rfc8941xml2.original.xml   rfc8941.xml 
<?xml version="1.0" encoding="UTF-8"?> <?xml version='1.0' encoding='utf-8'?>
<?xml-stylesheet type="text/xsl" href="rfc2629.xslt" ?>
<!-- generated by https://github.com/cabo/kramdown-rfc2629 version 1.3.8 -->
<!DOCTYPE rfc SYSTEM "rfc2629.dtd" [
]>
<?rfc toc="yes"?>
<?rfc tocindent="yes"?>
<?rfc sortrefs="yes"?>
<?rfc symrefs="yes"?>
<?rfc strict="yes"?>
<?rfc compact="yes"?>
<?rfc comments="yes"?>
<?rfc inline="yes"?>
<?rfc tocdepth="3"?>
<rfc ipr="trust200902" docName="draft-ietf-httpbis-header-structure-19" category
="std">
<!DOCTYPE rfc SYSTEM "rfc2629-xhtml.ent">
<rfc xmlns:xi="http://www.w3.org/2001/XInclude" ipr="trust200902" docName="draft
-ietf-httpbis-header-structure-19" number="8941" submissionType="IETF" category=
"std" consensus="true" obsoletes="" updates="" xml:lang="en" tocInclude="true" s
ortRefs="true" symRefs="true" tocDepth="3" version="3">
<!-- xml2rfc v2v3 conversion 2.46.0 -->
<front> <front>
<title>Structured Field Values for HTTP</title> <title>Structured Field Values for HTTP</title>
<seriesInfo name="RFC" value="8941"/>
<author initials="M." surname="Nottingham" fullname="Mark Nottingham"> <author initials="M." surname="Nottingham" fullname="Mark Nottingham">
<organization>Fastly</organization> <organization>Fastly</organization>
<address> <address>
<postal>
<city>Prahran</city>
<region>VIC</region>
<country>Australia</country>
</postal>
<email>mnot@mnot.net</email> <email>mnot@mnot.net</email>
<uri>https://www.mnot.net/</uri> <uri>https://www.mnot.net/</uri>
</address> </address>
</author> </author>
<author initials="P-H." surname="Kamp" fullname="Poul-Henning Kamp"> <author initials="P-H." surname="Kamp" fullname="Poul-Henning Kamp">
<organization>The Varnish Cache Project</organization> <organization>The Varnish Cache Project</organization>
<address> <address>
<email>phk@varnish-cache.org</email> <email>phk@varnish-cache.org</email>
</address> </address>
</author> </author>
<date month="February" year="2021"/>
<date />
<area>Applications and Real-Time</area> <area>Applications and Real-Time</area>
<workgroup>HTTP</workgroup> <workgroup>HTTP</workgroup>
<keyword>Internet-Draft</keyword>
<abstract> <abstract>
<t>This document describes a set of data types and associated algorithms
<t>This document describes a set of data types and associated algorithms that ar that are intended to make it easier and safer to define and handle HTTP
e intended to make it easier and safer to define and handle HTTP header and trai header and trailer fields, known as "Structured Fields", "Structured
ler fields, known as “Structured Fields”, “Structured Headers”, or “Structured T Headers", or "Structured Trailers". It is intended for use by
railers”. It is intended for use by specifications of new HTTP fields that wish specifications of new HTTP fields that wish to use a common syntax that
to use a common syntax that is more restrictive than traditional HTTP field valu is more restrictive than traditional HTTP field values.</t>
es.</t>
</abstract> </abstract>
<note title="Note to Readers">
<t><spanx style="emph">RFC EDITOR: please remove this section before publication
</spanx></t>
<t>Discussion of this draft takes place on the HTTP working group mailing list (
ietf-http-wg@w3.org), which is archived at <eref target="https://lists.w3.org/Ar
chives/Public/ietf-http-wg/">https://lists.w3.org/Archives/Public/ietf-http-wg/<
/eref>.</t>
<t>Working Group information can be found at <eref target="https://httpwg.github
.io/">https://httpwg.github.io/</eref>; source code and issues list for this dra
ft can be found at <eref target="https://github.com/httpwg/http-extensions/label
s/header-structure">https://github.com/httpwg/http-extensions/labels/header-stru
cture</eref>.</t>
<t>Tests for implementations are collected at <eref target="https://github.com/h
ttpwg/structured-field-tests">https://github.com/httpwg/structured-field-tests</
eref>.</t>
<t>Implementations are tracked at <eref target="https://github.com/httpwg/wiki/w
iki/Structured-Headers">https://github.com/httpwg/wiki/wiki/Structured-Headers</
eref>.</t>
</note>
</front> </front>
<middle> <middle>
<section anchor="introduction">
<name>Introduction</name>
<t>Specifying the syntax of new HTTP header (and trailer) fields is an
onerous task; even with the guidance in <xref target="RFC7231" section="8.
3.1"/>, there are many decisions -- and
pitfalls -- for a prospective HTTP field author.</t>
<t>Once a field is defined, bespoke parsers and serializers often need
to be written, because each field value has a slightly different handling
of what looks like common syntax.</t>
<t>This document introduces a set of common data structures for use in
definitions of new HTTP field values to address these problems. In
particular, it defines a generic, abstract model for them, along with a
concrete serialization for expressing that model in HTTP <xref target="RFC
7230"/> header and trailer fields.</t>
<t>An HTTP field that is defined as a "Structured Header" or "Structured
Trailer" (if the field can be either, it is a "Structured Field") uses
the types defined in this specification to define its syntax and basic
handling rules, thereby simplifying both its definition by specification
writers and handling by implementations.</t>
<t>Additionally, future versions of HTTP can define alternative
serializations of the abstract model of these structures, allowing
fields that use that model to be transmitted more efficiently without
being redefined.</t>
<t>Note that it is not a goal of this document to redefine the syntax of
existing HTTP fields; the mechanisms described herein are only intended
to be used with fields that explicitly opt into them.</t>
<t><xref target="specify"/> describes how to specify a
Structured Field.</t>
<t><xref target="types"/> defines a number of abstract
data types that can be used in Structured Fields.</t>
<t>Those abstract types can be serialized into and parsed from HTTP
field values using the algorithms described in <xref target="text"/>.</t>
<section anchor="strict">
<name>Intentionally Strict Processing</name>
<t>This specification intentionally defines strict parsing and
serialization behaviors using step-by-step algorithms; the only error
handling defined is to fail the operation altogether.</t>
<t>It is designed to encourage faithful implementation and
good interoperability. Therefore, an implementation that tried to be
helpful by being more tolerant of input would make interoperability
worse, since that would create pressure on other implementations to
implement similar (but likely subtly different) workarounds.</t>
<t>In other words, strict processing is an intentional feature of this
specification; it allows non-conformant input to be discovered and
corrected by the producer early and avoids both interoperability and
security issues that might otherwise result.</t>
<t>Note that as a result of this strictness, if a field is appended to
by multiple parties (e.g., intermediaries or different components in
the sender), an error in one party's value is likely to cause the
entire field value to fail parsing.</t>
</section>
<section anchor="notational-conventions">
<name>Notational Conventions</name>
<t>
The key words "<bcp14>MUST</bcp14>", "<bcp14>MUST NOT</bcp14>",
"<bcp14>REQUIRED</bcp14>", "<bcp14>SHALL</bcp14>", "<bcp14>SHALL
NOT</bcp14>", "<bcp14>SHOULD</bcp14>", "<bcp14>SHOULD NOT</bcp14>",
"<bcp14>RECOMMENDED</bcp14>", "<bcp14>NOT RECOMMENDED</bcp14>",
"<bcp14>MAY</bcp14>", and "<bcp14>OPTIONAL</bcp14>" in this document are
to be interpreted as
described in BCP 14 <xref target="RFC2119"/> <xref target="RFC8174"/>
when, and only when, they appear in all capitals, as shown here.
</t>
<t>This document uses algorithms to specify parsing and serialization
behaviors and the Augmented Backus-Naur Form (ABNF) notation of <xref tar
get="RFC5234"/> to illustrate expected syntax in
HTTP header fields. In doing so, it uses the VCHAR, SP, DIGIT, ALPHA,
and DQUOTE rules from <xref target="RFC5234"/>. It
also includes the tchar and OWS rules from <xref target="RFC7230"/>.</t>
<t>When parsing from HTTP fields, implementations <bcp14>MUST</bcp14> ha
ve behavior
that is indistinguishable from following the algorithms. If there is
disagreement between the parsing algorithms and ABNF, the specified
algorithms take precedence.</t>
<t>For serialization to HTTP fields, the ABNF illustrates their
expected wire representations, and the algorithms define the
recommended way to produce them. Implementations <bcp14>MAY</bcp14> vary
from the
specified behavior so long as the output is still correctly handled by
the parsing algorithm described in <xref target="text-parse"/>.</t>
</section>
</section>
<section anchor="specify">
<name>Defining New Structured Fields</name>
<t>To specify an HTTP field as a Structured Field, its authors need to:</t
>
<ul>
<li>Normatively reference this specification. Recipients and
generators of the field need to know that the requirements of this
document are in effect.</li>
<li>Identify whether the field is a Structured Header (i.e., it can
only be used in the header section -- the common case), a Structured
Trailer (only in the trailer section), or a Structured Field
(both).</li>
<li>Specify the type of the field value; either List (<xref target="list
"/>), Dictionary (<xref target="dictionary"/>), or Item (<xref target="item"/>).
</li>
<li>Define the semantics of the field value.</li>
<li>Specify any additional constraints upon the field value, as well
as the consequences when those constraints are violated.</li>
</ul>
<t>Typically, this means that a field definition will specify the
top-level type -- List, Dictionary, or Item -- and then define its
allowable types and constraints upon them. For example, a header
defined as a List might have all Integer members, or a mix of types; a
header defined as an Item might allow only Strings, and additionally
only strings beginning with the letter "Q", or strings in
lowercase. Likewise, Inner Lists (<xref target="inner-list"/>) are only va
lid when a field definition explicitly
allows them.</t>
<t>When parsing fails, the entire field is ignored (see <xref target="text
-parse"/>); in most situations, violating
field-specific constraints should have the same effect. Thus, if a
header is defined as an Item and required to be an Integer, but a String
is received, the field will by default be ignored. If the field requires
different error handling, this should be explicitly specified.</t>
<t>Both Items and Inner Lists allow parameters as an extensibility
mechanism; this means that values can later be extended to accommodate
more information, if need be. To preserve forward compatibility, field
specifications are discouraged from defining the presence of an
unrecognized parameter as an error condition.</t>
<t>To further assure that this extensibility is available in the future,
and to encourage consumers to use a complete parser implementation, a
field definition can specify that "grease" parameters be added by
senders. A specification could stipulate that all parameters that fit a
defined pattern are reserved for this use and then encourage them to be
sent on some portion of requests. This helps to discourage recipients
from writing a parser that does not account for Parameters.</t>
<t>Specifications that use Dictionaries can also allow for forward
compatibility by requiring that the presence of -- as well as value and
type associated with -- unknown members be ignored. Subsequent specificati
ons
can then add additional members, specifying constraints on them as
appropriate.</t>
<t>An extension to a Structured Field can then require that an entire
field value be ignored by a recipient that understands the extension if
constraints on the value it defines are not met.</t>
<t>A field definition cannot relax the requirements of this
specification because doing so would preclude handling by generic
software; they can only add additional constraints (for example, on the
numeric range of Integers and Decimals, the format of Strings and
Tokens, the types allowed in a Dictionary's values, or the number of
Items in a List). Likewise, field definitions can only use this
specification for the entire field value, not a portion thereof.</t>
<t>This specification defines minimums for the length or number of
various structures supported by implementations. It does not specify
maximum sizes in most cases, but authors should be aware that HTTP
implementations do impose various limits on the size of individual
fields, the total number of fields, and/or the size of the entire header
or trailer section.</t>
<t>Specifications can refer to a field name as a "structured header
name", "structured trailer name", or "structured field name" as
appropriate. Likewise, they can refer its field value as a "structured
header value", "structured trailer value", or "structured field value" as
necessary.
Field definitions are encouraged to use the ABNF rules
beginning with "sf-" defined in this specification; other rules in this
specification are not intended to be used in field definitions.</t>
<t>For example, a fictitious Foo-Example header field might be specified
as:</t>
<section anchor="introduction" title="Introduction"> <blockquote>
<t>42. Foo-Example Header</t>
<t>Specifying the syntax of new HTTP header (and trailer) fields is an onerous t
ask; even with the guidance in Section 8.3.1 of <xref target="RFC7231"/>, there
are many decisions – and pitfalls – for a prospective HTTP field author.</t>
<t>Once a field is defined, bespoke parsers and serializers often need to be wri
tten, because each field value has slightly different handling of what looks lik
e common syntax.</t>
<t>This document introduces a set of common data structures for use in definitio
ns of new HTTP field values to address these problems. In particular, it defines
a generic, abstract model for them, along with a concrete serialization for exp
ressing that model in HTTP <xref target="RFC7230"/> header and trailer fields.</
t>
<t>A HTTP field that is defined as a “Structured Header” or “Structured Trailer”
(if the field can be either, it is a “Structured Field”) uses the types defined
in this specification to define its syntax and basic handling rules, thereby si
mplifying both its definition by specification writers and handling by implement
ations.</t>
<t>Additionally, future versions of HTTP can define alternative serializations o
f the abstract model of these structures, allowing fields that use that model to
be transmitted more efficiently without being redefined.</t>
<t>Note that it is not a goal of this document to redefine the syntax of existin
g HTTP fields; the mechanisms described herein are only intended to be used with
fields that explicitly opt into them.</t>
<t><xref target="specify"/> describes how to specify a Structured Field.</t>
<t><xref target="types"/> defines a number of abstract data types that can be us
ed in Structured Fields.</t>
<t>Those abstract types can be serialized into and parsed from HTTP field values
using the algorithms described in <xref target="text"/>.</t>
<section anchor="strict" title="Intentionally Strict Processing">
<t>This specification intentionally defines strict parsing and serialization beh
aviors using step-by-step algorithms; the only error handling defined is to fail
the operation altogether.</t>
<t>It is designed to encourage faithful implementation and therefore good intero
perability. Therefore, an implementation that tried to be helpful by being more
tolerant of input would make interoperability worse, since that would create pre
ssure on other implementations to implement similar (but likely subtly different
) workarounds.</t>
<t>In other words, strict processing is an intentional feature of this specifica
tion; it allows non-conformant input to be discovered and corrected by the produ
cer early, and avoids both interoperability and security issues that might other
wise result.</t>
<t>Note that as a result of this strictness, if a field is appended to by multip
le parties (e.g., intermediaries, or different components in the sender), an err
or in one party’s value is likely to cause the entire field value to fail parsin
g.</t>
</section>
<section anchor="notational-conventions" title="Notational Conventions">
<t>The key words “MUST”, “MUST NOT”, “REQUIRED”, “SHALL”, “SHALL NOT”, “SHOULD”,
“SHOULD NOT”, “RECOMMENDED”, “NOT RECOMMENDED”, “MAY”, and “OPTIONAL” in this d
ocument are to be interpreted as described in BCP 14 <xref target="RFC2119"/> <x
ref target="RFC8174"/> when, and only when, they appear in all capitals, as show
n here.</t>
<t>This document uses algorithms to specify parsing and serialization behaviors,
and the Augmented Backus-Naur Form (ABNF) notation of <xref target="RFC5234"/>
to illustrate expected syntax in HTTP header fields. In doing so, it uses the VC
HAR, SP, DIGIT, ALPHA and DQUOTE rules from <xref target="RFC5234"/>. It also in
cludes the tchar and OWS rules from <xref target="RFC7230"/>.</t>
<t>When parsing from HTTP fields, implementations MUST have behavior that is ind
istinguishable from following the algorithms. If there is disagreement between t
he parsing algorithms and ABNF, the specified algorithms take precedence.</t>
<t>For serialization to HTTP fields, the ABNF illustrates their expected wire re
presentations, and the algorithms define the recommended way to produce them. Im
plementations MAY vary from the specified behavior so long as the output is stil
l correctly handled by the parsing algorithm.</t>
</section>
</section>
<section anchor="specify" title="Defining New Structured Fields">
<t>To specify a HTTP field as a Structured Field, its authors needs to:</t>
<t><list style="symbols">
<t>Normatively reference this specification. Recipients and generators of the
field need to know that the requirements of this document are in effect.</t>
<t>Identify whether the field is a Structured Header (i.e., it can only be use
d in the header section - the common case), a Structured Trailer (only in the tr
ailer section), or a Structured Field (both).</t>
<t>Specify the type of the field value; either List (<xref target="list"/>), D
ictionary (<xref target="dictionary"/>), or Item (<xref target="item"/>).</t>
<t>Define the semantics of the field value.</t>
<t>Specify any additional constraints upon the field value, as well as the con
sequences when those constraints are violated.</t>
</list></t>
<t>Typically, this means that a field definition will specify the top-level type
– List, Dictionary or Item – and then define its allowable types, and constrain
ts upon them. For example, a header defined as a List might have all Integer mem
bers, or a mix of types; a header defined as an Item might allow only Strings, a
nd additionally only strings beginning with the letter “Q”, or strings in lowerc
ase. Likewise, Inner Lists (<xref target="inner-list"/>) are only valid when a f
ield definition explicitly allows them.</t>
<t>When parsing fails, the entire field is ignored (see <xref target="text-parse
"/>); in most situations, violating field-specific constraints should have the s
ame effect. Thus, if a header is defined as an Item and required to be an Intege
r, but a String is received, the field will by default be ignored. If the field
requires different error handling, this should be explicitly specified.</t>
<t>Both Items and Inner Lists allow parameters as an extensibility mechanism; th
is means that values can later be extended to accommodate more information, if n
eed be. To preserve forward compatibility, field specifications are discouraged
from defining the presence of an unrecognized Parameter as an error condition.</
t>
<t>To further assure that this extensibility is available in the future, and to
encourage consumers to use a complete parser implementation, a field definition
can specify that “grease” Parameters be added by senders. A specification could
stipulate that all Parameters that fit a defined pattern are reserved for this u
se and then encourage them to be sent on some portion of requests. This helps to
discourage recipients from writing a parser that does not account for Parameter
s.</t>
<t>Specifications that use Dictionaries can also allow for forward compatibility
by requiring that the presence of – as well as value and type associated with –
unknown members be ignored. Later specifications can then add additional member
s, specifying constraints on them as appropriate.</t>
<t>An extension to a structured field can then require that an entire field valu
e be ignored by a recipient that understands the extension if constraints on the
value it defines are not met.</t>
<t>A field definition cannot relax the requirements of this specification becaus
e doing so would preclude handling by generic software; they can only add additi
onal constraints (for example, on the numeric range of Integers and Decimals, th
e format of Strings and Tokens, the types allowed in a Dictionary’s values, or t
he number of Items in a List). Likewise, field definitions can only use this spe
cification for the entire field value, not a portion thereof.</t>
<t>This specification defines minimums for the length or number of various struc
tures supported by implementations. It does not specify maximum sizes in most ca
ses, but authors should be aware that HTTP implementations do impose various lim
its on the size of individual fields, the total number of fields, and/or the siz
e of the entire header or trailer section.</t>
<t>Specifications can refer to a field name as a “structured header name”, “stru
ctured trailer name” or “structured field name” as appropriate. Likewise, they c
an refer its field value as a “structured header value”, “structured trailer val
ue” or “structured field value” as necessary. Field definitions are encouraged t
o use the ABNF rules beginning with “sf-“ defined in this specification; other r
ules in this specification are not intended for their use.</t>
<t>For example, a fictitious Foo-Example header field might be specified as:</t>
<figure><artwork type="example"><![CDATA[
42. Foo-Example Header
The Foo-Example HTTP header field conveys information about how <t>The Foo-Example HTTP header field conveys information about how
much Foo the message has. much Foo the message has.</t>
Foo-Example is a Item Structured Header [RFCxxxx]. Its value MUST be <t>Foo-Example is an Item Structured Header [RFC8941]. Its value MUST be
an Integer (Section Y.Y of [RFCxxxx]). Its ABNF is: an Integer (Section 3.3.1 of [RFC8941]). Its ABNF is:</t>
Foo-Example = sf-integer <artwork>
Foo-Example = sf-integer
</artwork>
Its value indicates the amount of Foo in the message, and MUST <t>Its value indicates the amount of Foo in the message, and it MUST
be between 0 and 10, inclusive; other values MUST cause be between 0 and 10, inclusive; other values MUST cause
the entire header field to be ignored. the entire header field to be ignored.</t>
The following parameters are defined: <t>The following parameter is defined:</t>
* A Parameter whose name is "foourl", and whose value is a String
(Section Y.Y of [RFCxxxx]), conveying the Foo URL
for the message. See below for processing requirements.
