Initial parser attempt added - seems to parse correctly

CMakeLists.txt

@@ -9,8 +9,8 @@ set(CMAKE_CXX_STANDARD 23)
 set(CMAKE_CXX_STANDARD_REQUIRED ON)
 
 
-set(HEADER_FILES src/include/lex.hpp)
+set(HEADER_FILES src/include/lex.hpp src/include/value.hpp src/include/parse.hpp)
-set(SOURCE_FILES src/lex.cpp)
+set(SOURCE_FILES src/lex.cpp src/parse.cpp src/value.cpp)
 set(CXX_WARNING_FLAGS -Wall -Wextra -Wpedantic -pedantic)
 
 # we're not actually shipping a library yet,

src/include/lex.hpp

@@ -1,12 +1,11 @@
 #pragma once
 #include <deque>
 #include <sstream>
-#include <vector>
 #include <stdint.h>
 #include <variant>
 #include <optional>
 
-enum TokenType {
+enum class TokenType {
     OpenParen,
     CloseParen,
     Dollar,

src/include/parse.hpp

@@ -0,0 +1,32 @@
+#pragma once
+#include <value.hpp>
+#include <lex.hpp>
+
+
+
+
+// The Parser produces a regular lisp value.
+// lisp code is made of lisp lists and atoms.
+class Parser {
+private:
+    // the token stream.
+    std::deque<Token> ts;
+    Token get_token();
+    void unget_token(Token);
+
+    // these may need to be interned later
+    String make_string(std::string);
+    Symbol make_symbol(std::string);
+
+
+    std::optional<LispValue> parse_one();
+    LispValue parse_list();
+
+public:
+    Parser(Lexer);
+
+    void feed(Lexer);
+
+    std::optional<LispValue> next();
+};
+

src/include/value.hpp

@@ -0,0 +1,33 @@
+#pragma once
+
+#include <concepts>
+#include <vector>
+#include <cstdint>
+#include <string>
+#include <variant>
+#include <iostream>
+
+// we're using a pure variant as our value type.
+struct Integer {int64_t value;};
+struct Double {double value;};
+struct String {std::string value;}; // might be a good idea to intern strings
+struct Symbol {std::string value;};
+struct List;
+struct Nil {};
+
+
+using LispValue = std::variant<Integer, Double, String, Symbol, List>;
+struct List {std::vector<LispValue> list;};
+// during compilation, we don't really care for cyclical lists etc.
+// during compilation we'll mostly be dealing with regular, flat lists
+// that form function calls.
+// We will have a different set of values during runtime
+// as the runtime will be a bytecode interpreter anyhow.
+
+
+
+void print_val(LispValue);
+
+String make_string(std::string);
+Symbol make_symbol(std::string);
+

src/lex.cpp

@@ -9,14 +9,14 @@ using namespace std;
 std::ostream &operator<<(std::ostream &os, Token const &t) { 
     os << "Token(";
     switch (t.type) {
-    case OpenParen: os << "OpenParen)"; break;
+    case TokenType::OpenParen: os << "OpenParen)"; break;
-    case CloseParen: os << "CloseParen)"; break;
+    case TokenType::CloseParen: os << "CloseParen)"; break;
-    case Dollar: os << "Dollar)"; break;
+    case TokenType::Dollar: os << "Dollar)"; break;
-    case Symbol: os << "Symbol, " << get<string>(*t.value) << ")"; break;
+    case TokenType::Symbol: os << "Symbol, " << get<string>(*t.value) << ")"; break;
-    case String: os << "String, \"" << get<string>(*t.value) << "\")"; break;
+    case TokenType::String: os << "String, \"" << get<string>(*t.value) << "\")"; break;
-    case Int: os << "Int, " << get<int64_t>(*t.value) << ")"; break;
+    case TokenType::Int: os << "Int, " << get<int64_t>(*t.value) << ")"; break;
-    case Double: os << "Double, " << get<double>(*t.value) << ")"; break;
+    case TokenType::Double: os << "Double, " << get<double>(*t.value) << ")"; break;
-    case End: os << "END)"; break;
+    case TokenType::End: os << "END)"; break;
     default:
         os << ")";
     }

src/main.cpp

@@ -1,5 +1,7 @@
+#include "value.hpp"
 #include <iostream>
 #include <lex.hpp>
+#include <parse.hpp>
 #include <string>
 
 using namespace std;
@@ -8,9 +10,9 @@ int main() {
     string s;
     getline(cin, s);
     cout << s << endl;
-    for (auto t : lex(s)) {
+    Parser p(s);
-        cout << t << " ";
+    print_val(*p.next());
-    }
+
     cout << endl;
     return 0;
 }

src/parse.cpp

@@ -0,0 +1,74 @@
+#include <lex.hpp>
+
+#include <parse.hpp>
+
+#include <iostream>
+
+using namespace std;
+
+Parser::Parser(Lexer l) : ts(l.collect()) {}
+
+void Parser::feed(Lexer l) {
+    ts.append_range(l.collect());
+}
+
+Token Parser::get_token() {
+    Token t = ts.front();
+    ts.pop_front();
+    return t;
+}
+void Parser::unget_token(Token t) {
+    ts.push_front(t);
+}
+
+String Parser::make_string(string s) {
+    return String {s};
+}
+
+Symbol Parser::make_symbol(string s) {
+    return Symbol {s};
+}
+
+optional<LispValue> Parser::parse_one() {
+    Token t = get_token();
+    switch (t.type) {
+    case TokenType::Int: return Integer {get<int64_t>(*t.value)};
+    case TokenType::Double: return Double {get<double>(*t.value)};
+    case TokenType::String: return make_string(get<string>(*t.value));
+    case TokenType::Symbol: return make_symbol(get<string>(*t.value));
+    case TokenType::OpenParen: return parse_list();
+    case TokenType::CloseParen: throw "whatever";
+
+    // I don't know what this will actually do, in theory maybe just like the OpenParen,
+    // but parses things in a different namespace? unimplemented for now.
+    case TokenType::Dollar: return parse_one();
+    case TokenType::End : return nullopt;
+    }
+    return nullopt;
+}
+
+LispValue Parser::parse_list() {
+    // assumes that we have read the OpenParen, and are reading elements until
+    // we find the CloseParen
+    List l;
+    Token t = get_token();
+
+    while (true) {
+        if (t.type == TokenType::End) {
+            // this is clearly an error!
+            cerr << "Parser::parse_list: Input ended before list ended." << endl;
+            break;
+        }
+        if (t.type == TokenType::CloseParen)
+            break;
+        unget_token(t);
+        l.list.push_back(*parse_one());
+        t = get_token();
+    }
+    return l;
+}
+
+
+optional<LispValue> Parser::next() {
+    return parse_one();
+}

src/value.cpp

@@ -0,0 +1,35 @@
+#include <concepts>
+#include <value.hpp>
+
+
+template <typename T>
+requires std::convertible_to<T, LispValue>
+void value_printer(T t) {
+    if constexpr (std::same_as<T, LispValue>) {
+        print_val(t);
+    } else if constexpr (requires { std::cout << t.value;}) {
+        std::cout << t.value;
+    } else {
+        std::cout << "{UNKNOWN}" << std::endl;
+    }
+}
+
+template <>
+void value_printer(List l) {
+    std::cout << "(";
+    for (auto i : l.list) {
+        value_printer(i);
+        std::cout << " ";
+    }
+    std::cout << ")";
+}
+
+template <>
+void value_printer(String s) {
+    std::cout << '"' << s.value << '"';
+}
+
+void print_val(LispValue v) {
+    std::visit([](auto arg) {value_printer(arg);}, v);
+}
+