Protocol -- HTTP/1.0", RFC 1945, 1951 DOI 10.17487/RFC1945, May 1996, 1952 . 1954 [RFC2045] Freed, N. and N.S. Borenstein, "Multipurpose Internet Mail 1955 Extensions (MIME) Part One: Format of Internet Message 1956 Bodies", RFC 2045, DOI 10.17487/RFC2045, November 1996, 1957 . 1959 [RFC2046] Freed, N. and N. Borenstein, "Multipurpose Internet Mail 1960 Extensions (MIME) Part Two: Media Types", RFC 2046, 1961 DOI 10.17487/RFC2046, November 1996, 1962 . 1964 [RFC2049] Freed, N. and N.S. Borenstein, "Multipurpose Internet Mail 1965 Extensions (MIME) Part Five: Conformance Criteria and 1966 Examples", RFC 2049, DOI 10.17487/RFC2049, November 1996, 1967 . 1969 [RFC2068] Fielding, R., Gettys, J., Mogul, J., Nielsen, H., and T. 1970 Berners-Lee, "Hypertext Transfer Protocol -- HTTP/1.1", 1971 RFC 2068, DOI 10.17487/RFC2068, January 1997, 1972 . 1974 [RFC2557] Palme, F., Hopmann, A., Shelness, N., and E. Stefferud, 1975 "MIME Encapsulation of Aggregate Documents, such as HTML 1976 (MHTML)", RFC 2557, DOI 10.17487/RFC2557, March 1999, 1977 . 1979 [RFC5322] Resnick, P., "Internet Message Format", RFC 5322, 1980 DOI 10.17487/RFC5322, October 2008, 1981 . 1983 [RFC7230] Fielding, R., Ed. and J. F. Reschke, Ed., "Hypertext 1984 Transfer Protocol (HTTP/1.1): Message Syntax and Routing", 1985 RFC 7230, DOI 10.17487/RFC7230, June 2014, 1986 . 1988 [RFC7231] Fielding, R., Ed. and J. F. Reschke, Ed., "Hypertext 1989 Transfer Protocol (HTTP/1.1): Semantics and Content", 1990 RFC 7231, DOI 10.17487/RFC7231, June 2014, 1991 . 1993 [RFC8126] Cotton, M., Leiba, B., and T. Narten, "Guidelines for 1994 Writing an IANA Considerations Section in RFCs", BCP 26, 1995 RFC 8126, DOI 10.17487/RFC8126, June 2017, 1996 . 1998 Appendix A. Collected ABNF 2000 In the collected ABNF below, list rules are expanded as per 2001 Section 5.5.1 of [Semantics]. 2003 BWS = 2005 HTTP-message = start-line CRLF *( field-line CRLF ) CRLF [ 2006 message-body ] 2007 HTTP-name = %x48.54.54.50 ; HTTP 2008 HTTP-version = HTTP-name "/" DIGIT "." DIGIT 2010 OWS = 2012 RWS = 2014 Transfer-Encoding = [ transfer-coding *( OWS "," OWS transfer-coding 2015 ) ] 2017 absolute-URI = 2018 absolute-form = absolute-URI 2019 absolute-path = 2020 asterisk-form = "*" 2021 authority = 2022 authority-form = authority 2024 chunk = chunk-size [ chunk-ext ] CRLF chunk-data CRLF 2025 chunk-data = 1*OCTET 2026 chunk-ext = *( BWS ";" BWS chunk-ext-name [ BWS "=" BWS chunk-ext-val 2027 ] ) 2028 chunk-ext-name = token 2029 chunk-ext-val = token / quoted-string 2030 chunk-size = 1*HEXDIG 2031 chunked-body = *chunk last-chunk trailer-section CRLF 2032 comment = 2034 field-line = field-name ":" OWS field-value OWS 2035 field-name = 2036 field-value = 2038 last-chunk = 1*"0" [ chunk-ext ] CRLF 2040 message-body = *OCTET 2041 method = token 2043 obs-fold = OWS CRLF RWS 2044 obs-text = 2045 origin-form = absolute-path [ "?" query ] 2047 port = 2049 query = 2050 quoted-string = 2052 reason-phrase = 1*( HTAB / SP / VCHAR / obs-text ) 2053 request-line = method SP request-target SP HTTP-version 2054 request-target = origin-form / absolute-form / authority-form / 2055 asterisk-form 2057 start-line = request-line / status-line 2058 status-code = 3DIGIT 2059 status-line = HTTP-version SP status-code SP [ reason-phrase ] 2061 token = 2062 trailer-section = *( field-line CRLF ) 2063 transfer-coding = token *( OWS ";" OWS transfer-parameter ) 2064 transfer-parameter = token BWS "=" BWS ( token / quoted-string ) 2066 uri-host = 2068 Appendix B. Differences between HTTP and MIME 2070 HTTP/1.1 uses many of the constructs defined for the Internet Message 2071 Format [RFC5322] and the Multipurpose Internet Mail Extensions (MIME) 2072 [RFC2045] to allow a message body to be transmitted in an open 2073 variety of representations and with extensible fields. However, RFC 2074 2045 is focused only on email; applications of HTTP have many 2075 characteristics that differ from email; hence, HTTP has features that 2076 differ from MIME. These differences were carefully chosen to 2077 optimize performance over binary connections, to allow greater 2078 freedom in the use of new media types, to make date comparisons 2079 easier, and to acknowledge the practice of some early HTTP servers 2080 and clients. 