Changeset - 814be01095d9
[Not reviewed]
0 5 3
mh - 4 years ago 2021-01-07 19:12:05
contact@maxhenger.nl
implement module/version pragma, WIP on visitor rewrite
8 files changed with 158 insertions and 24 deletions:
0 comments (0 inline, 0 general)
src/lib.rs
Show inline comments
 
#[macro_use]
 
mod macros;
 

	
 
mod common;
 
mod protocol;
 
mod runtime;
 

	
 
pub use common::{ConnectorId, EndpointPolarity, Payload, Polarity, PortId};
 
pub use protocol::{ProtocolDescription, TRIVIAL_PD};
 
pub use runtime::{error, Connector, DummyLogger, FileLogger, VecLogger};
 

	
 
// TODO: Remove when not benchmarking
 
pub use protocol::inputsource::InputSource;
 
pub use protocol::ast::Heap;
 
pub use protocol::lexer::Lexer;
 

	
 
#[cfg(feature = "ffi")]
 
pub mod ffi;
src/protocol/ast.rs
Show inline comments
 
@@ -1201,104 +1201,126 @@ impl Index<DeclarationId> for Heap {
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct Root {
 
    pub this: RootId,
 
    // Phase 1: parser
 
    pub position: InputPosition,
 
    pub pragmas: Vec<PragmaId>,
 
    pub imports: Vec<ImportId>,
 
    pub definitions: Vec<DefinitionId>,
 
    // Pase 2: linker
 
    pub declarations: Vec<DeclarationId>,
 
}
 

	
 
impl Root {
 
    pub fn get_definition(&self, h: &Heap, id: IdentifierId) -> Option<DefinitionId> {
 
        for &def in self.definitions.iter() {
 
            if h[h[def].identifier()] == h[id] {
 
                return Some(def);
 
            }
 
        }
 
        None
 
    }
 
    pub fn get_definition_ident(&self, h: &Heap, id: &[u8]) -> Option<DefinitionId> {
 
        for &def in self.definitions.iter() {
 
            if h[h[def].identifier()].ident() == id {
 
                return Some(def);
 
            }
 
        }
 
        None
 
    }
 
    pub fn get_declaration(&self, h: &Heap, id: IdentifierId) -> Option<DeclarationId> {
 
        for &decl in self.declarations.iter() {
 
            if h[h[decl].identifier()] == h[id] {
 
                return Some(decl);
 
            }
 
        }
 
        None
 
    }
 
}
 

	
 
impl SyntaxElement for Root {
 
    fn position(&self) -> InputPosition {
 
        self.position
 
    }
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct Pragma {
 
pub enum Pragma {
 
    Version(PragmaVersion),
 
    Module(PragmaModule)
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct PragmaVersion {
 
    pub this: PragmaId,
 
    // Phase 1: parser
 
    pub position: InputPosition,
 
    pub version: u64,
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct PragmaModule {
 
    pub this: PragmaId,
 
    // Phase 1: parser
 
    pub position: InputPosition,
 
    pub value: Vec<u8>,
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct PragmaOld {
 
    pub this: PragmaId,
 
    // Phase 1: parser
 
    pub position: InputPosition,
 
    pub value: Vec<u8>,
 
}
 

	
 
impl SyntaxElement for Pragma {
 
impl SyntaxElement for PragmaOld {
 
    fn position(&self) -> InputPosition {
 
        self.position
 
    }
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct Import {
 
    pub this: ImportId,
 
    // Phase 1: parser
 
    pub position: InputPosition,
 
    pub value: Vec<u8>,
 
}
 

	
 
impl SyntaxElement for Import {
 
    fn position(&self) -> InputPosition {
 
        self.position
 
    }
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub enum Identifier {
 
    External(ExternalIdentifier),
 
    Source(SourceIdentifier),
 
}
 

	
 
impl Identifier {
 
    pub fn as_source(&self) -> &SourceIdentifier {
 
        match self {
 
            Identifier::Source(result) => result,
 
            _ => panic!("Unable to cast `Identifier` to `SourceIdentifier`"),
 
        }
 
    }
 
    pub fn is_external(&self) -> bool {
 
        match self {
 
            Identifier::External(_) => true,
 
            _ => false,
 
        }
 
    }
 
    pub fn as_external(&self) -> &ExternalIdentifier {
 
        match self {
 
            Identifier::External(result) => result,
 
            _ => panic!("Unable to cast `Identifier` to `ExternalIdentifier`"),
 
        }
 
    }
 
    fn ident(&self) -> &[u8] {
 
        match self {
 
            Identifier::External(eid) => eid.ident(),
 
            Identifier::Source(sid) => sid.ident(),
 
@@ -1436,97 +1458,97 @@ impl Display for Type {
 
                write!(f, "out")?;
 
