1
0

Compare commits

...

3 Commits

Author SHA1 Message Date
e295aebf56 Add examples 2024-08-07 12:13:49 -04:00
f64babc546 Add more analysis to the analyzer 2024-08-07 11:57:15 -04:00
7328467e64 Add analysis step to run function 2024-08-07 11:38:08 -04:00
7 changed files with 403 additions and 195 deletions

View File

@ -1,5 +1,5 @@
use std::{
collections::HashMap,
collections::{HashMap, VecDeque},
fmt::{self, Display, Formatter},
};
@ -7,6 +7,11 @@ use serde::{Deserialize, Serialize};
use crate::{Identifier, ReservedIdentifier, Span, Type, Value};
#[derive(Debug, Clone, Eq, PartialEq, PartialOrd, Ord, Serialize, Deserialize)]
pub struct AbstractSyntaxTree {
pub nodes: VecDeque<Node>,
}
#[derive(Debug, Clone, Eq, PartialEq, PartialOrd, Ord, Serialize, Deserialize)]
pub struct Node {
pub statement: Statement,

View File

@ -1,22 +1,67 @@
use crate::{Node, Statement};
/// Tools for analyzing an abstract syntax tree and catch errors before running the virtual
/// machine.
///
/// This module provides to anlysis options, both of which borrow an abstract syntax tree and a
/// hash map of variables:
/// - `analyze` convenience function
/// - `Analyzer` struct
use std::collections::HashMap;
pub fn analyze(abstract_tree: Vec<Node>) -> Result<(), AnalyzerError> {
let analyzer = Analyzer::new(abstract_tree);
use crate::{AbstractSyntaxTree, Identifier, Node, Statement, Type, Value};
/// Analyzes the abstract syntax tree for errors.
///
/// # Examples
/// ```
/// # use std::collections::HashMap;
/// # use dust_lang::*;
/// let input = "x = 1 + false";
/// let abstract_tree = parse(input).unwrap();
/// let variables = HashMap::new();
/// let result = analyze(&abstract_tree, &variables);
///
/// assert!(result.is_err());
/// ```
pub fn analyze(
abstract_tree: &AbstractSyntaxTree,
variables: &HashMap<Identifier, Value>,
) -> Result<(), AnalyzerError> {
let analyzer = Analyzer::new(abstract_tree, variables);
analyzer.analyze()
}
pub struct Analyzer {
abstract_tree: Vec<Node>,
/// Static analyzer that checks for potential runtime errors.
///
/// # Examples
/// ```
/// # use std::collections::HashMap;
/// # use dust_lang::*;
/// let input = "x = 1 + false";
/// let abstract_tree = parse(input).unwrap();
/// let variables = HashMap::new();
/// let analyzer = Analyzer::new(&abstract_tree, &variables);
/// let result = analyzer.analyze();
///
/// assert!(result.is_err());
pub struct Analyzer<'a> {
abstract_tree: &'a AbstractSyntaxTree,
variables: &'a HashMap<Identifier, Value>,
}
impl Analyzer {
pub fn new(abstract_tree: Vec<Node>) -> Self {
Analyzer { abstract_tree }
impl<'a> Analyzer<'a> {
pub fn new(
abstract_tree: &'a AbstractSyntaxTree,
variables: &'a HashMap<Identifier, Value>,
) -> Self {
Self {
abstract_tree,
variables,
}
}
pub fn analyze(&self) -> Result<(), AnalyzerError> {
for node in &self.abstract_tree {
for node in &self.abstract_tree.nodes {
self.analyze_node(node)?;
}
@ -26,6 +71,24 @@ impl Analyzer {
fn analyze_node(&self, node: &Node) -> Result<(), AnalyzerError> {
match &node.statement {
Statement::Add(left, right) => {
if let Some(Type::Integer) | Some(Type::Float) =
left.statement.expected_type(self.variables)
{
} else {
return Err(AnalyzerError::ExpectedIntegerOrFloat {
actual: left.as_ref().clone(),
});
}
if let Some(Type::Integer) | Some(Type::Float) =
right.statement.expected_type(self.variables)
{
} else {
return Err(AnalyzerError::ExpectedIntegerOrFloat {
actual: right.as_ref().clone(),
});
}
self.analyze_node(left)?;
self.analyze_node(right)?;
}
@ -55,12 +118,32 @@ impl Analyzer {
}
}
Statement::Multiply(left, right) => {
