Something went wrong. Try again.
Markdown parser fork with extended syntax for personal use.
Something went wrong. Try again.
38 kB · 1144 lines
Rust
at hack
12345678910111213141516171819202122232425262728293031323334353637383940414243444546474849505152535455565758596061626364656667686970717273747576777879808182838485868788899091929394959697989910010110210310410510610710810911011111211311411511611711811912012112212312412512612712812913013113213313413513613713813914014114214314414514614714814915015115215315415515615715815916016116216316416516616716816917017117217317417517617717817918018118218318418518618718818919019119219319419519619719819920020120220320420520620720820921021121221321421521621721821922022122222322422522622722822923023123223323423523623723823924024124224324424524624724824925025125225325425525625725825926026126226326426526626726826927027127227327427527627727827928028128228328428528628728828929029129229329429529629729829930030130230330430530630730830931031131231331431531631731831932032132232332432532632732832933033133233333433533633733833934034134234334434534634734834935035135235335435535635735835936036136236336436536636736836937037137237337437537637737837938038138238338438538638738838939039139239339439539639739839940040140240340440540640740840941041141241341441541641741841942042142242342442542642742842943043143243343443543643743843944044144244344444544644744844945045145245345445545645745845946046146246346446546646746846947047147247347447547647747847948048148248348448548648748848949049149249349449549649749849950050150250350450550650750850951051151251351451551651751851952052152252352452552652752852953053153253353453553653753853954054154254354454554654754854955055155255355455555655755855956056156256356456556656756856957057157257357457557657757857958058158258358458558658758858959059159259359459559659759859960060160260360460560660760860961061161261361461561661761861962062162262362462562662762862963063163263363463563663763863964064164264364464564664764864965065165265365465565665765865966066166266366466566666766866967067167267367467567667767867968068168268368468568668768868969069169269369469569669769869970070170270370470570670770870971071171271371471571671771871972072172272372472572672772872973073173273373473573673773873974074174274374474574674774874975075175275375475575675775875976076176276376476576676776876977077177277377477577677777877978078178278378478578678778878979079179279379479579679779879980080180280380480580680780880981081181281381481581681781881982082182282382482582682782882983083183283383483583683783883984084184284384484584684784884985085185285385485585685785885986086186286386486586686786886987087187287387487587687787887988088188288388488588688788888989089189289389489589689789889990090190290390490590690790890991091191291391491591691791891992092192292392492592692792892993093193293393493593693793893994094194294394494594694794894995095195295395495595695795895996096196296396496596696796896997097197297397497597697797897998098198298398498598698798898999099199299399499599699799899910001001100210031004100510061007100810091010101110121013101410151016101710181019102010211022102310241025102610271028102910301031103210331034103510361037103810391040104110421043104410451046104710481049105010511052105310541055105610571058105910601061106210631064106510661067106810691070107110721073107410751076107710781079108010811082108310841085108610871088108910901091109210931094109510961097109810991100110111021103110411051106110711081109111011111112111311141115111611171118111911201121112211231124112511261127112811291130113111321133113411351136113711381139114011411142114311441145//! MDX JSX occurs in [MDX JSX (flow)][mdx_jsx_flow] and//! [MDX JSX (text)][mdx_jsx_text].//!//! ## Grammar//!//! MDX JSX forms with the following BNF//! (<small>see [construct][crate::construct] for character groups</small>)://!//! ```bnf//! ; constraint: markdown whitespace (`space_or_tab | eol`) is NOT//! ; allowed directly after `<` in order to allow `1 < 3` in markdown.//! mdx_jsx ::=//! '<' [closing]//! [*whitespace name [attributes_after_identifier] [closing]]//! *whitespace '>'//!