            }
 
            PrimitiveType::Message => {
 
                write!(f, "msg")?;
 
            }
 
            PrimitiveType::Boolean => {
 
                write!(f, "boolean")?;
 
            }
 
            PrimitiveType::Byte => {
 
                write!(f, "byte")?;
 
            }
 
            PrimitiveType::Short => {
 
                write!(f, "short")?;
 
            }
 
            PrimitiveType::Int => {
 
                write!(f, "int")?;
 
            }
 
            PrimitiveType::Long => {
 
                write!(f, "long")?;
 
            }
 
            PrimitiveType::Symbolic(data) => {
 
                // Type data is in ASCII range.
 
                write!(f, "{}", String::from_utf8_lossy(&data))?;
 
            }
 
        }
 
        if self.array {
 
            write!(f, "[]")
 
        } else {
 
            Ok(())
 
        }
 
    }
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub struct TypeAnnotation {
 
    pub this: TypeAnnotationId,
 
    // Phase 1: parser
 
    pub position: InputPosition,
 
    pub the_type: Type,
 
}
 

	
 
impl SyntaxElement for TypeAnnotation {
 
    fn position(&self) -> InputPosition {
 
        self.position
 
    }
 
}
 

	
 
type CharacterData = Vec<u8>;
 
type IntegerData = Vec<u8>;
 
type IntegerData = i64;
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub enum Constant {
 
    Null, // message
 
    True,
 
    False,
 
    Character(CharacterData),
 
    Integer(IntegerData),
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub enum Method {
 
    Get,
 
    Fires,
 
    Create,
 
    Symbolic(SourceIdentifierId),
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub enum Field {
 
    Length,
 
    Symbolic(SourceIdentifierId),
 
}
 
impl Field {
 
    pub fn is_length(&self) -> bool {
 
        match self {
 
            Field::Length => true,
 
            _ => false,
 
        }
 
    }
 
}
 

	
 
#[derive(Debug, Clone, Copy, serde::Serialize, serde::Deserialize)]
 
pub enum Scope {
 
    Definition(DefinitionId),
 
    Block(BlockStatementId),
 
    Synchronous(SynchronousStatementId),
 
}
 

	
 
impl Scope {
 
    pub fn to_block(&self) -> BlockStatementId {
 
        match &self {
 
            Scope::Block(id) => *id,
 
            _ => panic!("Unable to cast `Scope` to `BlockStatement`"),
 
        }
 
    }
 
}
 

	
src/protocol/eval.rs
Show inline comments
 
@@ -23,100 +23,99 @@ const ONE: Value = Value::Byte(ByteValue(1));
 

	
 
trait ValueImpl {
 
    fn exact_type(&self) -> Type;
 
    fn is_type_compatible(&self, t: &Type) -> bool;
 
}
 

	
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub enum Value {
 
    Input(InputValue),
 
    Output(OutputValue),
 
    Message(MessageValue),
 
    Boolean(BooleanValue),
 
    Byte(ByteValue),
 
    Short(ShortValue),
 
    Int(IntValue),
 
    Long(LongValue),
 
    InputArray(InputArrayValue),
 
    OutputArray(OutputArrayValue),
 
    MessageArray(MessageArrayValue),
 
    BooleanArray(BooleanArrayValue),
 
    ByteArray(ByteArrayValue),
 
    ShortArray(ShortArrayValue),
 
    IntArray(IntArrayValue),
 
    LongArray(LongArrayValue),
 
}
 
impl Value {
 
    pub fn receive_message(buffer: &Payload) -> Value {
 
        Value::Message(MessageValue(Some(buffer.clone())))
 
    }
 
    fn create_message(length: Value) -> Value {
 
        match length {
 
            Value::Byte(_) | Value::Short(_) | Value::Int(_) | Value::Long(_) => {
 
                let length: i64 = i64::from(length);
 
                if length < 0 || length > MESSAGE_MAX_LENGTH {
 
                    // Only messages within the expected length are allowed
 
                    Value::Message(MessageValue(None))
 
                } else {
 
                    Value::Message(MessageValue(Some(Payload::new(length as usize))))
 
                }
 
            }
 
            _ => unimplemented!(),
 
        }
 
    }
 
    fn from_constant(constant: &Constant) -> Value {
 
        match constant {
 
            Constant::Null => Value::Message(MessageValue(None)),
 
            Constant::True => Value::Boolean(BooleanValue(true)),
 
            Constant::False => Value::Boolean(BooleanValue(false)),
 
            Constant::Integer(data) => {
 
            Constant::Integer(val) => {
 
                // Convert raw ASCII data to UTF-8 string
 
                let raw = String::from_utf8_lossy(data);
 
                let val = raw.parse::<i64>().unwrap();
 
                let val = *val;
 
                if val >= BYTE_MIN && val <= BYTE_MAX {
 
                    Value::Byte(ByteValue(val as i8))
 