if let Some(Type::Integer) | Some(Type::Float) =
left.statement.expected_type(self.variables)
{
} else {
return Err(AnalyzerError::ExpectedIntegerOrFloat {
actual: left.as_ref().clone(),
});
}
if let Some(Type::Integer) | Some(Type::Float) =
right.statement.expected_type(self.variables)
{
} else {
return Err(AnalyzerError::ExpectedIntegerOrFloat {
actual: right.as_ref().clone(),
});
}
self.analyze_node(left)?;
self.analyze_node(right)?;
}
Statement::PropertyAccess(left, right) => {
if let Statement::Identifier(_) = &left.statement {
// Identifier is in the correct position
if let Statement::Identifier(_) | Statement::Constant(_) | Statement::List(_) =
&left.statement
{
// Left side is valid
} else {
return Err(AnalyzerError::ExpectedIdentifier {
actual: left.as_ref().clone(),
@ -79,6 +162,7 @@ impl Analyzer {
#[derive(Clone, Debug, PartialEq)]
pub enum AnalyzerError {
ExpectedIdentifier { actual: Node },
ExpectedIntegerOrFloat { actual: Node },
UnexpectedIdentifier { identifier: Node },
}
@ -89,16 +173,71 @@ mod tests {
use super::*;
#[test]
fn assignment_expect_identifier() {
let abstract_tree = vec![Node::new(
Statement::Assign(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(Statement::Constant(Value::integer(2)), (1, 2))),
),
(0, 2),
)];
fn multiply_expect_integer_or_float() {
let abstract_tree = AbstractSyntaxTree {
nodes: [Node::new(
Statement::Multiply(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(
Statement::Constant(Value::boolean(false)),
(1, 2),
)),
),
(0, 2),
)]
.into(),
};
let variables = HashMap::new();
let analyzer = Analyzer::new(&abstract_tree, &variables);
let analyzer = Analyzer::new(abstract_tree);
assert_eq!(
analyzer.analyze(),
Err(AnalyzerError::ExpectedIntegerOrFloat {
actual: Node::new(Statement::Constant(Value::boolean(false)), (1, 2))
})
)
}
#[test]
fn add_expect_integer_or_float() {
let abstract_tree = AbstractSyntaxTree {
nodes: [Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(
Statement::Constant(Value::boolean(false)),
(1, 2),
)),
),
(0, 2),
)]
.into(),
};
let variables = HashMap::new();
let analyzer = Analyzer::new(&abstract_tree, &variables);
assert_eq!(
analyzer.analyze(),
Err(AnalyzerError::ExpectedIntegerOrFloat {
actual: Node::new(Statement::Constant(Value::boolean(false)), (1, 2))
})
)
}
#[test]
fn assignment_expect_identifier() {
let abstract_tree = AbstractSyntaxTree {
nodes: [Node::new(
Statement::Assign(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(Statement::Constant(Value::integer(2)), (1, 2))),
),
(0, 2),
)]
.into(),
};
let variables = HashMap::new();
let analyzer = Analyzer::new(&abstract_tree, &variables);
assert_eq!(
analyzer.analyze(),
@ -109,13 +248,16 @@ mod tests {
}
#[test]
fn unexpected_identifier_simple() {
let abstract_tree = vec![Node::new(
Statement::Identifier(Identifier::new("x")),
(0, 1),
)];
let analyzer = Analyzer::new(abstract_tree);
fn unexpected_identifier() {
let abstract_tree = AbstractSyntaxTree {
nodes: [Node::new(
Statement::Identifier(Identifier::new("x")),
(0, 1),
)]
.into(),
};
let variables = HashMap::new();
let analyzer = Analyzer::new(&abstract_tree, &variables);
assert_eq!(
analyzer.analyze(),
@ -124,27 +266,4 @@ mod tests {
})
)
}
#[test]
fn unexpected_identifier_nested() {
let abstract_tree = vec![Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(
Statement::Identifier(Identifier::new("x")),
(1, 2),
)),
),
(0, 1),
)];
let analyzer = Analyzer::new(abstract_tree);
assert_eq!(
analyzer.analyze(),
Err(AnalyzerError::UnexpectedIdentifier {
identifier: Node::new(Statement::Identifier(Identifier::new("x")), (1, 2))
})
)
}
}