//! attributes_after_identifier ::=//! 1*whitespace (attributes_boolean | attributes_value) |//! *whitespace attributes_expression |//! attributes_after_value ::=//! *whitespace (attributes_boolean | attributes_expression | attributes_value)//! attributes_boolean ::= key [attributes_after_identifier]//! ; Note: in gnostic mode the value of the expression must instead be a single valid ES spread//! ; expression//! attributes_expression ::= expression [attributes_after_value]//! attributes_value ::= key initializer [attributes_after_value]//!//! closing ::= *whitespace '/'//!//! name ::= identifier [local | members]//! key ::= identifier [local]//! local ::= *whitespace ':' *whitespace identifier//! members ::= member *member//! member ::= *whitespace '.' *whitespace identifier//!//! identifier ::= identifier_start *identifier_part//! initializer ::= *whitespace '=' *whitespace value//! value ::= double_quoted | single_quoted | expression//! ; Note: in gnostic mode the value must instead be a single valid ES expression//! expression ::= '{' *(expression_text | expression) '}'//!//! double_quoted ::= '"' *double_quoted_text '"'//! single_quoted ::= "'" *single_quoted_text "'"//!//! text ::= char - '<' - '{'//! whitespace ::= es_whitespace//! double_quoted_text ::= char - '"'//! single_quoted_text ::= char - "'"//! expression_text ::= char - '{' - '}'//! identifier_start ::= es_identifier_start//! identifier_part ::= es_identifier_part | '-'//!//! ; ECMAScript//! ; See “Identifier_start”: <https://tc39.es/ecma262/#prod-IdentifierStart>//! es_identifier_start ::= ?//! ; See “Identifier_part”: <https://tc39.es/ecma262/#prod-IdentifierPart>//! es_identifier_part ::= ?//! ; See “Whitespace”: <https://tc39.es/ecma262/#prod-WhiteSpace>//! es_whitespace ::= ?//! ```//!//! The grammar for JSX in markdown is much stricter than that of HTML in//! markdown.//! The primary benefit of this is that tags are parsed into tokens, and thus//! can be processed.//! Another, arguable, benefit of this is that it comes with syntax errors: if//! an author types something that is nonsensical, an error is thrown with//! information about where it happened, what occurred, and what was expected//! instead.//!//! ## Tokens//!//! * [`LineEnding`][Name::LineEnding]//! * [`MdxJsxEsWhitespace`][Name::MdxJsxEsWhitespace]//! * [`MdxJsxTagMarker`][Name::MdxJsxTagMarker]//! * [`MdxJsxTagClosingMarker`][Name::MdxJsxTagClosingMarker]//! * [`MdxJsxTagName`][Name::MdxJsxTagName]//! * [`MdxJsxTagNamePrimary`][Name::MdxJsxTagNamePrimary]//! * [`MdxJsxTagNameMemberMarker`][Name::MdxJsxTagNameMemberMarker]//! * [`MdxJsxTagNamePrefixMarker`][Name::MdxJsxTagNamePrefixMarker]//! * [`MdxJsxTagNameMember`][Name::MdxJsxTagNameMember]//! * [`MdxJsxTagNameLocal`][Name::MdxJsxTagNameLocal]//! * [`MdxJsxTagAttribute`][Name::MdxJsxTagAttribute]//! * [`MdxJsxTagAttributeName`][Name::MdxJsxTagAttributeName]//! * [`MdxJsxTagAttributePrimaryName`][Name::MdxJsxTagAttributePrimaryName]//! * [`MdxJsxTagAttributeNamePrefixMarker`][Name::MdxJsxTagAttributeNamePrefixMarker]//! * [`MdxJsxTagAttributeNameLocal`][Name::MdxJsxTagAttributeNameLocal]//! * [`MdxJsxTagAttributeInitializerMarker`][Name::MdxJsxTagAttributeInitializerMarker]//! * [`MdxJsxTagAttributeValueLiteral`][Name::MdxJsxTagAttributeValueLiteral]//! * [`MdxJsxTagAttributeValueLiteralMarker`][Name::MdxJsxTagAttributeValueLiteralMarker]//! * [`MdxJsxTagAttributeValueLiteralValue`][Name::MdxJsxTagAttributeValueLiteralValue]//! * [`MdxJsxTagSelfClosingMarker`][Name::MdxJsxTagSelfClosingMarker]//!//! ## Recommendation//!//! When authoring markdown with JSX, keep in mind that MDX is a whitespace//! sensitive and line-based language, while JavaScript is insensitive to//! whitespace.//! This affects how markdown and JSX interleave with eachother in MDX.//! For more info on how it works, see [§ Interleaving][interleaving] on the//! MDX site.//!//! ###### Comments inside tags//!//! JavaScript comments in JSX are not supported.//!//! Incorrect://!//! ```jsx//! <hi/*comment!