updated compiler to support multiple frontends and backends

This commit is contained in:
2026-02-05 01:09:14 +00:00
parent 8d130a870c
commit a35cfbe864
14 changed files with 1737 additions and 324 deletions
+627
View File
@@ -0,0 +1,627 @@
use std::iter::Peekable;
use std::str::Chars;
#[derive(Debug, PartialEq, Clone)]
pub enum Token {
// Keywords
Fn,
Let,
If,
Else,
Loop,
While,
Break,
Return,
Continue,
Include,
Static,
Const,
// Identifiers and literals
Identifier(Name),
String(String),
Integer(u64),
Char(char),
// Symbols
LeftParen, // (
RightParen, // )
LeftBrace, // {
RightBrace, // }
Semicolon, // ;
Colon, // :
Comma, // ,
// Operators
Plus, // +
Minus, // -
Star, // *
Amphersand, // &
Slash, // /
Assign, // =
EqualEqual, // ==
Bang, // !
BangEqual, // !=
Less, // <
LessEqual, // <=
Greater, // >
GreaterEqual, // >=
RightArrow, // ->
// Special
Eof,
}
use std::fmt;
use crate::model::Name;
impl fmt::Display for Name {
fn fmt(&self, f: &mut fmt::Formatter<'_>) -> fmt::Result {
if let Some(ref ns) = self.namespace {
write!(f, "{}::{}", ns, self.name)
} else {
write!(f, "{}", self.name)
}
}
}
impl Token {
pub fn tt(&self) -> &str {
match self {
Token::Const => "Const",
Token::Static => "Static",
Token::Include => "Include",
Token::Fn => "Fn",
Token::If => "If",
Token::Let => "Let",
Token::Else => "Else",
Token::Loop => "Loop",
Token::While => "While",
Token::Break => "Break",
Token::Return => "Return",
Token::Continue => "Continue",
Token::Identifier(_) => "Identifier",
Token::String(_) => "String",
Token::Integer(_) => "UnsignedInt",
Token::Char(_) => "Char",
Token::LeftParen => "LeftParen",
Token::RightParen => "RightParen",
Token::LeftBrace => "LeftBrace",
Token::RightBrace => "RightBrace",
Token::Semicolon => "Semicolon",
Token::Colon => "Colon",
Token::Comma => "Comma",
Token::RightArrow => "RightArrow",
Token::Plus => "Plus",
Token::Minus => "Minus",
Token::Star => "Star",
Token::Amphersand => "Amphersand",
Token::Slash => "Slash",
Token::Assign => "Assign",
Token::EqualEqual => "EqualEqual",
Token::Bang => "Bang",
Token::BangEqual => "BangEqual",
Token::Less => "Less",
Token::LessEqual => "LessEqual",
Token::Greater => "Greater",
Token::GreaterEqual => "GreaterEqual",
Token::Eof => "Eof",
}
}
}
#[derive(Debug)]
pub struct Lexer<'a> {
chars: Peekable<Chars<'a>>,
current: Option<char>,
line: usize,
}
impl<'a> Lexer<'a> {
pub fn new(input: &'a str) -> Self {
let mut chars = input.chars().peekable();
let current = chars.next();
Lexer {
chars,
current,
line: 1,
}
}
fn advance(&mut self) -> Option<char> {
self.current = self.chars.next();
self.current
}
fn peek(&mut self) -> Option<&char> {
self.chars.peek()
}
fn skip_whitespace(&mut self) {
while let Some(c) = self.current {
if !c.is_whitespace() {
break;
}
if c == '\n' {
self.line += 1;
}
self.advance();
}
}
fn skip_line_comment(&mut self) {
// Skip the two slashes
self.advance(); // first /
self.advance(); // second /
// Skip until newline or EOF
while let Some(c) = self.current {
if c == '\n' {
self.line += 1;
self.advance();
break;
}
self.advance();
}
}
fn skip_block_comment(&mut self) -> Result<(), String> {
// Skip the /*
self.advance(); // /
self.advance(); // *
let start_line = self.line;
// Look for */
while let Some(c) = self.current {
if c == '\n' {
self.line += 1;
}
if c == '*' {
if let Some(&next) = self.peek() {
if next == '/' {
self.advance(); // *
self.advance(); // /
return Ok(());
}
}
}
self.advance();
}
Err(format!(
"Unterminated block comment starting at line {}",
start_line
))
}
fn skip_whitespace_and_comments(&mut self) {
loop {
self.skip_whitespace();
// Check for comments
if let Some('/') = self.current {
if let Some(&next) = self.peek() {
match next {
'/' => {
self.skip_line_comment();
continue;
}
'*' => {
if let Err(e) = self.skip_block_comment() {
eprintln!("Lexer error: {}", e);
}
continue;
}
_ => break,
}
}
}
break;
}
}
fn read_identifier(&mut self) -> String {
let mut ident = String::new();