2082 This appendix describes specific areas where HTTP differs from MIME. 2083 Proxies and gateways to and from strict MIME environments need to be 2084 aware of these differences and provide the appropriate conversions 2085 where necessary. 2087 B.1. MIME-Version 2089 HTTP is not a MIME-compliant protocol. However, messages can include 2090 a single MIME-Version header field to indicate what version of the 2091 MIME protocol was used to construct the message. Use of the MIME- 2092 Version header field indicates that the message is in full 2093 conformance with the MIME protocol (as defined in [RFC2045]). 2094 Senders are responsible for ensuring full conformance (where 2095 possible) when exporting HTTP messages to strict MIME environments. 2097 B.2. Conversion to Canonical Form 2099 MIME requires that an Internet mail body part be converted to 2100 canonical form prior to being transferred, as described in Section 4 2101 of [RFC2049]. Section 7.1.1.2 of [Semantics] describes the forms 2102 allowed for subtypes of the "text" media type when transmitted over 2103 HTTP. [RFC2046] requires that content with a type of "text" 2104 represent line breaks as CRLF and forbids the use of CR or LF outside 2105 of line break sequences. HTTP allows CRLF, bare CR, and bare LF to 2106 indicate a line break within text content. 2108 A proxy or gateway from HTTP to a strict MIME environment ought to 2109 translate all line breaks within text media types to the RFC 2049 2110 canonical form of CRLF. Note, however, this might be complicated by 2111 the presence of a Content-Encoding and by the fact that HTTP allows 2112 the use of some charsets that do not use octets 13 and 10 to 2113 represent CR and LF, respectively. 2115 Conversion will break any cryptographic checksums applied to the 2116 original content unless the original content is already in canonical 2117 form. Therefore, the canonical form is recommended for any content 2118 that uses such checksums in HTTP. 2120 B.3. Conversion of Date Formats 2122 HTTP/1.1 uses a restricted set of date formats (Section 5.4.1.5 of 2123 [Semantics]) to simplify the process of date comparison. Proxies and 2124 gateways from other protocols ought to ensure that any Date header 2125 field present in a message conforms to one of the HTTP/1.1 formats 2126 and rewrite the date if necessary. 2128 B.4. Conversion of Content-Encoding 2130 MIME does not include any concept equivalent to HTTP/1.1's Content- 2131 Encoding header field. Since this acts as a modifier on the media 2132 type, proxies and gateways from HTTP to MIME-compliant protocols 2133 ought to either change the value of the Content-Type header field or 2134 decode the representation before forwarding the message. (Some 2135 experimental applications of Content-Type for Internet mail have used 2136 a media-type parameter of ";conversions=" to perform 2137 a function equivalent to Content-Encoding. However, this parameter 2138 is not part of the MIME standards). 2140 B.5. Conversion of Content-Transfer-Encoding 2142 HTTP does not use the Content-Transfer-Encoding field of MIME. 2143 Proxies and gateways from MIME-compliant protocols to HTTP need to 2144 remove any Content-Transfer-Encoding prior to delivering the response 2145 message to an HTTP client. 2147 Proxies and gateways from HTTP to MIME-compliant protocols are 2148 responsible for ensuring that the message is in the correct format 2149 and encoding for safe transport on that protocol, where "safe 2150 transport" is defined by the limitations of the protocol being used. 2151 Such a proxy or gateway ought to transform and label the data with an 2152 appropriate Content-Transfer-Encoding if doing so will improve the 2153 likelihood of safe transport over the destination protocol. 2155 B.6. MHTML and Line Length Limitations 2157 HTTP implementations that share code with MHTML [RFC2557] 2158 implementations need to be aware of MIME line length limitations. 2159 Since HTTP does not have this limitation, HTTP does not fold long 2160 lines. MHTML messages being transported by HTTP follow all 2161 conventions of MHTML, including line length limitations and folding, 2162 canonicalization, etc., since HTTP transfers message-bodies as 2163 payload and, aside from the "multipart/byteranges" type 2164 (Section 7.3.5 of [Semantics]), does not interpret the content or any 2165 MIME header lines that might be contained therein. 