lol_html/parser/lexer/

actions.rs

use super::*;
use crate::parser::ActionError;
use crate::parser::state_machine::StateMachineActions;

use NonTagContentTokenOutline::*;
use TagTokenOutline::{EndTag, StartTag};

// NOTE: use macro instead of the function to make borrow
// checker happy with range construction inside match arm
// with a mutable borrow of lexer.
macro_rules! get_token_part_range {
    ($self:tt) => {
        Range {
            start: $self.token_part_start,
            end: $self.next_pos - 1,
        }
    };
}

impl<S: LexemeSink> Lexer<S> {
    fn emit_eof(&mut self, context: &mut ParserContext<S>, input: &[u8]) -> ActionResult {
        let lexeme = self.create_lexeme_with_raw_exclusive(
            context.previously_consumed_byte_count,
            input,
            Some(Eof),
        );

        self.emit_lexeme(context, &lexeme)
    }
}

impl<S: LexemeSink> StateMachineActions for Lexer<S> {
    type Context = ParserContext<S>;

    impl_common_sm_actions!();

    fn emit_text(&mut self, context: &mut ParserContext<S>, input: &[u8]) -> ActionResult {
        if self.pos() > self.lexeme_start {
            // NOTE: unlike any other tokens (except EOF), text tokens don't have
            // any lexical symbols that determine their bounds. Therefore,
            // representation of text token content is the raw slice.
            // Also, we always emit text if we encounter some other bounded
            // lexical structure and, thus, we use exclusive range for the raw slice.
            let lexeme = self.create_lexeme_with_raw_exclusive(
                context.previously_consumed_byte_count,
                input,
                Some(Text(self.last_text_type)),
            );

            self.emit_lexeme(context, &lexeme)?;
        }

        Ok(())
    }

    #[inline(never)]
    fn emit_text_and_eof(&mut self, context: &mut ParserContext<S>, input: &[u8]) -> ActionResult {
        self.emit_text(context, input)?;
        self.emit_eof(context, input)
    }

    #[inline(never)]
    fn emit_current_token(&mut self, context: &mut ParserContext<S>, input: &[u8]) -> ActionResult {
        let token = self.current_non_tag_content_token.take();
        let lexeme = self.create_lexeme_with_raw_inclusive(
            context.previously_consumed_byte_count,
            input,
            token,
        );

        self.emit_lexeme(context, &lexeme)
    }

    #[inline(never)]
    fn emit_tag(&mut self, context: &mut ParserContext<S>, input: &[u8]) -> ActionResult {
        let token = self
            .current_tag_token
            .take()
            .ok_or_else(|| ActionError::internal("Tag token should exist at this point"))?;

        let feedback = self.try_get_tree_builder_feedback(context, &token)?;

        let mut lexeme = self.create_lexeme_with_raw_inclusive(
            context.previously_consumed_byte_count,
            input,
            token,
        );

        // NOTE: exit from any non-initial text parsing mode always happens on tag emission
        // (except for CDATA, but there is a special action to take care of it).
        self.set_last_text_type(TextType::Data);

        if let Some(feedback) = feedback {
            self.handle_tree_builder_feedback(context, feedback, &lexeme);
        }

        if let StartTag {
            ref mut ns,
            name_hash,
            ..
        } = lexeme.token_outline
        {
            self.last_start_tag_name_hash = name_hash;
            *ns = context.tree_builder_simulator.current_ns();
        }

        match self.emit_tag_lexeme(context, &lexeme)? {
            ParserDirective::Lex => Ok(()),
            ParserDirective::WherePossibleScanForTagsOnly => self.change_parser_directive(
                self.lexeme_start,
                ParserDirective::WherePossibleScanForTagsOnly,
                FeedbackDirective::None,
            ),
        }
    }

    #[inline(never)]
    fn emit_current_token_and_eof(
        &mut self,
        context: &mut ParserContext<S>,
        input: &[u8],
    ) -> ActionResult {
        let token = self.current_non_tag_content_token.take();
        let lexeme = self.create_lexeme_with_raw_exclusive(
            context.previously_consumed_byte_count,
            input,
            token,
        );

        self.emit_lexeme(context, &lexeme)?;
        self.emit_eof(context, input)
    }

    /// Emits `<[CDATA[` and such.
    #[inline(never)]
    fn emit_raw_without_token(
        &mut self,
        context: &mut ParserContext<S>,
        input: &[u8],
    ) -> ActionResult {
        let lexeme = self.create_lexeme_with_raw_inclusive(
            context.previously_consumed_byte_count,
            input,
            None,
        );

        self.emit_lexeme(context, &lexeme)
    }

    #[inline(never)]
    fn emit_raw_without_token_and_eof(
        &mut self,
        context: &mut ParserContext<S>,
        input: &[u8],
    ) -> ActionResult {
        // NOTE: since we are at EOF we use exclusive range for token's raw.
        let lexeme = self.create_lexeme_with_raw_exclusive(
            context.previously_consumed_byte_count,
            input,
            None,
        );

        self.emit_lexeme(context, &lexeme)?;
        self.emit_eof(context, input)
    }

    #[inline]
    fn create_start_tag(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        self.current_tag_token = Some(StartTag {
            name: Range::default(),
            name_hash: LocalNameHash::new(),
            ns: Namespace::default(),
            attributes: Vec::new(),
            self_closing: false,
        });
    }

