Trait DataStack

Source

pub trait DataStack: Sized + Default {
    type Term;
    type NonTerm: NonTerminal;
    type UserData;
    type ReduceActionError;
    type StartType;
    type Location: Location;

    // Required methods
    fn pop_start(&mut self) -> Option<Self::StartType>;
    fn pop(&mut self);
    fn push_terminal(&mut self, term: Self::Term);
    fn push_empty(&mut self);
    fn clear(&mut self);
    fn reserve(&mut self, additional: usize);
    fn split_off(&mut self, at: usize) -> Self;
    fn append(&mut self, other: &mut Self);
    fn reduce_action(
        data_stack: &mut Self,
        location_stack: &mut Vec<Self::Location>,
        push_data: bool,
        rule_index: usize,
        shift: &mut bool,
        lookahead: &TerminalSymbol<Self::Term>,
        userdata: &mut Self::UserData,
        location0: &mut Self::Location,
    ) -> Result<(), Self::ReduceActionError>;

    // Provided method
    fn with_capacity(capacity: usize) -> Self { ... }
}

Expand description

A trait for data stack in the parser.

Since each non-terminal could have different ruletypes, this effectively handles those rule types into separated Vec stack, instead of using enum of rule types (since it would be costful at memory aspects if the size differs significantly). For people who is curious about the implementation details, you should see the actual generated DataStack structs, like GrammarDataStack in rusty_lr_parser/src/parser/parser_expanded.rs.