"foourl" contains a URI-reference (Section 4.1 of [RFC3986]). If <ul>
<li>A parameter whose key is "foourl", and whose value is a String
(Section 3.3.3 of [RFC8941]), conveying the Foo URL
for the message. See below for processing requirements.</li>
</ul>
<t>"foourl" contains a URI-reference (Section 4.1 of [RFC3986]). If
its value is not a valid URI-reference, the entire header field its value is not a valid URI-reference, the entire header field
MUST be ignored. If its value is a relative reference (Section 4.2 MUST be ignored. If its value is a relative reference (Section 4.2
of [RFC3986]), it MUST be resolved (Section 5 of [RFC3986]) before of [RFC3986]), it MUST be resolved (Section 5 of [RFC3986]) before
being used. being used.</t>
For example:
Foo-Example: 2; foourl="https://foo.example.com/"
]]></artwork></figure>
</section>
<section anchor="types" title="Structured Data Types">
<t>This section defines the abstract types for Structured Fields. The ABNF provi
ded represents the on-wire format in HTTP field values.</t>
<t>In summary:</t>
<t><list style="symbols">
<t>There are three top-level types that a HTTP field can be defined as: Lists,
Dictionaries, and Items.</t>
<t>Lists and Dictionaries are containers; their members can be Items or Inner
Lists (which are themselves arrays of Items).</t>
<t>Both Items and Inner Lists can be parameterized with key/value pairs.</t>
</list></t>
<section anchor="list" title="Lists">
<t>Lists are arrays of zero or more members, each of which can be an Item (<xref target="item"/>) or an Inner List (<xref target="inner-list"/>), both of which can be Parameterized (<xref target="param"/>).</t> <t>For example:</t>
<t>The ABNF for Lists in HTTP fields is:</t> <artwork>
Foo-Example: 2; foourl="https://foo.example.com/"
</artwork>
</blockquote>
<figure><artwork type="abnf"><![CDATA[ </section>
<section anchor="types">
<name>Structured Data Types</name>
<t>This section defines the abstract types for Structured Fields. The
ABNF provided represents the on-wire format in HTTP field values.</t>
<t>In summary:</t>
<ul>
<li>There are three top-level types that an HTTP field can be defined
as: Lists, Dictionaries, and Items.</li>
<li>Lists and Dictionaries are containers; their members can be Items
or Inner Lists (which are themselves arrays of Items).</li>
<li>Both Items and Inner Lists can be Parameterized with key/value pairs
.</li>
</ul>
<section anchor="list">
<name>Lists</name>
<t>Lists are arrays of zero or more members, each of which can be an
Item (<xref target="item"/>) or an Inner List (<xref target="inner-list"/
>), both of which can be
Parameterized (<xref target="param"/>).</t>
<t>The ABNF for Lists in HTTP fields is:</t>
<sourcecode type="abnf">
sf-list = list-member *( OWS "," OWS list-member ) sf-list = list-member *( OWS "," OWS list-member )
list-member = sf-item / inner-list list-member = sf-item / inner-list
]]></artwork></figure> </sourcecode>
<t>Each member is separated by a comma and optional whitespace. For
<t>Each member is separated by a comma and optional whitespace. For example, a f example, a field whose value is defined as a List of Tokens could
ield whose value is defined as a List of Strings could look like:</t> look like:</t>
<sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[ Example-List: sugar, tea, rum
Example-StrList: "foo", "bar", "It was the best of times." </sourcecode>
]]></artwork></figure> <t>An empty List is denoted by not serializing the field at all. This
implies that fields defined as Lists have a default empty value.</t>
<t>An empty List is denoted by not serializing the field at all. This implies th <t>Note that Lists can have their members split across multiple lines
at fields defined as Lists have a default empty value.</t> of the same header or trailer section, as per <xref target="RFC7230" sect
ion="3.2.2"/>; for example, the following are
<t>Note that Lists can have their members split across multiple lines inside a h equivalent:</t>
eader or trailer section, as per Section 3.2.2 of <xref target="RFC7230"/>; for <sourcecode type="http-message">
example, the following are equivalent:</t> Example-List: sugar, tea, rum
</sourcecode>
<figure><artwork type="example"><![CDATA[ <t>and</t>
Example-Hdr: foo, bar <sourcecode type="http-message">
]]></artwork></figure> Example-List: sugar, tea
Example-List: rum
<t>and</t> </sourcecode>
<t>However, individual members of a List cannot be safely split
<figure><artwork type="example"><![CDATA[ between lines; see <xref target="text-parse"/>
Example-Hdr: foo for details.</t>
Example-Hdr: bar <t>Parsers <bcp14>MUST</bcp14> support Lists containing at least 1024 me
]]></artwork></figure> mbers. Field
specifications can constrain the types and cardinality of individual
<t>However, individual members of a List cannot be safely split between across l List values as they require.</t>
ines; see <xref target="text-parse"/> for details.</t> <section anchor="inner-list">
<name>Inner Lists</name>
<t>Parsers MUST support Lists containing at least 1024 members. Field specificat <t>An Inner List is an array of zero or more Items (<xref target="item
ions can constrain the types and cardinality of individual List values as they r "/>). Both the individual Items and the
equire.</t> Inner List itself can be Parameterized (<xref target="param"/>).</t>
<t>The ABNF for Inner Lists is:</t>
<section anchor="inner-list" title="Inner Lists"> <sourcecode type="abnf">
<t>An Inner List is an array of zero or more Items (<xref target="item"/>). Both
the individual Items and the Inner List itself can be Parameterized (<xref targ
et="param"/>).</t>
<t>The ABNF for Inner Lists is:</t>
<figure><artwork type="abnf"><![CDATA[
inner-list = "(" *SP [ sf-item *( 1*SP sf-item ) *SP ] ")" inner-list = "(" *SP [ sf-item *( 1*SP sf-item ) *SP ] ")"
parameters parameters
]]></artwork></figure> </sourcecode>
<t>Inner Lists are denoted by surrounding parenthesis, and
<t>Inner Lists are denoted by surrounding parenthesis, and have their values del their values are delimited by one or more spaces. A field whose value i
imited by one or more spaces. A field whose value is defined as a List of Inner s
Lists of Strings could look like:</t> defined as a List of Inner Lists of Strings could look like:</t>
<sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[ Example-List: ("foo" "bar"), ("baz"), ("bat" "one"), ()
Example-StrListList: ("foo" "bar"), ("baz"), ("bat" "one"), () </sourcecode>
]]></artwork></figure> <t>Note that the last member in this example is an empty Inner List.</
t>
<t>Note that the last member in this example is an empty Inner List.</t> <t>A header field whose value is defined as a List of Inner Lists
with Parameters at both levels could look like:</t>
<t>A header field whose value is defined as a List of Inner Lists with Parameter <sourcecode type="http-message">
s at both levels could look like:</t> Example-List: ("foo"; a=1;b=2);lvl=5, ("bar" "baz");lvl=1
</sourcecode>
<figure><artwork type="example"><![CDATA[ <t>Parsers <bcp14>MUST</bcp14> support Inner Lists containing at least
Example-ListListParam: ("foo"; a=1;b=2);lvl=5, ("bar" "baz");lvl=1 256
]]></artwork></figure> members. Field specifications can constrain the types and
cardinality of individual Inner List members as they require.</t>
<t>Parsers MUST support Inner Lists containing at least 256 members. Field speci </section>
fications can constrain the types and cardinality of individual Inner List membe <section anchor="param">
rs as they require.</t> <name>Parameters</name>
<t>Parameters are an ordered map of key-value pairs that are
</section> associated with an Item (<xref target="item"/>) or
<section anchor="param" title="Parameters"> Inner List (<xref target="inner-list"/>).
The keys
<t>Parameters are an ordered map of key-value pairs that are associated with an are unique within the scope of the Parameters they occur within, and
Item (<xref target="item"/>) or Inner List (<xref target="inner-list"/>). The ke the values are bare items (i.e., they themselves cannot be
ys are unique within the scope the Parameters they occur within, and the values parameterized; see <xref target="item"/>).</t>
are bare items (i.e., they themselves cannot be parameterized; see <xref target= <t>Implementations <bcp14>MUST</bcp14> provide access to Parameters both by inde
"item"/>).</t> x and
by key. Specifications <bcp14>MAY</bcp14> use either means of accessing them
<t>The ABNF for Parameters is:</t> .</t>
<t>The ABNF for Parameters is:</t>
<figure><artwork type="abnf"><![CDATA[ <sourcecode type="abnf">
parameters = *( ";" *SP parameter ) parameters = *( ";" *SP parameter )
parameter = param-name [ "=" param-value ] parameter = param-key [ "=" param-value ]
param-name = key param-key = key
key = ( lcalpha / "*" ) key = ( lcalpha / "*" )
*( lcalpha / DIGIT / "_" / "-" / "." / "*" ) *( lcalpha / DIGIT / "_" / "-" / "." / "*" )
lcalpha = %x61-7A ; a-z lcalpha = %x61-7A ; a-z
param-value = bare-item param-value = bare-item
]]></artwork></figure> </sourcecode>
<t>Note that parameters are ordered as serialized, and parameter
<t>Note that Parameters are ordered as serialized, and Parameter keys cannot con keys cannot contain uppercase letters. A parameter is separated from
tain uppercase letters. A parameter is separated from its Item or Inner List and its Item or Inner List and other parameters by a semicolon. For
other parameters by a semicolon. For example:</t> example:</t>
<sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[ Example-List: abc;a=1;b=2; cde_456, (ghi;jk=4 l);q="9";r=w
Example-ParamList: abc;a=1;b=2; cde_456, (ghi;jk=4 l);q="9";r=w </sourcecode>
]]></artwork></figure> <t>Parameters whose value is Boolean (see <xref target="boolean"/>) tr
ue <bcp14>MUST</bcp14> omit that value when serialized. For
<t>Parameters whose value is Boolean (see <xref target="boolean"/>) true MUST om example, the "a" parameter here is true, while the "b" parameter is
it that value when serialized. For example, the “a” parameter here is true, whil false:</t>
e the “b” parameter is false:</t> <sourcecode type="http-message">
Example-Integer: 1; a; b=?0
<figure><artwork type="example"><![CDATA[ </sourcecode>
Example-Int: 1; a; b=?0 <t>Note that this requirement is only on serialization; parsers are
]]></artwork></figure> still required to correctly handle the true value when it appears in
a parameter.</t>
<t>Note that this requirement is only on serialization; parsers are still requir <t>Parsers <bcp14>MUST</bcp14> support at least 256 parameters on an I
ed to correctly handle the true value when it appears in a parameter.</t> tem or Inner
List, and support parameter keys with at least 64 characters. Field
<t>Parsers MUST support at least 256 parameters on an Item or Inner List, and su specifications can constrain the order of individual parameters, as
pport parameter keys with at least 64 characters. Field specifications can const well as their values' types as required.</t>
rain the order of individual Parameters, as well as their values’ types as requi </section>
red.</t> </section>
<section anchor="dictionary">
</section> <name>Dictionaries</name>
</section> <t>Dictionaries are ordered maps of key-value pairs, where the keys
<section anchor="dictionary" title="Dictionaries"> are short textual strings and the values are Items (<xref target="item"/>
) or arrays of Items, both of which
<t>Dictionaries are ordered maps of name-value pairs, where the names are short can be Parameterized (<xref target="param"/>). There
textual strings and the values are Items (<xref target="item"/>) or arrays of It can be zero or more members, and their keys are unique in the scope
ems, both of which can be Parameterized (<xref target="param"/>). There can be z of the Dictionary they occur within.</t>
ero or more members, and their names are unique in the scope of the Dictionary t <t>Implementations <bcp14>MUST</bcp14> provide access to Dictionaries bo
hey occur within.</t> th by index
and by key. Specifications <bcp14>MAY</bcp14> use either means of accessi
<t>Implementations MUST provide access to Dictionaries both by index and by name ng the
. Specifications MAY use either means of accessing the members.</t> members.</t>
<t>The ABNF for Dictionaries is:</t>
<t>The ABNF for Dictionaries is:</t> <sourcecode type="abnf">
<figure><artwork type="abnf"><![CDATA[
sf-dictionary = dict-member *( OWS "," OWS dict-member ) sf-dictionary = dict-member *( OWS "," OWS dict-member )
dict-member = member-name [ "=" member-value ] dict-member = member-key ( parameters / ( "=" member-value ))
member-name = key member-key = key
member-value = sf-item / inner-list member-value = sf-item / inner-list
]]></artwork></figure> </sourcecode>
<t>Members are ordered as serialized and separated by a comma with
<t>Members are ordered as serialized, and separated by a comma with optional whi optional whitespace. Member keys cannot contain uppercase
tespace. Member names cannot contain uppercase characters. Names and values are characters. Keys and values are separated by "=" (without
separated by “=” (without whitespace). For example:</t> whitespace). For example:</t>
<sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[
Example-Dict: en="Applepie", da=:w4ZibGV0w6ZydGU=: Example-Dict: en="Applepie", da=:w4ZibGV0w6ZydGU=:
]]></artwork></figure> </sourcecode>
<t>Note that in this example, the final "=" is due to the inclusion of
<t>Note that in this example, the final “=” is due to the inclusion of a Byte Se a Byte Sequence; see <xref target="binary"/>.</t>
quence; see <xref target="binary"/>.</t> <t>Members whose value is Boolean (see <xref target="boolean"/>) true <b
cp14>MUST</bcp14> omit that value when serialized. For
<t>Members whose value is Boolean (see <xref target="boolean"/>) true MUST omit example, here both "b" and "c" are true:</t>
that value when serialized. For example, here both “b” and “c” are true:</t> <sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[
Example-Dict: a=?0, b, c; foo=bar Example-Dict: a=?0, b, c; foo=bar
]]></artwork></figure> </sourcecode>
<t>Note that this requirement is only on serialization; parsers are
<t>Note that this requirement is only on serialization; parsers are still requir still required to correctly handle the true Boolean value when it
ed to correctly handle the true Boolean value when it appears in Dictionary valu appears in Dictionary values.</t>
es.</t> <t>A Dictionary with a member whose value is an Inner List of Tokens:</t
>
<t>A Dictionary with a member whose value is an Inner List of Tokens:</t> <sourcecode type="http-message">
Example-Dict: rating=1.5, feelings=(joy sadness)
<figure><artwork type="example"><![CDATA[ </sourcecode>
Example-DictList: rating=1.5, feelings=(joy sadness) <t>A Dictionary with a mix of Items and Inner Lists, some with parameter
]]></artwork></figure> s:</t>