                } else if val >= SHORT_MIN && val <= SHORT_MAX {
 
                    Value::Short(ShortValue(val as i16))
 
                } else if val >= INT_MIN && val <= INT_MAX {
 
                    Value::Int(IntValue(val as i32))
 
                } else {
 
                    Value::Long(LongValue(val))
 
                }
 
            }
 
            Constant::Character(_data) => unimplemented!(),
 
        }
 
    }
 
    fn set(&mut self, index: &Value, value: &Value) -> Option<Value> {
 
        // The index must be of integer type, and non-negative
 
        let the_index: usize;
 
        match index {
 
            Value::Byte(_) | Value::Short(_) | Value::Int(_) | Value::Long(_) => {
 
                let index = i64::from(index);
 
                if index < 0 || index >= MESSAGE_MAX_LENGTH {
 
                    // It is inconsistent to update out of bounds
 
                    return None;
 
                }
 
                the_index = index.try_into().unwrap();
 
            }
 
            _ => unreachable!(),
 
        }
 
        // The subject must be either a message or an array
 
        // And the value and the subject must be compatible
 
        match (self, value) {
 
            (Value::Message(MessageValue(None)), _) => {
 
                // It is inconsistent to update the null message
 
                None
 
            }
 
            (Value::Message(MessageValue(Some(payload))), Value::Byte(ByteValue(b))) => {
 
                if *b < 0 {
 
                    // It is inconsistent to update with a negative value
 
                    return None;
 
                }
 
                if let Some(slot) = payload.as_mut_vec().get_mut(the_index) {
 
                    *slot = (*b).try_into().unwrap();
 
                    Some(value.clone())
 
                } else {
 
                    // It is inconsistent to update out of bounds
 
                    None
 
                }
 
            }
 
            (Value::Message(MessageValue(Some(payload))), Value::Short(ShortValue(b))) => {
src/protocol/lexer.rs
Show inline comments
 
@@ -13,96 +13,97 @@ fn is_vchar(x: Option<u8>) -> bool {
 

	
 
fn is_wsp(x: Option<u8>) -> bool {
 
    if let Some(c) = x {
 
        c == b' ' || c == b'\t'
 
    } else {
 
        false
 
    }
 
}
 

	
 
fn is_ident_start(x: Option<u8>) -> bool {
 
    if let Some(c) = x {
 
        c >= b'A' && c <= b'Z' || c >= b'a' && c <= b'z'
 
    } else {
 
        false
 
    }
 
}
 

	
 
fn is_ident_rest(x: Option<u8>) -> bool {
 
    if let Some(c) = x {
 
        c >= b'A' && c <= b'Z' || c >= b'a' && c <= b'z' || c >= b'0' && c <= b'9' || c == b'_'
 
    } else {
 
        false
 
    }
 
}
 

	
 
fn is_constant(x: Option<u8>) -> bool {
 
    if let Some(c) = x {
 
        c >= b'0' && c <= b'9' || c == b'\''
 
    } else {
 
        false
 
    }
 
}
 

	
 
fn is_integer_start(x: Option<u8>) -> bool {
 
    if let Some(c) = x {
 
        c >= b'0' && c <= b'9'
 
    } else {
 
        false
 
    }
 
}
 

	
 
fn is_integer_rest(x: Option<u8>) -> bool {
 
    if let Some(c) = x {
 
        c >= b'0' && c <= b'9'
 
            || c >= b'a' && c <= b'f'
 
            || c >= b'A' && c <= b'F'
 
            || c == b'x'
 
            || c == b'X'
 
            || c == b'o'
 
    } else {
 
        false
 
    }
 
}
 

	
 
fn lowercase(x: u8) -> u8 {
 
    if x >= b'A' && x <= b'Z' {
 
        x - b'A' + b'a'
 
    } else {
 
        x
 
    }
 
}
 

	
 
pub struct Lexer<'a> {
 
    source: &'a mut InputSource,
 
    level: usize,
 
}
 

	
 
impl Lexer<'_> {
 
    pub fn new(source: &mut InputSource) -> Lexer {
 
        Lexer { source, level: 0 }
 
    }
 
    fn consume_line(&mut self) -> Result<Vec<u8>, ParseError> {
 
        let mut result: Vec<u8> = Vec::new();
 
        let mut next = self.source.next();
 
        while next.is_some() && next != Some(b'\n') && next != Some(b'\r') {
 
            if !(is_vchar(next) || is_wsp(next)) {
 
                return Err(self.source.error("Expected visible character or whitespace"));
 
            }
 
            result.push(next.unwrap());
 
            self.source.consume();
 
            next = self.source.next();
 
