2023-08-11 23:54:46 +00:00
|
|
|
use crate::scanner::{Marker, ScanError, Scanner, TScalarStyle, Token, TokenType};
|
2015-05-28 18:26:37 +00:00
|
|
|
use std::collections::HashMap;
|
2015-05-24 06:27:42 +00:00
|
|
|
|
|
|
|
#[derive(Clone, Copy, PartialEq, Debug, Eq)]
|
2015-05-31 09:59:43 +00:00
|
|
|
enum State {
|
2015-05-24 06:27:42 +00:00
|
|
|
StreamStart,
|
|
|
|
ImplicitDocumentStart,
|
|
|
|
DocumentStart,
|
|
|
|
DocumentContent,
|
|
|
|
DocumentEnd,
|
|
|
|
BlockNode,
|
2016-11-14 03:22:32 +00:00
|
|
|
// BlockNodeOrIndentlessSequence,
|
|
|
|
// FlowNode,
|
2015-05-24 06:27:42 +00:00
|
|
|
BlockSequenceFirstEntry,
|
|
|
|
BlockSequenceEntry,
|
|
|
|
IndentlessSequenceEntry,
|
|
|
|
BlockMappingFirstKey,
|
|
|
|
BlockMappingKey,
|
|
|
|
BlockMappingValue,
|
|
|
|
FlowSequenceFirstEntry,
|
|
|
|
FlowSequenceEntry,
|
|
|
|
FlowSequenceEntryMappingKey,
|
|
|
|
FlowSequenceEntryMappingValue,
|
|
|
|
FlowSequenceEntryMappingEnd,
|
|
|
|
FlowMappingFirstKey,
|
|
|
|
FlowMappingKey,
|
|
|
|
FlowMappingValue,
|
|
|
|
FlowMappingEmptyValue,
|
2018-09-15 16:49:04 +00:00
|
|
|
End,
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// An event generated by the YAML parser.
|
|
|
|
///
|
|
|
|
/// Events are used in the low-level event-based API (push parser). The API entrypoint is the
|
|
|
|
/// [`EventReceiver`] trait.
|
2015-05-24 06:27:42 +00:00
|
|
|
#[derive(Clone, PartialEq, Debug, Eq)]
|
|
|
|
pub enum Event {
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Reserved for internal use.
|
2016-06-22 08:05:40 +00:00
|
|
|
Nothing,
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Event generated at the very beginning of parsing.
|
2015-05-24 06:27:42 +00:00
|
|
|
StreamStart,
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Last event that will be generated by the parser. Signals EOF.
|
2015-05-24 06:27:42 +00:00
|
|
|
StreamEnd,
|
2023-08-17 00:17:40 +00:00
|
|
|
/// The YAML start document directive (`---`).
|
2015-05-24 06:27:42 +00:00
|
|
|
DocumentStart,
|
2023-08-17 00:17:40 +00:00
|
|
|
/// The YAML end document directive (`...`).
|
2015-05-24 06:27:42 +00:00
|
|
|
DocumentEnd,
|
2023-08-17 00:17:40 +00:00
|
|
|
/// A YAML Alias.
|
|
|
|
Alias(
|
|
|
|
/// The anchor ID the alias refers to.
|
|
|
|
usize,
|
|
|
|
),
|
2015-05-31 09:59:43 +00:00
|
|
|
/// Value, style, anchor_id, tag
|
2023-11-19 00:09:41 +00:00
|
|
|
Scalar(String, TScalarStyle, usize, Option<Tag>),
|
2023-08-17 00:17:40 +00:00
|
|
|
SequenceStart(
|
|
|
|
/// The anchor ID of the start of the squence.
|
|
|
|
usize,
|
2023-11-19 13:40:01 +00:00
|
|
|
/// An optional tag
|
|
|
|
Option<Tag>,
|
2023-08-17 00:17:40 +00:00
|
|
|
),
|
2015-05-24 06:27:42 +00:00
|
|
|
SequenceEnd,
|
2023-08-17 00:17:40 +00:00
|
|
|
MappingStart(
|
|
|
|
/// The anchor ID of the start of the mapping.
|
|
|
|
usize,
|
2023-11-19 00:09:41 +00:00
|
|
|
/// An optional tag
|
|
|
|
Option<Tag>,
|
2023-08-17 00:17:40 +00:00
|
|
|
),
|
2018-09-15 16:49:04 +00:00
|
|
|
MappingEnd,
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2023-11-19 00:09:41 +00:00
|
|
|
/// A YAML tag.
|
|
|
|
#[derive(Clone, PartialEq, Debug, Eq)]
|
|
|
|
pub struct Tag {
|
|
|
|
/// Handle of the tag (`!` included).
|
|
|
|
pub handle: String,
|
|
|
|
/// The suffix of the tag.
|
|
|
|
pub suffix: String,
|
|
|
|
}
|
|
|
|
|
2015-05-24 06:29:10 +00:00
|
|
|
impl Event {
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Create an empty scalar.
|
2015-05-24 06:29:10 +00:00
|
|
|
fn empty_scalar() -> Event {
|
2015-05-27 13:57:42 +00:00
|
|
|
// a null scalar
|
2016-02-28 00:30:13 +00:00
|
|
|
Event::Scalar("~".to_owned(), TScalarStyle::Plain, 0, None)
|
2015-05-30 14:39:50 +00:00
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Create an empty scalar with the given anchor.
|
2023-11-19 00:09:41 +00:00
|
|
|
fn empty_scalar_with_anchor(anchor: usize, tag: Option<Tag>) -> Event {
|
2023-08-11 23:54:46 +00:00
|
|
|
Event::Scalar(String::new(), TScalarStyle::Plain, anchor, tag)
|
2015-05-24 06:29:10 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// A YAML parser.
|
2015-05-24 06:27:42 +00:00
|
|
|
#[derive(Debug)]
|
|
|
|
pub struct Parser<T> {
|
|
|
|
scanner: Scanner<T>,
|
|
|
|
states: Vec<State>,
|
|
|
|
state: State,
|
|
|
|
token: Option<Token>,
|
2017-06-22 06:44:46 +00:00
|
|
|
current: Option<(Event, Marker)>,
|
2015-05-28 18:26:37 +00:00
|
|
|
anchors: HashMap<String, usize>,
|
|
|
|
anchor_id: usize,
|
2023-11-19 15:00:19 +00:00
|
|
|
/// The tag directives (`%TAG`) the parser has encountered.
|
|
|
|
///
|
|
|
|
/// Key is the handle, and value is the prefix.
|
|
|
|
tags: HashMap<String, String>,
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Trait to be implemented in order to use the low-level parsing API.
|
|
|
|
///
|
|
|
|
/// The low-level parsing API is event-based (a push parser), calling [`EventReceiver::on_event`]
|
|
|
|
/// for each YAML [`Event`] that occurs.