<sourcecode type="http-message">
<t>A Dictionary with a mix of Items and Inner Lists, some with Parameters:</t> Example-Dict: a=(1 2), b=3, c=4;aa=bb, d=(5 6);valid
</sourcecode>
<figure><artwork type="example"><![CDATA[ <t>As with Lists, an empty Dictionary is represented by omitting the
Example-MixDict: a=(1 2), b=3, c=4;aa=bb, d=(5 6);valid entire field. This implies that fields defined as Dictionaries have a
]]></artwork></figure> default empty value.</t>
<t>Typically, a field specification will define the semantics of
<t>As with lists, an empty Dictionary is represented by omitting the entire fiel Dictionaries by specifying the allowed type(s) for individual members
d. This implies that fields defined as Dictionaries have a default empty value.< by their keys, as well as whether their presence is required or
/t> optional. Recipients <bcp14>MUST</bcp14> ignore members whose keys that a
re undefined or unknown,
<t>Typically, a field specification will define the semantics of Dictionaries by unless the field's specification specifically disallows them.</t>
specifying the allowed type(s) for individual members by their names, as well a <t>Note that Dictionaries can have their members split across multiple
s whether their presence is required or optional. Recipients MUST ignore names t lines of the same header or trailer section; for example, the following
hat are undefined or unknown, unless the field’s specification specifically disa are equivalent:</t>
llows them.</t> <sourcecode type="http-message">
Example-Dict: foo=1, bar=2
<t>Note that Dictionaries can have their members split across multiple lines ins </sourcecode>
ide a header or trailer section; for example, the following are equivalent:</t> <t>and</t>
<sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[ Example-Dict: foo=1
Example-Hdr: foo=1, bar=2 Example-Dict: bar=2
]]></artwork></figure> </sourcecode>
<t>However, individual members of a Dictionary cannot be safely split
<t>and</t> between lines; see <xref target="text-parse"/> for
details.</t>
<figure><artwork type="example"><![CDATA[ <t>Parsers <bcp14>MUST</bcp14> support Dictionaries containing at least
Example-Hdr: foo=1 1024
Example-Hdr: bar=2 key/value pairs and keys with at least 64 characters. Field
]]></artwork></figure> specifications can constrain the order of individual Dictionary
members, as well as their values' types as required.</t>
<t>However, individual members of a Dictionary cannot be safely split between li </section>
nes; see <xref target="text-parse"/> for details.</t> <section anchor="item">
<name>Items</name>
<t>Parsers MUST support Dictionaries containing at least 1024 name/value pairs, <t>An Item can be an Integer (<xref target="integer"/>), a Decimal (<xre
and names with at least 64 characters. Field specifications can constrain the or f target="decimal"/>), a String (<xref target="string"/>), a Token (<xref target
der of individual Dictionary members, as well as their values’ types as required ="token"/>),
.</t> a Byte Sequence (<xref target="binary"/>), or a Boolean
(<xref target="boolean"/>). It can have associated
</section> parameters (<xref target="param"/>).</t>
<section anchor="item" title="Items"> <t>The ABNF for Items is:</t>
<sourcecode type="abnf">
<t>An Item can be a Integer (<xref target="integer"/>), Decimal (<xref target="d
ecimal"/>), String (<xref target="string"/>), Token (<xref target="token"/>), By
te Sequence (<xref target="binary"/>), or Boolean (<xref target="boolean"/>). It
can have associated Parameters (<xref target="param"/>).</t>
<t>The ABNF for Items is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-item = bare-item parameters sf-item = bare-item parameters
bare-item = sf-integer / sf-decimal / sf-string / sf-token bare-item = sf-integer / sf-decimal / sf-string / sf-token
/ sf-binary / sf-boolean / sf-binary / sf-boolean
]]></artwork></figure> </sourcecode>
<t>For example, a header field that is defined to be an Item that is
<t>For example, a header field that is defined to be an Item that is an Integer an Integer might look like:</t>
might look like:</t> <sourcecode type="http-message">
Example-Integer: 5
<figure><artwork type="example"><![CDATA[ </sourcecode>
Example-IntItemHeader: 5 <t>or with parameters:</t>
]]></artwork></figure> <sourcecode type="http-message">
Example-Integer: 5; foo=bar
<t>or with Parameters:</t> </sourcecode>
<section anchor="integer">
<figure><artwork type="example"><![CDATA[ <name>Integers</name>
Example-IntItem: 5; foo=bar <t>Integers have a range of -999,999,999,999,999 to
]]></artwork></figure> 999,999,999,999,999 inclusive (i.e., up to fifteen digits, signed),
for IEEE 754 compatibility <xref target="IEEE754"/>.</t>
<section anchor="integer" title="Integers"> <t>The ABNF for Integers is:</t>
<sourcecode type="abnf">
<t>Integers have a range of -999,999,999,999,999 to 999,999,999,999,999 inclusiv
e (i.e., up to fifteen digits, signed), for IEEE 754 compatibility (<xref target
="IEEE754"/>).</t>
<t>The ABNF for Integers is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-integer = ["-"] 1*15DIGIT sf-integer = ["-"] 1*15DIGIT
]]></artwork></figure> </sourcecode>
<t>For example:</t>
<t>For example:</t> <sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[
Example-Integer: 42 Example-Integer: 42
]]></artwork></figure> </sourcecode>
<t>Integers larger than 15 digits can be supported in a variety of
<t>Integers larger than 15 digits can be supported in a variety of ways; for exa ways; for example, by using a String (<xref target="string"/>), a Byte
mple, by using a String (<xref target="string"/>), Byte Sequence (<xref target=" Sequence (<xref target="binary"/>), or a parameter on an Integer that acts as a
binary"/>), or a parameter on an Integer that acts as a scaling factor.</t> scaling factor.</t>
<t>While it is possible to serialize Integers with leading zeros
<t>While it is possible to serialise Integers with leading zeros (e.g., “0002”, (e.g., "0002", "-01") and signed zero ("-0"), these distinctions may
“-01”) and signed zero (“-0”), these distinctions may not be preserved by implem not be preserved by implementations.</t>
entations.</t> <t>Note that commas in Integers are used in this section's prose
only for readability; they are not valid in the wire format.</t>
<t>Note that commas in Integers are used in this section’s prose only for readab </section>
ility; they are not valid in the wire format.</t> <section anchor="decimal">
<name>Decimals</name>
</section> <t>Decimals are numbers with an integer and a fractional
<section anchor="decimal" title="Decimals"> component. The integer component has at most 12 digits; the
fractional component has at most three digits.</t>
<t>Decimals are numbers with an integer and a fractional component. The integer <t>The ABNF for decimals is:</t>
component has at most 12 digits; the fractional component has at most three digi <sourcecode type="abnf">
ts.</t>
<t>The ABNF for decimals is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-decimal = ["-"] 1*12DIGIT "." 1*3DIGIT sf-decimal = ["-"] 1*12DIGIT "." 1*3DIGIT
]]></artwork></figure> </sourcecode>
<t>For example, a header whose value is defined as a Decimal could
<t>For example, a header whose value is defined as a Decimal could look like:</t look like:</t>
> <sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[
Example-Decimal: 4.5 Example-Decimal: 4.5
]]></artwork></figure> </sourcecode>
<t>While it is possible to serialize Decimals with leading zeros
<t>While it is possible to serialise Decimals with leading zeros (e.g., “0002.5” (e.g., "0002.5", "-01.334"), trailing zeros (e.g., "5.230",
, “-01.334”), trailing zeros (e.g., “5.230”, “-0.40”), and signed zero (e.g., “- "-0.40"), and signed zero (e.g., "-0.0"), these distinctions may not
0.0”), these distinctions may not be preserved by implementations.</t> be preserved by implementations.</t>
<t>Note that the serialization algorithm (<xref target="ser-decimal"/>
<t>Note that the serialisation algorithm (<xref target="ser-decimal"/>) rounds i ) rounds input with more than three digits of
nput with more than three digits of precision in the fractional component. If an precision in the fractional component. If an alternative rounding
alternative rounding strategy is desired, this should be specified by the heade strategy is desired, this should be specified by the header
r definition to occur before serialisation.</t> definition to occur before serialization.</t>
</section>
</section> <section anchor="string">
<section anchor="string" title="Strings"> <name>Strings</name>
<t>Strings are zero or more printable ASCII <xref target="RFC0020"/> c
<t>Strings are zero or more printable ASCII <xref target="RFC0020"/> characters haracters (i.e., the range %x20 to %x7E). Note
(i.e., the range %x20 to %x7E). Note that this excludes tabs, newlines, carriage that this excludes tabs, newlines, carriage returns, etc.</t>
returns, etc.</t> <t>The ABNF for Strings is:</t>
<sourcecode type="abnf">
<t>The ABNF for Strings is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-string = DQUOTE *chr DQUOTE sf-string = DQUOTE *chr DQUOTE
chr = unescaped / escaped chr = unescaped / escaped
unescaped = %x20-21 / %x23-5B / %x5D-7E unescaped = %x20-21 / %x23-5B / %x5D-7E
escaped = "\" ( DQUOTE / "\" ) escaped = "\" ( DQUOTE / "\" )
]]></artwork></figure> </sourcecode>
<t>Strings are delimited with double quotes, using a backslash ("\")
<t>Strings are delimited with double quotes, using a backslash (“\”) to escape d to escape double quotes and backslashes. For example:</t>
ouble quotes and backslashes. For example:</t> <sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[
Example-String: "hello world" Example-String: "hello world"
]]></artwork></figure> </sourcecode>
<t>Note that Strings only use DQUOTE as a delimiter; single quotes
<t>Note that Strings only use DQUOTE as a delimiter; single quotes do not delimi do not delimit Strings. Furthermore, only DQUOTE and "\" can be
t Strings. Furthermore, only DQUOTE and “\” can be escaped; other characters aft escaped; other characters after "\" <bcp14>MUST</bcp14> cause parsing t
er “\” MUST cause parsing to fail.</t> o fail.</t>
<t>Unicode is not directly supported in Strings, because it causes a
<t>Unicode is not directly supported in Strings, because it causes a number of i number of interoperability issues, and -- with few exceptions -- field
nteroperability issues, and – with few exceptions – field values do not require values do not require it.</t>
it.</t> <t>When it is necessary for a field value to convey non-ASCII
content, a Byte Sequence (<xref target="binary"/>)
<t>When it is necessary for a field value to convey non-ASCII content, a Byte Se can be specified, along with a character encoding (preferably UTF-8 <xr
quence (<xref target="binary"/>) can be specified, along with a character encodi ef target="STD63"/>).</t>
ng (preferably <xref target="UTF-8"/>).</t> <t>Parsers <bcp14>MUST</bcp14> support Strings (after any decoding) wi
th at least
<t>Parsers MUST support Strings (after any decoding) with at least 1024 characte 1024 characters.</t>
rs.</t> </section>
<section anchor="token">
</section> <name>Tokens</name>
<section anchor="token" title="Tokens"> <t>Tokens are short textual words; their abstract model is identical
to their expression in the HTTP field value serialization.</t>
<t>Tokens are short textual words; their abstract model is identical to their ex <t>The ABNF for Tokens is:</t>
pression in the HTTP field value serialization.</t> <sourcecode type="abnf">
<t>The ABNF for Tokens is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-token = ( ALPHA / "*" ) *( tchar / ":" / "/" ) sf-token = ( ALPHA / "*" ) *( tchar / ":" / "/" )
]]></artwork></figure> </sourcecode>
<t>For example:</t>
<t>For example:</t> <sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[
Example-Token: foo123/456 Example-Token: foo123/456
]]></artwork></figure> </sourcecode>
<t>Parsers <bcp14>MUST</bcp14> support Tokens with at least 512 charac
<t>Parsers MUST support Tokens with at least 512 characters.</t> ters.</t>
<t>Note that Token allows the same characters as the "token" ABNF
<t>Note that Token allows the same characters as the “token” ABNF rule defined i rule defined in <xref target="RFC7230"/>, with the
n <xref target="RFC7230"/>, with the exceptions that the first character is requ exceptions that the first character is required to be either ALPHA
ired to be either ALPHA or “*”, and “:” and “/” are also allowed in subsequent c or "*", and ":" and "/" are also allowed in subsequent
haracters.</t> characters.</t>
</section>
</section> <section anchor="binary">
<section anchor="binary" title="Byte Sequences"> <name>Byte Sequences</name>
<t>Byte Sequences can be conveyed in Structured Fields.</t>
<t>Byte Sequences can be conveyed in Structured Fields.</t> <t>The ABNF for a Byte Sequence is:</t>
<sourcecode type="abnf">
<t>The ABNF for a Byte Sequence is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-binary = ":" *(base64) ":" sf-binary = ":" *(base64) ":"
base64 = ALPHA / DIGIT / "+" / "/" / "=" base64 = ALPHA / DIGIT / "+" / "/" / "="
]]></artwork></figure> </sourcecode>
<t>A Byte Sequence is delimited with colons and encoded using base64
<t>A Byte Sequence is delimited with colons and encoded using base64 (<xref targ (<xref target="RFC4648" sectionFormat="comma" section="4"/>). For
et="RFC4648"/>, Section 4). For example:</t> example:</t>
<sourcecode type="http-message">
<figure><artwork type="example"><![CDATA[ Example-ByteSequence: :cHJldGVuZCB0aGlzIGlzIGJpbmFyeSBjb250ZW50Lg==:
Example-Binary: :cHJldGVuZCB0aGlzIGlzIGJpbmFyeSBjb250ZW50Lg==: </sourcecode>
]]></artwork></figure> <t>Parsers <bcp14>MUST</bcp14> support Byte Sequences with at least 16
384 octets
<t>Parsers MUST support Byte Sequences with at least 16384 octets after decoding after decoding.</t>
.</t> </section>
<section anchor="boolean">
</section> <name>Booleans</name>
<section anchor="boolean" title="Booleans"> <t>Boolean values can be conveyed in Structured Fields.</t>
<t>The ABNF for a Boolean is:</t>
<t>Boolean values can be conveyed in Structured Fields.</t> <sourcecode type="abnf">
<t>The ABNF for a Boolean is:</t>
<figure><artwork type="abnf"><![CDATA[
sf-boolean = "?" boolean sf-boolean = "?" boolean
boolean = "0" / "1" boolean = "0" / "1"
]]></artwork></figure> </sourcecode>
<t>A Boolean is indicated with a leading "?" character followed by a
<t>A Boolean is indicated with a leading “?” character followed by a “1” for a t "1" for a true value or "0" for false. For example:</t>
rue value or “0” for false. For example:</t> <sourcecode type="http-message">
Example-Boolean: ?1
<figure><artwork type="example"><![CDATA[ </sourcecode>
Example-Bool: ?1 <t>Note that in Dictionary (<xref target="dictionary"/>) and Parameter
