aboutsummaryrefslogtreecommitdiffstats
path: root/src
diff options
context:
space:
mode:
Diffstat (limited to '')
-rw-r--r--src/compiler.rs29
-rw-r--r--src/construct/block_quote.rs58
-rw-r--r--src/construct/heading_setext.rs27
-rw-r--r--src/construct/mod.rs1
-rw-r--r--src/construct/paragraph.rs28
-rw-r--r--src/content/document.rs439
-rw-r--r--src/content/flow.rs49
-rw-r--r--src/content/mod.rs6
-rw-r--r--src/parser.rs4
-rw-r--r--src/tokenizer.rs38
-rw-r--r--src/util/edit_map.rs1
-rw-r--r--src/util/mod.rs1
-rw-r--r--src/util/skip.rs44
13 files changed, 645 insertions, 80 deletions
diff --git a/src/compiler.rs b/src/compiler.rs
index 7e47f95..f27c0de 100644
--- a/src/compiler.rs
+++ b/src/compiler.rs
@@ -193,9 +193,7 @@ pub struct Options {
/// // micromark is safe by default:
/// assert_eq!(
/// micromark("> a"),
- /// // To do: block quote
- /// // "<blockquote>\n<p>a</p>\n</blockquote>"
- /// "<p>&gt; a</p>"
+ /// "<blockquote>\n<p>a</p>\n</blockquote>"
/// );
///
/// // Define `default_line_ending` to configure the default:
@@ -209,9 +207,7 @@ pub struct Options {
///
/// }
/// ),
- /// // To do: block quote
- /// // "<blockquote>\r\n<p>a</p>\r\n</blockquote>"
- /// "<p>&gt; a</p>"
+ /// "<blockquote>\r\n<p>a</p>\r\n</blockquote>"
/// );
/// ```
pub default_line_ending: Option<LineEnding>,
@@ -418,6 +414,7 @@ pub fn compile(events: &[Event], codes: &[Code], options: &Options) -> String {
enter_map.insert(TokenType::HeadingSetextText, on_enter_buffer);
enter_map.insert(TokenType::Label, on_enter_buffer);
enter_map.insert(TokenType::ResourceTitleString, on_enter_buffer);
+ enter_map.insert(TokenType::BlockQuote, on_enter_block_quote);
enter_map.insert(TokenType::CodeIndented, on_enter_code_indented);
enter_map.insert(TokenType::CodeFenced, on_enter_code_fenced);
enter_map.insert(TokenType::CodeText, on_enter_code_text);
@@ -491,6 +488,7 @@ pub fn compile(events: &[Event], codes: &[Code], options: &Options) -> String {
exit_map.insert(TokenType::CodeFlowChunk, on_exit_code_flow_chunk);
exit_map.insert(TokenType::CodeText, on_exit_code_text);
exit_map.insert(TokenType::CodeTextLineEnding, on_exit_code_text_line_ending);
+ exit_map.insert(TokenType::BlockQuote, on_exit_block_quote);
exit_map.insert(TokenType::HardBreakEscape, on_exit_break);
exit_map.insert(TokenType::HardBreakTrailing, on_exit_break);
exit_map.insert(TokenType::HeadingAtx, on_exit_heading_atx);
@@ -607,6 +605,13 @@ fn on_enter_buffer(context: &mut CompileContext) {
context.buffer();
}
+/// Handle [`Enter`][EventType::Enter]:[`BlockQuote`][TokenType::BlockQuote].
+fn on_enter_block_quote(context: &mut CompileContext) {
+ // tightStack.push(false)
+ context.line_ending_if_needed();
+ context.tag("<blockquote>".to_string());
+}
+
/// Handle [`Enter`][EventType::Enter]:[`CodeIndented`][TokenType::CodeIndented].
fn on_enter_code_indented(context: &mut CompileContext) {
context.code_flow_seen_data = Some(false);
@@ -695,6 +700,7 @@ fn on_enter_link(context: &mut CompileContext) {
/// Handle [`Enter`][EventType::Enter]:[`Paragraph`][TokenType::Paragraph].
fn on_enter_paragraph(context: &mut CompileContext) {
+ context.line_ending_if_needed();
context.tag("<p>".to_string());
}
@@ -756,6 +762,14 @@ fn on_exit_break(context: &mut CompileContext) {
context.tag("<br />".to_string());
}
+/// Handle [`Exit`][EventType::Exit]:[`BlockQuote`][TokenType::BlockQuote].