|
|
|
|
/// The [`EventReceiver`] trait only receives events. In order to receive both events and their
|
|
|
|
/// location in the source, use [`MarkedEventReceiver`]. Note that [`EventReceiver`]s implement
|
|
|
|
/// [`MarkedEventReceiver`] automatically.
|
|
|
|
///
|
|
|
|
/// # Event hierarchy
|
|
|
|
/// The event stream starts with an [`Event::StreamStart`] event followed by an
|
|
|
|
/// [`Event::DocumentStart`] event. If the YAML document starts with a mapping (an object), an
|
|
|
|
/// [`Event::MappingStart`] event is emitted. If it starts with a sequence (an array), an
|
|
|
|
/// [`Event::SequenceStart`] event is emitted. Otherwise, an [`Event::Scalar`] event is emitted.
|
|
|
|
///
|
|
|
|
/// In a mapping, key-values are sent as consecutive events. The first event after an
|
|
|
|
/// [`Event::MappingStart`] will be the key, and following its value. If the mapping contains no
|
|
|
|
/// sub-mapping or sub-sequence, then even events (starting from 0) will always be keys and odd
|
|
|
|
/// ones will always be values. The mapping ends when an [`Event::MappingEnd`] event is received.
|
|
|
|
///
|
|
|
|
/// In a sequence, values are sent consecutively until the [`Event::SequenceEnd`] event.
|
|
|
|
///
|
|
|
|
/// If a value is a sub-mapping or a sub-sequence, an [`Event::MappingStart`] or
|
|
|
|
/// [`Event::SequenceStart`] event will be sent respectively. Following events until the associated
|
|
|
|
/// [`Event::MappingStart`] or [`Event::SequenceEnd`] (beware of nested mappings or sequences) will
|
|
|
|
/// be part of the value and not another key-value pair or element in the sequence.
|
|
|
|
///
|
|
|
|
/// For instance, the following yaml:
|
|
|
|
/// ```yaml
|
|
|
|
/// a: b
|
|
|
|
/// c:
|
|
|
|
/// d: e
|
|
|
|
/// f:
|
|
|
|
/// - g
|
|
|
|
/// - h
|
|
|
|
/// ```
|
|
|
|
/// will emit (indented and commented for lisibility):
|
|
|
|
/// ```text
|
|
|
|
/// StreamStart, DocumentStart, MappingStart,
|
|
|
|
/// Scalar("a", ..), Scalar("b", ..)
|
|
|
|
/// Scalar("c", ..), MappingStart, Scalar("d", ..), Scalar("e", ..), MappingEnd,
|
|
|
|
/// Scalar("f", ..), SequenceStart, Scalar("g", ..), Scalar("h", ..), SequenceEnd,
|
|
|
|
/// MappingEnd, DocumentEnd, StreamEnd
|
|
|
|
/// ```
|
|
|
|
///
|
|
|
|
/// # Example
|
|
|
|
/// ```
|
|
|
|
/// # use yaml_rust::parser::{Event, EventReceiver, Parser};
|
|
|
|
/// #
|
|
|
|
/// /// Sink of events. Collects them into an array.
|
|
|
|
/// struct EventSink {
|
|
|
|
/// events: Vec<Event>,
|
|
|
|
/// }
|
|
|
|
///
|
|
|
|
/// /// Implement `on_event`, pushing into `self.events`.
|
|
|
|
/// impl EventReceiver for EventSink {
|
|
|
|
/// fn on_event(&mut self, ev: Event) {
|
|
|
|
/// self.events.push(ev);
|
|
|
|
/// }
|
|
|
|
/// }
|
|
|
|
///
|
|
|
|
/// /// Load events from a yaml string.
|
|
|
|
/// fn str_to_events(yaml: &str) -> Vec<Event> {
|
|
|
|
/// let mut sink = EventSink { events: Vec::new() };
|
|
|
|
/// let mut parser = Parser::new(yaml.chars());
|
|
|
|
/// // Load events using our sink as the receiver.
|
|
|
|
/// parser.load(&mut sink, true).unwrap();
|
|
|
|
/// sink.events
|
|
|
|
/// }
|
|
|
|
/// ```
|
2015-05-26 16:29:40 +00:00
|
|
|
pub trait EventReceiver {
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Handler called for each YAML event that is emitted by the parser.
|
2017-06-19 13:41:26 +00:00
|
|
|
fn on_event(&mut self, ev: Event);
|
2016-11-10 10:53:28 +00:00
|
|
|
}
|
2016-11-09 12:04:48 +00:00
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Trait to be implemented for using the low-level parsing API.
|
|
|
|
///
|
|
|
|
/// Functionally similar to [`EventReceiver`], but receives a [`Marker`] as well as the event.
|
2016-11-10 10:53:28 +00:00
|
|
|
pub trait MarkedEventReceiver {
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Handler called for each event that occurs.
|
2017-06-19 13:41:26 +00:00
|
|
|
fn on_event(&mut self, ev: Event, _mark: Marker);
|
2016-11-10 10:53:28 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
impl<R: EventReceiver> MarkedEventReceiver for R {
|
2017-06-19 13:41:26 +00:00
|
|
|
fn on_event(&mut self, ev: Event, _mark: Marker) {
|
2023-08-11 23:54:46 +00:00
|
|
|
self.on_event(ev);
|
2016-11-09 12:04:48 +00:00
|
|
|
}
|
2015-05-26 16:29:40 +00:00
|
|
|
}
|
|
|
|
|
2016-11-09 12:04:48 +00:00
|
|
|
pub type ParseResult = Result<(Event, Marker), ScanError>;
|
2015-05-24 06:27:42 +00:00
|
|
|
|
2018-09-15 16:49:04 +00:00
|
|
|
impl<T: Iterator<Item = char>> Parser<T> {
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Crate a new instance of a parser from the given input of characters.
|
2015-05-24 06:27:42 +00:00
|
|
|
pub fn new(src: T) -> Parser<T> {
|
|
|
|
Parser {
|
|
|
|
scanner: Scanner::new(src),
|
|
|
|
states: Vec::new(),
|
|
|
|
state: State::StreamStart,
|
2015-05-24 17:34:18 +00:00
|
|
|
token: None,
|
2017-06-22 06:44:46 +00:00
|
|
|
current: None,
|
2015-05-28 18:26:37 +00:00
|
|
|
|
|
|
|
anchors: HashMap::new(),
|
|
|
|
// valid anchor_id starts from 1
|
|
|
|
anchor_id: 1,
|
2023-11-19 15:00:19 +00:00
|
|
|
tags: HashMap::new(),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Try to load the next event and return it, but do not consuming it from `self`.