]]></artwork></figure> (<xref target="param"/>) values, Boolean true is indicated by omitting
the value.</t>
<t>Note that in Dictionary (<xref target="dictionary"/>) and Parameter (<xref ta </section>
rget="param"/>) values, Boolean true is indicated by omitting the value.</t> </section>
</section>
</section> <section anchor="text">
</section> <name>Working with Structured Fields in HTTP</name>
</section> <t>This section defines how to serialize and parse Structured Fields in
<section anchor="text" title="Working With Structured Fields in HTTP"> textual HTTP field values and other encodings compatible with them
(e.g., in HTTP/2 <xref target="RFC7540"/> before
<t>This section defines how to serialize and parse Structured Fields in textual compression with HPACK <xref target="RFC7541"/>).</t>
HTTP field values and other encodings compatible with them (e.g., in HTTP/2 <xre <section anchor="text-serialize">
f target="RFC7540"/> before compression with HPACK <xref target="RFC7541"/>).</t <name>Serializing Structured Fields</name>
> <t>Given a structure defined in this specification, return an ASCII
string suitable for use in an HTTP field value.</t>
<section anchor="text-serialize" title="Serializing Structured Fields"> <ol>
<li>If the structure is a Dictionary or List and its value is empty
<t>Given a structure defined in this specification, return an ASCII string suita (i.e., it has no members), do not serialize the field at all (i.e.,
ble for use in a HTTP field value.</t> omit both the field-name and field-value).</li>
<li>If the structure is a List, let output_string be the result of
<t><list style="numbers"> running Serializing a List (<xref target="ser-list"/>) with the structu
<t>If the structure is a Dictionary or List and its value is empty (i.e., it h re.</li>
as no members), do not serialize the field at all (i.e., omit both the field-nam <li>Else, if the structure is a Dictionary, let output_string be the
e and field-value).</t> result of running Serializing a Dictionary (<xref target="ser-dictionar
<t>If the structure is a List, let output_string be the result of running Seri y"/>) with the structure.</li>
alizing a List (<xref target="ser-list"/>) with the structure.</t> <li>Else, if the structure is an Item, let output_string be the
<t>Else if the structure is a Dictionary, let output_string be the result of r result of running Serializing an Item (<xref target="ser-item"/>) with
unning Serializing a Dictionary (<xref target="ser-dictionary"/>) with the struc the structure.</li>
ture.</t> <li>Else, fail serialization.</li>
<t>Else if the structure is an Item, let output_string be the result of runnin <li>Return output_string converted into an array of bytes, using
g Serializing an Item (<xref target="ser-item"/>) with the structure.</t> ASCII encoding <xref target="RFC0020"/>.</li>
<t>Else, fail serialization.</t> </ol>
<t>Return output_string converted into an array of bytes, using ASCII encoding <section anchor="ser-list">
<xref target="RFC0020"/>.</t> <name>Serializing a List</name>
</list></t> <t>Given an array of (member_value, parameters) tuples as
input_list, return an ASCII string suitable for use in an HTTP field
<section anchor="ser-list" title="Serializing a List"> value.</t>
<ol>
<t>Given an array of (member_value, parameters) tuples as input_list, return an <li>Let output be an empty string.</li>
ASCII string suitable for use in a HTTP field value.</t> <li>
<t>For each (member_value, parameters) of input_list:
<t><list style="numbers"> </t>
<t>Let output be an empty string.</t> <ol>
<t>For each (member_value, parameters) of input_list: <li>If member_value is an array, append the result of running
<list style="numbers"> Serializing an Inner List (<xref target="ser-innerlist"/>) with (
<t>If member_value is an array, append the result of running Serializing a member_value, parameters) to
n Inner List (<xref target="ser-innerlist"/>) with (member_value, parameters) to output.</li>
output.</t> <li>Otherwise, append the result of running Serializing an
<t>Otherwise, append the result of running Serializing an Item (<xref targ Item (<xref target="ser-item"/>) with
et="ser-item"/>) with (member_value, parameters) to output.</t> (member_value, parameters) to output.</li>
<t>If more member_values remain in input_list: <li>
<list style="numbers"> <t>If more member_values remain in input_list:
<t>Append “,” to output.</t> </t>
<t>Append a single SP to output.</t> <ol>
</list></t> <li>Append "," to output.</li>
</list></t> <li>Append a single SP to output.</li>
<t>Return output.</t> </ol>
</list></t> </li>
</ol>
<section anchor="ser-innerlist" title="Serializing an Inner List"> </li>
<li>Return output.</li>
<t>Given an array of (member_value, parameters) tuples as inner_list, and parame </ol>
ters as list_parameters, return an ASCII string suitable for use in a HTTP field <section anchor="ser-innerlist">
value.</t> <name>Serializing an Inner List</name>
<t>Given an array of (member_value, parameters) tuples as
<t><list style="numbers"> inner_list, and parameters as list_parameters, return an ASCII
<t>Let output be the string “(“.</t> string suitable for use in an HTTP field value.</t>
<t>For each (member_value, parameters) of inner_list: <ol>
<list style="numbers"> <li>Let output be the string "(".</li>
<t>Append the result of running Serializing an Item (<xref target="ser-ite <li>
m"/>) with (member_value, parameters) to output.</t> <t>For each (member_value, parameters) of inner_list:
<t>If more values remain in inner_list, append a single SP to output.</t> </t>
</list></t> <ol>
<t>Append “)” to output.</t> <li>Append the result of running Serializing an Item (<xref ta
<t>Append the result of running Serializing Parameters (<xref target="ser-para rget="ser-item"/>) with (member_value,
ms"/>) with list_parameters to output.</t> parameters) to output.</li>
<t>Return output.</t> <li>If more values remain in inner_list, append a single SP to
</list></t> output.</li>
</ol>
</section> </li>
<section anchor="ser-params" title="Serializing Parameters"> <li>Append ")" to output.</li>
<li>Append the result of running Serializing Parameters (<xref tar
<t>Given an ordered Dictionary as input_parameters (each member having a param_n get="ser-params"/>) with list_parameters to
ame and a param_value), return an ASCII string suitable for use in a HTTP field output.</li>
value.</t> <li>Return output.</li>
</ol>
<t><list style="numbers"> </section>
<t>Let output be an empty string.</t> <section anchor="ser-params">
<t>For each param_name with a value of param_value in input_parameters: <name>Serializing Parameters</name>
<list style="numbers"> <t>Given an ordered Dictionary as input_parameters (each member
<t>Append “;” to output.</t> having a param_key and a param_value), return an ASCII string
<t>Append the result of running Serializing a Key (<xref target="ser-key"/ suitable for use in an HTTP field value.</t>
>) with param_name to output.</t> <ol>
<t>If param_value is not Boolean true: <li>Let output be an empty string.</li>
<list style="numbers"> <li>
<t>Append “=” to output.</t> <t>For each param_key with a value of param_value in input_param
<t>Append the result of running Serializing a bare Item (<xref target= eters:
"ser-bare-item"/>) with param_value to output.</t> </t>
</list></t> <ol>
</list></t> <li>Append ";" to output.</li>
<t>Return output.</t> <li>Append the result of running Serializing a Key (<xref targ
</list></t> et="ser-key"/>) with param_key to
output.</li>
</section> <li>
<section anchor="ser-key" title="Serializing a Key"> <t>If param_value is not Boolean true:
</t>
<t>Given a key as input_key, return an ASCII string suitable for use in a HTTP f <ol>
ield value.</t> <li>Append "=" to output.</li>
<li>Append the result of running Serializing a bare Item
<t><list style="numbers"> (<xref target="ser-bare-item"/>) with
<t>Convert input_key into a sequence of ASCII characters; if conversion fails, param_value to output.</li>
fail serialization.</t> </ol>
<t>If input_key contains characters not in lcalpha, DIGIT, “_”, “-“, “.”, or “ </li>
*” fail serialization.</t> </ol>
<t>If the first character of input_key is not lcalpha or “*”, fail serializati </li>
on.</t> <li>Return output.</li>
<t>Let output be an empty string.</t> </ol>
<t>Append input_key to output.</t> </section>
<t>Return output.</t> <section anchor="ser-key">
</list></t> <name>Serializing a Key</name>
<t>Given a key as input_key, return an ASCII string suitable for
</section> use in an HTTP field value.</t>
</section> <ol>
<section anchor="ser-dictionary" title="Serializing a Dictionary"> <li>Convert input_key into a sequence of ASCII characters; if
conversion fails, fail serialization.</li>
<t>Given an ordered Dictionary as input_dictionary (each member having a member_ <li>If input_key contains characters not in lcalpha, DIGIT, "_",
name and a tuple value of (member_value, parameters)), return an ASCII string su "-", ".", or "*", fail serialization.</li>
itable for use in a HTTP field value.</t> <li>If the first character of input_key is not lcalpha or "*",
fail serialization.</li>
<t><list style="numbers"> <li>Let output be an empty string.</li>
<t>Let output be an empty string.</t> <li>Append input_key to output.</li>
<t>For each member_name with a value of (member_value, parameters) in input_di <li>Return output.</li>
ctionary: </ol>
<list style="numbers"> </section>
<t>Append the result of running Serializing a Key (<xref target="ser-key"/ </section>
>) with member’s member_name to output.</t> <section anchor="ser-dictionary">
<t>If member_value is Boolean true: <name>Serializing a Dictionary</name>
<list style="numbers"> <t>Given an ordered Dictionary as input_dictionary (each member
<t>Append the result of running Serializing Parameters (<xref target=" having a member_key and a tuple value of (member_value,
ser-params"/>) with parameters to output.</t> parameters)), return an ASCII string suitable for use in an HTTP
</list></t> field value.</t>
<t>Otherwise: <ol>
<list style="numbers"> <li>Let output be an empty string.</li>
<t>Append “=” to output.</t> <li>
<t>If member_value is an array, append the result of running Serializi <t>For each member_key with a value of (member_value, parameters)
ng an Inner List (<xref target="ser-innerlist"/>) with (member_value, parameters in input_dictionary:
) to output.</t> </t>
<t>Otherwise, append the result of running Serializing an Item (<xref <ol>
target="ser-item"/>) with (member_value, parameters) to output.</t> <li>Append the result of running Serializing a Key (<xref target
</list></t> ="ser-key"/>) with member's member_key
<t>If more members remain in input_dictionary: to output.</li>
<list style="numbers"> <li>
<t>Append “,” to output.</t> <t>If member_value is Boolean true:
<t>Append a single SP to output.</t> </t>
</list></t> <ol>
</list></t> <li>Append the result of running Serializing Parameters
<t>Return output.</t> (<xref target="ser-params"/>) with
</list></t> parameters to output.</li>
</ol>
</section> </li>
<section anchor="ser-item" title="Serializing an Item"> <li>
<t>Otherwise:
<t>Given an Item as bare_item and Parameters as item_parameters, return an ASCII </t>
string suitable for use in a HTTP field value.</t> <ol>
<li>Append "=" to output.</li>
<t><list style="numbers"> <li>If member_value is an array, append the result of
<t>Let output be an empty string.</t> running Serializing an Inner List (<xref target="ser-innerlis
<t>Append the result of running Serializing a Bare Item <xref target="ser-bare t"/>) with
-item"/> with bare_item to output.</t> (member_value, parameters) to output.</li>
<t>Append the result of running Serializing Parameters <xref target="ser-param <li>Otherwise, append the result of running Serializing an
s"/> with item_parameters to output.</t> Item (<xref target="ser-item"/>) with
<t>Return output.</t> (member_value, parameters) to output.</li>
</list></t> </ol>
</li>
<section anchor="ser-bare-item" title="Serializing a Bare Item"> <li>
<t>If more members remain in input_dictionary:
<t>Given an Item as input_item, return an ASCII string suitable for use in a HTT </t>
P field value.</t> <ol>
<li>Append "," to output.</li>
<t><list style="numbers"> <li>Append a single SP to output.</li>
<t>If input_item is an Integer, return the result of running Serializing an In </ol>
teger (<xref target="ser-integer"/>) with input_item.</t> </li>
<t>If input_item is a Decimal, return the result of running Serializing a Deci </ol>
mal (<xref target="ser-decimal"/>) with input_item.</t> </li>
<t>If input_item is a String, return the result of running Serializing a Strin <li>Return output.</li>
g (<xref target="ser-string"/>) with input_item.</t> </ol>
<t>If input_item is a Token, return the result of running Serializing a Token </section>
(<xref target="ser-token"/>) with input_item.</t> <section anchor="ser-item">
<t>If input_item is a Boolean, return the result of running Serializing a Bool <name>Serializing an Item</name>
ean (<xref target="ser-boolean"/>) with input_item.</t> <t>Given an Item as bare_item and Parameters as item_parameters,
<t>If input_item is a Byte Sequence, return the result of running Serializing return an ASCII string suitable for use in an HTTP field value.</t>
a Byte Sequence (<xref target="ser-binary"/>) with input_item.</t> <ol>
<t>Otherwise, fail serialization.</t> <li>Let output be an empty string.</li>
</list></t> <li>Append the result of running Serializing a Bare Item (<xref targ
et="ser-bare-item"/>) with bare_item to
</section> output.</li>
</section> <li>Append the result of running Serializing Parameters (<xref targe
<section anchor="ser-integer" title="Serializing an Integer"> t="ser-params"/>) with item_parameters to
output.</li>
<t>Given an Integer as input_integer, return an ASCII string suitable for use in <li>Return output.</li>
a HTTP field value.</t> </ol>
<section anchor="ser-bare-item">
<t><list style="numbers"> <name>Serializing a Bare Item</name>
<t>If input_integer is not an integer in the range of -999,999,999,999,999 to <t>Given an Item as input_item, return an ASCII string suitable
999,999,999,999,999 inclusive, fail serialization.</t> for use in an HTTP field value.</t>
<t>Let output be an empty string.</t> <ol>
<t>If input_integer is less than (but not equal to) 0, append “-“ to output.</ <li>If input_item is an Integer, return the result of running
t> Serializing an Integer (<xref target="ser-integer"/>) with input_it
<t>Append input_integer’s numeric value represented in base 10 using only deci em.</li>
mal digits to output.</t> <li>If input_item is a Decimal, return the result of running
<t>Return output.</t> Serializing a Decimal (<xref target="ser-decimal"/>) with input_ite
</list></t> m.</li>
<li>If input_item is a String, return the result of running
</section> Serializing a String (<xref target="ser-string"/>) with input_item.
