//! Heading (setext) occurs in the [flow][] content type.
//!
//! ## Grammar
//!
//! Heading (setext) forms with the following BNF
//! (see [construct][crate::construct] for character groups):
//!
//! ```bnf
//! heading_setext ::= paragraph eol *space_or_tab (1*'-' | 1*'=') *space_or_tab
//!
//! ; See the `paragraph` construct for the BNF of that part.
//! ```
//!
//! As this construct occurs in flow, like all flow constructs, it must be
//! followed by an eol (line ending) or eof (end of file).
//!
//! See [`paragraph`][paragraph] for grammar, notes, and recommendations on
//! that part.
//!
//! In markdown, it is also possible to create headings with a
//! [heading (atx)][heading_atx] construct.
//! The benefit of setext headings is that their text can include line endings,
//! and by extensions also hard breaks (e.g., with
//! [hard break (escape)][hard_break_escape]).
//! However, their limit is that they cannot form `
` through `
`
//! headings.
//!
//! [Thematic breaks][thematic_break] formed with dashes and without whitespace
//! could be interpreted as a heading (setext).
//! Which one forms depends on whether there is text directly in fron of the
//! sequence.
//!
//! > 🏛 **Background**: the word *setext* originates from a small markup
//! > language by Ian Feldman from 1991.
//! > See [*§ Setext* on Wikipedia][wiki_setext] for more info.
//! > The word *atx* originates from a tiny markup language by Aaron Swartz
//! > from 2002.
//! > See [*§ atx, the true structured text format* on `aaronsw.com`][atx] for
//! > more info.
//!
//! ## HTML
//!
//! Heading (setext) in markdown relates to the `
` and `
` elements in
//! HTML.
//! See [*§ 4.3.6 The `h1`, `h2`, `h3`, `h4`, `h5`, and `h6` elements* in the
//! HTML spec][html] for more info.
//!
//! ## Recommendation
//!
//! Always use heading (atx), never heading (setext).
//!
//! ## Tokens
//!
//! * [`HeadingSetext`][Name::HeadingSetext]
//! * [`HeadingSetextText`][Name::HeadingSetextText]
//! * [`HeadingSetextUnderline`][Name::HeadingSetextUnderline]
//! * [`HeadingSetextUnderlineSequence`][Name::HeadingSetextUnderlineSequence]
//!
//! ## References
//!
//! * [`setext-underline.js` in `micromark`](https://github.com/micromark/micromark/blob/main/packages/micromark-core-commonmark/dev/lib/setext-underline.js)
//! * [*§ 4.3 Setext headings* in `CommonMark`](https://spec.commonmark.org/0.30/#setext-headings)
//!
//! [flow]: crate::construct::flow
//! [paragraph]: crate::construct::paragraph
//! [heading_atx]: crate::construct::heading_atx
//! [thematic_break]: crate::construct::thematic_break
//! [hard_break_escape]: crate::construct::hard_break_escape
//! [html]: https://html.spec.whatwg.org/multipage/sections.html#the-h1,-h2,-h3,-h4,-h5,-and-h6-elements
//! [wiki_setext]: https://en.wikipedia.org/wiki/Setext
//! [atx]: http://www.aaronsw.com/2002/atx/
use crate::construct::partial_space_or_tab::{space_or_tab, space_or_tab_min_max};
use crate::event::{Content, Event, Kind, Link, Name};
use crate::resolve::Name as ResolveName;
use crate::state::{Name as StateName, State};
use crate::subtokenize::Subresult;
use crate::tokenizer::Tokenizer;
use crate::util::{constant::TAB_SIZE, skip};
use alloc::{string::String, vec};
/// At start of heading (setext) underline.
///
/// ```markdown
/// | aa
/// > | ==
/// ^
/// ```
pub fn start(tokenizer: &mut Tokenizer) -> State {
if tokenizer.parse_state.options.constructs.heading_setext
&& !tokenizer.lazy
&& !tokenizer.pierce
// Require a paragraph before.
&& (!tokenizer.events.is_empty()
&& tokenizer.events[skip::opt_back(
&tokenizer.events,
tokenizer.events.len() - 1,
&[Name::LineEnding, Name::SpaceOrTab],
)]
.name
== Name::Content)
{
tokenizer.enter(Name::HeadingSetextUnderline);
if matches!(tokenizer.current, Some(b'\t' | b' ')) {
tokenizer.attempt(State::Next(StateName::HeadingSetextBefore), State::Nok);
State::Retry(space_or_tab_min_max(
tokenizer,
0,
if tokenizer.parse_state.options.constructs.code_indented {
TAB_SIZE - 1
} else {
usize::MAX
},
))
} else {
State::Retry(StateName::HeadingSetextBefore)
}
} else {
State::Nok
}
}
/// After optional whitespace, at `-` or `=`.
///
/// ```markdown
/// | aa
/// > | ==
/// ^
/// ```
pub fn before(tokenizer: &mut Tokenizer) -> State {
match tokenizer.current {
Some(b'-' | b'=') => {
tokenizer.tokenize_state.marker = tokenizer.current.unwrap();
tokenizer.enter(Name::HeadingSetextUnderlineSequence);
State::Retry(StateName::HeadingSetextInside)
}
_ => State::Nok,
}
}
/// In sequence.
///
/// ```markdown
/// | aa
/// > | ==
/// ^
/// ```
pub fn inside(tokenizer: &mut Tokenizer) -> State {
if tokenizer.current == Some(tokenizer.tokenize_state.marker) {
tokenizer.consume();
State::Next(StateName::HeadingSetextInside)
} else {
tokenizer.tokenize_state.marker = 0;
tokenizer.exit(Name::HeadingSetextUnderlineSequence);
if matches!(tokenizer.current, Some(b'\t' | b' ')) {
tokenizer.attempt(State::Next(StateName::HeadingSetextAfter), State::Nok);
State::Retry(space_or_tab(tokenizer))
} else {
State::Retry(StateName::HeadingSetextAfter)
}
}
}
/// After sequence, after optional whitespace.
///
/// ```markdown
/// | aa
/// > | ==
/// ^
/// ```
pub fn after(tokenizer: &mut Tokenizer) -> State {
match tokenizer.current {
None | Some(b'\n') => {
// Feel free to interrupt.
tokenizer.interrupt = false;
tokenizer.register_resolver(ResolveName::HeadingSetext);
tokenizer.exit(Name::HeadingSetextUnderline);
State::Ok
}
_ => State::Nok,
}
}
/// Resolve heading (setext).
pub fn resolve(tokenizer: &mut Tokenizer) -> Result