*//>//! <hello// comment!//! />//! ```//!//! Correct://!//! ```jsx//! <hi/>//! <hello//! />//! ```//!//! A PR that adds support for them would be accepted.//!//! ###### Element or fragment attribute values//!//! JSX elements or JSX fragments as attribute values are not supported.//! The reason for this change is that it would be confusing whether markdown//! would work.//!//! Incorrect://!//! ```jsx//! <welcome name=<>Venus</> />//! <welcome name=<span>Pluto</span> />//! ```//!//! Correct://!//! ```jsx//! <welcome name='Mars' />//! <welcome name={<span>Jupiter</span>} />//! ```//!//! ###### Greater than (`>`) and right curly brace (`}`)//!//! JSX does not allow U+003E GREATER THAN (`>`) or U+007D RIGHT CURLY BRACE//! (`}`) literally in text, they need to be encoded as character references//! (or expressions).//! There is no good reason for this (some JSX parsers agree with us and don’t//! crash either).//! Therefore, in MDX, U+003E GREATER THAN (`>`) and U+007D RIGHT CURLY BRACE//! (`}`) are fine literally and don’t need to be encoded.//!//! ## References//!//! * [`jsx-flow.js` in `micromark-extension-mdx-jsx`](https://github.com/micromark/micromark-extension-mdx-jsx/blob/main/dev/lib/jsx-flow.js)//! * [`mdxjs.com`](https://mdxjs.com)//!//! [mdx_jsx_flow]: crate::construct::mdx_jsx_flow//! [mdx_jsx_text]: crate::construct::mdx_jsx_text//! [interleaving]: https://mdxjs.com/docs/what-is-mdx/#interleaving
use crate::event::Name;use crate::message;use crate::state::{Name as StateName, State};use crate::tokenizer::Tokenizer;use crate::util::{ char::{ after_index as char_after_index, format_byte, format_opt as format_char_opt, kind_after_index, Kind as CharacterKind, }, identifier::{id_cont, id_start},};use alloc::{boxed::Box, format};use core::str;
/// Start of MDX: JSX.////// ```markdown/// > | a <B /> c/// ^/// ```pub fn start(tokenizer: &mut Tokenizer) -> State { debug_assert_eq!(tokenizer.current, Some(b'<')); tokenizer.enter(tokenizer.tokenize_state.token_1.clone()); tokenizer.enter(Name::MdxJsxTagMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagMarker); State::Next(StateName::MdxJsxStartAfter)}
/// After `<`.////// ```markdown/// > | a <B /> c/// ^/// ```pub fn start_after(tokenizer: &mut Tokenizer) -> State { // Deviate from JSX, which allows arbitrary whitespace. // See: <https://github.com/micromark/micromark-extension-mdx-jsx/issues/7>. if let Some(b'\t' | b'\n' | b' ') = tokenizer.current { State::Nok } else { tokenizer.attempt(State::Next(StateName::MdxJsxNameBefore), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart) }}
/// Before name, self slash, or end of tag for fragments.////// ```markdown/// > | a <B> c/// ^/// > | a </B> c/// ^/// > | a <> b/// ^/// ```pub fn name_before(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Closing tag. Some(b'/') => { tokenizer.enter(Name::MdxJsxTagClosingMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagClosingMarker); tokenizer.attempt( State::Next(StateName::MdxJsxClosingTagNameBefore), State::Nok, ); State::Next(StateName::MdxJsxEsWhitespaceStart) } // Fragment opening tag. Some(b'>') => State::Retry(StateName::MdxJsxTagEnd), _ => { if id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.enter(Name::MdxJsxTagName); tokenizer.enter(Name::MdxJsxTagNamePrimary); tokenizer.consume(); State::Next(StateName::MdxJsxPrimaryName) } else { crash( tokenizer, "before name", &format!( "a character that can start a name, such as a letter, `$`, or `_`{}", if tokenizer.current == Some(b'!') { " (note: to create a comment in MDX, use `{/* text */}`)" } else { "" } ), ) } } }}
/// Before name of closing tag or end of closing fragment tag.////// ```markdown/// > | a </> b/// ^/// > | a </B> c/// ^/// ```pub fn closing_tag_name_before(tokenizer: &mut Tokenizer) -> State { // Fragment closing tag. if let Some(b'>') = tokenizer.current { State::Retry(StateName::MdxJsxTagEnd) } // Start of a closing tag name. else if id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.enter(Name::MdxJsxTagName); tokenizer.enter(Name::MdxJsxTagNamePrimary); tokenizer.consume(); State::Next(StateName::MdxJsxPrimaryName) } else { crash( tokenizer, "before name", &format!( "a character that can start a name, such as a letter, `$`, or `_`{}", if tokenizer.current == Some(b'*' | b'/') { " (note: JS comments in JSX tags are not supported in MDX)" } else { "" } ), ) }}