        }
 
        if next.is_some() {
 
            self.source.consume();
 
        }
 
        if next == Some(b'\r') && self.source.next() == Some(b'\n') {
 
            self.source.consume();
 
        }
 
        Ok(result)
 
    }
 
    fn consume_whitespace(&mut self, expected: bool) -> Result<(), ParseError> {
 
        let mut found = false;
 
        let mut next = self.source.next();
 
        while next.is_some() {
 
            if next == Some(b' ')
 
                || next == Some(b'\t')
 
                || next == Some(b'\r')
 
@@ -174,96 +175,138 @@ impl Lexer<'_> {
 
                    .source
 
                    .error(format!("Expected keyword: {}", String::from_utf8_lossy(keyword))));
 
            }
 
            self.source.consume();
 
        }
 
        if let Some(next) = self.source.next() {
 
            if next >= b'A' && next <= b'Z' || next >= b'a' && next <= b'z' {
 
                return Err(self.source.error(format!(
 
                    "Expected word boundary after keyword: {}",
 
                    String::from_utf8_lossy(keyword)
 
                )));
 
            }
 
        }
 
        Ok(())
 
    }
 
    fn has_string(&self, string: &[u8]) -> bool {
 
        self.source.has(string)
 
    }
 
    fn consume_string(&mut self, string: &[u8]) -> Result<(), ParseError> {
 
        let len = string.len();
 
        for i in 0..len {
 
            let expected = Some(string[i]);
 
            let next = self.source.next();
 
            if next != expected {
 
                return Err(self
 
                    .source
 
                    .error(format!("Expected {}", String::from_utf8_lossy(string))));
 
            }
 
            self.source.consume();
 
        }
 
        Ok(())
 
    }
 
    fn consume_ident(&mut self) -> Result<Vec<u8>, ParseError> {
 
        if !self.has_identifier() {
 
            return Err(self.source.error("Expected identifier"));
 
        }
 
        let mut result = Vec::new();
 
        let mut next = self.source.next();
 
        result.push(next.unwrap());
 
        self.source.consume();
 
        next = self.source.next();
 
        while is_ident_rest(next) {
 
            result.push(next.unwrap());
 
            self.source.consume();
 
            next = self.source.next();
 
        }
 
        Ok(result)
 
    }
 
    fn has_integer(&mut self) -> bool {
 
        is_integer_start(self.source.next())
 
    }
 
    fn consume_integer(&mut self) -> Result<i64, ParseError> {
 
        let position = self.source.pos();
 
        let mut data = Vec::new();
 
        let mut next = self.source.next();
 
        while is_integer_rest(next) {
 
            data.push(next.unwrap());
 
            self.source.consume();
 
            next = self.source.next();
 
        }
 

	
 
        let data_len = data.len();
 
        debug_assert_ne!(data_len, 0);
 
        if data_len == 1 {
 
            debug_assert!(data[0] >= b'0' && data[0] <= b'9');
 
            return Ok((data[0] - b'0') as i64);
 
        } else {
 
            // TODO: Fix, u64 should be supported as well
 
            let parsed = if data[1] == b'b' {
 
                let data = String::from_utf8_lossy(&data[2..]);
 
                i64::from_str_radix(&data, 2)
 
            } else if data[1] == b'o' {
 
                let data = String::from_utf8_lossy(&data[2..]);
 
                i64::from_str_radix(&data, 8)
 
            } else if data[1] == b'x' {
 
                let data = String::from_utf8_lossy(&data[2..]);
 
                i64::from_str_radix(&data, 16)
 
            } else {
 
                // Assume decimal
 
                let data = String::from_utf8_lossy(&data);
 
                i64::from_str_radix(&data, 10)
 
            };
 

	
 
            if let Err(_err) = parsed {
 
                return Err(ParseError::new(position, "Invalid integer constant"));
 
            }
 

	
 
            Ok(parsed.unwrap())
 
        }
 
    }
 

	
 
    // Statement keywords
 

	
 
    fn has_statement_keyword(&self) -> bool {
 
        self.has_keyword(b"channel")
 
            || self.has_keyword(b"skip")
 
            || self.has_keyword(b"if")
 
            || self.has_keyword(b"while")
 
            || self.has_keyword(b"break")
 
            || self.has_keyword(b"continue")
 
            || self.has_keyword(b"synchronous")
 
            || self.has_keyword(b"return")
 
            || self.has_keyword(b"assert")
 
            || self.has_keyword(b"goto")
 
            || self.has_keyword(b"new")
 
            || self.has_keyword(b"put")
 