View File

@ -8,6 +8,25 @@ use std::num::{ParseFloatError, ParseIntError};
use crate::{Identifier, ReservedIdentifier, Span, Token};
/// Lex the input and return a vector of tokens and their positions.
///
/// # Examples
/// ```
/// # use dust_lang::*;
/// let input = "x = 1 + 2";
/// let tokens = lex(input).unwrap();
///
/// assert_eq!(
/// tokens,
/// [
/// (Token::Identifier(Identifier::new("x")), (0, 1)),
/// (Token::Equal, (2, 3)),
/// (Token::Integer(1), (4, 5)),
/// (Token::Plus, (6, 7)),
/// (Token::Integer(2), (8, 9)),
/// (Token::Eof, (9, 9)),
/// ]
/// );
/// ```
pub fn lex(input: &str) -> Result<Vec<(Token, Span)>, LexError> {
let mut lexer = Lexer::new(input);
let mut tokens = Vec::new();
@ -28,6 +47,37 @@ pub fn lex(input: &str) -> Result<Vec<(Token, Span)>, LexError> {
#[derive(Debug, Clone)]
/// Low-level tool for lexing a single token at a time.
///
/// # Examples
/// ```
/// # use dust_lang::*;
/// let input = "x = 1 + 2";
/// let mut lexer = Lexer::new(input);
/// let mut tokens = Vec::new();
///
/// loop {
/// let (token, span) = lexer.next_token().unwrap();
/// let is_eof = matches!(token, Token::Eof);
///
/// tokens.push((token, span));
///
/// if is_eof {
/// break;
/// }
/// }
///
/// assert_eq!(
/// tokens,
/// [
/// (Token::Identifier(Identifier::new("x")), (0, 1)),
/// (Token::Equal, (2, 3)),
/// (Token::Integer(1), (4, 5)),
/// (Token::Plus, (6, 7)),
/// (Token::Integer(2), (8, 9)),
/// (Token::Eof, (9, 9)),
/// ]
/// )
/// ```
pub struct Lexer<'a> {
source: &'a str,
position: usize,

View File

@ -15,7 +15,7 @@ pub mod r#type;
pub mod value;
pub mod vm;
pub use abstract_tree::{Node, Statement};
pub use abstract_tree::{AbstractSyntaxTree, Node, Statement};
pub use analyzer::{analyze, Analyzer, AnalyzerError};
pub use identifier::Identifier;
pub use lex::{lex, LexError, Lexer};