+fn on_exit_block_quote(context: &mut CompileContext) {
+ // tightStack.pop()
+ context.line_ending_if_needed();
+ context.tag("</blockquote>".to_string());
+ // let mut slurp_all_line_endings = false;
+}
+
/// Handle [`Exit`][EventType::Exit]:[`CharacterReferenceMarker`][TokenType::CharacterReferenceMarker].
fn on_exit_character_reference_marker(context: &mut CompileContext) {
context.character_reference_kind = Some(CharacterReferenceKind::Named);
@@ -971,6 +985,7 @@ fn on_exit_heading_atx_sequence(context: &mut CompileContext) {
false,
)
.len();
+ context.line_ending_if_needed();
context.atx_opening_sequence_size = Some(rank);
context.tag(format!("<h{}>", rank));
}
@@ -1001,6 +1016,7 @@ fn on_exit_heading_setext_underline(context: &mut CompileContext) {
)[0];
let level: usize = if head == Code::Char('-') { 2 } else { 1 };
+ context.line_ending_if_needed();
context.tag(format!("<h{}>", level));
context.push(text);
context.tag(format!("</h{}>", level));
@@ -1157,5 +1173,6 @@ fn on_exit_strong(context: &mut CompileContext) {
/// Handle [`Exit`][EventType::Exit]:[`ThematicBreak`][TokenType::ThematicBreak].
fn on_exit_thematic_break(context: &mut CompileContext) {
+ context.line_ending_if_needed();
context.tag("<hr />".to_string());
}
diff --git a/src/construct/block_quote.rs b/src/construct/block_quote.rs
new file mode 100644
index 0000000..cd5b872
--- /dev/null
+++ b/src/construct/block_quote.rs
@@ -0,0 +1,58 @@
+//! To do.
+
+use crate::constant::TAB_SIZE;
+use crate::construct::partial_space_or_tab::space_or_tab_min_max;
+use crate::tokenizer::{Code, State, StateFnResult, TokenType, Tokenizer};
+
+pub fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+ // To do: allow arbitrary when code (indented) is turned off.
+ tokenizer.go(space_or_tab_min_max(0, TAB_SIZE - 1), before)(tokenizer, code)
+}
+
+fn before(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+ match code {
+ Code::Char('>') => {
+ tokenizer.enter(TokenType::BlockQuote);
+ cont_before(tokenizer, code)
+ }
+ _ => cont_before(tokenizer, code),
+ }
+}
+
+pub fn cont(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+ // To do: allow arbitrary when code (indented) is turned off.
+ tokenizer.go(space_or_tab_min_max(0, TAB_SIZE - 1), cont_before)(tokenizer, code)
+}
+
+fn cont_before(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+ match code {
+ Code::Char('>') => {
+ tokenizer.enter(TokenType::BlockQuotePrefix);
+ tokenizer.enter(TokenType::BlockQuoteMarker);
+ tokenizer.consume(code);
+ tokenizer.exit(TokenType::BlockQuoteMarker);
+ (State::Fn(Box::new(cont_after)), None)
+ }
+ _ => (State::Nok, None),
+ }
+}
+
+fn cont_after(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+ match code {
+ Code::VirtualSpace | Code::Char('\t' | ' ') => {
+ tokenizer.enter(TokenType::BlockQuotePrefixWhitespace);
+ tokenizer.consume(code);
+ tokenizer.exit(TokenType::BlockQuotePrefixWhitespace);
+ tokenizer.exit(TokenType::BlockQuotePrefix);
+ (State::Ok, None)
+ }
+ _ => {
+ tokenizer.exit(TokenType::BlockQuotePrefix);
+ (State::Ok, Some(vec![code]))
+ }
+ }
+}
+
+pub fn end() -> Vec<TokenType> {
+ vec![TokenType::BlockQuote]
+}
diff --git a/src/construct/heading_setext.rs b/src/construct/heading_setext.rs
index 211434f..440baa8 100644
--- a/src/construct/heading_setext.rs
+++ b/src/construct/heading_setext.rs
@@ -60,7 +60,7 @@
use crate::constant::TAB_SIZE;
use crate::construct::partial_space_or_tab::{space_or_tab, space_or_tab_min_max};
use crate::tokenizer::{Code, Event, EventType, State, StateFnResult, TokenType, Tokenizer};
-use crate::util::edit_map::EditMap;
+use crate::util::{edit_map::EditMap, skip::opt_back as skip_opt_back};
/// Kind of underline.