    }
 
    fn has_type_keyword(&self) -> bool {
 
        self.has_keyword(b"in")
 
            || self.has_keyword(b"out")
 
            || self.has_keyword(b"msg")
 
            || self.has_keyword(b"boolean")
 
            || self.has_keyword(b"byte")
 
            || self.has_keyword(b"short")
 
            || self.has_keyword(b"int")
 
            || self.has_keyword(b"long")
 
    }
 
    fn has_builtin_keyword(&self) -> bool {
 
        self.has_keyword(b"get")
 
            || self.has_keyword(b"fires")
 
            || self.has_keyword(b"create")
 
            || self.has_keyword(b"length")
 
    }
 

	
 
    // Identifiers
 

	
 
    fn has_identifier(&self) -> bool {
 
        if self.has_statement_keyword() || self.has_type_keyword() || self.has_builtin_keyword() {
 
            return false;
 
        }
 
        let next = self.source.next();
 
        is_ident_start(next)
 
    }
 
    fn consume_identifier(&mut self, h: &mut Heap) -> Result<SourceIdentifierId, ParseError> {
 
        if self.has_statement_keyword() || self.has_type_keyword() || self.has_builtin_keyword() {
 
            return Err(self.source.error("Expected identifier"));
 
        }
 
        let position = self.source.pos();
 
@@ -980,107 +1023,101 @@ impl Lexer<'_> {
 
        self.consume_whitespace(false)?;
 
        if !self.has_string(b"}") {
 
            while self.source.next().is_some() {
 
                elements.push(self.consume_expression(h)?);
 
                self.consume_whitespace(false)?;
 
                if self.has_string(b"}") {
 
                    break;
 
                }
 
                self.consume_string(b",")?;
 
                self.consume_whitespace(false)?;
 
            }
 
        }
 
        self.consume_string(b"}")?;
 
        Ok(h.alloc_array_expression(|this| ArrayExpression { this, position, elements }))
 
    }
 
    fn has_constant(&self) -> bool {
 
        is_constant(self.source.next())
 
    }
 
    fn consume_constant_expression(
 
        &mut self,
 
        h: &mut Heap,
 
    ) -> Result<ConstantExpressionId, ParseError> {
 
        let position = self.source.pos();
 
        let value;
 
        if self.has_keyword(b"null") {
 
            self.consume_keyword(b"null")?;
 
            value = Constant::Null;
 
        } else if self.has_keyword(b"true") {
 
            self.consume_keyword(b"true")?;
 
            value = Constant::True;
 
        } else if self.has_keyword(b"false") {
 
            self.consume_keyword(b"false")?;
 
            value = Constant::False;
 
        } else if self.source.next() == Some(b'\'') {
 
            self.source.consume();
 
            let mut data = Vec::new();
 
            let mut next = self.source.next();
 
            while next != Some(b'\'') && (is_vchar(next) || next == Some(b' ')) {
 
                data.push(next.unwrap());
 
                self.source.consume();
 
                next = self.source.next();
 
            }
 
            if next != Some(b'\'') || data.is_empty() {
 
                return Err(self.source.error("Expected character constant"));
 
            }
 
            self.source.consume();
 
            value = Constant::Character(data);
 
        } else {
 
            let mut data = Vec::new();
 
            let mut next = self.source.next();
 
            if !is_integer_start(next) {
 
            if !self.has_integer() {
 
                return Err(self.source.error("Expected integer constant"));
 
            }
 
            while is_integer_rest(next) {
 
                data.push(next.unwrap());
 
                self.source.consume();
 
                next = self.source.next();
 
            }
 
            value = Constant::Integer(data);
 

	
 
            value = Constant::Integer(self.consume_integer()?);
 
        }
 
        Ok(h.alloc_constant_expression(|this| ConstantExpression { this, position, value }))
 
    }
 
    fn has_call_expression(&mut self) -> bool {
 
        /* We prevent ambiguity with variables, by looking ahead
 
        the identifier to see if we can find an opening
 
        parenthesis: this signals a call expression. */
 
        if self.has_builtin_keyword() {
 
            return true;
 
        }
 
        let backup_pos = self.source.pos();
 
        let mut result = false;
 
        match self.consume_identifier_spilled() {
 
            Ok(_) => match self.consume_whitespace(false) {
 
                Ok(_) => {
 
                    result = self.has_string(b"(");
 
                }
 
                Err(_) => {}
 
            },
 
            Err(_) => {}
 
        }
 
        self.source.seek(backup_pos);
 
        return result;
 
    }
 
    fn consume_call_expression(&mut self, h: &mut Heap) -> Result<CallExpressionId, ParseError> {
 
        let position = self.source.pos();
 
        let method;
 
        if self.has_keyword(b"get") {
 
            self.consume_keyword(b"get")?;
 
            method = Method::Get;
 
        } else if self.has_keyword(b"fires") {
 
            self.consume_keyword(b"fires")?;
 
            method = Method::Fires;
 