|
|
|
|
///
|
|
|
|
/// Any subsequent call to [`Parser::peek`] will return the same value, until a call to
|
|
|
|
/// [`Parser::next`] or [`Parser::load`].
|
2017-06-22 06:44:46 +00:00
|
|
|
pub fn peek(&mut self) -> Result<&(Event, Marker), ScanError> {
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Some(ref x) = self.current {
|
|
|
|
Ok(x)
|
|
|
|
} else {
|
|
|
|
self.current = Some(self.next()?);
|
|
|
|
self.peek()
|
2017-06-22 06:44:46 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Try to load the next event and return it, consuming it from `self`.
|
2017-06-22 06:44:46 +00:00
|
|
|
pub fn next(&mut self) -> ParseResult {
|
2023-08-17 00:17:40 +00:00
|
|
|
match self.current.take() {
|
2017-06-22 06:44:46 +00:00
|
|
|
None => self.parse(),
|
2023-08-17 00:17:40 +00:00
|
|
|
Some(v) => Ok(v),
|
2017-06-22 06:44:46 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn peek_token(&mut self) -> Result<&Token, ScanError> {
|
2017-06-20 13:47:19 +00:00
|
|
|
match self.token {
|
2018-09-15 16:49:04 +00:00
|
|
|
None => {
|
2018-09-16 06:58:48 +00:00
|
|
|
self.token = Some(self.scan_next_token()?);
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok(self.token.as_ref().unwrap())
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
Some(ref tok) => Ok(tok),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
fn scan_next_token(&mut self) -> Result<Token, ScanError> {
|
|
|
|
let token = self.scanner.next();
|
|
|
|
match token {
|
2018-09-15 16:49:04 +00:00
|
|
|
None => match self.scanner.get_error() {
|
|
|
|
None => Err(ScanError::new(self.scanner.mark(), "unexpected eof")),
|
|
|
|
Some(e) => Err(e),
|
|
|
|
},
|
|
|
|
Some(tok) => Ok(tok),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2017-06-20 13:47:19 +00:00
|
|
|
fn fetch_token(&mut self) -> Token {
|
2018-09-15 16:49:04 +00:00
|
|
|
self.token
|
|
|
|
.take()
|
|
|
|
.expect("fetch_token needs to be preceded by peek_token")
|
2017-06-20 13:47:19 +00:00
|
|
|
}
|
|
|
|
|
2015-05-24 06:27:42 +00:00
|
|
|
fn skip(&mut self) {
|
|
|
|
self.token = None;
|
2017-06-22 06:44:46 +00:00
|
|
|
//self.peek_token();
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
fn pop_state(&mut self) {
|
2023-08-11 23:54:46 +00:00
|
|
|
self.state = self.states.pop().unwrap();
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
fn push_state(&mut self, state: State) {
|
|
|
|
self.states.push(state);
|
|
|
|
}
|
|
|
|
|
2017-06-19 13:41:26 +00:00
|
|
|
fn parse(&mut self) -> ParseResult {
|
2016-05-25 05:19:36 +00:00
|
|
|
if self.state == State::End {
|
2017-06-14 08:29:27 +00:00
|
|
|
return Ok((Event::StreamEnd, self.scanner.mark()));
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.state_machine()?;
|
2015-05-30 14:39:50 +00:00
|
|
|
// println!("EV {:?}", ev);
|
2017-06-14 08:29:27 +00:00
|
|
|
Ok((ev, mark))
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2023-08-17 00:17:40 +00:00
|
|
|
/// Load the YAML from the stream in `self`, pushing events into `recv`.
|
|
|
|
///
|
|
|
|
/// The contents of the stream are parsed and the corresponding events are sent into the
|
|
|
|
/// recveiver. For detailed explanations about how events work, see [`EventReceiver`].
|
|
|
|
///
|
|
|
|
/// If `multi` is set to `true`, the parser will allow parsing of multiple YAML documents
|
|
|
|
/// inside the stream.
|
|
|
|
///
|
|
|
|
/// Note that any [`EventReceiver`] is also a [`MarkedEventReceiver`], so implementing the
|
|
|
|
/// former is enough to call this function.
|
2018-09-15 16:49:04 +00:00
|
|
|
pub fn load<R: MarkedEventReceiver>(
|
|
|
|
&mut self,
|
|
|
|
recv: &mut R,
|
|
|
|
multi: bool,
|
|
|
|
) -> Result<(), ScanError> {
|
2015-05-24 06:27:42 +00:00
|
|
|
if !self.scanner.stream_started() {
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.next()?;
|
2015-05-24 06:27:42 +00:00
|
|
|
assert_eq!(ev, Event::StreamStart);
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(ev, mark);
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
if self.scanner.stream_ended() {
|
2015-05-26 08:41:35 +00:00
|
|
|
// XXX has parsed?
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(Event::StreamEnd, self.scanner.mark());
|
2015-05-26 18:50:51 +00:00
|
|
|
return Ok(());
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2015-05-26 08:41:35 +00:00
|
|
|
loop {
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.next()?;
|
2015-05-26 08:41:35 +00:00
|
|
|
if ev == Event::StreamEnd {
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(ev, mark);
|
2015-05-26 18:50:51 +00:00
|
|
|
return Ok(());
|
|
|
|
}
|
2015-05-28 18:26:37 +00:00
|
|
|
// clear anchors before a new document
|
|
|
|
self.anchors.clear();
|
2018-09-16 06:58:48 +00:00
|
|
|
self.load_document(ev, mark, recv)?;
|
2015-05-26 18:50:51 +00:00
|
|
|
if !multi {
|
|
|
|
break;
|
2015-05-26 08:41:35 +00:00
|
|
|
}
|
|
|
|
}
|
2015-05-26 18:50:51 +00:00
|
|
|
Ok(())
|
2015-05-26 08:41:35 +00:00
|
|
|
}
|
|
|
|
|
2018-09-15 16:49:04 +00:00
|
|
|
fn load_document<R: MarkedEventReceiver>(
|
|
|
|
&mut self,
|
|
|
|
first_ev: Event,
|
|
|
|
mark: Marker,
|
|
|
|
recv: &mut R,
|
|
|
|
) -> Result<(), ScanError> {
|
2017-06-14 08:29:27 +00:00
|
|
|
assert_eq!(first_ev, Event::DocumentStart);
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(first_ev, mark);
|
2015-05-24 06:27:42 +00:00
|
|
|
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.next()?;
|
|
|
|
self.load_node(ev, mark, recv)?;
|
2015-05-26 18:50:51 +00:00
|
|
|
|
2015-05-26 08:41:35 +00:00
|
|
|
// DOCUMENT-END is expected.