    #[inline]
    fn create_end_tag(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        self.current_tag_token = Some(EndTag {
            name: Range::default(),
            name_hash: LocalNameHash::new(),
        });
    }

    #[cold]
    fn create_doctype(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        self.current_non_tag_content_token = Some(Doctype(Box::new(DoctypeTokenOutline {
            name: None,
            public_id: None,
            system_id: None,
            force_quirks: false,
        })));
    }

    #[inline]
    fn create_comment(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        self.current_non_tag_content_token = Some(Comment(Range::default()));
    }

    #[inline]
    fn start_token_part(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        self.token_part_start = self.pos();
    }

    #[inline]
    fn mark_comment_text_end(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(Comment(ref mut text)) = self.current_non_tag_content_token {
            *text = get_token_part_range!(self);
        }
    }

    #[inline]
    fn shift_comment_text_end_by(
        &mut self,
        _context: &mut ParserContext<S>,
        _input: &[u8],
        offset: usize,
    ) {
        if let Some(Comment(ref mut text)) = self.current_non_tag_content_token {
            text.end += offset;
        }
    }

    #[inline]
    fn set_force_quirks(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(Doctype(doctype)) = &mut self.current_non_tag_content_token {
            doctype.force_quirks = true;
        }
    }

    #[inline]
    fn finish_doctype_name(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(Doctype(doctype)) = &mut self.current_non_tag_content_token {
            doctype.name = Some(get_token_part_range!(self));
        }
    }

    #[inline]
    fn finish_doctype_public_id(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(Doctype(doctype)) = &mut self.current_non_tag_content_token {
            doctype.public_id = Some(get_token_part_range!(self));
        }
    }

    #[inline]
    fn finish_doctype_system_id(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(Doctype(doctype)) = &mut self.current_non_tag_content_token {
            doctype.system_id = Some(get_token_part_range!(self));
        }
    }

    #[inline]
    fn finish_tag_name(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) -> ActionResult {
        match self.current_tag_token {
            Some(StartTag { ref mut name, .. } | EndTag { ref mut name, .. }) => {
                *name = get_token_part_range!(self);
            }
            _ => return Err(ActionError::internal("Tag should exist at this point")),
        }

        Ok(())
    }

    #[inline]
    fn update_tag_name_hash(&mut self, _context: &mut ParserContext<S>, input: &[u8]) {
        if let Some(ch) = input.get(self.pos()).copied() {
            match self.current_tag_token {
                Some(
                    StartTag {
                        ref mut name_hash, ..
                    }
                    | EndTag {
                        ref mut name_hash, ..
                    },
                ) => name_hash.update(ch),
                _ => debug_assert!(false, "Tag should exist at this point"),
            }
        }
    }

    #[inline]
    fn mark_as_self_closing(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(StartTag {
            ref mut self_closing,
            ..
        }) = self.current_tag_token
        {
            *self_closing = true;
        }
    }

    #[inline]
    fn start_attr(&mut self, context: &mut ParserContext<S>, input: &[u8]) {
        // NOTE: create attribute only if we are parsing a start tag
        if let Some(StartTag { .. }) = self.current_tag_token {
            self.current_attr = Some(AttributeOutline::default());

            self.start_token_part(context, input);
        }
    }

    #[inline]
    fn finish_attr_name(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(AttributeOutline {
            ref mut name,
            ref mut raw_range,
            ..
        }) = self.current_attr
        {
            *name = get_token_part_range!(self);
            *raw_range = *name;
        }
    }

    #[inline]
    fn finish_attr_value(&mut self, _context: &mut ParserContext<S>, input: &[u8]) {
        if let Some(AttributeOutline {
            ref mut value,
            ref mut raw_range,
            ..
        }) = self.current_attr
        {
            *value = get_token_part_range!(self);

            // NOTE: include closing quote into the raw value if it's present
            raw_range.end = match input.get(self.next_pos - 1).copied() {
                Some(ch) if ch == self.closing_quote => value.end + 1,
                _ => value.end,
            };
        }
    }

    #[inline]
    fn finish_attr(&mut self, _context: &mut ParserContext<S>, _input: &[u8]) {
        if let Some(attr) = self.current_attr.take() {
            if let Some(StartTag { attributes, .. }) = self.current_tag_token.as_mut() {
                attributes.push(attr);
            }
        }
    }

    noop_action!(mark_tag_start, unmark_tag_start);
}