/// In primary name.////// ```markdown/// > | a <Bc> d/// ^/// ```pub fn primary_name(tokenizer: &mut Tokenizer) -> State { // End of name. if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace || matches!(tokenizer.current, Some(b'.' | b'/' | b':' | b'>' | b'{')) { tokenizer.exit(Name::MdxJsxTagNamePrimary); tokenizer.attempt(State::Next(StateName::MdxJsxPrimaryNameAfter), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart) } // Continuation of name: remain. // Allow continuation bytes. else if matches!(tokenizer.current, Some(0x80..=0xBF)) || id_cont_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.consume(); State::Next(StateName::MdxJsxPrimaryName) } else { crash( tokenizer, "in name", &format!( "a name character such as letters, digits, `$`, or `_`; whitespace before attributes; or the end of the tag{}", if tokenizer.current == Some(b'@') { " (note: to create a link in MDX, use `[text](url)`)" } else { "" } ), ) }}
/// After primary name.////// ```markdown/// > | a <b.c> d/// ^/// > | a <b:c> d/// ^/// ```pub fn primary_name_after(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Start of a member name. Some(b'.') => { tokenizer.enter(Name::MdxJsxTagNameMemberMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagNameMemberMarker); tokenizer.attempt(State::Next(StateName::MdxJsxMemberNameBefore), State::Nok); State::Next(StateName::MdxJsxEsWhitespaceStart) } // Start of a local name. Some(b':') => { tokenizer.enter(Name::MdxJsxTagNamePrefixMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagNamePrefixMarker); tokenizer.attempt(State::Next(StateName::MdxJsxLocalNameBefore), State::Nok); State::Next(StateName::MdxJsxEsWhitespaceStart) } // End of name. _ => { if matches!(tokenizer.current, Some(b'/' | b'>' | b'{')) || id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.exit(Name::MdxJsxTagName); State::Retry(StateName::MdxJsxAttributeBefore) } else { crash( tokenizer, "after name", "a character that can start an attribute name, such as a letter, `$`, or `_`; whitespace before attributes; or the end of the tag" ) } } }}
/// Before member name.////// ```markdown/// > | a <b.c> d/// ^/// ```pub fn member_name_before(tokenizer: &mut Tokenizer) -> State { // Start of a member name. if id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.enter(Name::MdxJsxTagNameMember); tokenizer.consume(); State::Next(StateName::MdxJsxMemberName) } else { crash( tokenizer, "before member name", "a character that can start an attribute name, such as a letter, `$`, or `_`; whitespace before attributes; or the end of the tag" ) }}
/// In member name.////// ```markdown/// > | a <b.cd> e/// ^/// ```pub fn member_name(tokenizer: &mut Tokenizer) -> State { // End of name. // Note: no `:` allowed here. if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace || matches!(tokenizer.current, Some(b'.' | b'/' | b'>' | b'{')) { tokenizer.exit(Name::MdxJsxTagNameMember); tokenizer.attempt(State::Next(StateName::MdxJsxMemberNameAfter), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart) } // Continuation of name: remain. // Allow continuation bytes. else if matches!(tokenizer.current, Some(0x80..=0xBF)) || id_cont_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.consume(); State::Next(StateName::MdxJsxMemberName) } else { crash( tokenizer, "in member name", &format!( "a name character such as letters, digits, `$`, or `_`; whitespace before attributes; or the end of the tag{}", if tokenizer.current == Some(b'@') { " (note: to create a link in MDX, use `[text](url)`)" } else { "" } ), ) }}