View File

@ -1,11 +1,8 @@
use std::collections::VecDeque;
use crate::{
lex::{LexError, Lexer},
Node, Span, Statement, Token, Value,
};
use crate::{AbstractSyntaxTree, LexError, Lexer, Node, Span, Statement, Token, Value};
pub fn parse(input: &str) -> Result<VecDeque<Node>, ParseError> {
pub fn parse(input: &str) -> Result<AbstractSyntaxTree, ParseError> {
let lexer = Lexer::new(input);
let mut parser = Parser::new(lexer);
let mut nodes = VecDeque::new();
@ -20,7 +17,7 @@ pub fn parse(input: &str) -> Result<VecDeque<Node>, ParseError> {
}
}
Ok(nodes)
Ok(AbstractSyntaxTree { nodes })
}
pub struct Parser<'src> {
@ -227,7 +224,9 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(Statement::Constant(Value::boolean(true)), (0, 4))].into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(Statement::Constant(Value::boolean(true)), (0, 4))].into()
})
);
}
@ -237,21 +236,23 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::PropertyAccess(
Box::new(Node::new(
Statement::List(vec![
Node::new(Statement::Constant(Value::integer(1)), (1, 2)),
Node::new(Statement::Constant(Value::integer(2)), (4, 5)),
Node::new(Statement::Constant(Value::integer(3)), (7, 8)),
]),
(0, 9)
)),
Box::new(Node::new(Statement::Constant(Value::integer(0)), (10, 11))),
),
(0, 11),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::PropertyAccess(
Box::new(Node::new(
Statement::List(vec![
Node::new(Statement::Constant(Value::integer(1)), (1, 2)),
Node::new(Statement::Constant(Value::integer(2)), (4, 5)),
Node::new(Statement::Constant(Value::integer(3)), (7, 8)),
]),
(0, 9)
)),
Box::new(Node::new(Statement::Constant(Value::integer(0)), (10, 11))),
),
(0, 11),
)]
.into()
})
);
}
@ -261,20 +262,22 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::PropertyAccess(
Box::new(Node::new(
Statement::Identifier(Identifier::new("a")),
(0, 1)
)),
Box::new(Node::new(
Statement::Identifier(Identifier::new("b")),
(2, 3)
)),
),
(0, 3),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::PropertyAccess(
Box::new(Node::new(
Statement::Identifier(Identifier::new("a")),
(0, 1)
)),
Box::new(Node::new(
Statement::Identifier(Identifier::new("b")),
(2, 3)
)),
),
(0, 3),
)]
.into()
})
);
}
@ -284,39 +287,44 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::List(vec![
Node::new(Statement::Constant(Value::integer(1)), (1, 2)),
Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (4, 5))),
Box::new(Node::new(Statement::Constant(Value::integer(1)), (8, 9))),
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::List(vec![
Node::new(Statement::Constant(Value::integer(1)), (1, 2)),
Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (4, 5))),
Box::new(Node::new(Statement::Constant(Value::integer(1)), (8, 9))),
),
(4, 9),
),
(4, 9),
),
Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(2)), (11, 12))),
Box::new(Node::new(
Statement::Multiply(
Box::new(Node::new(
Statement::Constant(Value::integer(4)),
(16, 17)
)),
Box::new(Node::new(
Statement::Constant(Value::integer(10)),
(20, 22)
)),
),
(15, 23),
),),
Node::new(
Statement::Add(
Box::new(Node::new(
Statement::Constant(Value::integer(2)),
(11, 12)
)),
Box::new(Node::new(
Statement::Multiply(
Box::new(Node::new(
Statement::Constant(Value::integer(4)),
(16, 17)
)),
Box::new(Node::new(
Statement::Constant(Value::integer(10)),
(20, 22)
)),
),
(15, 23),
),),
),
(11, 23),
),
(11, 23),
),
]),
(0, 24),
)]
.into())
]),
(0, 24),
)]
.into()
})
);
}
@ -326,14 +334,16 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::List(vec![
Node::new(Statement::Constant(Value::integer(1)), (1, 2)),
Node::new(Statement::Constant(Value::integer(2)), (4, 5)),
]),
(0, 6),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::List(vec![
Node::new(Statement::Constant(Value::integer(1)), (1, 2)),
Node::new(Statement::Constant(Value::integer(2)), (4, 5)),
]),
(0, 6),
)]
.into()
})
);
}
@ -343,7 +353,9 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(Statement::List(vec![]), (0, 2))].into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(Statement::List(vec![]), (0, 2))].into()
})
);
}
@ -353,7 +365,9 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(Statement::Constant(Value::float(42.0)), (0, 4))].into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(Statement::Constant(Value::float(42.0)), (0, 4))].into()
})
);
}
@ -363,14 +377,16 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(Statement::Constant(Value::integer(2)), (4, 5))),
),
(0, 5),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(Statement::Constant(Value::integer(2)), (4, 5))),
),
(0, 5),
)]
.into()
})
);
}
@ -380,14 +396,16 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::Multiply(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(Statement::Constant(Value::integer(2)), (4, 5))),
),
(0, 5),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::Multiply(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(Statement::Constant(Value::integer(2)), (4, 5))),
),
(0, 5),
)]
.into()
})
);
}
@ -397,20 +415,22 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(
Statement::Multiply(
Box::new(Node::new(Statement::Constant(Value::integer(2)), (4, 5))),
Box::new(Node::new(Statement::Constant(Value::integer(3)), (8, 9))),
),
(4, 9),
)),
),
(0, 9),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (0, 1))),
Box::new(Node::new(
Statement::Multiply(
Box::new(Node::new(Statement::Constant(Value::integer(2)), (4, 5))),
Box::new(Node::new(Statement::Constant(Value::integer(3)), (8, 9))),
),
(4, 9),
)),
),
(0, 9),
)]
.into()
})
);
}
@ -420,35 +440,37 @@ mod tests {
assert_eq!(
parse(input),
Ok([Node::new(
Statement::Assign(
Box::new(Node::new(
Statement::Identifier(Identifier::new("a")),
(0, 1)
)),
Box::new(Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (4, 5))),
Box::new(Node::new(
Statement::Multiply(
Box::new(Node::new(
Statement::Constant(Value::integer(2)),
(8, 9)
)),
Box::new(Node::new(
Statement::Constant(Value::integer(3)),
(12, 13)
)),
),
(8, 13),
)),
),
(4, 13),
)),
),
(0, 13),
)]
.into())
Ok(AbstractSyntaxTree {
nodes: [Node::new(
Statement::Assign(
Box::new(Node::new(
Statement::Identifier(Identifier::new("a")),
(0, 1)
)),
Box::new(Node::new(
Statement::Add(
Box::new(Node::new(Statement::Constant(Value::integer(1)), (4, 5))),
Box::new(Node::new(
Statement::Multiply(
Box::new(Node::new(
Statement::Constant(Value::integer(2)),
(8, 9)
)),
Box::new(Node::new(
Statement::Constant(Value::integer(3)),
(12, 13)
)),
),
(8, 13),
)),
),
(4, 13),
)),
),
(0, 13),
)]
.into()
})
);
}
}