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.next()?;
|
2015-05-26 08:41:35 +00:00
|
|
|
assert_eq!(ev, Event::DocumentEnd);
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(ev, mark);
|
2015-05-26 08:41:35 +00:00
|
|
|
|
2015-05-26 18:50:51 +00:00
|
|
|
Ok(())
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2018-09-15 16:49:04 +00:00
|
|
|
fn load_node<R: MarkedEventReceiver>(
|
|
|
|
&mut self,
|
|
|
|
first_ev: Event,
|
|
|
|
mark: Marker,
|
|
|
|
recv: &mut R,
|
|
|
|
) -> Result<(), ScanError> {
|
2017-06-14 08:29:27 +00:00
|
|
|
match first_ev {
|
2016-02-28 00:30:13 +00:00
|
|
|
Event::Alias(..) | Event::Scalar(..) => {
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(first_ev, mark);
|
2015-05-28 17:56:03 +00:00
|
|
|
Ok(())
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-11-19 13:40:01 +00:00
|
|
|
Event::SequenceStart(..) => {
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(first_ev, mark);
|
2017-06-14 08:29:27 +00:00
|
|
|
self.load_sequence(recv)
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-11-19 00:09:41 +00:00
|
|
|
Event::MappingStart(..) => {
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(first_ev, mark);
|
2017-06-14 08:29:27 +00:00
|
|
|
self.load_mapping(recv)
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
_ => {
|
2023-08-11 23:54:46 +00:00
|
|
|
println!("UNREACHABLE EVENT: {first_ev:?}");
|
2018-09-15 16:49:04 +00:00
|
|
|
unreachable!();
|
|
|
|
}
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-09-15 16:49:04 +00:00
|
|
|
fn load_mapping<R: MarkedEventReceiver>(&mut self, recv: &mut R) -> Result<(), ScanError> {
|
2018-09-16 06:58:48 +00:00
|
|
|
let (mut key_ev, mut key_mark) = self.next()?;
|
2017-06-14 08:29:27 +00:00
|
|
|
while key_ev != Event::MappingEnd {
|
2015-05-24 06:27:42 +00:00
|
|
|
// key
|
2018-09-16 06:58:48 +00:00
|
|
|
self.load_node(key_ev, key_mark, recv)?;
|
2015-05-24 06:27:42 +00:00
|
|
|
|
|
|
|
// value
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.next()?;
|
|
|
|
self.load_node(ev, mark, recv)?;
|
2015-05-24 06:27:42 +00:00
|
|
|
|
|
|
|
// next event
|
2018-09-16 06:58:48 +00:00
|
|
|
let (ev, mark) = self.next()?;
|
2017-06-14 08:29:27 +00:00
|
|
|
key_ev = ev;
|
|
|
|
key_mark = mark;
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(key_ev, key_mark);
|
2015-05-26 18:50:51 +00:00
|
|
|
Ok(())
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2018-09-15 16:49:04 +00:00
|
|
|
fn load_sequence<R: MarkedEventReceiver>(&mut self, recv: &mut R) -> Result<(), ScanError> {
|
2018-09-16 06:58:48 +00:00
|
|
|
let (mut ev, mut mark) = self.next()?;
|
2015-05-24 06:27:42 +00:00
|
|
|
while ev != Event::SequenceEnd {
|
2018-09-16 06:58:48 +00:00
|
|
|
self.load_node(ev, mark, recv)?;
|
2015-05-24 06:27:42 +00:00
|
|
|
|
|
|
|
// next event
|
2018-09-16 06:58:48 +00:00
|
|
|
let (next_ev, next_mark) = self.next()?;
|
2017-06-14 08:29:27 +00:00
|
|
|
ev = next_ev;
|
|
|
|
mark = next_mark;
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2017-06-19 13:41:26 +00:00
|
|
|
recv.on_event(ev, mark);
|
2015-05-26 18:50:51 +00:00
|
|
|
Ok(())
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
fn state_machine(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
// let next_tok = self.peek_token()?;
|
2015-05-30 14:39:50 +00:00
|
|
|
// println!("cur_state {:?}, next tok: {:?}", self.state, next_tok);
|
2015-05-24 06:27:42 +00:00
|
|
|
match self.state {
|
|
|
|
State::StreamStart => self.stream_start(),
|
2015-05-26 08:41:35 +00:00
|
|
|
|
2015-05-24 06:27:42 +00:00
|
|
|
State::ImplicitDocumentStart => self.document_start(true),
|
|
|
|
State::DocumentStart => self.document_start(false),
|
|
|
|
State::DocumentContent => self.document_content(),
|
2015-05-26 08:41:35 +00:00
|
|
|
State::DocumentEnd => self.document_end(),
|
2015-05-24 06:27:42 +00:00
|
|
|
|
2015-05-24 17:34:18 +00:00
|
|
|
State::BlockNode => self.parse_node(true, false),
|
2016-11-14 03:22:32 +00:00
|
|
|
// State::BlockNodeOrIndentlessSequence => self.parse_node(true, true),
|
|
|
|
// State::FlowNode => self.parse_node(false, false),
|
2015-05-24 06:27:42 +00:00
|
|
|
State::BlockMappingFirstKey => self.block_mapping_key(true),
|
|
|
|
State::BlockMappingKey => self.block_mapping_key(false),
|
|
|
|
State::BlockMappingValue => self.block_mapping_value(),
|
|
|
|
|
|
|
|
State::BlockSequenceFirstEntry => self.block_sequence_entry(true),
|
|
|
|
State::BlockSequenceEntry => self.block_sequence_entry(false),
|
|
|
|
|
|
|
|
State::FlowSequenceFirstEntry => self.flow_sequence_entry(true),
|
|
|
|
State::FlowSequenceEntry => self.flow_sequence_entry(false),
|
|
|
|
|
2015-05-27 13:34:52 +00:00
|
|
|
State::FlowMappingFirstKey => self.flow_mapping_key(true),
|
|
|
|
State::FlowMappingKey => self.flow_mapping_key(false),
|
|
|
|
State::FlowMappingValue => self.flow_mapping_value(false),
|
|
|
|
|
2015-05-27 08:35:13 +00:00
|
|
|
State::IndentlessSequenceEntry => self.indentless_sequence_entry(),
|
|
|
|
|
2015-05-27 15:18:38 +00:00
|
|
|
State::FlowSequenceEntryMappingKey => self.flow_sequence_entry_mapping_key(),
|
|
|
|
State::FlowSequenceEntryMappingValue => self.flow_sequence_entry_mapping_value(),
|
|
|
|
State::FlowSequenceEntryMappingEnd => self.flow_sequence_entry_mapping_end(),
|
2015-06-04 08:10:43 +00:00
|
|
|
State::FlowMappingEmptyValue => self.flow_mapping_value(true),
|
2015-05-27 15:18:38 +00:00
|
|
|
|
2015-06-04 08:10:43 +00:00
|
|
|
/* impossible */
|
|
|
|
State::End => unreachable!(),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn stream_start(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::StreamStart(_)) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::ImplicitDocumentStart;
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::StreamStart, mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
Token(mark, _) => Err(ScanError::new(mark, "did not find expected <stream-start>")),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn document_start(&mut self, implicit: bool) -> ParseResult {
|
|
|
|
if !implicit {
|
2018-09-16 06:58:48 +00:00
|
|
|
while let TokenType::DocumentEnd = self.peek_token()?.1 {
|
2016-02-28 00:30:13 +00:00
|
|
|
self.skip();
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::StreamEnd) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::End;
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::StreamEnd, mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-08-11 23:54:46 +00:00
|
|
|
Token(
|
|
|
|
_,
|
|
|
|
TokenType::VersionDirective(..)