// Include the current character if it's valid
if let Some(c) = self.current {
if c.is_alphabetic() || c == '_' {
ident.push(c);
}
}
// Read remaining characters
while let Some(&c) = self.peek() {
if c.is_alphanumeric() || c == '_' {
self.advance();
ident.push(c);
} else {
break;
}
}
ident
}
fn keyword_or_identifier(&mut self) -> Token {
let first_ident = self.read_identifier();
// Check if it's a keyword first (keywords can't have namespaces)
let keyword = match first_ident.as_str() {
"fn" => Some(Token::Fn),
"if" => Some(Token::If),
"else" => Some(Token::Else),
"while" => Some(Token::While),
"loop" => Some(Token::Loop),
"break" => Some(Token::Break),
"return" => Some(Token::Return),
"continue" => Some(Token::Continue),
"include" => Some(Token::Include),
"let" => Some(Token::Let),
"const" => Some(Token::Const),
"static" => Some(Token::Static),
_ => None,
};
if let Some(kw) = keyword {
return kw;
}
// Not a keyword - check for namespace separator (::)
// We need to peek TWO characters ahead without consuming anything
if let Some(&':') = self.peek() {
// We see one colon, but we need to check if there's another one after it
// We can't peek two ahead directly, so we need a different approach
// Save the current position by using a temporary peekable iterator
// Actually, we can't do that easily. Instead, let's just check:
// If we see ':', temporarily advance and check the next char
// Create a temporary check
let mut temp_chars = self.chars.clone();
let _ = temp_chars.next(); // This is the ':' we already saw
let second_peek = temp_chars.peek();
if let Some(&':') = second_peek {
// It's :: - consume both colons
self.advance(); // consume first :
self.advance(); // consume second :
// Read the second identifier (the actual name)
let second_ident = self.read_identifier();
// Return namespaced identifier
return Token::Identifier(Name {
namespace: Some(first_ident),
name: second_ident,
});
}
// else: It's a single colon (type annotation) - DON'T consume it
// Just fall through and return the identifier
}
// No namespace separator - just a regular identifier
Token::Identifier(Name {
namespace: None,
name: first_ident,
})
}
fn read_number(&mut self) -> Result<u64, String> {
let current = self.current.unwrap();
// Check for hex (0x) or binary (0b) prefix
if current == '0' {
if let Some(&next_char) = self.peek() {
match next_char {
'x' | 'X' => {
self.advance(); // consume '0'
self.advance(); // consume 'x'
return self.read_hex_number();
}
'b' | 'B' => {
self.advance(); // consume '0'
self.advance(); // consume 'b'
return self.read_binary_number();
}
_ => {}
}
}
}
// Read decimal number
self.read_decimal_number()
}
fn read_decimal_number(&mut self) -> Result<u64, String> {
let mut num_str = String::new();
if let Some(c) = self.current {
num_str.push(c);
}
while let Some(&c) = self.peek() {
if c.is_ascii_digit() {
self.advance();
num_str.push(c);
} else {
break;
}
}
num_str
.parse::<u64>()
.map_err(|_| format!("Invalid decimal number: {}", num_str))
}
fn read_hex_number(&mut self) -> Result<u64, String> {
let mut num_str = String::new();
// Read current character if it's a hex digit
if let Some(c) = self.current {
if c.is_ascii_hexdigit() {
num_str.push(c);
}
}
while let Some(&c) = self.peek() {
if c.is_ascii_hexdigit() {
self.advance();
num_str.push(c);
} else {
break;
}
}
if num_str.is_empty() {
return Err("Invalid hexadecimal number: no digits after 0x".to_string());
}
u64::from_str_radix(&num_str, 16)
.map_err(|_| format!("Invalid hexadecimal number: {}", num_str))
}
fn read_binary_number(&mut self) -> Result<u64, String> {
let mut num_str = String::new();
// Read current character if it's a binary digit
if let Some(c) = self.current {
if c == '0' || c == '1' {
num_str.push(c);
}
}
while let Some(&c) = self.peek() {
if c == '0' || c == '1' {
self.advance();
num_str.push(c);
} else {
break;
}
}
if num_str.is_empty() {
return Err("Invalid binary number: no digits after 0b".to_string());
}
u64::from_str_radix(&num_str, 2)