#[derive(Debug, Clone, PartialEq)]
@@ -116,11 +116,26 @@ impl Kind {
/// ```
pub fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
let index = tokenizer.events.len();
- let paragraph_before = index > 3
- && tokenizer.events[index - 1].token_type == TokenType::LineEnding
- && tokenizer.events[index - 3].token_type == TokenType::Paragraph;
-
- if paragraph_before {
+ let previous = if index > 1 {
+ skip_opt_back(
+ &tokenizer.events,
+ index - 1,
+ &[TokenType::SpaceOrTab, TokenType::BlockQuotePrefix],
+ )
+ } else {
+ 0
+ };
+ let previous = skip_opt_back(&tokenizer.events, previous, &[TokenType::LineEnding]);
+ let paragraph_before =
+ previous > 1 && tokenizer.events[previous].token_type == TokenType::Paragraph;
+
+ println!(
+ "setext-start: {:?} {:?} {:?}",
+ tokenizer.interrupt, tokenizer.lazy, paragraph_before
+ );
+
+ // Require a paragraph before and do not allow on a lazy line.
+ if paragraph_before && !tokenizer.lazy {
// To do: allow arbitrary when code (indented) is turned off.
tokenizer.go(space_or_tab_min_max(0, TAB_SIZE - 1), before)(tokenizer, code)
} else {
diff --git a/src/construct/mod.rs b/src/construct/mod.rs
index 66b2a3c..936ecf6 100644
--- a/src/construct/mod.rs
+++ b/src/construct/mod.rs
@@ -64,6 +64,7 @@
pub mod attention;
pub mod autolink;
pub mod blank_line;
+pub mod block_quote;
pub mod character_escape;
pub mod character_reference;
pub mod code_fenced;
diff --git a/src/construct/paragraph.rs b/src/construct/paragraph.rs
index 4f5e662..ace174f 100644
--- a/src/construct/paragraph.rs
+++ b/src/construct/paragraph.rs
@@ -35,7 +35,7 @@
use crate::tokenizer::{
Code, ContentType, Event, EventType, State, StateFnResult, TokenType, Tokenizer,
};
-use crate::util::edit_map::EditMap;
+use crate::util::{edit_map::EditMap, skip::opt as skip_opt};
/// Before a paragraph.
///
@@ -90,19 +90,27 @@ pub fn resolve(tokenizer: &mut Tokenizer) -> Vec<Event> {
if event.event_type == EventType::Enter && event.token_type == TokenType::Paragraph {
// Exit:Paragraph
let mut exit_index = index + 3;
+ let mut enter_next_index =
+ skip_opt(&tokenizer.events, exit_index + 1, &[TokenType::LineEnding]);
// Enter:Paragraph
- let mut enter_next_index = exit_index + 3;
+ enter_next_index = skip_opt(
+ &tokenizer.events,
+ enter_next_index,
+ &[TokenType::SpaceOrTab, TokenType::BlockQuotePrefix],
+ );
// Find future `Paragraphs`.
- // There will be `LineEnding` between.
- while enter_next_index < len
+ while enter_next_index < tokenizer.events.len()
&& tokenizer.events[enter_next_index].token_type == TokenType::Paragraph
{
// Remove Exit:Paragraph, Enter:LineEnding, Exit:LineEnding, Enter:Paragraph.
- edit_map.add(exit_index, 4, vec![]);
+ edit_map.add(exit_index, 3, vec![]);
+
+ // Remove Enter:Paragraph.
+ edit_map.add(enter_next_index, 1, vec![]);
// Add Exit:LineEnding position info to Exit:Data.
- let line_ending_exit = &tokenizer.events[enter_next_index - 1];
+ let line_ending_exit = &tokenizer.events[exit_index + 2];
let line_ending_point = line_ending_exit.point.clone();
let line_ending_index = line_ending_exit.index;
let data_exit = &mut tokenizer.events[exit_index - 1];
@@ -117,7 +125,13 @@ pub fn resolve(tokenizer: &mut Tokenizer) -> Vec<Event> {
// Potential next start.
exit_index = enter_next_index + 3;
- enter_next_index = exit_index + 3;
+ enter_next_index =
+ skip_opt(&tokenizer.events, exit_index + 1, &[TokenType::LineEnding]);
+ enter_next_index = skip_opt(
+ &tokenizer.events,
+ enter_next_index,
+ &[TokenType::SpaceOrTab, TokenType::BlockQuotePrefix],
+ );
}
// Move to `Exit:Paragraph`.
diff --git a/src/content/document.rs b/src/content/document.rs
new file mode 100644
index 0000000..dd5038f
--- /dev/null
+++ b/src/content/document.rs
@@ -0,0 +1,439 @@
+//! The document content type.
+//!
+//! **Document** represents the containers, such as block quotes and lists,
+//! which structure the document and contain other sections.