        } else if self.has_keyword(b"create") {
 
            self.consume_keyword(b"create")?;
 
            method = Method::Create;
 
        } else {
 
            let identifier = self.consume_identifier(h)?;
 
            method = Method::Symbolic(identifier)
 
        }
 
        self.consume_whitespace(false)?;
 
        let mut arguments = Vec::new();
 
        self.consume_string(b"(")?;
 
        self.consume_whitespace(false)?;
 
        if !self.has_string(b")") {
 
            while self.source.next().is_some() {
 
                arguments.push(self.consume_expression(h)?);
 
                self.consume_whitespace(false)?;
 
@@ -1529,161 +1566,218 @@ impl Lexer<'_> {
 
    fn consume_primitive_definition(&mut self, h: &mut Heap) -> Result<PrimitiveId, ParseError> {
 
        let position = self.source.pos();
 
        self.consume_keyword(b"primitive")?;
 
        self.consume_whitespace(true)?;
 
        let identifier = self.consume_identifier(h)?;
 
        self.consume_whitespace(false)?;
 
        let mut parameters = Vec::new();
 
        self.consume_parameters(h, &mut parameters)?;
 
        self.consume_whitespace(false)?;
 
        let body = self.consume_block_statement(h)?;
 
        Ok(h.alloc_primitive(|this| Primitive { this, position, identifier, parameters, body }))
 
    }
 
    fn consume_function_definition(&mut self, h: &mut Heap) -> Result<FunctionId, ParseError> {
 
        let position = self.source.pos();
 
        let return_type = self.consume_type_annotation(h)?;
 
        self.consume_whitespace(true)?;
 
        let identifier = self.consume_identifier(h)?;
 
        self.consume_whitespace(false)?;
 
        let mut parameters = Vec::new();
 
        self.consume_parameters(h, &mut parameters)?;
 
        self.consume_whitespace(false)?;
 
        let body = self.consume_block_statement(h)?;
 
        Ok(h.alloc_function(|this| Function {
 
            this,
 
            position,
 
            return_type,
 
            identifier,
 
            parameters,
 
            body,
 
        }))
 
    }
 
    fn has_pragma(&self) -> bool {
 
        if let Some(c) = self.source.next() {
 
            c == b'#'
 
        } else {
 
            false
 
        }
 
    }
 
    fn consume_pragma(&mut self, h: &mut Heap) -> Result<PragmaId, ParseError> {
 
        let position = self.source.pos();
 
        let next = self.source.next();
 
        if next != Some(b'#') {
 
            return Err(self.source.error("Expected pragma"));
 
        }
 
        self.source.consume();
 
        if !is_vchar(self.source.next()) {
 
            return Err(self.source.error("Expected pragma"));
 
        }
 
        let value = self.consume_line()?;
 
        Ok(h.alloc_pragma(|this| Pragma { this, position, value }))
 
        if self.has_string(b"version") {
 
            self.consume_string(b"version")?;
 
            self.consume_whitespace(true)?;
 
            if !self.has_integer() {
 
                return Err(self.source.error("Expected integer constant"));
 
            }
 
            let version = self.consume_integer()?;
 
            debug_assert!(version >= 0);
 
            return Ok(h.alloc_pragma(|this| Pragma::Version(PragmaVersion{
 
                this, position, version: version as u64
 
            })))
 
        } else if self.has_string(b"module") {
 
            self.consume_string(b"module")?;
 
            self.consume_whitespace(true)?;
 
            if !self.has_identifier() {
 
                return Err(self.source.error("Expected identifier"));
 
            }
 
            let mut value = Vec::new();
 
            let mut ident = self.consume_ident()?;
 
            value.append(&mut ident);
 
            while self.has_string(b".") {
 
                self.consume_string(b".")?;
 
                value.push(b'.');
 
                ident = self.consume_ident()?;
 
                value.append(&mut ident);
 
            }
 
            return Ok(h.alloc_pragma(|this| Pragma::Module(PragmaModule{
 
                this, position, value
 
            })));
 
        } else {
 
            return Err(self.source.error("Unknown pragma"));
 
        }
 
    }
 
    fn has_import(&self) -> bool {
 
        self.has_keyword(b"import")
 
    }
 
    fn consume_import(&mut self, h: &mut Heap) -> Result<ImportId, ParseError> {
 
        let position = self.source.pos();
 
        self.consume_keyword(b"import")?;
 
        self.consume_whitespace(true)?;
 
        let mut value = Vec::new();
 
        let mut ident = self.consume_ident()?;
 
        value.append(&mut ident);
 
        while self.has_string(b".") {
 
            self.consume_string(b".")?;
 
            value.push(b'.');
 
            ident = self.consume_ident()?;
 
            value.append(&mut ident);
 