|
|
|
|
| TokenType::TagDirective(..)
|
|
|
|
| TokenType::DocumentStart,
|
|
|
|
) => {
|
2017-06-20 13:47:19 +00:00
|
|
|
// explicit document
|
2023-11-19 13:52:21 +00:00
|
|
|
self.explicit_document_start()
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, _) if implicit => {
|
2018-09-16 06:58:48 +00:00
|
|
|
self.parser_process_directives()?;
|
2015-05-24 06:27:42 +00:00
|
|
|
self.push_state(State::DocumentEnd);
|
|
|
|
self.state = State::BlockNode;
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::DocumentStart, mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2015-05-24 06:27:42 +00:00
|
|
|
_ => {
|
|
|
|
// explicit document
|
2023-11-19 13:52:21 +00:00
|
|
|
self.explicit_document_start()
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-05-28 14:07:59 +00:00
|
|
|
fn parser_process_directives(&mut self) -> Result<(), ScanError> {
|
|
|
|
loop {
|
2023-11-19 15:00:19 +00:00
|
|
|
let mut tags = HashMap::new();
|
|
|
|
match self.peek_token()? {
|
|
|
|
Token(_, TokenType::VersionDirective(_, _)) => {
|
2015-05-28 14:07:59 +00:00
|
|
|
// XXX parsing with warning according to spec
|
|
|
|
//if major != 1 || minor > 2 {
|
|
|
|
// return Err(ScanError::new(tok.0,
|
|
|
|
// "found incompatible YAML document"));
|
|
|
|
//}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-11-19 15:00:19 +00:00
|
|
|
Token(mark, TokenType::TagDirective(handle, prefix)) => {
|
|
|
|
if tags.contains_key(handle) {
|
|
|
|
return Err(ScanError::new(*mark, "the TAG directive must only be given at most once per handle in the same document"));
|
|
|
|
}
|
|
|
|
tags.insert(handle.to_string(), prefix.to_string());
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
_ => break,
|
2015-05-28 14:07:59 +00:00
|
|
|
}
|
2023-11-19 15:00:19 +00:00
|
|
|
self.tags = tags;
|
2015-05-28 14:07:59 +00:00
|
|
|
self.skip();
|
|
|
|
}
|
|
|
|
// TODO tag directive
|
|
|
|
Ok(())
|
|
|
|
}
|
|
|
|
|
2023-11-19 13:52:21 +00:00
|
|
|
fn explicit_document_start(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
self.parser_process_directives()?;
|
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::DocumentStart) => {
|
|
|
|
self.push_state(State::DocumentEnd);
|
|
|
|
self.state = State::DocumentContent;
|
|
|
|
self.skip();
|
|
|
|
Ok((Event::DocumentStart, mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
Token(mark, _) => Err(ScanError::new(
|
|
|
|
mark,
|
|
|
|
"did not find expected <document start>",
|
|
|
|
)),
|
|
|
|
}
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
fn document_content(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2023-08-11 23:54:46 +00:00
|
|
|
Token(
|
|
|
|
mark,
|
|
|
|
TokenType::VersionDirective(..)
|
|
|
|
| TokenType::TagDirective(..)
|
|
|
|
| TokenType::DocumentStart
|
|
|
|
| TokenType::DocumentEnd
|
|
|
|
| TokenType::StreamEnd,
|
|
|
|
) => {
|
2017-06-20 13:47:19 +00:00
|
|
|
self.pop_state();
|
|
|
|
// empty scalar
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
_ => self.parse_node(true, false),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-05-26 08:41:35 +00:00
|
|
|
fn document_end(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
let marker: Marker = match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::DocumentEnd) => {
|
|
|
|
self.skip();
|
|
|
|
mark
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
Token(mark, _) => mark,
|
2017-06-20 13:47:19 +00:00
|
|
|
};
|
2018-09-15 16:49:04 +00:00
|
|
|
|
2015-05-26 08:41:35 +00:00
|
|
|
// TODO tag handling
|
|
|
|
self.state = State::DocumentStart;
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::DocumentEnd, marker))
|
2015-05-26 08:41:35 +00:00
|
|
|
}
|
|
|
|
|
2023-08-11 23:54:46 +00:00
|
|
|
fn register_anchor(&mut self, name: String, _: &Marker) -> usize {
|
2020-06-01 12:59:27 +00:00
|
|
|
// anchors can be overridden/reused
|
2015-05-28 18:26:37 +00:00
|
|
|
// if self.anchors.contains_key(name) {
|
|
|
|
// return Err(ScanError::new(*mark,
|
|
|
|
// "while parsing anchor, found duplicated anchor"));
|
|
|
|
// }
|
|
|
|
let new_id = self.anchor_id;
|
|
|
|
self.anchor_id += 1;
|
2017-06-20 13:47:19 +00:00
|
|
|
self.anchors.insert(name, new_id);
|
2023-08-11 23:54:46 +00:00
|
|
|
new_id
|
2015-05-28 18:26:37 +00:00
|
|
|
}
|
|
|
|
|
2015-05-24 06:27:42 +00:00
|
|
|
fn parse_node(&mut self, block: bool, indentless_sequence: bool) -> ParseResult {
|
2015-05-28 18:26:37 +00:00
|
|
|
let mut anchor_id = 0;
|
2015-05-30 14:39:50 +00:00
|
|
|
let mut tag = None;
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::Alias(_)) => {
|
2015-05-28 17:56:03 +00:00
|
|
|
self.pop_state();
|
2017-06-20 13:47:19 +00:00
|
|
|
if let Token(mark, TokenType::Alias(name)) = self.fetch_token() {
|
|
|
|
match self.anchors.get(&name) {
|
2018-09-15 16:49:04 +00:00
|
|
|
None => {
|
|
|
|
return Err(ScanError::new(
|
|
|
|
mark,
|
|
|
|
"while parsing node, found unknown anchor",
|
|
|
|
))
|
|
|
|
}
|
|
|
|
Some(id) => return Ok((Event::Alias(*id), mark)),
|
2017-06-20 13:47:19 +00:00
|
|
|
}
|
|
|
|
} else {
|
|
|
|
unreachable!()
|
2015-05-28 18:26:37 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::Anchor(_)) => {
|
|
|
|
if let Token(mark, TokenType::Anchor(name)) = self.fetch_token() {
|
2023-08-11 23:54:46 +00:00
|
|
|
anchor_id = self.register_anchor(name, &mark);
|
2018-09-16 06:58:48 +00:00
|
|
|
if let TokenType::Tag(..) = self.peek_token()?.1 {
|
2023-11-19 00:09:41 +00:00
|
|
|
if let TokenType::Tag(handle, suffix) = self.fetch_token().1 {
|
2023-11-19 15:00:19 +00:00
|
|
|
tag = Some(self.resolve_tag(mark, &handle, suffix)?);
|
2017-06-20 13:47:19 +00:00
|
|
|
} else {
|
|
|
|
unreachable!()
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
unreachable!()
|
2015-05-28 17:56:03 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-11-19 15:00:19 +00:00
|
|
|
Token(mark, TokenType::Tag(..)) => {
|
2023-11-19 00:09:41 +00:00
|
|
|
if let TokenType::Tag(handle, suffix) = self.fetch_token().1 {
|
2023-11-19 15:00:19 +00:00
|
|
|
tag = Some(self.resolve_tag(mark, &handle, suffix)?);
|
2023-11-19 00:09:41 +00:00
|
|
|