<section anchor="ser-decimal" title="Serializing a Decimal"> </li>
<li>If input_item is a Token, return the result of running
<t>Given a decimal number as input_decimal, return an ASCII string suitable for Serializing a Token (<xref target="ser-token"/>) with input_item.</
use in a HTTP field value.</t> li>
<li>If input_item is a Byte Sequence, return the result of
<t><list style="numbers"> running Serializing a Byte Sequence (<xref target="ser-binary"/>) w
<t>If input_decimal is not a decimal number, fail serialization.</t> ith input_item.</li>
<t>If input_decimal has more than three significant digits to the right of the <li>If input_item is a Boolean, return the result of running
decimal point, round it to three decimal places, rounding the final digit to th Serializing a Boolean (<xref target="ser-boolean"/>) with
e nearest value, or to the even value if it is equidistant.</t> input_item.</li>
<t>If input_decimal has more than 12 significant digits to the left of the dec <li>Otherwise, fail serialization.</li>
imal point after rounding, fail serialization.</t>
<t>Let output be an empty string.</t>
<t>If input_decimal is less than (but not equal to) 0, append “-“ to output.</
t>
<t>Append input_decimal’s integer component represented in base 10 (using only
decimal digits) to output; if it is zero, append “0”.</t>
<t>Append “.” to output.</t>
<t>If input_decimal’s fractional component is zero, append “0” to output.</t>
<t>Otherwise, append the significant digits of input_decimal’s fractional comp
onent represented in base 10 (using only decimal digits) to output.</t>
<t>Return output.</t>
</list></t>
</section>
<section anchor="ser-string" title="Serializing a String">
<t>Given a String as input_string, return an ASCII string suitable for use in a
HTTP field value.</t>
<t><list style="numbers">
<t>Convert input_string into a sequence of ASCII characters; if conversion fai
ls, fail serialization.</t>
<t>If input_string contains characters in the range %x00-1f or %x7f (i.e., not
in VCHAR or SP), fail serialization.</t>
<t>Let output be the string DQUOTE.</t>
<t>For each character char in input_string:
<list style="numbers">
<t>If char is “\” or DQUOTE:
<list style="numbers">
<t>Append “\” to output.</t>
</list></t>
<t>Append char to output.</t>
</list></t>
<t>Append DQUOTE to output.</t>
<t>Return output.</t>
</list></t>
</section>
<section anchor="ser-token" title="Serializing a Token">
<t>Given a Token as input_token, return an ASCII string suitable for use in a HT
TP field value.</t>
<t><list style="numbers">
<t>Convert input_token into a sequence of ASCII characters; if conversion fail
s, fail serialization.</t>
<t>If the first character of input_token is not ALPHA or “*”, or the remaining
portion contains a character not in tchar, “:” or “/”, fail serialization.</t>
<t>Let output be an empty string.</t>
<t>Append input_token to output.</t>
<t>Return output.</t>
</list></t>
</section>
<section anchor="ser-binary" title="Serializing a Byte Sequence">
<t>Given a Byte Sequence as input_bytes, return an ASCII string suitable for use
in a HTTP field value.</t>
<t><list style="numbers">
<t>If input_bytes is not a sequence of bytes, fail serialization.</t>
<t>Let output be an empty string.</t>
<t>Append “:” to output.</t>
<t>Append the result of base64-encoding input_bytes as per <xref target="RFC46
48"/>, Section 4, taking account of the requirements below.</t>
<t>Append “:” to output.</t>
<t>Return output.</t>
</list></t>
<t>The encoded data is required to be padded with “=”, as per <xref target="RFC4
648"/>, Section 3.2.</t>
<t>Likewise, encoded data SHOULD have pad bits set to zero, as per <xref target=
"RFC4648"/>, Section 3.5, unless it is not possible to do so due to implementati
on constraints.</t>
</section>
<section anchor="ser-boolean" title="Serializing a Boolean">
<t>Given a Boolean as input_boolean, return an ASCII string suitable for use in
a HTTP field value.</t>
<t><list style="numbers">
<t>If input_boolean is not a boolean, fail serialization.</t>
<t>Let output be an empty string.</t>
<t>Append “?” to output.</t>
<t>If input_boolean is true, append “1” to output.</t>
<t>If input_boolean is false, append “0” to output.</t>
<t>Return output.</t>
</list></t>
</section>
</section>
<section anchor="text-parse" title="Parsing Structured Fields">
<t>When a receiving implementation parses HTTP fields that are known to be Struc
tured Fields, it is important that care be taken, as there are a number of edge
cases that can cause interoperability or even security problems. This section sp
ecifies the algorithm for doing so.</t>
<t>Given an array of bytes input_bytes that represents the chosen field’s field-
value (which is empty if that field is not present), and field_type (one of “dic
tionary”, “list”, or “item”), return the parsed header value.</t>
<t><list style="numbers">
<t>Convert input_bytes into an ASCII string input_string; if conversion fails,
fail parsing.</t>
<t>Discard any leading SP characters from input_string.</t>
<t>If field_type is “list”, let output be the result of running Parsing a List
(<xref target="parse-list"/>) with input_string.</t>
<t>If field_type is “dictionary”, let output be the result of running Parsing
a Dictionary (<xref target="parse-dictionary"/>) with input_string.</t>
<t>If field_type is “item”, let output be the result of running Parsing an Ite
m (<xref target="parse-item"/>) with input_string.</t>
<t>Discard any leading SP characters from input_string.</t>
<t>If input_string is not empty, fail parsing.</t>
<t>Otherwise, return output.</t>
</list></t>
<t>When generating input_bytes, parsers MUST combine all field lines in the same
section (header or trailer) that case-insensitively match the field name into o
ne comma-separated field-value, as per <xref target="RFC7230"/>, Section 3.2.2;
this assures that the entire field value is processed correctly.</t>
<t>For Lists and Dictionaries, this has the effect of correctly concatenating al
l of the field’s lines, as long as individual members of the top-level data stru
cture are not split across multiple header instances. The parsing algorithms for
both types allow tab characters, since these might
be used to combine field lines by some implementations.</t>
<t>Strings split across multiple field lines will have unpredictable results, be
cause comma(s) and whitespace inserted upon combination will become part of the
string output by the parser. Since concatenation might be done by an upstream in
termediary, the results are not under the control of the serializer or the parse
r, even when they are both under the control of the same party.</t>
<t>Tokens, Integers, Decimals and Byte Sequences cannot be split across multiple
field lines because the inserted commas will cause parsing to fail.</t>
<t>Parsers MAY fail when processing a field value spread across multiple field l
ines, when one of those lines does not parse as that field. For example, a parsi
ng handling an Example-String field that’s defined as a sf-string is allowed to
fail when processing this field section:</t>
<figure><artwork type="example"><![CDATA[ </ol>
</section>
</section>
<section anchor="ser-integer">
<name>Serializing an Integer</name>
<t>Given an Integer as input_integer, return an ASCII string
suitable for use in an HTTP field value.</t>
<ol>
<li>If input_integer is not an integer in the range of
-999,999,999,999,999 to 999,999,999,999,999 inclusive, fail
serialization.</li>
<li>Let output be an empty string.</li>
<li>If input_integer is less than (but not equal to) 0, append "-"
to output.</li>
<li>Append input_integer's numeric value represented in base 10
using only decimal digits to output.</li>
<li>Return output.</li>
</ol>
</section>
<section anchor="ser-decimal">
<name>Serializing a Decimal</name>
<t>Given a decimal number as input_decimal, return an ASCII string
suitable for use in an HTTP field value.</t>
<ol>
<li>If input_decimal is not a decimal number, fail serialization.</l
i>
<li>If input_decimal has more than three significant digits to the
right of the decimal point, round it to three decimal places,
rounding the final digit to the nearest value, or to the even
value if it is equidistant.</li>
<li>If input_decimal has more than 12 significant digits to the
left of the decimal point after rounding, fail serialization.</li>
<li>Let output be an empty string.</li>
<li>If input_decimal is less than (but not equal to) 0, append "-"
to output.</li>
<li>Append input_decimal's integer component represented in base
10 (using only decimal digits) to output; if it is zero, append
"0".</li>
<li>Append "." to output.</li>
<li>If input_decimal's fractional component is zero, append "0" to
output.</li>
<li>Otherwise, append the significant digits of input_decimal's
fractional component represented in base 10 (using only decimal
digits) to output.</li>
<li>Return output.</li>
</ol>
</section>
<section anchor="ser-string">
<name>Serializing a String</name>
<t>Given a String as input_string, return an ASCII string suitable
for use in an HTTP field value.</t>
<ol>
<li>Convert input_string into a sequence of ASCII characters; if
conversion fails, fail serialization.</li>
<li>If input_string contains characters in the range %x00-1f or
%x7f-ff (i.e., not in VCHAR or SP), fail serialization.</li>
<li>Let output be the string DQUOTE.</li>
<li>
<t>For each character char in input_string:
</t>
<ol>
<li>
<t>If char is "\" or DQUOTE:
</t>
<ol>
<li>Append "\" to output.</li>
</ol>
</li>
<li>Append char to output.</li>
</ol>
</li>
<li>Append DQUOTE to output.</li>
<li>Return output.</li>
</ol>
</section>
<section anchor="ser-token">
<name>Serializing a Token</name>
<t>Given a Token as input_token, return an ASCII string suitable for
use in an HTTP field value.</t>
<ol>
<li>Convert input_token into a sequence of ASCII characters; if
conversion fails, fail serialization.</li>
<li>If the first character of input_token is not ALPHA or "*", or
the remaining portion contains a character not in tchar, ":", or
"/", fail serialization.</li>
<li>Let output be an empty string.</li>
<li>Append input_token to output.</li>
<li>Return output.</li>
</ol>
</section>
<section anchor="ser-binary">
<name>Serializing a Byte Sequence</name>
<t>Given a Byte Sequence as input_bytes, return an ASCII string
suitable for use in an HTTP field value.</t>
<ol>
<li>If input_bytes is not a sequence of bytes, fail serialization.</
li>
<li>Let output be an empty string.</li>
<li>Append ":" to output.</li>
<li>Append the result of base64-encoding input_bytes as per <xref ta
rget="RFC4648" sectionFormat="comma" section="4"/>, taking account of
the requirements below.</li>
<li>Append ":" to output.</li>
<li>Return output.</li>
</ol>
<t>The encoded data is required to be padded with "=", as per <xref ta
rget="RFC4648" sectionFormat="comma" section="3.2"/>.</t>
<t>Likewise, encoded data <bcp14>SHOULD</bcp14> have pad bits set to z
ero, as per
<xref target="RFC4648" sectionFormat="comma" section="3.5"/>, unless it
is
not possible to do so due to implementation constraints.</t>
</section>
<section anchor="ser-boolean">
<name>Serializing a Boolean</name>
<t>Given a Boolean as input_boolean, return an ASCII string suitable
for use in an HTTP field value.</t>
<ol>
<li>If input_boolean is not a boolean, fail serialization.</li>
<li>Let output be an empty string.</li>
<li>Append "?" to output.</li>
<li>If input_boolean is true, append "1" to output.</li>
<li>If input_boolean is false, append "0" to output.</li>
<li>Return output.</li>
</ol>
</section>
</section>
<section anchor="text-parse">
<name>Parsing Structured Fields</name>
<t>When a receiving implementation parses HTTP fields that are known
to be Structured Fields, it is important that care be taken, as there
are a number of edge cases that can cause interoperability or even
security problems. This section specifies the algorithm for doing
so.</t>
<t>Given an array of bytes as input_bytes that represent the chosen
field's field-value (which is empty if that field is not present) and
field_type (one of "dictionary", "list", or "item"), return the parsed
header value.</t>
<ol>
<li>Convert input_bytes into an ASCII string input_string; if
conversion fails, fail parsing.</li>
<li>Discard any leading SP characters from input_string.</li>
<li>If field_type is "list", let output be the result of running
Parsing a List (<xref target="parse-list"/>) with
input_string.</li>
<li>If field_type is "dictionary", let output be the result of
running Parsing a Dictionary (<xref target="parse-dictionary"/>) with i
nput_string.</li>
<li>If field_type is "item", let output be the result of running
Parsing an Item (<xref target="parse-item"/>) with
input_string.</li>
<li>Discard any leading SP characters from input_string.</li>
<li>If input_string is not empty, fail parsing.</li>
<li>Otherwise, return output.</li>
</ol>
<t>When generating input_bytes, parsers <bcp14>MUST</bcp14> combine all
field lines
in the same section (header or trailer) that case-insensitively match
the field name into one comma-separated field-value, as per <xref target=
"RFC7230" sectionFormat="comma" section="3.2.2"/>; this assures that
the entire field value is processed correctly.</t>
<t>For Lists and Dictionaries, this has the effect of correctly
concatenating all of the field's lines, as long as individual members
of the top-level data structure are not split across multiple header
instances. The parsing algorithms for both types allow tab characters,
since these might be used to combine field lines by some
implementations.</t>
<t>Strings split across multiple field lines will have unpredictable
results, because one or more commas (with optional whitespace)
will become part of the string output by the parser. Since
concatenation might be done by an upstream intermediary, the results
are not under the control of the serializer or the parser, even when
they are both under the control of the same party.</t>
<t>Tokens, Integers, Decimals, and Byte Sequences cannot be split
across multiple field lines because the inserted commas will cause
parsing to fail.</t>
<t>Parsers <bcp14>MAY</bcp14> fail when processing a field value spread
across
multiple field lines, when one of those lines does not parse as that
field. For example, a parsing handling an Example-String field that's
defined as an sf-string is allowed to fail when processing this field
section:</t>
<sourcecode type="http-message">
Example-String: "foo Example-String: "foo
Example-String: bar" Example-String: bar"
]]></artwork></figure> </sourcecode>
<t>If parsing fails -- including when calling another algorithm -- the
<t>If parsing fails – including when calling another algorithm – the entire fiel entire field value <bcp14>MUST</bcp14> be ignored (i.e., treated as if th
d value MUST be ignored (i.e., treated as if the field were not present in the s e field were
ection). This is intentionally strict, to improve interoperability and safety, a not present in the section). This is intentionally strict, to improve
nd specifications referencing this document are not allowed to loosen this requi interoperability and safety, and specifications referencing this
rement.</t> document are not allowed to loosen this requirement.</t>
<t>Note that this requirement does not apply to an implementation that
<t>Note that this requirement does not apply to an implementation that is not pa is not parsing the field; for example, an intermediary is not required
rsing the field; for example, an intermediary is not required to strip a failing to strip a failing field from a message before forwarding it.</t>
field from a message before forwarding it.</t> <section anchor="parse-list">
<name>Parsing a List</name>
<section anchor="parse-list" title="Parsing a List"> <t>Given an ASCII string as input_string, return an array of
(item_or_inner_list, parameters) tuples. input_string is modified to
<t>Given an ASCII string as input_string, return an array of (item_or_inner_list remove the parsed value.</t>
, parameters) tuples. input_string is modified to remove the parsed value.</t> <ol>
<li>Let members be an empty array.</li>
<t><list style="numbers"> <li>
<t>Let members be an empty array.</t> <t>While input_string is not empty:
<t>While input_string is not empty: </t>
<list style="numbers"> <ol>
<t>Append the result of running Parsing an Item or Inner List (<xref targe <li>Append the result of running Parsing an Item or Inner List
t="parse-item-or-list"/>) with input_string to members.</t> (<xref target="parse-item-or-list"/>) with
<t>Discard any leading OWS characters from input_string.</t> input_string to members.</li>
<t>If input_string is empty, return members.</t> <li>Discard any leading OWS characters from input_string.</li>
<t>Consume the first character of input_string; if it is not “,”, fail par <li>If input_string is empty, return members.</li>
sing.</t> <li>Consume the first character of input_string; if it is not
<t>Discard any leading OWS characters from input_string.</t> ",", fail parsing.</li>
<t>If input_string is empty, there is a trailing comma; fail parsing.</t> <li>Discard any leading OWS characters from input_string.</li>
</list></t> <li>If input_string is empty, there is a trailing comma; fail pa
<t>No structured data has been found; return members (which is empty).</t> rsing.</li>
</list></t> </ol>
</li>
<section anchor="parse-item-or-list" title="Parsing an Item or Inner List"> <li>No structured data has been found; return members (which is empt
y).</li>
<t>Given an ASCII string as input_string, return the tuple (item_or_inner_list, </ol>
parameters), where item_or_inner_list can be either a single bare item, or an ar <section anchor="parse-item-or-list">
ray of (bare_item, parameters) tuples. input_string is modified to remove the pa <name>Parsing an Item or Inner List</name>
rsed value.</t> <t>Given an ASCII string as input_string, return the tuple
(item_or_inner_list, parameters), where item_or_inner_list can be
<t><list style="numbers"> either a single bare item or an array of (bare_item, parameters)
<t>If the first character of input_string is “(“, return the result of running tuples. input_string is modified to remove the parsed value.</t>
Parsing an Inner List (<xref target="parse-innerlist"/>) with input_string.</t> <ol>
<t>Return the result of running Parsing an Item (<xref target="parse-item"/>) <li>If the first character of input_string is "(", return the