/// After member name.////// ```markdown/// > | a <b.c> d/// ^/// > | a <b.c.d> e/// ^/// ```pub fn member_name_after(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Start of another member name. Some(b'.') => { tokenizer.enter(Name::MdxJsxTagNameMemberMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagNameMemberMarker); tokenizer.attempt(State::Next(StateName::MdxJsxMemberNameBefore), State::Nok); State::Next(StateName::MdxJsxEsWhitespaceStart) } // End of name. _ => { if matches!(tokenizer.current, Some(b'/' | b'>' | b'{')) || id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.exit(Name::MdxJsxTagName); State::Retry(StateName::MdxJsxAttributeBefore) } else { crash( tokenizer, "after member name", "a character that can start an attribute name, such as a letter, `$`, or `_`; whitespace before attributes; or the end of the tag" ) } } }}
/// Local member name.////// ```markdown/// > | a <b:c> d/// ^/// ```pub fn local_name_before(tokenizer: &mut Tokenizer) -> State { // Start of a local name. if id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.enter(Name::MdxJsxTagNameLocal); tokenizer.consume(); State::Next(StateName::MdxJsxLocalName) } else { crash( tokenizer, "before local name", &format!( "a character that can start a name, such as a letter, `$`, or `_`{}", if matches!(tokenizer.current, Some(b'+' | b'/'..=b'9')) { " (note: to create a link in MDX, use `[text](url)`)" } else { "" } ), ) }}
/// In local name.////// ```markdown/// > | a <b:cd> e/// ^/// ```pub fn local_name(tokenizer: &mut Tokenizer) -> State { // End of local name (note that we don’t expect another colon, or a member). if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace || matches!(tokenizer.current, Some(b'/' | b'>' | b'{')) { tokenizer.exit(Name::MdxJsxTagNameLocal); tokenizer.attempt(State::Next(StateName::MdxJsxLocalNameAfter), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart) } // Continuation of name: remain. // Allow continuation bytes. else if matches!(tokenizer.current, Some(0x80..=0xBF)) || id_cont_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.consume(); State::Next(StateName::MdxJsxLocalName) } else { crash( tokenizer, "in local name", "a name character such as letters, digits, `$`, or `_`; whitespace before attributes; or the end of the tag" ) }}
/// After local name.////// This is like as `primary_name_after`, but we don’t expect colons or/// periods.////// ```markdown/// > | a <b.c> d/// ^/// > | a <b.c.d> e/// ^/// ```pub fn local_name_after(tokenizer: &mut Tokenizer) -> State { // End of name. if matches!(tokenizer.current, Some(b'/' | b'>' | b'{')) || id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.exit(Name::MdxJsxTagName); State::Retry(StateName::MdxJsxAttributeBefore) } else { crash( tokenizer, "after local name", "a character that can start an attribute name, such as a letter, `$`, or `_`; whitespace before attributes; or the end of the tag" ) }}
/// Before attribute.////// ```markdown/// > | a <b /> c/// ^/// > | a <b > c/// ^/// > | a <b {...c}> d/// ^/// > | a <b c> d/// ^/// ```pub fn attribute_before(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Self-closing. Some(b'/') => { tokenizer.enter(Name::MdxJsxTagSelfClosingMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagSelfClosingMarker); tokenizer.attempt(State::Next(StateName::MdxJsxSelfClosing), State::Nok); State::Next(StateName::MdxJsxEsWhitespaceStart) } // End of tag. Some(b'>') => State::Retry(StateName::MdxJsxTagEnd), // Attribute expression. Some(b'{') => { tokenizer.tokenize_state.token_2 = tokenizer.tokenize_state.token_1.clone(); tokenizer.tokenize_state.token_1 = Name::MdxJsxTagAttributeExpression; tokenizer.attempt( State::Next(StateName::MdxJsxAttributeExpressionAfter), State::Nok, ); State::Retry(StateName::MdxExpressionStart) } _ => { // Start of an attribute name. if id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.enter(Name::MdxJsxTagAttribute); tokenizer.enter(Name::MdxJsxTagAttributeName); tokenizer.enter(Name::MdxJsxTagAttributePrimaryName); tokenizer.consume(); State::Next(StateName::MdxJsxAttributePrimaryName) } else { crash( tokenizer, "before attribute name", "a character that can start an attribute name, such as a letter, `$`, or `_`; whitespace before attributes; or the end of the tag" ) } } }}