if let TokenType::Anchor(_) = &self.peek_token()?.1 {
|
2017-06-20 13:47:19 +00:00
|
|
|
if let Token(mark, TokenType::Anchor(name)) = self.fetch_token() {
|
2023-08-11 23:54:46 +00:00
|
|
|
anchor_id = self.register_anchor(name, &mark);
|
2017-06-20 13:47:19 +00:00
|
|
|
} else {
|
|
|
|
unreachable!()
|
|
|
|
}
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
unreachable!()
|
2015-05-28 17:56:03 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2015-05-28 17:56:03 +00:00
|
|
|
_ => {}
|
|
|
|
}
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::BlockEntry) if indentless_sequence => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::IndentlessSequenceEntry;
|
2023-11-19 13:40:01 +00:00
|
|
|
Ok((Event::SequenceStart(anchor_id, tag), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::Scalar(..)) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.pop_state();
|
2017-06-20 13:47:19 +00:00
|
|
|
if let Token(mark, TokenType::Scalar(style, v)) = self.fetch_token() {
|
|
|
|
Ok((Event::Scalar(v, style, anchor_id, tag), mark))
|
|
|
|
} else {
|
|
|
|
unreachable!()
|
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::FlowSequenceStart) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::FlowSequenceFirstEntry;
|
2023-11-19 13:40:01 +00:00
|
|
|
Ok((Event::SequenceStart(anchor_id, tag), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::FlowMappingStart) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::FlowMappingFirstKey;
|
2023-11-19 00:09:41 +00:00
|
|
|
Ok((Event::MappingStart(anchor_id, tag), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::BlockSequenceStart) if block => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::BlockSequenceFirstEntry;
|
2023-11-19 13:40:01 +00:00
|
|
|
Ok((Event::SequenceStart(anchor_id, tag), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::BlockMappingStart) if block => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::BlockMappingFirstKey;
|
2023-11-19 00:09:41 +00:00
|
|
|
Ok((Event::MappingStart(anchor_id, tag), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2015-05-30 14:39:50 +00:00
|
|
|
// ex 7.2, an empty scalar can follow a secondary tag
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, _) if tag.is_some() || anchor_id > 0 => {
|
2015-05-30 14:39:50 +00:00
|
|
|
self.pop_state();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::empty_scalar_with_anchor(anchor_id, tag), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
Token(mark, _) => Err(ScanError::new(
|
|
|
|
mark,
|
|
|
|
"while parsing a node, did not find expected node content",
|
|
|
|
)),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn block_mapping_key(&mut self, first: bool) -> ParseResult {
|
2016-02-27 23:29:34 +00:00
|
|
|
// skip BlockMappingStart
|
2015-05-24 06:27:42 +00:00
|
|
|
if first {
|
2018-09-16 06:58:48 +00:00
|
|
|
let _ = self.peek_token()?;
|
2015-05-24 06:27:42 +00:00
|
|
|
//self.marks.push(tok.0);
|
|
|
|
self.skip();
|
|
|
|
}
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::Key) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.skip();
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Token(mark, TokenType::Key | TokenType::Value | TokenType::BlockEnd) =
|
|
|
|
*self.peek_token()?
|
|
|
|
{
|
|
|
|
self.state = State::BlockMappingValue;
|
|
|
|
// empty scalar
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
|
|
|
} else {
|
|
|
|
self.push_state(State::BlockMappingValue);
|
|
|
|
self.parse_node(true, true)
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2015-05-27 15:04:03 +00:00
|
|
|
// XXX(chenyh): libyaml failed to parse spec 1.2, ex8.18
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::Value) => {
|
2015-05-27 15:04:03 +00:00
|
|
|
self.state = State::BlockMappingValue;
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::empty_scalar(), mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::BlockEnd) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.pop_state();
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::MappingEnd, mark))
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
Token(mark, _) => Err(ScanError::new(
|
|
|
|
mark,
|
|
|
|
"while parsing a block mapping, did not find expected key",
|
|
|
|
)),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn block_mapping_value(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::Value) => {
|
|
|
|
self.skip();
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Token(mark, TokenType::Key | TokenType::Value | TokenType::BlockEnd) =
|
|
|
|
*self.peek_token()?
|
|
|
|
{
|
|
|
|
self.state = State::BlockMappingKey;
|
|
|
|
// empty scalar
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
|
|
|
} else {
|
|
|
|
self.push_state(State::BlockMappingKey);
|
|
|
|
self.parse_node(true, true)
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, _) => {
|
|
|
|
self.state = State::BlockMappingKey;
|
|
|
|
// empty scalar
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
}
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
|
2015-05-27 13:34:52 +00:00
|
|
|
fn flow_mapping_key(&mut self, first: bool) -> ParseResult {
|
|
|
|
if first {
|
2018-09-16 06:58:48 +00:00
|
|
|
let _ = self.peek_token()?;
|
2015-05-27 13:34:52 +00:00
|
|
|
self.skip();
|
|
|
|
}
|
2023-08-11 23:54:46 +00:00
|
|
|
let marker: Marker = {
|
|
|
|
match *self.peek_token()? {
|
|
|
|
Token(mark, TokenType::FlowMappingEnd) => mark,
|
|
|
|
Token(mark, _) => {
|
|
|
|
if !first {
|
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::FlowEntry) => self.skip(),
|
2023-08-11 23:54:46 +00:00
|
|
|
Token(mark, _) => return Err(ScanError::new(
|
|
|
|
mark,
|
|
|
|
"while parsing a flow mapping, did not find expected ',' or '}'",
|
|
|
|
)),
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-08-11 23:54:46 +00:00
|
|
|
}
|
2015-05-27 13:34:52 +00:00
|
|
|
|
2023-08-11 23:54:46 +00:00
|
|
|
match *self.peek_token()? {
|
|
|
|
Token(_, TokenType::Key) => {
|
|
|
|
self.skip();
|
|
|
|
if let Token(
|
|
|
|
mark,
|
|
|
|
TokenType::Value | TokenType::FlowEntry | TokenType::FlowMappingEnd,
|
|
|
|
) = *self.peek_token()?