+//!
+//! The constructs found in flow are:
+//!
+//! * [Block quote][crate::construct::block_quote]
+//! * List
+
+use crate::construct::block_quote::{
+ cont as block_quote_cont, end as block_quote_end, start as block_quote,
+};
+use crate::content::flow::start as flow;
+use crate::parser::ParseState;
+use crate::subtokenize::subtokenize;
+use crate::tokenizer::{
+ Code, Event, EventType, Point, State, StateFn, StateFnResult, TokenType, Tokenizer,
+};
+use crate::util::edit_map::EditMap;
+use crate::util::{
+ normalize_identifier::normalize_identifier,
+ span::{from_exit_event, serialize},
+};
+use std::collections::HashSet;
+
+struct DocumentInfo {
+ continued: usize,
+ stack: Vec<String>,
+ next: Box<StateFn>,
+ last_line_ending_index: Option<usize>,
+ map: EditMap,
+}
+
+/// Turn `codes` as the document content type into events.
+pub fn document(parse_state: &mut ParseState, point: Point, index: usize) -> Vec<Event> {
+ let mut tokenizer = Tokenizer::new(point, index, parse_state);
+
+ tokenizer.push(&parse_state.codes, Box::new(start), true);
+
+ let mut index = 0;
+ let mut next_definitions: HashSet<String> = HashSet::new();
+
+ while index < tokenizer.events.len() {
+ let event = &tokenizer.events[index];
+
+ if event.event_type == EventType::Exit
+ && event.token_type == TokenType::DefinitionLabelString
+ {
+ next_definitions.insert(normalize_identifier(
+ serialize(
+ &parse_state.codes,
+ &from_exit_event(&tokenizer.events, index),
+ false,
+ )
+ .as_str(),
+ ));
+ }
+
+ index += 1;
+ }
+
+ let mut result = (tokenizer.events, false);
+
+ parse_state.definitions = next_definitions;
+
+ while !result.1 {
+ result = subtokenize(result.0, parse_state);
+ }
+
+ result.0
+}
+
+fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+ let info = DocumentInfo {
+ continued: 0,
+ stack: vec![],
+ next: Box::new(flow),
+ last_line_ending_index: None,
+ map: EditMap::new(),
+ };
+ before(tokenizer, code, info)
+}
+
+fn before(tokenizer: &mut Tokenizer, code: Code, info: DocumentInfo) -> StateFnResult {
+ println!("before: check existing open containers");
+ // First we iterate through the open blocks, starting with the root
+ // document, and descending through last children down to the last open
+ // block.
+ // Each block imposes a condition that the line must satisfy if the block
+ // is to remain open.
+ // For example, a block quote requires a `>` character.
+ // A paragraph requires a non-blank line.
+ // In this phase we may match all or just some of the open blocks.
+ // But we cannot close unmatched blocks yet, because we may have a lazy
+ // continuation line.
+ if info.continued < info.stack.len() {
+ let name = &info.stack[info.continued];
+ // To do: list.
+ let cont = if name == "blockquote" {
+ block_quote_cont
+ } else {
+ unreachable!("todo: cont construct {:?}", name)
+ };
+
+ // To do: state?
+
+ tokenizer.attempt(cont, move |ok| {
+ if ok {
+ Box::new(|t, c| document_continue(t, c, info))
+ } else {
+ Box::new(|t, c| check_new_containers(t, c, info))
+ }
+ })(tokenizer, code)
+ } else {
+ // Done.
+ check_new_containers(tokenizer, code, info)
+ }
+}
+
+fn document_continue(
+ tokenizer: &mut Tokenizer,
+ code: Code,
+ mut info: DocumentInfo,
+) -> StateFnResult {
+ println!("document_continue");
+ info.continued += 1;
+
+ println!(" to do: close flow sometimes?");
+ // // Note: this field is called `_closeFlow` but it also closes containers.
+ // // Perhaps a good idea to rename it but it’s already used in the wild by
+ // // extensions.
+ // if (self.containerState._closeFlow) {
+ // self.containerState._closeFlow = undefined
+
+ // if (childFlow) {
+ // closeFlow()
+ // }
+
+ // // Note: this algorithm for moving events around is similar to the
+ // // algorithm when dealing with lazy lines in `writeToChild`.
+ // const indexBeforeExits = self.events.length
+ // let indexBeforeFlow = indexBeforeExits
+ // /** @type {Point|undefined} */
+ // let point
+
+ // // Find the flow chunk.