/// After attribute expression.////// ```markdown/// > | a <b {c} d/> e/// ^/// ```pub fn attribute_expression_after(tokenizer: &mut Tokenizer) -> State { tokenizer.tokenize_state.token_1 = tokenizer.tokenize_state.token_2.clone(); tokenizer.tokenize_state.token_2 = Name::Data; tokenizer.attempt(State::Next(StateName::MdxJsxAttributeBefore), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart)}
/// In primary attribute name.////// ```markdown/// > | a <b cd/> e/// ^/// > | a <b c:d> e/// ^/// > | a <b c=d> e/// ^/// ```pub fn attribute_primary_name(tokenizer: &mut Tokenizer) -> State { // End of attribute name or tag. if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace || matches!(tokenizer.current, Some(b'/' | b':' | b'=' | b'>' | b'{')) { tokenizer.exit(Name::MdxJsxTagAttributePrimaryName); tokenizer.attempt( State::Next(StateName::MdxJsxAttributePrimaryNameAfter), State::Nok, ); State::Retry(StateName::MdxJsxEsWhitespaceStart) } // Continuation of name: remain. // Allow continuation bytes. else if matches!(tokenizer.current, Some(0x80..=0xBF)) || id_cont_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.consume(); State::Next(StateName::MdxJsxAttributePrimaryName) } else { crash( tokenizer, "in attribute name", "an attribute name character such as letters, digits, `$`, or `_`; `=` to initialize a value; whitespace before attributes; or the end of the tag" ) }}
/// After primary attribute name.////// ```markdown/// > | a <b c/> d/// ^/// > | a <b c:d> e/// ^/// > | a <b c=d> e/// ^/// ```pub fn attribute_primary_name_after(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Start of a local name. Some(b':') => { tokenizer.enter(Name::MdxJsxTagAttributeNamePrefixMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagAttributeNamePrefixMarker); tokenizer.attempt( State::Next(StateName::MdxJsxAttributeLocalNameBefore), State::Nok, ); State::Next(StateName::MdxJsxEsWhitespaceStart) } // Initializer: start of an attribute value. Some(b'=') => { tokenizer.exit(Name::MdxJsxTagAttributeName); tokenizer.enter(Name::MdxJsxTagAttributeInitializerMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagAttributeInitializerMarker); tokenizer.attempt( State::Next(StateName::MdxJsxAttributeValueBefore), State::Nok, ); State::Retry(StateName::MdxJsxEsWhitespaceStart) } _ => { // End of tag / new attribute. if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace || matches!(tokenizer.current, Some(b'/' | b'>' | b'{')) || id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.exit(Name::MdxJsxTagAttributeName); tokenizer.exit(Name::MdxJsxTagAttribute); tokenizer.attempt(State::Next(StateName::MdxJsxAttributeBefore), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart) } else { crash( tokenizer, "after attribute name", "a character that can start an attribute name, such as a letter, `$`, or `_`; `=` to initialize a value; or the end of the tag" ) } } }}
/// Before local attribute name.////// ```markdown/// > | a <b c:d/> e/// ^/// ```pub fn attribute_local_name_before(tokenizer: &mut Tokenizer) -> State { // Start of a local name. if id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.enter(Name::MdxJsxTagAttributeNameLocal); tokenizer.consume(); State::Next(StateName::MdxJsxAttributeLocalName) } else { crash( tokenizer, "before local attribute name", "a character that can start an attribute name, such as a letter, `$`, or `_`; `=` to initialize a value; or the end of the tag" ) }}
/// In local attribute name.////// ```markdown/// > | a <b c:de/> f/// ^/// > | a <b c:d=e/> f/// ^/// ```pub fn attribute_local_name(tokenizer: &mut Tokenizer) -> State { // End of local name (note that we don’t expect another colon). if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace || matches!(tokenizer.current, Some(b'/' | b'=' | b'>' | b'{')) { tokenizer.exit(Name::MdxJsxTagAttributeNameLocal); tokenizer.exit(Name::MdxJsxTagAttributeName); tokenizer.attempt( State::Next(StateName::MdxJsxAttributeLocalNameAfter), State::Nok, ); State::Retry(StateName::MdxJsxEsWhitespaceStart) } // Continuation of name: remain. // Allow continuation bytes. else if matches!(tokenizer.current, Some(0x80..=0xBF)) || id_cont_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.consume(); State::Next(StateName::MdxJsxAttributeLocalName) } else { crash( tokenizer, "in local attribute name", "an attribute name character such as letters, digits, `$`, or `_`; `=` to initialize a value; whitespace before attributes; or the end of the tag" ) }}