|
|
|
|
{
|
2018-09-15 16:49:04 +00:00
|
|
|
self.state = State::FlowMappingValue;
|
2023-08-11 23:54:46 +00:00
|
|
|
return Ok((Event::empty_scalar(), mark));
|
|
|
|
} else {
|
|
|
|
self.push_state(State::FlowMappingValue);
|
2018-09-15 16:49:04 +00:00
|
|
|
return self.parse_node(false, false);
|
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
}
|
2023-08-11 23:54:46 +00:00
|
|
|
Token(marker, TokenType::Value) => {
|
|
|
|
self.state = State::FlowMappingValue;
|
|
|
|
return Ok((Event::empty_scalar(), marker));
|
|
|
|
}
|
|
|
|
Token(_, TokenType::FlowMappingEnd) => (),
|
|
|
|
_ => {
|
|
|
|
self.push_state(State::FlowMappingEmptyValue);
|
|
|
|
return self.parse_node(false, false);
|
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2023-08-11 23:54:46 +00:00
|
|
|
|
|
|
|
mark
|
2015-05-27 13:34:52 +00:00
|
|
|
}
|
2023-08-11 23:54:46 +00:00
|
|
|
}
|
|
|
|
};
|
2015-05-27 13:34:52 +00:00
|
|
|
|
|
|
|
self.pop_state();
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::MappingEnd, marker))
|
2015-05-27 13:34:52 +00:00
|
|
|
}
|
|
|
|
|
|
|
|
fn flow_mapping_value(&mut self, empty: bool) -> ParseResult {
|
2017-06-20 13:47:19 +00:00
|
|
|
let mark: Marker = {
|
|
|
|
if empty {
|
2018-09-16 06:58:48 +00:00
|
|
|
let Token(mark, _) = *self.peek_token()?;
|
2017-06-20 13:47:19 +00:00
|
|
|
self.state = State::FlowMappingKey;
|
|
|
|
return Ok((Event::empty_scalar(), mark));
|
|
|
|
} else {
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(marker, TokenType::Value) => {
|
|
|
|
self.skip();
|
2018-09-16 06:58:48 +00:00
|
|
|
match self.peek_token()?.1 {
|
2018-09-15 16:49:04 +00:00
|
|
|
TokenType::FlowEntry | TokenType::FlowMappingEnd => {}
|
2017-06-20 13:47:19 +00:00
|
|
|
_ => {
|
|
|
|
self.push_state(State::FlowMappingKey);
|
|
|
|
return self.parse_node(false, false);
|
|
|
|
}
|
|
|
|
}
|
|
|
|
marker
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
Token(marker, _) => marker,
|
2015-05-27 13:34:52 +00:00
|
|
|
}
|
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
};
|
2018-09-15 16:49:04 +00:00
|
|
|
|
2015-05-27 13:34:52 +00:00
|
|
|
self.state = State::FlowMappingKey;
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::empty_scalar(), mark))
|
2015-05-27 13:34:52 +00:00
|
|
|
}
|
|
|
|
|
2015-05-24 06:27:42 +00:00
|
|
|
fn flow_sequence_entry(&mut self, first: bool) -> ParseResult {
|
2016-02-27 23:29:34 +00:00
|
|
|
// skip FlowMappingStart
|
2015-05-24 06:27:42 +00:00
|
|
|
if first {
|
2018-09-16 06:58:48 +00:00
|
|
|
let _ = self.peek_token()?;
|
2015-05-24 06:27:42 +00:00
|
|
|
//self.marks.push(tok.0);
|
|
|
|
self.skip();
|
|
|
|
}
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::FlowSequenceEnd) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.pop_state();
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
return Ok((Event::SequenceEnd, mark));
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::FlowEntry) if !first => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.skip();
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, _) if !first => {
|
2018-09-15 16:49:04 +00:00
|
|
|
return Err(ScanError::new(
|
|
|
|
mark,
|
2020-06-01 12:59:27 +00:00
|
|
|
"while parsing a flow sequence, expected ',' or ']'",
|
2018-09-15 16:49:04 +00:00
|
|
|
));
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
_ => { /* next */ }
|
|
|
|
}
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::FlowSequenceEnd) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.pop_state();
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::SequenceEnd, mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::Key) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.state = State::FlowSequenceEntryMappingKey;
|
|
|
|
self.skip();
|
2023-11-19 00:09:41 +00:00
|
|
|
Ok((Event::MappingStart(0, None), mark))
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
_ => {
|
|
|
|
self.push_state(State::FlowSequenceEntry);
|
|
|
|
self.parse_node(false, false)
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-05-27 08:35:13 +00:00
|
|
|
fn indentless_sequence_entry(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::BlockEntry) => (),
|
|
|
|
Token(mark, _) => {
|
|
|
|
self.pop_state();
|
|
|
|
return Ok((Event::SequenceEnd, mark));
|
|
|
|
}
|
2015-05-27 08:35:13 +00:00
|
|
|
}
|
|
|
|
self.skip();
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Token(
|
|
|
|
mark,
|
|
|
|
TokenType::BlockEntry | TokenType::Key | TokenType::Value | TokenType::BlockEnd,
|
|
|
|
) = *self.peek_token()?
|
|
|
|
{
|
|
|
|
self.state = State::IndentlessSequenceEntry;
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
|
|
|
} else {
|
|
|
|
self.push_state(State::IndentlessSequenceEntry);
|
|
|
|
self.parse_node(true, false)
|
2015-05-27 08:35:13 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-05-24 06:27:42 +00:00
|
|
|
fn block_sequence_entry(&mut self, first: bool) -> ParseResult {
|
|
|
|
// BLOCK-SEQUENCE-START
|
|
|
|
if first {
|
2018-09-16 06:58:48 +00:00
|
|
|
let _ = self.peek_token()?;
|
2015-05-24 06:27:42 +00:00
|
|
|
//self.marks.push(tok.0);
|
|
|
|
self.skip();
|
|
|
|
}
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, TokenType::BlockEnd) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.pop_state();
|
|
|
|
self.skip();
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::SequenceEnd, mark))
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::BlockEntry) => {
|
2015-05-24 06:27:42 +00:00
|
|
|
self.skip();
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Token(mark, TokenType::BlockEntry | TokenType::BlockEnd) =
|
|
|
|
*self.peek_token()?