+ // while (indexBeforeFlow--) {
+ // if (
+ // self.events[indexBeforeFlow][0] === 'exit' &&
+ // self.events[indexBeforeFlow][1].type === types.chunkFlow
+ // ) {
+ // point = self.events[indexBeforeFlow][1].end
+ // break
+ // }
+ // }
+
+ // assert(point, 'could not find previous flow chunk')
+
+ let size = info.continued;
+ exit_containers(tokenizer, &mut info, size);
+
+ // // Fix positions.
+ // let index = indexBeforeExits
+
+ // while (index < self.events.length) {
+ // self.events[index][1].end = Object.assign({}, point)
+ // index++
+ // }
+
+ // // Inject the exits earlier (they’re still also at the end).
+ // splice(
+ // self.events,
+ // indexBeforeFlow + 1,
+ // 0,
+ // self.events.slice(indexBeforeExits)
+ // )
+
+ // // Discard the duplicate exits.
+ // self.events.length = index
+
+ // return checkNewContainers(code)
+ // }
+
+ before(tokenizer, code, info)
+}
+// documentContinue
+
+fn check_new_containers(
+ tokenizer: &mut Tokenizer,
+ code: Code,
+ info: DocumentInfo,
+) -> StateFnResult {
+ println!("check_new_containers");
+ // Next, after consuming the continuation markers for existing blocks, we
+ // look for new block starts (e.g. `>` for a block quote).
+ // If we encounter a new block start, we close any blocks unmatched in
+ // step 1 before creating the new block as a child of the last matched
+ // block.
+ if info.continued == info.stack.len() {
+ println!(" to do: concrete? interrupt?");
+ // // No need to `check` whether there’s a container, of `exitContainers`
+ // // would be moot.
+ // // We can instead immediately `attempt` to parse one.
+ // if (!childFlow) {
+ // return documentContinued(code)
+ // }
+
+ // // If we have concrete content, such as block HTML or fenced code,
+ // // we can’t have containers “pierce” into them, so we can immediately
+ // // start.
+ // if (childFlow.currentConstruct && childFlow.currentConstruct.concrete) {
+ // return flowStart(code)
+ // }
+
+ // // If we do have flow, it could still be a blank line,
+ // // but we’d be interrupting it w/ a new container if there’s a current
+ // // construct.
+ // self.interrupt = Boolean(
+ // childFlow.currentConstruct && !childFlow._gfmTableDynamicInterruptHack
+ // )
+ }
+
+ // Check if there is a new container.
+ // To do: list.
+ tokenizer.attempt(block_quote, move |ok| {
+ if ok {
+ Box::new(|t, c| there_is_a_new_container(t, c, info, "blockquote".to_string()))
+ } else {
+ Box::new(|t, c| there_is_no_new_container(t, c, info))
+ }
+ })(tokenizer, code)
+}
+
+fn there_is_a_new_container(
+ tokenizer: &mut Tokenizer,
+ code: Code,
+ mut info: DocumentInfo,
+ name: String,
+) -> StateFnResult {
+ println!("there_is_a_new_container");
+ println!(" todo: close_flow");
+ // if (childFlow) closeFlow()
+ let size = info.continued;
+ exit_containers(tokenizer, &mut info, size);
+ info.stack.push(name);
+ info.continued += 1;
+ document_continued(tokenizer, code, info)
+}
+
+/// Exit open containers.
+fn exit_containers(tokenizer: &mut Tokenizer, info: &mut DocumentInfo, size: usize) {
+ while info.stack.len() > size {
+ let name = info.stack.pop().unwrap();
+
+ // To do: list.
+ let end = if name == "blockquote" {
+ block_quote_end
+ } else {
+ unreachable!("todo: cont {:?}", name)
+ };
+
+ // To do: improve below code.