/// After local attribute name.////// ```markdown/// > | a <b c:d/> f/// ^/// > | a <b c:d=e/> f/// ^/// ```pub fn attribute_local_name_after(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Start of an attribute value. Some(b'=') => { tokenizer.enter(Name::MdxJsxTagAttributeInitializerMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagAttributeInitializerMarker); tokenizer.attempt( State::Next(StateName::MdxJsxAttributeValueBefore), State::Nok, ); State::Next(StateName::MdxJsxEsWhitespaceStart) } _ => { // End of name. if matches!(tokenizer.current, Some(b'/' | b'>' | b'{')) || id_start_opt(char_after_index( tokenizer.parse_state.bytes, tokenizer.point.index, )) { tokenizer.exit(Name::MdxJsxTagAttribute); State::Retry(StateName::MdxJsxAttributeBefore) } else { crash( tokenizer, "after local attribute name", "a character that can start an attribute name, such as a letter, `$`, or `_`; `=` to initialize a value; or the end of the tag" ) } } }}
/// After `=`, before value.////// ```markdown/// > | a <b c="d"/> e/// ^/// > | a <b c={d}/> e/// ^/// ```pub fn attribute_value_before(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { // Start of double- or single quoted value. Some(b'"' | b'\'') => { tokenizer.tokenize_state.marker = tokenizer.current.unwrap(); tokenizer.enter(Name::MdxJsxTagAttributeValueLiteral); tokenizer.enter(Name::MdxJsxTagAttributeValueLiteralMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagAttributeValueLiteralMarker); State::Next(StateName::MdxJsxAttributeValueQuotedStart) } // Attribute value expression. Some(b'{') => { tokenizer.tokenize_state.token_2 = tokenizer.tokenize_state.token_1.clone(); tokenizer.tokenize_state.token_1 = Name::MdxJsxTagAttributeValueExpression; tokenizer.attempt( State::Next(StateName::MdxJsxAttributeValueExpressionAfter), State::Nok, ); State::Retry(StateName::MdxExpressionStart) } _ => crash( tokenizer, "before attribute value", &format!( "a character that can start an attribute value, such as `\"`, `'`, or `{{`{}", if tokenizer.current == Some(b'<') { " (note: to use an element or fragment as a prop value in MDX, use `{<element />}`)" } else { "" } ), ), }}
/// After attribute value expression.////// ```markdown/// > | a <b c={d} e/> f/// ^/// ```pub fn attribute_value_expression_after(tokenizer: &mut Tokenizer) -> State { tokenizer.tokenize_state.token_1 = tokenizer.tokenize_state.token_2.clone(); tokenizer.tokenize_state.token_2 = Name::Data; tokenizer.exit(Name::MdxJsxTagAttribute); tokenizer.attempt(State::Next(StateName::MdxJsxAttributeBefore), State::Nok); State::Retry(StateName::MdxJsxEsWhitespaceStart)}
/// Before quoted literal attribute value.////// ```markdown/// > | a <b c="d"/> e/// ^/// ```pub fn attribute_value_quoted_start(tokenizer: &mut Tokenizer) -> State { if let Some(byte) = tokenizer.current { if byte == tokenizer.tokenize_state.marker { tokenizer.tokenize_state.marker = 0; tokenizer.enter(Name::MdxJsxTagAttributeValueLiteralMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagAttributeValueLiteralMarker); tokenizer.exit(Name::MdxJsxTagAttributeValueLiteral); tokenizer.exit(Name::MdxJsxTagAttribute); tokenizer.attempt(State::Next(StateName::MdxJsxAttributeBefore), State::Nok); State::Next(StateName::MdxJsxEsWhitespaceStart) } else if byte == b'\n' { tokenizer.attempt( State::Next(StateName::MdxJsxAttributeValueQuotedStart), State::Nok, ); State::Retry(StateName::MdxJsxEsWhitespaceStart) } else { tokenizer.enter(Name::MdxJsxTagAttributeValueLiteralValue); State::Retry(StateName::MdxJsxAttributeValueQuoted) } } else { crash( tokenizer, "in attribute value", &format!( "a corresponding closing quote {}", format_byte(tokenizer.tokenize_state.marker) ), ) }}