        }
 
        self.consume_whitespace(false)?;
 
        self.consume_string(b";")?;
 
        Ok(h.alloc_import(|this| Import { this, position, value }))
 
    }
 
    pub fn consume_protocol_description(&mut self, h: &mut Heap) -> Result<RootId, ParseError> {
 
        let position = self.source.pos();
 
        let mut pragmas = Vec::new();
 
        let mut imports = Vec::new();
 
        let mut definitions = Vec::new();
 
        self.consume_whitespace(false)?;
 
        while self.has_pragma() {
 
            let pragma = self.consume_pragma(h)?;
 
            pragmas.push(pragma);
 
            self.consume_whitespace(false)?;
 
        }
 
        while self.has_import() {
 
            let import = self.consume_import(h)?;
 
            imports.push(import);
 
            self.consume_whitespace(false)?;
 
        }
 
        // do-while block
 
        while {
 
            let def = self.consume_symbol_definition(h)?;
 
            definitions.push(def);
 
            self.consume_whitespace(false)?;
 
            self.has_symbol_definition()
 
        } {}
 
        // end of file
 
        if !self.source.is_eof() {
 
            return Err(self.source.error("Expected end of file"));
 
        }
 
        Ok(h.alloc_protocol_description(|this| Root {
 
            this,
 
            position,
 
            pragmas,
 
            imports,
 
            definitions,
 
            declarations: Vec::new(),
 
        }))
 
    }
 
}
 

	
 
// #[cfg(test)]
 
// mod tests {
 
//     use crate::protocol::ast::Expression::*;
 
//     use crate::protocol::{ast, lexer::*};
 
#[cfg(test)]
 
mod tests {
 
    use crate::protocol::ast::Expression::*;
 
    use crate::protocol::{ast, lexer::*};
 

	
 
    #[test]
 
    fn test_pragmas() {
 
        let mut h = Heap::new();
 
        let mut input = InputSource::from_string("
 
        #version 0o7777
 
        #module something.dot.separated
 
        ").expect("new InputSource");
 
        let mut lex = Lexer::new(&mut input);
 
        let lexed = lex.consume_protocol_description(&mut h)
 
            .expect("lex input source");
 
        let root = &h[lexed];
 
        assert_eq!(root.pragmas.len(), 2);
 
        let pv = &h[root.pragmas[0]];
 
        let pm = &h[root.pragmas[1]];
 

	
 
        if let Pragma::Version(v) = pv {
 
            assert_eq!(v.version, 0o7777)
 
        } else {
 
            assert!(false, "first pragma not version");
 
        }
 
        if let Pragma::Module(m) = pm {
 
            assert_eq!(m.value, b"something.dot.separated");
 
        } else {
 
            assert!(false, "second pragma not version");
 
        }
 
    }
 

	
 
//     #[test]
 
//     fn test_lowercase() {
 
//         assert_eq!(lowercase(b'a'), b'a');
 
//         assert_eq!(lowercase(b'A'), b'a');
 
//         assert_eq!(lowercase(b'z'), b'z');
 
//         assert_eq!(lowercase(b'Z'), b'z');
 
//     }
 

	
 
//     #[test]
 
//     fn test_basic_expression() {
 
//         let mut h = Heap::new();
 
//         let mut is = InputSource::from_string("a+b;").unwrap();
 
//         let mut lex = Lexer::new(&mut is);
 
//         match lex.consume_expression(&mut h) {
 
//             Ok(expr) => {
 
//                 println!("{:?}", expr);
 
//                 if let Binary(bin) = &h[expr] {
 
//                     if let Variable(left) = &h[bin.left] {
 
//                         if let Variable(right) = &h[bin.right] {
 
//                             assert_eq!("a", format!("{}", h[left.identifier]));
 
//                             assert_eq!("b", format!("{}", h[right.identifier]));
 
//                             assert_eq!(Some(b';'), is.next());
 
//                             return;
 
//                         }
 
//                     }
 
//                 }
 
//                 assert!(false);
 
//             }
 
//             Err(err) => {
 
//                 err.print(&is);
 
//                 assert!(false);
 
//             }
 
//         }
 
//     }
 

	
 
//     #[test]
 
//     fn test_paren_expression() {
 
//         let mut h = Heap::new();
 
//         let mut is = InputSource::from_string("(true)").unwrap();
 
//         let mut lex = Lexer::new(&mut is);
 
//         match lex.consume_paren_expression(&mut h) {
 
//             Ok(expr) => {
 
//                 println!("{:#?}", expr);
 
//                 if let Constant(con) = &h[expr] {
 
//                     if let ast::Constant::True = con.value {
 
//                         return;
 
//                     }
 
@@ -1710,49 +1804,49 @@ impl Lexer<'_> {
 
//                 err.print(&is);
 