+ let insert_index = if let Some(index) = info.last_line_ending_index {
+ index
+ } else {
+ tokenizer.events.len()
+ };
+ let eol_point = if let Some(index) = info.last_line_ending_index {
+ tokenizer.events[index].point.clone()
+ } else {
+ tokenizer.point.clone()
+ };
+ let eol_index = if let Some(index) = info.last_line_ending_index {
+ tokenizer.events[index].index
+ } else {
+ tokenizer.index
+ };
+
+ let token_types = end();
+
+ let mut index = 0;
+ while index < token_types.len() {
+ let token_type = &token_types[index];
+
+ info.map.add(
+ insert_index,
+ 0,
+ vec![Event {
+ event_type: EventType::Exit,
+ token_type: token_type.clone(),
+ point: eol_point.clone(),
+ index: eol_index,
+ previous: None,
+ next: None,
+ content_type: None,
+ }],
+ );
+
+ let mut stack_index = tokenizer.stack.len();
+
+ while stack_index > 0 {
+ stack_index -= 1;
+
+ if tokenizer.stack[stack_index] == *token_type {
+ break;
+ }
+ }
+
+ assert_eq!(
+ tokenizer.stack[stack_index], *token_type,
+ "expected token type"
+ );
+ tokenizer.stack.remove(stack_index);
+
+ index += 1;
+ }
+ }
+}
+
+fn there_is_no_new_container(
+ tokenizer: &mut Tokenizer,
+ code: Code,
+ info: DocumentInfo,
+) -> StateFnResult {
+ let lazy = info.continued != info.stack.len();
+ tokenizer.lazy = lazy;
+ println!("there is no new container");
+ if lazy {
+ println!(
+ " This line will be lazy. Depending on what is parsed now, we need to close containers before?"
+ );
+ }
+ // lineStartOffset = self.now().offset
+ flow_start(tokenizer, code, info)
+}
+
+fn document_continued(tokenizer: &mut Tokenizer, code: Code, info: DocumentInfo) -> StateFnResult {
+ println!("document_continued");
+
+ // Try new containers.
+ // To do: list.
+ tokenizer.attempt(block_quote, |ok| {
+ if ok {
+ Box::new(|t, c| container_continue(t, c, info))
+ } else {
+ Box::new(|t, c| {
+ // To do: this looks like a bug?
+ t.lazy = false;
+ flow_start(t, c, info)
+ })
+ }
+ })(tokenizer, code)
+}
+
+fn container_continue(
+ tokenizer: &mut Tokenizer,
+ code: Code,
+ mut info: DocumentInfo,
+) -> StateFnResult {
+ println!("container_continue");
+ // assert(
+ // self.currentConstruct,
+ // 'expected `currentConstruct` to be defined on tokenizer'
+ // )
+ // assert(
+ // self.containerState,
+ // 'expected `containerState` to be defined on tokenizer'
+ // )
+ info.continued += 1;
+ // To do: add to stack?
+ // stack.push([self.currentConstruct, self.containerState])
+ // Try another.
+ document_continued(tokenizer, code, info)
+}
+
+fn flow_start(tokenizer: &mut Tokenizer, code: Code, mut info: DocumentInfo) -> StateFnResult {
+ println!("flow_start");
+ let next = info.next;
+ info.next = Box::new(flow); // This is weird but Rust needs a function there.
+
+ let size = info.continued;
+ exit_containers(tokenizer, &mut info, size);
+
+ tokenizer.go_until(next, eof_eol, move |(state, remainder)| {
+ (
+ State::Fn(Box::new(move |t, c| flow_end(t, c, info, state))),
+ remainder,
+ )
+ })(tokenizer, code)
+}
+
+fn flow_end(
+ tokenizer: &mut Tokenizer,
+ code: Code,
+ mut info: DocumentInfo,
+ result: State,
+) -> StateFnResult {
+ println!("flow_end");
+ let was_lazy = tokenizer.lazy;
+
+ if was_lazy {
+ println!(
+ "this line was lazy. Depeding on what was parsed, we need to exit containers after it?"
+ );
+ }
+
+ info.continued = 0;
+
+ // To do: blank lines? Other things?
+ if tokenizer.events.len() > 2
+ && tokenizer.events[tokenizer.events.len() - 1].token_type == TokenType::LineEnding
+ {
+ info.last_line_ending_index = Some(tokenizer.events.len() - 2);
+ } else {
+ info.last_line_ending_index = None;
+ }
+
+ match result {
+ State::Ok => {
+ println!("State::Ok");
+ exit_containers(tokenizer, &mut info, 0);
+ tokenizer.events = info.map.consume(&mut tokenizer.events);
+ (State::Ok, Some(vec![code]))
+ }
+ State::Nok => unreachable!("handle nok in `flow`?"),
+ State::Fn(func) => {
+ info.next = func;
+ before(tokenizer, code, info)
+ }
+ }
+}
+
+fn eof_eol(code: Code) -> bool {
+ matches!(
+ code,
+ Code::None | Code::CarriageReturnLineFeed | Code::Char('\n' | '\r')
+ )
+}
diff --git a/src/content/flow.rs b/src/content/flow.rs
index 74c6a62..f406685 100644
--- a/src/content/flow.rs
+++ b/src/content/flow.rs
@@ -26,52 +26,7 @@ use crate::construct::{
html_flow::start as html_flow, paragraph::start as paragraph,
thematic_break::start as thematic_break,
};
-use crate::parser::ParseState;
-use crate::subtokenize::subtokenize;
-use crate::tokenizer::{Code, Event, EventType, Point, State, StateFnResult, TokenType, Tokenizer};
-use crate::util::{
- normalize_identifier::normalize_identifier,
- span::{from_exit_event, serialize},
-};
-use std::collections::HashSet;
-
-/// Turn `codes` as the flow content type into events.