/// In quoted literal attribute value.////// ```markdown/// > | a <b c="d"/> e/// ^/// ```pub fn attribute_value_quoted(tokenizer: &mut Tokenizer) -> State { if tokenizer.current == Some(tokenizer.tokenize_state.marker) || matches!(tokenizer.current, None | Some(b'\n')) { tokenizer.exit(Name::MdxJsxTagAttributeValueLiteralValue); State::Retry(StateName::MdxJsxAttributeValueQuotedStart) } else { tokenizer.consume(); State::Next(StateName::MdxJsxAttributeValueQuoted) }}
/// After self-closing slash.////// ```markdown/// > | a <b/> c/// ^/// ```pub fn self_closing(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { Some(b'>') => State::Retry(StateName::MdxJsxTagEnd), _ => crash( tokenizer, "after self-closing slash", &format!( "`>` to end the tag{}", if tokenizer.current == Some(b'*' | b'/') { " (note: JS comments in JSX tags are not supported in MDX)" } else { "" } ), ), }}
/// At final `>`.////// ```markdown/// > | a <b> c/// ^/// ```pub fn tag_end(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { Some(b'>') => { tokenizer.enter(Name::MdxJsxTagMarker); tokenizer.consume(); tokenizer.exit(Name::MdxJsxTagMarker); tokenizer.exit(tokenizer.tokenize_state.token_1.clone()); State::Ok } _ => unreachable!("expected `>`"), }}
/// Before optional ECMAScript whitespace.////// ```markdown/// > | a <a b> c/// ^/// ```pub fn es_whitespace_start(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { Some(b'\n') => { tokenizer.enter(Name::LineEnding); tokenizer.consume(); tokenizer.exit(Name::LineEnding); State::Next(StateName::MdxJsxEsWhitespaceEolAfter) } _ => { if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace { tokenizer.enter(Name::MdxJsxEsWhitespace); State::Retry(StateName::MdxJsxEsWhitespaceInside) } else { State::Ok } } }}
/// In ECMAScript whitespace.////// ```markdown/// > | a <a b> c/// ^/// ```pub fn es_whitespace_inside(tokenizer: &mut Tokenizer) -> State { match tokenizer.current { Some(b'\n') => { tokenizer.exit(Name::MdxJsxEsWhitespace); State::Retry(StateName::MdxJsxEsWhitespaceStart) } // Allow continuation bytes. Some(0x80..=0xBF) => { tokenizer.consume(); State::Next(StateName::MdxJsxEsWhitespaceInside) } Some(_) if kind_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) == CharacterKind::Whitespace => { tokenizer.consume(); State::Next(StateName::MdxJsxEsWhitespaceInside) } Some(_) => { tokenizer.exit(Name::MdxJsxEsWhitespace); State::Ok } // Handle EOF. None => State::Nok, }}
/// After eol in whitespace.////// ```markdown/// > | a <a\nb> c/// ^/// ```pub fn es_whitespace_eol_after(tokenizer: &mut Tokenizer) -> State { // Lazy continuation in a flow tag is a syntax error. if tokenizer.tokenize_state.token_1 == Name::MdxJsxFlowTag && tokenizer.lazy { State::Error( message::Message { place: Some(Box::new(message::Place::Point(tokenizer.point.to_unist()))), reason: "Unexpected lazy line in jsx in container, expected line to be prefixed with `>` when in a block quote, whitespace when in a list, etc".into(), rule_id: Box::new("unexpected-lazy".into()), source: Box::new("markdown-rs".into()), } ) } else { State::Retry(StateName::MdxJsxEsWhitespaceStart) }}
/// Check if a character can start a JSX identifier.fn id_start_opt(code: Option<char>) -> bool { code.map_or(false, id_start)}
/// Check if a character can continue a JSX identifier.fn id_cont_opt(code: Option<char>) -> bool { code.map_or(false, |c| id_cont(c, true))}
/// Crash because something happened `at`, with info on what was `expect`ed/// instead.fn crash(tokenizer: &Tokenizer, at: &str, expect: &str) -> State { State::Error(message::Message { place: Some(Box::new(message::Place::Point(tokenizer.point.to_unist()))), reason: format!( "Unexpected {} {}, expected {}", format_char_opt(if tokenizer.current.is_none() { None } else { char_after_index(tokenizer.parse_state.bytes, tokenizer.point.index) }), at, expect ), rule_id: Box::new(format!( "unexpected-{}", if tokenizer.current.is_none() { "eof" } else { "character" } )), source: Box::new("markdown-rs".into()), })}