with input_string.</t> result of running Parsing an Inner List (<xref target="parse-innerl
</list></t> ist"/>) with
input_string.</li>
</section> <li>Return the result of running Parsing an Item (<xref target="pa
<section anchor="parse-innerlist" title="Parsing an Inner List"> rse-item"/>) with input_string.</li>
</ol>
<t>Given an ASCII string as input_string, return the tuple (inner_list, paramete </section>
rs), where inner_list is an array of (bare_item, parameters) tuples. input_strin <section anchor="parse-innerlist">
g is modified to remove the parsed value.</t> <name>Parsing an Inner List</name>
<t>Given an ASCII string as input_string, return the tuple
<t><list style="numbers"> (inner_list, parameters), where inner_list is an array of
<t>Consume the first character of input_string; if it is not “(“, fail parsing (bare_item, parameters) tuples. input_string is modified to remove
.</t> the parsed value.</t>
<t>Let inner_list be an empty array.</t> <ol>
<t>While input_string is not empty: <li>Consume the first character of input_string; if it is not
<list style="numbers"> "(", fail parsing.</li>
<t>Discard any leading SP characters from input_string.</t> <li>Let inner_list be an empty array.</li>
<t>If the first character of input_string is “)”: <li>
<list style="numbers"> <t>While input_string is not empty:
<t>Consume the first character of input_string.</t> </t>
<t>Let parameters be the result of running Parsing Parameters (<xref t <ol>
arget="parse-param"/>) with input_string.</t> <li>Discard any leading SP characters from input_string.</li>
<t>Return the tuple (inner_list, parameters).</t> <li>
</list></t> <t>If the first character of input_string is ")":
<t>Let item be the result of running Parsing an Item (<xref target="parse- </t>
item"/>) with input_string.</t> <ol>
<t>Append item to inner_list.</t> <li>Consume the first character of input_string.</li>
<t>If the first character of input_string is not SP or “)”, fail parsing.< <li>Let parameters be the result of running Parsing
/t> Parameters (<xref target="parse-param"/>) with input_string
</list></t> .</li>
<t>The end of the inner list was not found; fail parsing.</t> <li>Return the tuple (inner_list, parameters).</li>
</list></t> </ol>
</li>
</section> <li>Let item be the result of running Parsing an Item (<xref t
</section> arget="parse-item"/>) with
<section anchor="parse-dictionary" title="Parsing a Dictionary"> input_string.</li>
<li>Append item to inner_list.</li>
<t>Given an ASCII string as input_string, return an ordered map whose values are <li>If the first character of input_string is not SP or ")",
(item_or_inner_list, parameters) tuples. input_string is modified to remove the fail parsing.</li>
parsed value.</t> </ol>
</li>
<t><list style="numbers"> <li>The end of the Inner List was not found; fail parsing.</li>
<t>Let dictionary be an empty, ordered map.</t> </ol>
<t>While input_string is not empty: </section>
<list style="numbers"> </section>
<t>Let this_key be the result of running Parsing a Key (<xref target="pars <section anchor="parse-dictionary">
e-key"/>) with input_string.</t> <name>Parsing a Dictionary</name>
<t>If the first character of input_string is “=”: <t>Given an ASCII string as input_string, return an ordered map
<list style="numbers"> whose values are (item_or_inner_list, parameters)
<t>Consume the first character of input_string.</t> tuples. input_string is modified to remove the parsed value.</t>
<t>Let member be the result of running Parsing an Item or Inner List ( <ol>
<xref target="parse-item-or-list"/>) with input_string.</t> <li>Let dictionary be an empty, ordered map.</li>
</list></t> <li>
<t>Otherwise: <t>While input_string is not empty:
<list style="numbers"> </t>
<t>Let value be Boolean true.</t> <ol>
<t>Let parameters be the result of running Parsing Parameters <xref ta <li>Let this_key be the result of running Parsing a Key (<xref t
rget="parse-param"/> with input_string.</t> arget="parse-key"/>) with input_string.</li>
<t>Let member be the tuple (value, parameters).</t> <li>
</list></t> <t>If the first character of input_string is "=":
<t>Add name this_key with value member to dictionary. If dictionary alread </t>
y contains a name this_key (comparing character-for-character), overwrite its va <ol>
lue.</t> <li>Consume the first character of input_string.</li>
<t>Discard any leading OWS characters from input_string.</t> <li>Let member be the result of running Parsing an Item or
<t>If input_string is empty, return dictionary.</t> Inner List (<xref target="parse-item-or-list"/>) with input_s
<t>Consume the first character of input_string; if it is not “,”, fail par tring.</li>
sing.</t> </ol>
<t>Discard any leading OWS characters from input_string.</t> </li>
<t>If input_string is empty, there is a trailing comma; fail parsing.</t> <li>
</list></t> <t>Otherwise:
<t>No structured data has been found; return dictionary (which is empty).</t> </t>
</list></t> <ol>
<li>Let value be Boolean true.</li>
<t>Note that when duplicate Dictionary keys are encountered, this has the effect <li>Let parameters be the result of running Parsing
of ignoring all but the last instance.</t> Parameters (<xref target="parse-param"/>)
with input_string.</li>
</section> <li>Let member be the tuple (value, parameters).</li>
<section anchor="parse-item" title="Parsing an Item"> </ol>
</li>
<t>Given an ASCII string as input_string, return a (bare_item, parameters) tuple <li>If dictionary already contains a key this_key (comparing cha
. input_string is modified to remove the parsed value.</t> racter for character), overwrite its value with member.</li>
<li>Otherwise, append key this_key with value member to dictionar
<t><list style="numbers"> y.</li>
<t>Let bare_item be the result of running Parsing a Bare Item (<xref target="p <li>Discard any leading OWS characters from input_string.</li>
arse-bare-item"/>) with input_string.</t> <li>If input_string is empty, return dictionary.</li>
<t>Let parameters be the result of running Parsing Parameters (<xref target="p <li>Consume the first character of input_string; if it is not
arse-param"/>) with input_string.</t> ",", fail parsing.</li>
<t>Return the tuple (bare_item, parameters).</t> <li>Discard any leading OWS characters from input_string.</li>
</list></t> <li>If input_string is empty, there is a trailing comma; fail pa
rsing.</li>
<section anchor="parse-bare-item" title="Parsing a Bare Item"> </ol>
</li>
<t>Given an ASCII string as input_string, return a bare Item. input_string is mo <li>No structured data has been found; return dictionary (which is e
dified to remove the parsed value.</t> mpty).</li>
</ol>
<t><list style="numbers"> <t>Note that when duplicate Dictionary keys are encountered, all but
<t>If the first character of input_string is a “-“ or a DIGIT, return the resu the last instance are ignored.</t>
lt of running Parsing an Integer or Decimal (<xref target="parse-number"/>) with </section>
input_string.</t> <section anchor="parse-item">
<t>If the first character of input_string is a DQUOTE, return the result of ru <name>Parsing an Item</name>
nning Parsing a String (<xref target="parse-string"/>) with input_string.</t> <t>Given an ASCII string as input_string, return a (bare_item,
<t>If the first character of input_string is “:”, return the result of running parameters) tuple. input_string is modified to remove the parsed
Parsing a Byte Sequence (<xref target="parse-binary"/>) with input_string.</t> value.</t>
<t>If the first character of input_string is “?”, return the result of running <ol>
Parsing a Boolean (<xref target="parse-boolean"/>) with input_string.</t> <li>Let bare_item be the result of running Parsing a Bare Item
<t>If the first character of input_string is an ALPHA or “*”, return the resul (<xref target="parse-bare-item"/>) with
t of running Parsing a Token (<xref target="parse-token"/>) with input_string.</ input_string.</li>
t> <li>Let parameters be the result of running Parsing Parameters
<t>Otherwise, the item type is unrecognized; fail parsing.</t> (<xref target="parse-param"/>) with
</list></t> input_string.</li>
<li>Return the tuple (bare_item, parameters).</li>
</section> </ol>
<section anchor="parse-param" title="Parsing Parameters"> <section anchor="parse-bare-item">
<name>Parsing a Bare Item</name>
<t>Given an ASCII string as input_string, return an ordered map whose values are <t>Given an ASCII string as input_string, return a bare
bare Items. input_string is modified to remove the parsed value.</t> Item. input_string is modified to remove the parsed value.</t>
<ol>
<t><list style="numbers"> <li>If the first character of input_string is a "-" or a DIGIT,
<t>Let parameters be an empty, ordered map.</t> return the result of running Parsing an Integer or Decimal
<t>While input_string is not empty: (<xref target="parse-number"/>) with
<list style="numbers"> input_string.</li>
<t>If the first character of input_string is not “;”, exit the loop.</t> <li>If the first character of input_string is a DQUOTE, return
<t>Consume a “;” character from the beginning of input_string.</t> the result of running Parsing a String (<xref target="parse-string"
<t>Discard any leading SP characters from input_string.</t> />) with
<t>let param_name be the result of running Parsing a Key (<xref target="pa input_string.</li>
rse-key"/>) with input_string.</t> <li>If the first character of input_string is an ALPHA or "*",
<t>Let param_value be Boolean true.</t> return the result of running Parsing a Token (<xref target="parse-t
<t>If the first character of input_string is “=”: oken"/>) with input_string.</li>
<list style="numbers"> <li>If the first character of input_string is ":", return the
<t>Consume the “=” character at the beginning of input_string.</t> result of running Parsing a Byte Sequence (<xref target="parse-bina
<t>Let param_value be the result of running Parsing a Bare Item (<xref ry"/>) with
target="parse-bare-item"/>) with input_string.</t> input_string.</li>
</list></t> <li>If the first character of input_string is "?", return the
<t>Append key param_name with value param_value to parameters. If paramete result of running Parsing a Boolean (<xref target="parse-boolean"/>
rs already contains a name param_name (comparing character-for-character), overw ) with
rite its value.</t> input_string.</li>
</list></t> <li>Otherwise, the item type is unrecognized; fail parsing.</li>
<t>Return parameters.</t> </ol>
</list></t> </section>
<section anchor="parse-param">
<t>Note that when duplicate Parameter keys are encountered, this has the effect <name>Parsing Parameters</name>
of ignoring all but the last instance.</t> <t>Given an ASCII string as input_string, return an ordered map
whose values are bare Items. input_string is modified to remove
</section> the parsed value.</t>
<section anchor="parse-key" title="Parsing a Key"> <ol>
<li>Let parameters be an empty, ordered map.</li>
<t>Given an ASCII string as input_string, return a key. input_string is modified <li>
to remove the parsed value.</t> <t>While input_string is not empty:
</t>
<t><list style="numbers"> <ol>
<t>If the first character of input_string is not lcalpha or “*”, fail parsing. <li>If the first character of input_string is not ";", exit th
</t> e loop.</li>
<t>Let output_string be an empty string.</t> <li>Consume the ";" character from the beginning of input_stri
<t>While input_string is not empty: ng.</li>
<list style="numbers"> <li>Discard any leading SP characters from input_string.</li>
<t>If the first character of input_string is not one of lcalpha, DIGIT, “_ <li>Let param_key be the result of running Parsing a Key
”, “-“, “.”, or “*”, return output_string.</t> (<xref target="parse-key"/>) with
<t>Let char be the result of consuming the first character of input_string input_string.</li>
.</t> <li>Let param_value be Boolean true.</li>
<t>Append char to output_string.</t> <li>
</list></t> <t>If the first character of input_string is "=":
<t>Return output_string.</t> </t>
</list></t> <ol>
<li>Consume the "=" character at the beginning of input_st
</section> ring.</li>
</section> <li>Let param_value be the result of running Parsing a
<section anchor="parse-number" title="Parsing an Integer or Decimal"> Bare Item (<xref target="parse-bare-item"/>) with input_str
ing.</li>
<t>Given an ASCII string as input_string, return an Integer or Decimal. input_st </ol>
ring is modified to remove the parsed value.</t> </li>
<li>If parameters already contains a key param_key (comparing
<t>NOTE: This algorithm parses both Integers (<xref target="integer"/>) and Deci character for character), overwrite its value with param_value.</li>
mals (<xref target="decimal"/>), and returns the corresponding structure.</t> <li>Otherwise, append key param_key with value param_value to p
arameters.</li>
<t><list style="numbers"> </ol>
<t>Let type be “integer”.</t> </li>
<t>Let sign be 1.</t> <li>Return parameters.</li>
<t>Let input_number be an empty string.</t> </ol>
<t>If the first character of input_string is “-“, consume it and set sign to - <t>Note that when duplicate parameter keys are encountered,
1.</t> all but the last instance are ignored.</t>
<t>If input_string is empty, there is an empty integer; fail parsing.</t> </section>
<t>If the first character of input_string is not a DIGIT, fail parsing.</t> <section anchor="parse-key">
<t>While input_string is not empty: <name>Parsing a Key</name>
<list style="numbers"> <t>Given an ASCII string as input_string, return a
<t>Let char be the result of consuming the first character of input_string key. input_string is modified to remove the parsed value.</t>
.</t> <ol>
<t>If char is a DIGIT, append it to input_number.</t> <li>If the first character of input_string is not lcalpha or
<t>Else, if type is “integer” and char is “.”: "*", fail parsing.</li>
<list style="numbers"> <li>Let output_string be an empty string.</li>
<t>If input_number contains more than 12 characters, fail parsing.</t> <li>
<t>Otherwise, append char to input_number and set type to “decimal”.</ <t>While input_string is not empty:
t> </t>
</list></t> <ol>
<t>Otherwise, prepend char to input_string, and exit the loop.</t> <li>If the first character of input_string is not one of
<t>If type is “integer” and input_number contains more than 15 characters, lcalpha, DIGIT, "_", "-", ".", or "*", return
fail parsing.</t> output_string.</li>
<t>If type is “decimal” and input_number contains more than 16 characters, <li>Let char be the result of consuming the first character of
fail parsing.</t> input_string.</li>
</list></t> <li>Append char to output_string.</li>
<t>If type is “integer”: </ol>
<list style="numbers"> </li>
<t>Parse input_number as an integer and let output_number be the product o <li>Return output_string.</li>
f the result and sign.</t> </ol>
<t>If output_number is outside the range -999,999,999,999,999 to 999,999,9 </section>
99,999,999 inclusive, fail parsing.</t> </section>
</list></t> <section anchor="parse-number">
<t>Otherwise: <name>Parsing an Integer or Decimal</name>
<list style="numbers"> <t>Given an ASCII string as input_string, return an Integer or
<t>If the final character of input_number is “.”, fail parsing.</t> Decimal. input_string is modified to remove the parsed value.</t>
<t>If the number of characters after “.” in input_number is greater than t <t>NOTE: This algorithm parses both Integers (<xref target="integer"/>
hree, fail parsing.</t> ) and Decimals (<xref target="decimal"/>), and returns the corresponding structu
<t>Parse input_number as a decimal number and let output_number be the pro re.</t>
duct of the result and sign.</t> <ol>
</list></t> <li>Let type be "integer".</li>
<t>Return output_number.</t> <li>Let sign be 1.</li>
</list></t> <li>Let input_number be an empty string.</li>
<li>If the first character of input_string is "-", consume it and
</section> set sign to -1.</li>
<section anchor="parse-string" title="Parsing a String"> <li>If input_string is empty, there is an empty integer; fail
parsing.</li>
<t>Given an ASCII string as input_string, return an unquoted String. input_strin <li>If the first character of input_string is not a DIGIT, fail
g is modified to remove the parsed value.</t> parsing.</li>
<li>
<t><list style="numbers"> <t>While input_string is not empty:
<t>Let output_string be an empty string.</t> </t>
<t>If the first character of input_string is not DQUOTE, fail parsing.</t> <ol>
<t>Discard the first character of input_string.</t> <li>Let char be the result of consuming the first character of
<t>While input_string is not empty: input_string.</li>
<list style="numbers"> <li>If char is a DIGIT, append it to input_number.</li>
<t>Let char be the result of consuming the first character of input_string <li>
.</t> <t>Else, if type is "integer" and char is ".":
<t>If char is a backslash (“\”): </t>
<list style="numbers"> <ol>
<t>If input_string is now empty, fail parsing.</t> <li>If input_number contains more than 12 characters, fail p
<t>Let next_char be the result of consuming the first character of inp arsing.</li>
ut_string.</t> <li>Otherwise, append char to input_number and set type to "
<t>If next_char is not DQUOTE or “\”, fail parsing.</t> decimal".</li>
<t>Append next_char to output_string.</t> </ol>
</list></t> </li>
<t>Else, if char is DQUOTE, return output_string.</t> <li>Otherwise, prepend char to input_string, and exit the loop.<
<t>Else, if char is in the range %x00-1f or %x7f (i.e., is not in VCHAR or /li>
SP), fail parsing.</t> <li>If type is "integer" and input_number contains more than
<t>Else, append char to output_string.</t> 15 characters, fail parsing.</li>
</list></t> <li>If type is "decimal" and input_number contains more than
<t>Reached the end of input_string without finding a closing DQUOTE; fail pars 16 characters, fail parsing.</li>
ing.</t> </ol>
</list></t> </li>
<li>
</section> <t>If type is "integer":
<section anchor="parse-token" title="Parsing a Token"> </t>
<ol>
<t>Given an ASCII string as input_string, return a Token. input_string is modifi <li>Parse input_number as an integer and let output_number be
ed to remove the parsed value.</t> the product of the result and sign.</li>
</ol>
<t><list style="numbers"> </li>
<t>If the first character of input_string is not ALPHA or “*”, fail parsing.</ <li>
t> <t>Otherwise:
<t>Let output_string be an empty string.</t> </t>
<t>While input_string is not empty: <ol>
<list style="numbers"> <li>If the final character of input_number is ".", fail parsing.