|
|
|
|
{
|
|
|
|
self.state = State::BlockSequenceEntry;
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
|
|
|
} else {
|
|
|
|
self.push_state(State::BlockSequenceEntry);
|
|
|
|
self.parse_node(true, false)
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
2018-09-15 16:49:04 +00:00
|
|
|
Token(mark, _) => Err(ScanError::new(
|
|
|
|
mark,
|
|
|
|
"while parsing a block collection, did not find expected '-' indicator",
|
|
|
|
)),
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2015-05-27 15:18:38 +00:00
|
|
|
fn flow_sequence_entry_mapping_key(&mut self) -> ParseResult {
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Token(mark, TokenType::Value | TokenType::FlowEntry | TokenType::FlowSequenceEnd) =
|
|
|
|
*self.peek_token()?
|
|
|
|
{
|
|
|
|
self.skip();
|
|
|
|
self.state = State::FlowSequenceEntryMappingValue;
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
|
|
|
} else {
|
|
|
|
self.push_state(State::FlowSequenceEntryMappingValue);
|
|
|
|
self.parse_node(false, false)
|
2015-05-27 15:18:38 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
|
|
|
|
fn flow_sequence_entry_mapping_value(&mut self) -> ParseResult {
|
2018-09-16 06:58:48 +00:00
|
|
|
match *self.peek_token()? {
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(_, TokenType::Value) => {
|
2018-09-15 16:49:04 +00:00
|
|
|
self.skip();
|
|
|
|
self.state = State::FlowSequenceEntryMappingValue;
|
2023-08-11 23:54:46 +00:00
|
|
|
if let Token(mark, TokenType::FlowEntry | TokenType::FlowSequenceEnd) =
|
|
|
|
*self.peek_token()?
|
|
|
|
{
|
|
|
|
self.state = State::FlowSequenceEntryMappingEnd;
|
|
|
|
Ok((Event::empty_scalar(), mark))
|
|
|
|
} else {
|
|
|
|
self.push_state(State::FlowSequenceEntryMappingEnd);
|
|
|
|
self.parse_node(false, false)
|
2018-09-15 16:49:04 +00:00
|
|
|
}
|
|
|
|
}
|
2017-06-20 13:47:19 +00:00
|
|
|
Token(mark, _) => {
|
2015-05-27 15:18:38 +00:00
|
|
|
self.state = State::FlowSequenceEntryMappingEnd;
|
2017-06-20 13:47:19 +00:00
|
|
|
Ok((Event::empty_scalar(), mark))
|
2015-05-27 15:18:38 +00:00
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
|
2023-08-11 23:54:46 +00:00
|
|
|
#[allow(clippy::unnecessary_wraps)]
|
2015-05-27 15:18:38 +00:00
|
|
|
fn flow_sequence_entry_mapping_end(&mut self) -> ParseResult {
|
|
|
|
self.state = State::FlowSequenceEntry;
|
2016-11-09 12:04:48 +00:00
|
|
|
Ok((Event::MappingEnd, self.scanner.mark()))
|
2015-05-27 15:18:38 +00:00
|
|
|
}
|
2023-11-19 15:00:19 +00:00
|
|
|
|
|
|
|
/// Resolve a tag from the handle and the suffix.
|
|
|
|
fn resolve_tag(&self, mark: Marker, handle: &str, suffix: String) -> Result<Tag, ScanError> {
|
|
|
|
if handle == "!!" {
|
|
|
|
// "!!" is a shorthand for "tag:yaml.org,2002:".
|
|
|
|
Ok(Tag {
|
|
|
|
handle: "tag:yaml.org,2002:".to_string(),
|
|
|
|
suffix,
|
|
|
|
})
|
|
|
|
} else if handle.is_empty() && suffix == "!" {
|
|
|
|
// "!" is a shorthand for "whatever would be the default". However, that
|
|
|
|
// default can be overridden.
|
|
|
|
match self.tags.get("") {
|
|
|
|
Some(prefix) => Ok(Tag {
|
|
|
|
handle: prefix.to_string(),
|
|
|
|
suffix,
|
|
|
|
}),
|
|
|
|
None => Ok(Tag {
|
|
|
|
handle: String::new(),
|
|
|
|
suffix,
|
|
|
|
}),
|
|
|
|
}
|
|
|
|
} else {
|
|
|
|
// Lookup handle in our tag directives.
|
|
|
|
let prefix = self.tags.get(handle);
|
|
|
|
if let Some(prefix) = prefix {
|
|
|
|
Ok(Tag {
|
|
|
|
handle: prefix.to_string(),
|
|
|
|
suffix,
|
|
|
|
})
|
|
|
|
} else {
|
|
|
|
// Otherwise, it may be a local handle. With a local handle, the handle is set to
|
|
|
|
// "!" and the suffix to whatever follows it ("!foo" -> ("!", "foo")).
|
|
|
|
// If the handle is of the form "!foo!", this cannot be a local handle and we need
|
|
|
|
// to error.
|
|
|
|
if handle.len() >= 2 && handle.starts_with('!') && handle.ends_with('!') {
|
|
|
|
Err(ScanError::new(mark, "the handle wasn't declared"))
|
|
|
|
} else {
|
|
|
|
Ok(Tag {
|
|
|
|
handle: handle.to_string(),
|
|
|
|
suffix,
|
|
|
|
})
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
|
|
|
}
|
2015-05-24 06:27:42 +00:00
|
|
|
}
|
2017-06-22 06:44:46 +00:00
|
|
|
|
|
|
|
#[cfg(test)]
|
|
|
|
mod test {
|
|
|
|
use super::{Event, Parser};
|
|
|
|
|
|
|
|
#[test]
|
|
|
|
fn test_peek_eq_parse() {
|
|
|
|
let s = "
|
|
|
|
a0 bb: val
|
|
|
|
a1: &x
|
|
|
|
b1: 4
|
|
|
|
b2: d
|
|
|
|
a2: 4
|
|
|
|
a3: [1, 2, 3]
|
|
|
|
a4:
|
|
|
|
- [a1, a2]
|
|
|
|
- 2
|
|
|
|
a5: *x
|
|
|
|
";
|
|
|
|
let mut p = Parser::new(s.chars());
|
|
|
|
while {
|
|
|
|
let event_peek = p.peek().unwrap().clone();
|
|
|
|
let event = p.next().unwrap();
|
|
|
|
assert_eq!(event, event_peek);
|
|
|
|
event.0 != Event::StreamEnd
|
|
|
|
} {}
|
|
|
|
}
|
|
|
|
}
|