2167 Appendix C. HTTP Version History 2169 HTTP has been in use since 1990. The first version, later referred 2170 to as HTTP/0.9, was a simple protocol for hypertext data transfer 2171 across the Internet, using only a single request method (GET) and no 2172 metadata. HTTP/1.0, as defined by [RFC1945], added a range of 2173 request methods and MIME-like messaging, allowing for metadata to be 2174 transferred and modifiers placed on the request/response semantics. 2175 However, HTTP/1.0 did not sufficiently take into consideration the 2176 effects of hierarchical proxies, caching, the need for persistent 2177 connections, or name-based virtual hosts. The proliferation of 2178 incompletely implemented applications calling themselves "HTTP/1.0" 2179 further necessitated a protocol version change in order for two 2180 communicating applications to determine each other's true 2181 capabilities. 2183 HTTP/1.1 remains compatible with HTTP/1.0 by including more stringent 2184 requirements that enable reliable implementations, adding only those 2185 features that can either be safely ignored by an HTTP/1.0 recipient 2186 or only be sent when communicating with a party advertising 2187 conformance with HTTP/1.1. 2189 HTTP/1.1 has been designed to make supporting previous versions easy. 2190 A general-purpose HTTP/1.1 server ought to be able to understand any 2191 valid request in the format of HTTP/1.0, responding appropriately 2192 with an HTTP/1.1 message that only uses features understood (or 2193 safely ignored) by HTTP/1.0 clients. Likewise, an HTTP/1.1 client 2194 can be expected to understand any valid HTTP/1.0 response. 2196 Since HTTP/0.9 did not support header fields in a request, there is 2197 no mechanism for it to support name-based virtual hosts (selection of 2198 resource by inspection of the Host header field). Any server that 2199 implements name-based virtual hosts ought to disable support for 2200 HTTP/0.9. Most requests that appear to be HTTP/0.9 are, in fact, 2201 badly constructed HTTP/1.x requests caused by a client failing to 2202 properly encode the request-target. 2204 C.1. Changes from HTTP/1.0 2206 This section summarizes major differences between versions HTTP/1.0 2207 and HTTP/1.1. 2209 C.1.1. Multihomed Web Servers 2211 The requirements that clients and servers support the Host header 2212 field (Section 6.5 of [Semantics]), report an error if it is missing 2213 from an HTTP/1.1 request, and accept absolute URIs (Section 3.2) are 2214 among the most important changes defined by HTTP/1.1. 2216 Older HTTP/1.0 clients assumed a one-to-one relationship of IP 2217 addresses and servers; there was no other established mechanism for 2218 distinguishing the intended server of a request than the IP address 2219 to which that request was directed. The Host header field was 2220 introduced during the development of HTTP/1.1 and, though it was 2221 quickly implemented by most HTTP/1.0 browsers, additional 2222 requirements were placed on all HTTP/1.1 requests in order to ensure 2223 complete adoption. At the time of this writing, most HTTP-based 2224 services are dependent upon the Host header field for targeting 2225 requests. 2227 C.1.2. Keep-Alive Connections 2229 In HTTP/1.0, each connection is established by the client prior to 2230 the request and closed by the server after sending the response. 2231 However, some implementations implement the explicitly negotiated 2232 ("Keep-Alive") version of persistent connections described in 2233 Section 19.7.1 of [RFC2068]. 2235 Some clients and servers might wish to be compatible with these 2236 previous approaches to persistent connections, by explicitly 2237 negotiating for them with a "Connection: keep-alive" request header 2238 field. However, some experimental implementations of HTTP/1.0 2239 persistent connections are faulty; for example, if an HTTP/1.0 proxy 2240 server doesn't understand Connection, it will erroneously forward 2241 that header field to the next inbound server, which would result in a 2242 hung connection. 