-pub fn flow(parse_state: &mut ParseState, point: Point, index: usize) -> Vec<Event> {
- let mut tokenizer = Tokenizer::new(point, index, parse_state);
- tokenizer.push(&parse_state.codes, Box::new(start), true);
- let mut next_definitions: HashSet<String> = HashSet::new();
-
- let mut index = 0;
-
- while index < tokenizer.events.len() {
- let event = &tokenizer.events[index];
-
- if event.event_type == EventType::Exit
- && event.token_type == TokenType::DefinitionLabelString
- {
- next_definitions.insert(normalize_identifier(
- serialize(
- &parse_state.codes,
- &from_exit_event(&tokenizer.events, index),
- false,
- )
- .as_str(),
- ));
- }
-
- index += 1;
- }
-
- let mut result = (tokenizer.events, false);
-
- parse_state.definitions = next_definitions;
-
- while !result.1 {
- result = subtokenize(result.0, parse_state);
- }
-
- result.0
-}
+use crate::tokenizer::{Code, State, StateFnResult, TokenType, Tokenizer};
/// Before flow.
///
@@ -83,7 +38,7 @@ pub fn flow(parse_state: &mut ParseState, point: Point, index: usize) -> Vec<Eve
/// | bravo
/// |***
/// ```
-fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
+pub fn start(tokenizer: &mut Tokenizer, code: Code) -> StateFnResult {
match code {
Code::None => (State::Ok, None),
_ => tokenizer.attempt(blank_line, |ok| {
diff --git a/src/content/mod.rs b/src/content/mod.rs
index ae8ad83..af40cc0 100644
--- a/src/content/mod.rs
+++ b/src/content/mod.rs
@@ -1,5 +1,11 @@
//! Content types found in markdown.
+//!
+//! * [document][document]
+//! * [flow][flow]
+//! * [string][string]
+//! * [text][text]
+pub mod document;
pub mod flow;
pub mod string;
pub mod text;
diff --git a/src/parser.rs b/src/parser.rs
index 69dd355..b1fd4fd 100644
--- a/src/parser.rs
+++ b/src/parser.rs
@@ -2,7 +2,7 @@
use std::collections::HashSet;
// To do: this should start with `containers`, when they’re done.
-use crate::content::flow::flow;
+use crate::content::document::document;
use crate::tokenizer::{Code, Event, Point};
use crate::util::codes::parse as parse_codes;
@@ -27,7 +27,7 @@ pub fn parse(value: &str) -> (Vec<Event>, ParseState) {
definitions: HashSet::new(),
};
- let events = flow(
+ let events = document(
&mut parse_state,
Point {
line: 1,
diff --git a/src/tokenizer.rs b/src/tokenizer.rs
index 8c11a68..cbcc464 100644
--- a/src/tokenizer.rs
+++ b/src/tokenizer.rs
@@ -1702,6 +1702,10 @@ pub enum TokenType {
///
/// > 👉 **Note**: this is used while parsing but compiled away.
AttentionSequence,
+ BlockQuote,
+ BlockQuoteMarker,
+ BlockQuotePrefix,
+ BlockQuotePrefixWhitespace,
}
/// Embedded content type.
@@ -1841,6 +1845,7 @@ struct InternalState {
// #[derive(Debug)]
/// A tokenizer itself.
+#[allow(clippy::struct_excessive_bools)]
pub struct Tokenizer<'a> {
column_start: HashMap<usize, usize>,
/// Track whether a character is expected to be consumed, and whether it’s
@@ -1855,15 +1860,15 @@ pub struct Tokenizer<'a> {
/// Hierarchy of semantic labels.
///
/// Tracked to make sure everything’s valid.
- stack: Vec<TokenType>,
+ pub stack: Vec<TokenType>,
/// Previous character code.
pub previous: Code,
/// Current character code.
current: Code,
/// `index` in codes of the current code.
- index: usize,
+ pub index: usize,
/// Current relative and absolute place in the file.