<t>If the first character of input_string is not in tchar, “:” or “/”, ret </li>
urn output_string.</t> <li>If the number of characters after "." in input_number is
<t>Let char be the result of consuming the first character of input_string greater than three, fail parsing.</li>
.</t> <li>Parse input_number as a decimal number and let
<t>Append char to output_string.</t> output_number be the product of the result and sign.</li>
</list></t> </ol>
<t>Return output_string.</t> </li>
</list></t> <li>Return output_number.</li>
</ol>
</section> </section>
<section anchor="parse-binary" title="Parsing a Byte Sequence"> <section anchor="parse-string">
<name>Parsing a String</name>
<t>Given an ASCII string as input_string, return a Byte Sequence. input_string i <t>Given an ASCII string as input_string, return an unquoted
s modified to remove the parsed value.</t> String. input_string is modified to remove the parsed value.</t>
<ol>
<t><list style="numbers"> <li>Let output_string be an empty string.</li>
<t>If the first character of input_string is not “:”, fail parsing.</t> <li>If the first character of input_string is not DQUOTE, fail parsi
<t>Discard the first character of input_string.</t> ng.</li>
<t>If there is not a “:” character before the end of input_string, fail parsin <li>Discard the first character of input_string.</li>
g.</t> <li>
<t>Let b64_content be the result of consuming content of input_string up to bu <t>While input_string is not empty:
t not including the first instance of the character “:”.</t> </t>
<t>Consume the “:” character at the beginning of input_string.</t> <ol>
<t>If b64_content contains a character not included in ALPHA, DIGIT, “+”, “/” <li>Let char be the result of consuming the first character of i
and “=”, fail parsing.</t> nput_string.</li>
<t>Let binary_content be the result of Base 64 Decoding <xref target="RFC4648" <li>
/> b64_content, synthesizing padding if necessary (note the requirements about r <t>If char is a backslash ("\"):
ecipient behavior below).</t> </t>
<t>Return binary_content.</t> <ol>
</list></t> <li>If input_string is now empty, fail parsing.</li>
<li>Let next_char be the result of consuming the first
<t>Because some implementations of base64 do not allow rejection of encoded data character of input_string.</li>
that is not properly “=” padded (see <xref target="RFC4648"/>, Section 3.2), pa <li>If next_char is not DQUOTE or "\", fail parsing.</li>
rsers SHOULD NOT fail when “=” padding is not present, unless they cannot be con <li>Append next_char to output_string.</li>
figured to do so.</t> </ol>
</li>
<t>Because some implementations of base64 do not allow rejection of encoded data <li>Else, if char is DQUOTE, return output_string.</li>
that has non-zero pad bits (see <xref target="RFC4648"/>, Section 3.5), parsers <li>Else, if char is in the range %x00-1f or %x7f-ff (i.e., it i
SHOULD NOT fail when non-zero pad bits are present, unless they cannot be confi s
gured to do so.</t> not in VCHAR or SP), fail parsing.</li>
<li>Else, append char to output_string.</li>
<t>This specification does not relax the requirements in <xref target="RFC4648"/ </ol>
>, Section 3.1 and 3.3; therefore, parsers MUST fail on characters outside the b </li>
ase64 alphabet, and on line feeds in encoded data.</t> <li>Reached the end of input_string without finding a closing
DQUOTE; fail parsing.</li>
</section> </ol>
<section anchor="parse-boolean" title="Parsing a Boolean"> </section>
<section anchor="parse-token">
<t>Given an ASCII string as input_string, return a Boolean. input_string is modi <name>Parsing a Token</name>
fied to remove the parsed value.</t> <t>Given an ASCII string as input_string, return a
Token. input_string is modified to remove the parsed value.</t>
<t><list style="numbers"> <ol>
<t>If the first character of input_string is not “?”, fail parsing.</t> <li>If the first character of input_string is not ALPHA or "*",
<t>Discard the first character of input_string.</t> fail parsing.</li>
<t>If the first character of input_string matches “1”, discard the first chara <li>Let output_string be an empty string.</li>
cter, and return true.</t> <li>
<t>If the first character of input_string matches “0”, discard the first chara <t>While input_string is not empty:
cter, and return false.</t> </t>
<t>No value has matched; fail parsing.</t> <ol>
</list></t> <li>If the first character of input_string is not in tchar,
":", or "/", return output_string.</li>
</section> <li>Let char be the result of consuming the first character of
</section> input_string.</li>
</section> <li>Append char to output_string.</li>
<section anchor="iana-considerations" title="IANA Considerations"> </ol>
</li>
<t>This document has no actions for IANA.</t> <li>Return output_string.</li>
</ol>
</section> </section>
<section anchor="security-considerations" title="Security Considerations"> <section anchor="parse-binary">
<name>Parsing a Byte Sequence</name>
<t>The size of most types defined by Structured Fields is not limited; as a resu <t>Given an ASCII string as input_string, return a Byte
lt, extremely large fields could be an attack vector (e.g., for resource consump Sequence. input_string is modified to remove the parsed value.</t>
tion). Most HTTP implementations limit the sizes of individual fields as well as <ol>
the overall header or trailer section size to mitigate such attacks.</t> <li>If the first character of input_string is not ":", fail parsing.
</li>
<t>It is possible for parties with the ability to inject new HTTP fields to chan <li>Discard the first character of input_string.</li>
ge the meaning <li>If there is not a ":" character before the end of input_string,
of a Structured Field. In some circumstances, this will cause parsing to fail, b fail parsing.</li>
ut it is not possible to reliably fail in all such circumstances.</t> <li>Let b64_content be the result of consuming content of
input_string up to but not including the first instance of the
</section> character ":".</li>
<li>Consume the ":" character at the beginning of input_string.</li>
<li>If b64_content contains a character not included in ALPHA,
DIGIT, "+", "/", and "=", fail parsing.</li>
<li>Let binary_content be the result of base64-decoding <xref target
="RFC4648"/> b64_content, synthesizing
padding if necessary (note the requirements about recipient
behavior below). If base64 decoding fails, parsing fails.</li>
<li>Return binary_content.</li>
</ol>
<t>Because some implementations of base64 do not allow rejection of
encoded data that is not properly "=" padded (see <xref target="RFC4648
" sectionFormat="comma" section="3.2"/>), parsers <bcp14>SHOULD
NOT</bcp14> fail when "=" padding is not present, unless they cannot be
configured to do so.</t>
<t>Because some implementations of base64 do not allow rejection of
encoded data that has non-zero pad bits (see <xref target="RFC4648" sec
tionFormat="comma" section="3.5"/>), parsers <bcp14>SHOULD NOT</bcp14> fail when
non-zero pad bits are present, unless they cannot be configured to
do so.</t>
<t>This specification does not relax the requirements in <xref target=
"RFC4648"/>, Sections <xref target="RFC4648" sectionFormat="bare" section="3.1"/
> and <xref target="RFC4648" sectionFormat="bare" section="3.3"/>; therefore,
parsers <bcp14>MUST</bcp14> fail on characters outside the base64 alpha
bet and on line
feeds in encoded data.</t>
</section>
<section anchor="parse-boolean">
<name>Parsing a Boolean</name>
<t>Given an ASCII string as input_string, return a
Boolean. input_string is modified to remove the parsed value.</t>
<ol>
<li>If the first character of input_string is not "?", fail parsing.
</li>
<li>Discard the first character of input_string.</li>
<li>If the first character of input_string matches "1", discard
the first character, and return true.</li>
<li>If the first character of input_string matches "0", discard
the first character, and return false.</li>
<li>No value has matched; fail parsing.</li>
</ol>
</section>
</section>
</section>
<section anchor="iana-considerations">
<name>IANA Considerations</name>
<t>This document has no IANA actions.</t>
</section>
<section anchor="security-considerations">
<name>Security Considerations</name>
<t>The size of most types defined by Structured Fields is not limited;
as a result, extremely large fields could be an attack vector (e.g., for
resource consumption). Most HTTP implementations limit the sizes of
individual fields as well as the overall header or trailer section size
to mitigate such attacks.</t>
<t>It is possible for parties with the ability to inject new HTTP fields
to change the meaning of a Structured Field. In some circumstances, this
will cause parsing to fail, but it is not possible to reliably fail in
all such circumstances.</t>
</section>
</middle> </middle>
<back> <back>
<references>
<name>References</name>
<references>
<name>Normative References</name>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.7230.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.2119.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.8174.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.5234.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.0020.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.4648.xml"/>
</references>
<references>
<name>Informative References</name>
<reference anchor="IEEE754" target="https://ieeexplore.ieee.org/document
/8766229">
<front>
<title>IEEE Standard for Floating-Point Arithmetic</title>
<seriesInfo name="DOI" value="10.1109/IEEESTD.2019.8766229"/>
<seriesInfo name="IEEE" value="754-2019"/>
<author>
<organization>IEEE</organization>
</author>
<date year="2019" month="July"/>
</front>
</reference>
<references title='Normative References'> <reference anchor='STD63' target='https://www.rfc-editor.org/info/std63'>
<reference anchor="RFC7230" target='https://www.rfc-editor.org/info/rfc7230'>
<front>
<title>Hypertext Transfer Protocol (HTTP/1.1): Message Syntax and Routing</title
>
<author initials='R.' surname='Fielding' fullname='R. Fielding' role='editor'><o
rganization /></author>
<author initials='J.' surname='Reschke' fullname='J. Reschke' role='editor'><org
anization /></author>
<date year='2014' month='June' />
<abstract><t>The Hypertext Transfer Protocol (HTTP) is a stateless application-l
evel protocol for distributed, collaborative, hypertext information systems. Th
is document provides an overview of HTTP architecture and its associated termino
logy, defines the &quot;http&quot; and &quot;https&quot; Uniform Resource Identi
fier (URI) schemes, defines the HTTP/1.1 message syntax and parsing requirements
, and describes related security concerns for implementations.</t></abstract>
</front>
<seriesInfo name='RFC' value='7230'/>
<seriesInfo name='DOI' value='10.17487/RFC7230'/>
</reference>
<reference anchor="RFC2119" target='https://www.rfc-editor.org/info/rfc2119'>
<front>
<title>Key words for use in RFCs to Indicate Requirement Levels</title>
<author initials='S.' surname='Bradner' fullname='S. Bradner'><organization /></
author>
<date year='1997' month='March' />
<abstract><t>In many standards track documents several words are used to signify
the requirements in the specification. These words are often capitalized. This
document defines these words as they should be interpreted in IETF documents.
This document specifies an Internet Best Current Practices for the Internet Comm
unity, and requests discussion and suggestions for improvements.</t></abstract>
</front>
<seriesInfo name='BCP' value='14'/>
<seriesInfo name='RFC' value='2119'/>
<seriesInfo name='DOI' value='10.17487/RFC2119'/>
</reference>
<reference anchor="RFC8174" target='https://www.rfc-editor.org/info/rfc8174'>
<front>
<title>Ambiguity of Uppercase vs Lowercase in RFC 2119 Key Words</title>
<author initials='B.' surname='Leiba' fullname='B. Leiba'><organization /></auth
or>
<date year='2017' month='May' />
<abstract><t>RFC 2119 specifies common key words that may be used in protocol s
pecifications. This document aims to reduce the ambiguity by clarifying that on
ly UPPERCASE usage of the key words have the defined special meanings.</t></abs
tract>
</front>
<seriesInfo name='BCP' value='14'/>
<seriesInfo name='RFC' value='8174'/>
<seriesInfo name='DOI' value='10.17487/RFC8174'/>
</reference>
<reference anchor="RFC5234" target='https://www.rfc-editor.org/info/rfc5234'>
<front>
<title>Augmented BNF for Syntax Specifications: ABNF</title>
<author initials='D.' surname='Crocker' fullname='D. Crocker' role='editor'><org
anization /></author>
<author initials='P.' surname='Overell' fullname='P. Overell'><organization /></
author>
<date year='2008' month='January' />
<abstract><t>Internet technical specifications often need to define a formal syn
tax. Over the years, a modified version of Backus-Naur Form (BNF), called Augme
nted BNF (ABNF), has been popular among many Internet specifications. The curre
nt specification documents ABNF. It balances compactness and simplicity with rea
sonable representational power. The differences between standard BNF and ABNF i
nvolve naming rules, repetition, alternatives, order-independence, and value ran
ges. This specification also supplies additional rule definitions and encoding
for a core lexical analyzer of the type common to several Internet specification
s. [STANDARDS-TRACK]</t></abstract>
</front>
<seriesInfo name='STD' value='68'/>
<seriesInfo name='RFC' value='5234'/>
<seriesInfo name='DOI' value='10.17487/RFC5234'/>
</reference>
<reference anchor="RFC0020" target='https://www.rfc-editor.org/info/rfc20'>
<front>
<title>ASCII format for network interchange</title>
<author initials='V.G.' surname='Cerf' fullname='V.G. Cerf'><organization /></au
thor>
<date year='1969' month='October' />
</front>
<seriesInfo name='STD' value='80'/>
<seriesInfo name='RFC' value='20'/>
<seriesInfo name='DOI' value='10.17487/RFC0020'/>
</reference>
<reference anchor="RFC4648" target='https://www.rfc-editor.org/info/rfc4648'>
<front>
<title>The Base16, Base32, and Base64 Data Encodings</title>
<author initials='S.' surname='Josefsson' fullname='S. Josefsson'><organization
/></author>
<date year='2006' month='October' />
<abstract><t>This document describes the commonly used base 64, base 32, and bas
e 16 encoding schemes. It also discusses the use of line-feeds in encoded data,
use of padding in encoded data, use of non-alphabet characters in encoded data,
use of different encoding alphabets, and canonical encodings. [STANDARDS-TRACK
]</t></abstract>
</front>
<seriesInfo name='RFC' value='4648'/>
<seriesInfo name='DOI' value='10.17487/RFC4648'/>
</reference>
</references>
<references title='Informative References'>
<reference anchor="IEEE754" target="https://ieeexplore.ieee.org/document/8766229
">
<front>
<title>IEEE Standard for Floating-Point Arithmetic</title>
<author >
<organization>IEEE</organization>
</author>
<date year="2019" month="July"/>
</front>
<seriesInfo name="IEEE" value="754-2019"/>
<seriesInfo name="DOI" value="10.1109/IEEESTD.2019.8766229"/>
<seriesInfo name="ISBN" value="978-1-5044-5924-2"/>
</reference>
<reference anchor="UTF-8" target="http://www.rfc-editor.org/info/std63">
<front>
<title>UTF-8, a transformation format of ISO 10646</title>
<author initials="F." surname="Yergeau" fullname="F. Yergeau">
<organization></organization>
</author>
<date year="2003" month="November"/>
</front>
<seriesInfo name="STD" value="63"/>
<seriesInfo name="RFC" value="3629"/>
<seriesInfo name="DOI" value="10.17487/RFC3629"/>
</reference>
<reference anchor="RFC7231" target='https://www.rfc-editor.org/info/rfc7231'>
<front>
<title>Hypertext Transfer Protocol (HTTP/1.1): Semantics and Content</title>
<author initials='R.' surname='Fielding' fullname='R. Fielding' role='editor'><o
rganization /></author>
<author initials='J.' surname='Reschke' fullname='J. Reschke' role='editor'><org
anization /></author>
<date year='2014' month='June' />
<abstract><t>The Hypertext Transfer Protocol (HTTP) is a stateless \%application
- level protocol for distributed, collaborative, hypertext information systems.