//                 assert!(false);
 
//             }
 
//         }
 
//     }
 

	
 
//     #[test]
 
//     fn test_basic_statement() {
 
//         let mut h = Heap::new();
 
//         let mut is = InputSource::from_string("while (true) { skip; }").unwrap();
 
//         let mut lex = Lexer::new(&mut is);
 
//         match lex.consume_statement(&mut h) {
 
//             Ok(stmt) => {
 
//                 println!("{:#?}", stmt);
 
//                 if let Statement::While(w) = &h[stmt] {
 
//                     if let Expression::Constant(_) = h[w.test] {
 
//                         if let Statement::Block(_) = h[w.body] {
 
//                             return;
 
//                         }
 
//                     }
 
//                 }
 
//                 assert!(false);
 
//             }
 
//             Err(err) => {
 
//                 err.print(&is);
 
//                 assert!(false);
 
//             }
 
//         }
 
//     }
 

	
 
//     #[test]
 
//     fn test_statement() {
 
//         let mut h = Heap::new();
 
//         let mut is = InputSource::from_string(
 
//             "label: while (true) { if (x++ > y[0]) break label; else continue; }\n",
 
//         )
 
//         .unwrap();
 
//         let mut lex = Lexer::new(&mut is);
 
//         match lex.consume_statement(&mut h) {
 
//             Ok(stmt) => {
 
//                 println!("{:#?}", stmt);
 
//             }
 
//             Err(err) => {
 
//                 err.print(&is);
 
//                 assert!(false);
 
//             }
 
//         }
 
//     }
 
// }
 
}
src/protocol/mod.rs
Show inline comments
 
mod arena;
 
mod ast;
 
// mod ast;
 
mod eval;
 
pub(crate) mod inputsource;
 
mod lexer;
 
// mod lexer;
 
mod library;
 
mod parser;
 

	
 
// TODO: Remove when not benchmarking
 
pub(crate) mod ast;
 
pub(crate) mod lexer;
 

	
 
lazy_static::lazy_static! {
 
    /// Conveniently-provided protocol description initialized with a zero-length PDL string.
 
    /// Exposed to minimize repeated initializations of this common protocol description.
 
    pub static ref TRIVIAL_PD: std::sync::Arc<ProtocolDescription> = {
 
        std::sync::Arc::new(ProtocolDescription::parse(b"").unwrap())
 
    };
 
}
 

	
 
use crate::common::*;
 
use crate::protocol::ast::*;
 
use crate::protocol::eval::*;
 
use crate::protocol::inputsource::*;
 
use crate::protocol::parser::*;
 

	
 
/// Description of a protocol object, used to configure new connectors.
 
/// (De)serializable.
 
#[derive(serde::Serialize, serde::Deserialize)]
 
#[repr(C)]
 
pub struct ProtocolDescription {
 
    heap: Heap,
 
    source: InputSource,
 
    root: RootId,
 
}
 
#[derive(Debug, Clone, serde::Serialize, serde::Deserialize)]
 
pub(crate) struct ComponentState {
 
    prompt: Prompt,
 
}
 
pub(crate) enum EvalContext<'a> {
 
    Nonsync(&'a mut NonsyncProtoContext<'a>),
 
    Sync(&'a mut SyncProtoContext<'a>),
 
    // None,
 
}
 
//////////////////////////////////////////////
 

	
 
impl std::fmt::Debug for ProtocolDescription {
 
    fn fmt(&self, f: &mut std::fmt::Formatter) -> std::fmt::Result {
 
        write!(f, "(An opaque protocol description)")
 
    }
 
}
 
impl ProtocolDescription {
 
    pub fn parse(buffer: &[u8]) -> Result<Self, String> {
 
        let mut heap = Heap::new();
 
        let mut source = InputSource::from_buffer(buffer).unwrap();
 
        let mut parser = Parser::new(&mut source);
 
        match parser.parse(&mut heap) {
 
            Ok(root) => {
 
                return Ok(ProtocolDescription { heap, source, root });
 
            }
src/protocol/parser/depth_visitor.rs
Show inline comments
 
new file 100644
src/protocol/parser/shallow_visitor.rs
Show inline comments
 
new file 100644
 
use crate::protocol::ast::*;
 
use crate::protocol::input_source::*;
 
use crate::protocol::lexer::*;
 

	
 
type Unit = ();
 
type VisitorResult = Result<Unit, ParseError>;
 

	
 
trait ShallowVisitor: Sized {
 
    fn visit_protocol_description(&mut self, h: &mut Heap, pd: RootId) -> 
 
}
 
\ No newline at end of file
src/protocol/parser/visitor.rs
Show inline comments
 
new file 100644
0 comments (0 inline, 0 general)