- point: Point,
+ pub point: Point,
/// List of attached resolvers, which will be called when done feeding,
/// to clean events.
resolvers: Vec<Box<Resolver>>,
@@ -1887,6 +1892,7 @@ pub struct Tokenizer<'a> {
///
/// Used when tokenizing [flow content][crate::content::flow].
pub interrupt: bool,
+ pub lazy: bool,
}
impl<'a> Tokenizer<'a> {
@@ -1907,6 +1913,7 @@ impl<'a> Tokenizer<'a> {
label_start_list_loose: vec![],
media_list: vec![],
interrupt: false,
+ lazy: false,
resolvers: vec![],
resolver_ids: vec![],
}
@@ -2120,7 +2127,8 @@ impl<'a> Tokenizer<'a> {
state_fn,
until,
vec![],
- |result: (Vec<Code>, Vec<Code>), _ok, _tokenizer: &mut Tokenizer, state| {
+ |result: (Vec<Code>, Vec<Code>), _ok, tokenizer: &mut Tokenizer, state| {
+ tokenizer.consumed = true;
done(check_statefn_result((state, Some(result.1))))
},
)
@@ -2262,6 +2270,20 @@ fn attempt_impl(
done: impl FnOnce((Vec<Code>, Vec<Code>), bool, &mut Tokenizer, State) -> StateFnResult + 'static,
) -> Box<StateFn> {
Box::new(|tokenizer, code| {
+ // To do: `pause` is currently used after the code.
+ // Should it be before?
+ // How to match `eof`?
+ if !codes.is_empty() && pause(tokenizer.previous) {
+ tokenizer.consumed = true;
+ println!("pause!: {:?}", (codes.clone(), vec![code]));
+ return done(
+ (codes, vec![code]),
+ false,
+ tokenizer,
+ State::Fn(Box::new(state)),
+ );
+ }
+
let (next, remainder) = check_statefn_result(state(tokenizer, code));
match code {
@@ -2278,14 +2300,6 @@ fn attempt_impl(
);
}
- // To do: `pause` is currently used after the code.
- // Should it be before?
- if pause(code) {
- tokenizer.consumed = true;
- let remaining = if let Some(x) = remainder { x } else { vec![] };
- return done((codes, remaining), false, tokenizer, next);
- }
-
match next {
State::Ok => {
let remaining = if let Some(x) = remainder { x } else { vec![] };
diff --git a/src/util/edit_map.rs b/src/util/edit_map.rs
index ae627c1..f67a8b9 100644
--- a/src/util/edit_map.rs
+++ b/src/util/edit_map.rs
@@ -48,6 +48,7 @@ fn shift_links(events: &mut [Event], jumps: &[(usize, isize)]) {
/// Make it easy to insert and remove things while being performant and keeping
/// links in check.
+#[derive(Debug)]
pub struct EditMap {
/// Whether this map was consumed already.
consumed: bool,
diff --git a/src/util/mod.rs b/src/util/mod.rs
index d1a0e01..ae1add6 100644
--- a/src/util/mod.rs
+++ b/src/util/mod.rs
@@ -6,4 +6,5 @@ pub mod edit_map;
pub mod encode;
pub mod normalize_identifier;
pub mod sanitize_uri;
+pub mod skip;
pub mod span;
diff --git a/src/util/skip.rs b/src/util/skip.rs
new file mode 100644
index 0000000..2c4198a
--- /dev/null
+++ b/src/util/skip.rs
@@ -0,0 +1,44 @@
+use crate::tokenizer::{Event, TokenType};
+
+/// To do.
+pub fn opt(events: &[Event], index: usize, token_types: &[TokenType]) -> usize {
+ skip_opt_with_direction(events, index, token_types, true)
+}
+
+/// To do.
+pub fn opt_back(events: &[Event], index: usize, token_types: &[TokenType]) -> usize {
+ skip_opt_with_direction(events, index, token_types, false)
+}
+
+/// To do.
+fn skip_opt_with_direction(
+ events: &[Event],
+ index: usize,
+ token_types: &[TokenType],
+ forward: bool,
+) -> usize {
+ let mut index = index;
+
+ while index < events.len() {
+ let current = &events[index].token_type;
+
+ if !token_types.contains(current) {
+ break;
+ }
+
+ // assert_eq!(events[index].event_type, EventType::Enter);
+ index = if forward { index + 1 } else { index - 1 };
+
+ loop {
+ if events[index].token_type == *current {
+ // assert_eq!(events[index].event_type, EventType::Exit);
+ index = if forward { index + 1 } else { index - 1 };
+ break;
+ }
+
+ index = if forward { index + 1 } else { index - 1 };
+ }
+ }
+
+ index
+}
generated by cgit (git 2.34.1) at 2024-11-26 00:15:02 +0000