This document defines the semantics of HTTP/1.1 messages, as expressed by reque
st methods, request header fields, response status codes, and response header fi
elds, along with the payload of messages (metadata and body content) and mechani
sms for content negotiation.</t></abstract>
</front>
<seriesInfo name='RFC' value='7231'/>
<seriesInfo name='DOI' value='10.17487/RFC7231'/>
</reference>
<reference anchor="RFC7540" target='https://www.rfc-editor.org/info/rfc7540'>
<front>
<title>Hypertext Transfer Protocol Version 2 (HTTP/2)</title>
<author initials='M.' surname='Belshe' fullname='M. Belshe'><organization /></au
thor>
<author initials='R.' surname='Peon' fullname='R. Peon'><organization /></author
>
<author initials='M.' surname='Thomson' fullname='M. Thomson' role='editor'><org
anization /></author>
<date year='2015' month='May' />
<abstract><t>This specification describes an optimized expression of the semanti
cs of the Hypertext Transfer Protocol (HTTP), referred to as HTTP version 2 (HTT
P/2). HTTP/2 enables a more efficient use of network resources and a reduced pe
rception of latency by introducing header field compression and allowing multipl
e concurrent exchanges on the same connection. It also introduces unsolicited p
ush of representations from servers to clients.</t><t>This specification is an a
lternative to, but does not obsolete, the HTTP/1.1 message syntax. HTTP's exist
ing semantics remain unchanged.</t></abstract>
</front>
<seriesInfo name='RFC' value='7540'/>
<seriesInfo name='DOI' value='10.17487/RFC7540'/>
</reference>
<reference anchor="RFC7541" target='https://www.rfc-editor.org/info/rfc7541'>
<front>
<title>HPACK: Header Compression for HTTP/2</title>
<author initials='R.' surname='Peon' fullname='R. Peon'><organization /></author
>
<author initials='H.' surname='Ruellan' fullname='H. Ruellan'><organization /></
author>
<date year='2015' month='May' />
<abstract><t>This specification defines HPACK, a compression format for efficien
tly representing HTTP header fields, to be used in HTTP/2.</t></abstract>
</front>
<seriesInfo name='RFC' value='7541'/>
<seriesInfo name='DOI' value='10.17487/RFC7541'/>
</reference>
<reference anchor="RFC8259" target='https://www.rfc-editor.org/info/rfc8259'>
<front>
<title>The JavaScript Object Notation (JSON) Data Interchange Format</title>
<author initials='T.' surname='Bray' fullname='T. Bray' role='editor'><organizat
ion /></author>
<date year='2017' month='December' />
<abstract><t>JavaScript Object Notation (JSON) is a lightweight, text-based, lan
guage-independent data interchange format. It was derived from the ECMAScript P
rogramming Language Standard. JSON defines a small set of formatting rules for
the portable representation of structured data.</t><t>This document removes inco
nsistencies with other specifications of JSON, repairs specification errors, and
offers experience-based interoperability guidance.</t></abstract>
</front>
<seriesInfo name='STD' value='90'/>
<seriesInfo name='RFC' value='8259'/>
<seriesInfo name='DOI' value='10.17487/RFC8259'/>
</reference>
<reference anchor="RFC7493" target='https://www.rfc-editor.org/info/rfc7493'>
<front> <front>
<title>The I-JSON Message Format</title> <title>UTF-8, a transformation format of ISO 10646</title>
<author initials='T.' surname='Bray' fullname='T. Bray' role='editor'><organizat <author initials='F.' surname='Yergeau' fullname='F. Yergeau'><organization /></
ion /></author> author>
<date year='2015' month='March' /> <date year='2003' month='November' />
<abstract><t>I-JSON (short for &quot;Internet JSON&quot;) is a restricted profil
e of JSON designed to maximize interoperability and increase confidence that sof
tware can process it successfully with predictable results.</t></abstract>
</front> </front>
<seriesInfo name='RFC' value='7493'/> <seriesInfo name='STD' value='63'/>
<seriesInfo name='DOI' value='10.17487/RFC7493'/> <seriesInfo name='RFC' value='3629'/>
</reference> </reference>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.7231.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.7540.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.7541.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.8259.xml"/>
<xi:include href="https://xml2rfc.tools.ietf.org/public/rfc/bibxml/refer
ence.RFC.7493.xml"/>
</references>
</references> </references>
<section anchor="faq">
<section anchor="faq" title="Frequently Asked Questions"> <name>Frequently Asked Questions</name>
<section anchor="why-not-json">
<section anchor="why-not-json" title="Why not JSON?"> <name>Why Not JSON?</name>
<t>Earlier proposals for Structured Fields were based upon JSON <xref ta
<t>Earlier proposals for Structured Fields were based upon JSON <xref target="RF rget="RFC8259"/>. However, constraining its use to
C8259"/>. However, constraining its use to make it suitable for HTTP header fiel make it suitable for HTTP header fields required senders and
ds required senders and recipients to implement specific additional handling.</t recipients to implement specific additional handling.</t>
> <t>For example, JSON has specification issues around large numbers and
objects with duplicate members. Although advice for avoiding these
<t>For example, JSON has specification issues around large numbers and objects w issues is available (e.g., <xref target="RFC7493"/>),
ith duplicate members. Although advice for avoiding these issues is available (e it cannot be relied upon.</t>
.g., <xref target="RFC7493"/>), it cannot be relied upon.</t> <t>Likewise, JSON strings are by default Unicode strings, which have a
number of potential interoperability issues (e.g., in
<t>Likewise, JSON strings are by default Unicode strings, which have a number of comparison). Although implementers can be advised to avoid non-ASCII
potential interoperability issues (e.g., in comparison). Although implementers content where unnecessary, this is difficult to enforce.</t>
can be advised to avoid non-ASCII content where unnecessary, this is difficult t <t>Another example is JSON's ability to nest content to arbitrary
o enforce.</t> depths. Since the resulting memory commitment might be unsuitable
(e.g., in embedded and other limited server deployments), it's
<t>Another example is JSON’s ability to nest content to arbitrary depths. Since necessary to limit it in some fashion; however, existing JSON
the resulting memory commitment might be unsuitable (e.g., in embedded and other implementations have no such limits, and even if a limit is specified,
limited server deployments), it’s necessary to limit it in some fashion; howeve it's likely that some field definition will find a need to violate
r, existing JSON implementations have no such limits, and even if a limit is spe it.</t>
cified, it’s likely that some field definition will find a need to violate it.</ <t>Because of JSON's broad adoption and implementation, it is
t> difficult to impose such additional constraints across all
implementations; some deployments would fail to enforce them, thereby
<t>Because of JSON’s broad adoption and implementation, it is difficult to impos harming interoperability. In short, if it looks like JSON, people will
e such additional constraints across all implementations; some deployments would be tempted to use a JSON parser/serializer on field values.</t>
fail to enforce them, thereby harming interoperability. In short, if it looks l <t>Since a major goal for Structured Fields is to improve
ike JSON, people will be tempted to use a JSON parser / serializer on field valu interoperability and simplify implementation, these concerns led to a
es.</t> format that requires a dedicated parser and serializer.</t>
<t>Additionally, there were widely shared feelings that JSON doesn't
<t>Since a major goal for Structured Fields is to improve interoperability and s "look right" in HTTP fields.</t>
implify implementation, these concerns led to a format that requires a dedicated </section>
parser and serializer.</t> </section>
<section anchor="implementation-notes">
<t>Additionally, there were widely shared feelings that JSON doesn’t “look right <name>Implementation Notes</name>
” in HTTP fields.</t> <t>A generic implementation of this specification should expose the
top-level serialize (<xref target="text-serialize"/>)
</section> and parse (<xref target="text-parse"/>) functions. They
</section> need not be functions; for example, it could be implemented as an
<section anchor="implementation-notes" title="Implementation Notes"> object, with methods for each of the different top-level types.</t>
<t>For interoperability, it's important that generic implementations be
<t>A generic implementation of this specification should expose the top-level se complete and follow the algorithms closely; see <xref target="strict"/>. T
rialize (<xref target="text-serialize"/>) and parse (<xref target="text-parse"/> o aid this, a common test suite is being maintained
) functions. They need not be functions; for example, it could be implemented as by the community at <eref brackets="angle" target="https://github.com/http
an object, with methods for each of the different top-level types.</t> wg/structured-field-tests"/>.</t>
<t>Implementers should note that Dictionaries and Parameters are
<t>For interoperability, it’s important that generic implementations be complete order-preserving maps. Some fields may not convey meaning in the
and follow the algorithms closely; see <xref target="strict"/>. To aid this, a ordering of these data types, but it should still be exposed so
common test suite is being maintained by the community at <eref target="https:// that it will be available to applications that need to use it.</t>
github.com/httpwg/structured-field-tests">https://github.com/httpwg/structured-f <t>Likewise, implementations should note that it's important to preserve
ield-tests</eref>.</t> the distinction between Tokens and Strings. While most programming
languages have native types that map to the other types well, it may be
<t>Implementers should note that Dictionaries and Parameters are order-preservin necessary to create a wrapper "token" object or use a parameter on
g maps. Some fields may not convey meaning in the ordering of these data types, functions to assure that these types remain separate.</t>
but it should still be exposed so that applications which need to use it will ha <t>The serialization algorithm is defined in a way that it is not
ve it available.</t> strictly limited to the data types defined in <xref target="types"/> in ev
ery case. For example, Decimals are designed to
<t>Likewise, implementations should note that it’s important to preserve the dis take broader input and round to allowed values.</t>
tinction between Tokens and Strings. While most programming languages have nativ <t>Implementations are allowed to limit the size of different
e types that map to the other types well, it may be necessary to create a wrappe structures, subject to the minimums defined for each type. When a
r “token” object or use a parameter on functions to assure that these types rema structure exceeds an implementation limit, that structure fails parsing
in separate.</t> or serialization.</t>
</section>
<t>The serialization algorithm is defined in a way that it is not strictly limit <section numbered="false" anchor="acknowledgements">
ed to the data types defined in <xref target="types"/> in every case. For exampl <name>Acknowledgements</name>
e, Decimals are designed to take broader input and round to allowed values.</t> <t>Many thanks to <contact fullname="Matthew Kerwin"/> for his detailed fe
edback and careful
<t>Implementations are allowed to limit the allowed size of different structures consideration during the development of this specification.</t>
, subject to the minimums defined for each type. When a structure exceeds an imp <t>Thanks also to <contact fullname="Ian Clelland"/>, <contact fullname="R
lementation limit, that structure fails parsing or serialisation.</t> oy Fielding"/>, <contact fullname="Anne van Kesteren"/>,
<contact fullname="Kazuho Oku"/>, <contact fullname="Evert Pot"/>,
</section> <contact fullname="Julian Reschke"/>, <contact fullname="Martin Thom
<section anchor="changes" title="Changes"> son"/>, <contact fullname="Mike West"/>, and <contact fullname="Jeffrey Yasskin"
/> for their contributions.</t>
<t><spanx style="emph">RFC Editor: Please remove this section before publication </section>
.</spanx></t>
<section anchor="since-draft-ietf-httpbis-header-structure-18" title="Since draf
t-ietf-httpbis-header-structure-18">
<t><list style="symbols">
<t>Use “sf-“ prefix for ABNF, not “sh-“.</t>
<t>Fix indentation in Dictionary serialisation (#1164).</t>
<t>Add example for Token; tweak example field names (#1147).</t>
<t>Editorial improvements.</t>
<t>Note that exceeding implementation limits implies failure.</t>
<t>Talk about specifying order of Dictionary members and Parameters, not cardi
nality.</t>
<t>Allow (but don’t require) parsers to fail when a single field line isn’t va
lid.</t>
<t>Note that some aspects of Integers and Decimals are not necessarily preserv
ed.</t>
<t>Allow Lists and Dictionaries to be delimited by OWS, rather than *SP, to ma
ke parsing more robust.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-17" title="Since draf
t-ietf-httpbis-header-structure-17">
<t><list style="symbols">
<t>Editorial improvements.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-16" title="Since draf
t-ietf-httpbis-header-structure-16">
<t><list style="symbols">
<t>Editorial improvements.</t>
<t>Discussion on forwards compatibility.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-15" title="Since draf
t-ietf-httpbis-header-structure-15">
<t><list style="symbols">
<t>Editorial improvements.</t>
<t>Use HTTP field terminology more consistently, in line with recent changes t
o HTTP-core.</t>
<t>String length requirements apply to decoded strings (#1051).</t>
<t>Correctly round decimals in serialisation (#1043).</t>
<t>Clarify input to serialisation algorithms (#1055).</t>
<t>Omitted True dictionary value can have parameters (#1083).</t>
<t>Keys can now start with ‘*’ (#1068).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-14" title="Since draf
t-ietf-httpbis-header-structure-14">
<t><list style="symbols">
<t>Editorial improvements.</t>
<t>Allow empty dictionary values (#992).</t>
<t>Change value of omitted parameter value to True (#995).</t>
<t>Explain more about splitting dictionaries and lists across header instances
(#997).</t>
<t>Disallow HTAB, replace OWS with spaces (#998).</t>
<t>Change byte sequence delimiters from “*” to “:” (#991).</t>
<t>Allow tokens to start with “*” (#991).</t>
<t>Change Floats to fixed-precision Decimals (#982).</t>
<t>Round the fractional component of decimal, rather than truncating it (#982)
.</t>
<t>Handle duplicate dictionary and parameter keys by overwriting their values,
rather than failing (#997).</t>
<t>Allow “.” in key (#1027).</t>
<t>Check first character of key in serialisation (#1037).</t>
<t>Talk about greasing headers (#1015).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-13" title="Since draf
t-ietf-httpbis-header-structure-13">
<t><list style="symbols">
<t>Editorial improvements.</t>
<t>Define “structured header name” and “structured header value” terms (#908).
</t>
<t>Corrected text about valid characters in strings (#931).</t>
<t>Removed most instances of the word “textual”, as it was redundant (#915).</
t>
<t>Allowed parameters on Items and Inner Lists (#907).</t>
<t>Expand the range of characters in token (#961).</t>
<t>Disallow OWS before “;” delimiter in parameters (#961).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-12" title="Since draf
t-ietf-httpbis-header-structure-12">
<t><list style="symbols">
<t>Editorial improvements.</t>
<t>Reworked float serialisation (#896).</t>
<t>Don’t add a trailing space in inner-list (#904).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-11" title="Since draf
t-ietf-httpbis-header-structure-11">
<t><list style="symbols">
<t>Allow * in key (#844).</t>
<t>Constrain floats to six digits of precision (#848).</t>
<t>Allow dictionary members to have parameters (#842).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-10" title="Since draf
t-ietf-httpbis-header-structure-10">
<t><list style="symbols">
<t>Update abstract (#799).</t>
<t>Input and output are now arrays of bytes (#662).</t>
<t>Implementations need to preserve difference between token and string (#790)
.</t>
<t>Allow empty dictionaries and lists (#781).</t>
<t>Change parameterized lists to have primary items (#797).</t>
<t>Allow inner lists in both dictionaries and lists; removes lists of lists (#
816).</t>
<t>Subsume Parameterised Lists into Lists (#839).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-09" title="Since draf
t-ietf-httpbis-header-structure-09">
<t><list style="symbols">
<t>Changed Boolean from T/F to 1/0 (#784).</t>
<t>Parameters are now ordered maps (#765).</t>
<t>Clamp integers to 15 digits (#737).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-08" title="Since draf
t-ietf-httpbis-header-structure-08">
<t><list style="symbols">
<t>Disallow whitespace before items properly (#703).</t>
<t>Created “key” for use in dictionaries and parameters, rather than relying o
n identifier (#702). Identifiers have a separate minimum supported size.</t>
<t>Expanded the range of special characters allowed in identifier to include a
ll of ALPHA, “.”, “:”, and “%” (#702).</t>
<t>Use “?” instead of “!” to indicate a Boolean (#719).</t>
<t>Added “Intentionally Strict Processing” (#684).</t>
<t>Gave better names for referring specs to use in Parameterised Lists (#720).
</t>
<t>Added Lists of Lists (#721).</t>
<t>Rename Identifier to Token (#725).</t>
<t>Add implementation guidance (#727).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-07" title="Since draf
t-ietf-httpbis-header-structure-07">
<t><list style="symbols">
<t>Make Dictionaries ordered mappings (#659).</t>
<t>Changed “binary content” to “byte sequence” to align with Infra specificati
on (#671).</t>
<t>Changed “mapping” to “map” for #671.</t>
<t>Don’t fail if byte sequences aren’t “=” padded (#658).</t>
<t>Add Booleans (#683).</t>
<t>Allow identifiers in items again (#629).</t>
<t>Disallowed whitespace before items (#703).</t>
<t>Explain the consequences of splitting a string across multiple headers (#68
6).</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-06" title="Since draf
t-ietf-httpbis-header-structure-06">
<t><list style="symbols">
<t>Add a FAQ.</t>
<t>Allow non-zero pad bits.</t>
<t>Explicitly check for integers that violate constraints.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-05" title="Since draf
t-ietf-httpbis-header-structure-05">
<t><list style="symbols">
<t>Reorganise specification to separate parsing out.</t>
<t>Allow referencing specs to use ABNF.</t>
<t>Define serialisation algorithms.</t>
<t>Refine relationship between ABNF, parsing and serialisation algorithms.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-04" title="Since draf
t-ietf-httpbis-header-structure-04">
<t><list style="symbols">
<t>Remove identifiers from item.</t>
<t>Remove most limits on sizes.</t>
<t>Refine number parsing.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-03" title="Since draf
t-ietf-httpbis-header-structure-03">
<t><list style="symbols">
<t>Strengthen language around failure handling.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-02" title="Since draf
t-ietf-httpbis-header-structure-02">
<t><list style="symbols">
<t>Split Numbers into Integers and Floats.</t>
<t>Define number parsing.</t>
<t>Tighten up binary parsing and give it an explicit end delimiter.</t>
<t>Clarify that mappings are unordered.</t>
<t>Allow zero-length strings.</t>
<t>Improve string parsing algorithm.</t>
<t>Improve limits in algorithms.</t>
<t>Require parsers to combine header fields before processing.</t>
<t>Throw an error on trailing garbage.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-01" title="Since draf
t-ietf-httpbis-header-structure-01">
<t><list style="symbols">
<t>Replaced with draft-nottingham-structured-headers.</t>
</list></t>
</section>
<section anchor="since-draft-ietf-httpbis-header-structure-00" title="Since draf
t-ietf-httpbis-header-structure-00">
<t><list style="symbols">
<t>Added signed 64bit integer type.</t>
<t>Drop UTF8, and settle on BCP137 ::EmbeddedUnicodeChar for h1-unicode-string
.</t>
<t>Change h1_blob delimiter to “:” since “’” is valid t_char</t>
</list></t>
</section>
</section>
<section numbered="false" anchor="acknowledgements" title="Acknowledgements">
<t>Many thanks to Matthew Kerwin for his detailed feedback and careful considera
tion during the development of this specification.</t>
<t>Thanks also to Ian Clelland, Roy Fielding, Anne van Kesteren, Kazuho Oku, Eve
rt Pot, Julian Reschke, Martin Thomson, Mike West, and Jeffrey Yasskin for their
contributions.</t>
</section>
</back> </back>
<!-- ##markdown-source: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</rfc> </rfc>
 End of changes. 51 change blocks. 
2313 lines changed or deleted 1512 lines changed or added

This html diff was produced by rfcdiff 1.48. The latest version is available from http://tools.ietf.org/tools/rfcdiff/