.map_err(|_| format!("Invalid binary number: {}", num_str))
}
fn read_string(&mut self) -> Result<String, String> {
self.advance(); // Skip the opening quote
let mut s = String::new();
while let Some(c) = self.current {
if c == '"' {
return Ok(s);
}
// Handle escape sequences
if c == '\\' {
self.advance();
if let Some(escaped) = self.current {
let escaped_char = match escaped {
'n' => '\n',
't' => '\t',
'r' => '\r',
'\\' => '\\',
'"' => '"',
_ => escaped, // For now, just use the character as-is
};
s.push(escaped_char);
} else {
return Err("Unexpected end of string after escape".to_string());
}
} else {
s.push(c);
}
self.advance();
}
Err("Unterminated string literal".to_string())
}
fn match_next(&mut self, expected: char) -> bool {
match self.peek() {
Some(&c) if c == expected => {
self.advance();
true
}
_ => false,
}
}
fn scan_single_char_token(&mut self, c: char) -> Option<Token> {
match c {
'(' => Some(Token::LeftParen),
')' => Some(Token::RightParen),
'{' => Some(Token::LeftBrace),
'}' => Some(Token::RightBrace),
';' => Some(Token::Semicolon),
',' => Some(Token::Comma),
'&' => Some(Token::Amphersand),
'+' => Some(Token::Plus),
'*' => Some(Token::Star),
_ => None,
}
}
fn scan_operator(&mut self, c: char) -> Option<Token> {
match c {
'-' => Some(if self.match_next('>') {
Token::RightArrow
} else {
Token::Minus
}),
'!' => Some(if self.match_next('=') {
Token::BangEqual
} else {
Token::Bang
}),
'=' => Some(if self.match_next('=') {
Token::EqualEqual
} else {
Token::Assign
}),
'<' => Some(if self.match_next('=') {
Token::LessEqual
} else {
Token::Less
}),
'>' => Some(if self.match_next('=') {
Token::GreaterEqual
} else {
Token::Greater
}),
':' => {
// Single colon (for type annotations)
// Note: :: is handled in keyword_or_identifier for namespaces
Some(Token::Colon)
}
'/' => {
// Check if it's a comment or division
if let Some(&next) = self.peek() {
if next == '/' || next == '*' {
// It's a comment, don't consume it here
// Let skip_whitespace_and_comments handle it
None
} else {
Some(Token::Slash)
}
} else {
Some(Token::Slash)
}
}
_ => None,
}
}
pub fn next_token(&mut self) -> Token {
self.skip_whitespace_and_comments();
let Some(c) = self.current else {
return Token::Eof;
};
// Try single-character tokens first
if let Some(token) = self.scan_single_char_token(c) {
self.advance();
return token;
}
// Try operators (may be multi-character)
if let Some(token) = self.scan_operator(c) {
self.advance();
return token;
}
// Char literals
if c == '\'' {
let mut value = ' ';
self.advance();
if let Some(ch) = self.current {
value = ch;
self.advance();
}
if self.current == Some('\'') {
self.advance();
return Token::Char(value);
}
eprintln!("Lexer error on line {}: Invalid char literal", self.line);
}
// String literals
if c == '"' {
let token = match self.read_string() {
Ok(s) => Token::String(s),
Err(e) => {
eprintln!("Lexer error on line {}: {}", self.line, e);
// Skip to next quote or end
while let Some(ch) = self.current {
if ch == '"' || ch == '\n' {
break;
}
self.advance();
}
Token::String(String::new())
}
};
self.advance();
return token;
}
// Identifiers and keywords (including namespaced identifiers)
if c.is_alphabetic() || c == '_' {
let token = self.keyword_or_identifier();
self.advance();
return token;
}
// Numbers (decimal, hex, binary)
if c.is_ascii_digit() {
let token = match self.read_number() {
Ok(num) => Token::Integer(num),
Err(e) => {
eprintln!("Lexer error on line {}: {}", self.line, e);
// Skip invalid number
while let Some(&ch) = self.peek() {
if !ch.is_alphanumeric() {
break;
}
self.advance();
}
Token::Integer(0)
}
};
self.advance();
return token;
}
// Unknown character - skip it
eprintln!(
"Lexer warning on line {}: Skipping unknown character '{}'",
self.line, c
);
self.advance();
self.next_token()
}
}
impl<'a> Iterator for Lexer<'a> {
type Item = Token;
fn next(&mut self) -> Option<Self::Item> {
match self.next_token() {
Token::Eof => None,
token => Some(token),
}
}
}