2244 One attempted solution was the introduction of a Proxy-Connection 2245 header field, targeted specifically at proxies. In practice, this 2246 was also unworkable, because proxies are often deployed in multiple 2247 layers, bringing about the same problem discussed above. 2249 As a result, clients are encouraged not to send the Proxy-Connection 2250 header field in any requests. 2252 Clients are also encouraged to consider the use of Connection: keep- 2253 alive in requests carefully; while they can enable persistent 2254 connections with HTTP/1.0 servers, clients using them will need to 2255 monitor the connection for "hung" requests (which indicate that the 2256 client ought stop sending the header field), and this mechanism ought 2257 not be used by clients at all when a proxy is being used. 2259 C.1.3. Introduction of Transfer-Encoding 2261 HTTP/1.1 introduces the Transfer-Encoding header field (Section 6.1). 2262 Transfer codings need to be decoded prior to forwarding an HTTP 2263 message over a MIME-compliant protocol. 2265 C.2. Changes from RFC 7230 2267 Most of the sections introducing HTTP's design goals, history, 2268 architecture, conformance criteria, protocol versioning, URIs, 2269 message routing, and header fields have been moved to [Semantics]. 2270 This document has been reduced to just the messaging syntax and 2271 connection management requirements specific to HTTP/1.1. 2273 Prohibited generation of bare CRs outside of payload body. 2274 (Section 2.2) 2276 In the ABNF for chunked extensions, re-introduced (bad) whitespace 2277 around ";" and "=". Whitespace was removed in [RFC7230], but that 2278 change was found to break existing implementations (see [Err4667]). 2279 (Section 7.1.1) 2281 Trailer field semantics now transcend the specifics of chunked 2282 encoding. The decoding algorithm for chunked (Section 7.1.3) has 2283 been updated to encourage storage/forwarding of trailer fields 2284 separately from the header section, to only allow merging into the 2285 header section if the recipient knows the corresponding field 2286 definition permits and defines how to merge, and otherwise to discard 2287 the trailer fields instead of merging. The trailer part is now 2288 called the trailer section to be more consistent with the header 2289 section and more distinct from a body part. (Section 7.1.2) 2291 Disallowed transfer coding parameters called "q" in order to avoid 2292 conflicts with the use of ranks in the TE header field. 2293 (Section 7.3) 2295 Appendix D. Change Log 2297 This section is to be removed before publishing as an RFC. 2299 D.1. Between RFC7230 and draft 00 2301 The changes were purely editorial: 2303 o Change boilerplate and abstract to indicate the "draft" status, 2304 and update references to ancestor specifications. 2306 o Adjust historical notes. 2308 o Update links to sibling specifications. 2310 o Replace sections listing changes from RFC 2616 by new empty 2311 sections referring to RFC 723x. 2313 o Remove acknowledgements specific to RFC 723x. 2315 o Move "Acknowledgements" to the very end and make them unnumbered. 2317 D.2. Since draft-ietf-httpbis-messaging-00 2319 The changes in this draft are editorial, with respect to HTTP as a 2320 whole, to move all core HTTP semantics into [Semantics]: 2322 o Moved introduction, architecture, conformance, and ABNF extensions 2323 from RFC 7230 (Messaging) to semantics [Semantics]. 2325 o Moved discussion of MIME differences from RFC 7231 (Semantics) to 2326 Appendix B since they mostly cover transforming 1.1 messages. 2328 o Moved all extensibility tips, registration procedures, and 2329 registry tables from the IANA considerations to normative 2330 sections, reducing the IANA considerations to just instructions 2331 that will be removed prior to publication as an RFC. 2333 D.3. Since draft-ietf-httpbis-messaging-01 2335 o Cite RFC 8126 instead of RFC 5226 () 2338 o Resolved erratum 4779, no change needed here 2339 (, 2340 ) 2342 o In Section 7, fixed prose claiming transfer parameters allow bare 2343 names (, 2344 ) 2346 o Resolved erratum 4225, no change needed here 2347 (, 2348 ) 2350 o Replace "response code" with "response status code" 2351 (, 2352 ) 2354 o In Section 9.3, clarify statement about HTTP/1.0 keep-alive 2355 (, 2356 ) 2358 o In Section 7.1.1, re-introduce (bad) whitespace around ";" and "=" 2359 (, 2360 , ) 2363 o In Section 7.3, state that transfer codings should not use 2364 parameters named "q" (, ) 2367 o In Section 7, mark coding name "trailers" as reserved in the IANA 2368 registry () 2370 D.4. Since draft-ietf-httpbis-messaging-02 2372 o In Section 4, explain why the reason phrase should be ignored by 2373 clients (). 