View File

@ -18,7 +18,7 @@ use std::{
use serde::{Deserialize, Serialize};
use crate::identifier::Identifier;
use crate::Identifier;
#[derive(Clone, Debug, Eq, PartialEq, Serialize, Deserialize)]
pub struct TypeConflict {

View File

@ -1,7 +1,8 @@
use std::collections::{HashMap, VecDeque};
use std::collections::HashMap;
use crate::{
parse, Identifier, Node, ParseError, ReservedIdentifier, Span, Statement, Value, ValueError,
parse, AbstractSyntaxTree, Analyzer, AnalyzerError, Identifier, Node, ParseError,
ReservedIdentifier, Span, Statement, Value, ValueError,
};
pub fn run(
@ -9,18 +10,22 @@ pub fn run(
variables: &mut HashMap<Identifier, Value>,
) -> Result<Option<Value>, VmError> {
let abstract_syntax_tree = parse(input)?;
let analyzer = Analyzer::new(&abstract_syntax_tree, variables);
analyzer.analyze()?;
let mut vm = Vm::new(abstract_syntax_tree);
vm.run(variables)
}
pub struct Vm {
statement_nodes: VecDeque<Node>,
abstract_tree: AbstractSyntaxTree,
}
impl Vm {
pub fn new(statement_nodes: VecDeque<Node>) -> Self {
Vm { statement_nodes }
pub fn new(abstract_tree: AbstractSyntaxTree) -> Self {
Self { abstract_tree }
}
pub fn run(
@ -29,7 +34,7 @@ impl Vm {
) -> Result<Option<Value>, VmError> {
let mut previous_value = None;
while let Some(node) = self.statement_nodes.pop_front() {
while let Some(node) = self.abstract_tree.nodes.pop_front() {
previous_value = self.run_node(node, variables)?;
}
@ -166,6 +171,7 @@ impl Vm {
#[derive(Clone, Debug, PartialEq)]
pub enum VmError {
AnaylyzerError(AnalyzerError),
ParseError(ParseError),
ValueError(ValueError),
@ -178,6 +184,12 @@ pub enum VmError {
ExpectedValue { position: Span },
}
impl From<AnalyzerError> for VmError {
fn from(error: AnalyzerError) -> Self {
Self::AnaylyzerError(error)
}
}
impl From<ParseError> for VmError {
fn from(error: ParseError) -> Self {
Self::ParseError(error)