2375 o Add Section 9.2 to explain how request/response correlation is 2376 performed () 2378 D.5. Since draft-ietf-httpbis-messaging-03 2380 o In Section 9.2, caution against treating data on a connection as 2381 part of a not-yet-issued request () 2384 o In Section 7, remove the predefined codings from the ABNF and make 2385 it generic instead () 2388 o Use RFC 7405 ABNF notation for case-sensitive string constants 2389 () 2391 D.6. Since draft-ietf-httpbis-messaging-04 2393 o In Section 6.7 of [Semantics], clarify that protocol-name is to be 2394 matched case-insensitively () 2397 o In Section 5.2, add leading optional whitespace to obs-fold ABNF 2398 (, 2399 ) 2401 o In Section 4, add clarifications about empty reason phrases 2402 () 2404 o Move discussion of retries from Section 9.3.1 into [Semantics] 2405 () 2407 D.7. Since draft-ietf-httpbis-messaging-05 2409 o In Section 7.1.2, the trailer part has been renamed the trailer 2410 section (for consistency with the header section) and trailers are 2411 no longer merged as header fields by default, but rather can be 2412 discarded, kept separate from header fields, or merged with header 2413 fields only if understood and defined as being mergeable 2414 () 2416 o In Section 2.1 and related Sections, move the trailing CRLF from 2417 the line grammars into the message format 2418 () 2420 o Moved Section 2.3 down () 2423 o In Section 6.7 of [Semantics], use 'websocket' instead of 2424 'HTTP/2.0' in examples () 2427 o Move version non-specific text from Section 6 into semantics as 2428 "payload body" () 2430 o In Section 9.8, add text from RFC 2818 2431 () 2433 D.8. Since draft-ietf-httpbis-messaging-06 2435 o In Section 12.5, update the APLN protocol id for HTTP/1.1 2436 () 2438 o In Section 5, align with updates to field terminology in semantics 2439 () 2441 o In Section 6.8 of [Semantics], clarify that new connection options 2442 indeed need to be registered () 2445 o In Section 1.1, reference RFC 8174 as well 2446 () 2448 D.9. Since draft-ietf-httpbis-messaging-07 2450 o Move TE: trailers into [Semantics] () 2453 o In Section 6.3, adjust requirements for handling multiple content- 2454 length values () 2456 o Throughout, replace "effective request URI" with "target URI" 2457 () 2459 o In Section 6.1, don't claim Transfer-Encoding is supported by 2460 HTTP/2 or later () 2462 D.10. Since draft-ietf-httpbis-messaging-08 2464 o In Section 2.2, disallow bare CRs () 2467 o Appendix A now uses the sender variant of the "#" list expansion 2468 () 2470 o In Section 5, adjust IANA "Close" entry for new registry format 2471 () 2473 D.11. Since draft-ietf-httpbis-messaging-09 2475 o Switch to xml2rfc v3 mode for draft generation 2476 () 2478 D.12. Since draft-ietf-httpbis-messaging-10 2480 o In Section 6.3, note that TCP half-close does not delimit a 2481 request; talk about corresponding server-side behaviour in 2482 Section 9.6 () 2484 o Moved requirements specific to HTTP/1.1 from [Semantics] into 2485 Section 3.2 () 2487 o In Section 6.1 (Transfer-Encoding), adjust ABNF to allow empty 2488 lists () 2490 o In Section 9.7, add text from RFC 2818 2491 () 2493 o Moved definitions of "TE" and "Upgrade" into [Semantics] 2494 () 2496 o Moved definition of "Connection" into [Semantics] 2497 () 2499 Acknowledgments 2501 See Appendix "Acknowledgments" of [Semantics]. 2503 Authors' Addresses 2505 Roy T. Fielding (editor) 2506 Adobe 2507 345 Park Ave 2508 San Jose, CA 95110 2509 United States of America 2511 Email: fielding@gbiv.com 2512 URI: https://roy.gbiv.com/ 2514 Mark Nottingham (editor) 2515 Fastly 2516 Prahran VIC 2517 Australia 2519 Email: mnot@mnot.net 2520 URI: https://www.mnot.net/ 2522 Julian F. Reschke (editor) 2523 greenbytes GmbH 2524 Hafenweg 16 2525 48155 Münster 2526 Germany 2528 Email: julian.reschke@greenbytes.de 2529 URI: https://greenbytes.de/tech/webdav/