Add string type

Rakefile

@@ -1,3 +1,14 @@
+# MacOS:
+# ---
+# CC=gcc-14 CXX=g++-14 \
+# CFLAGS="-I/opt/homebrew/Cellar/flex/2.6.4_2/include" \
+# CXXFLAGS="-I/opt/homebrew/Cellar/flex/2.6.4_2/include" \
+# ../gcc-14.2.0/configure \
+#   --disable-bootstrap \
+#   --enable-languages=c,c++,elna \
+#   --with-sysroot=/Library/Developer/CommandLineTools/SDKs/MacOSX15.2.sdk \
+#   --prefix=$(realpath ../gcc-install)
+
 task :default do
   sh 'make -C build'
   sh './build/bin/elna'

gcc/elna-diagnostic.cc

@@ -32,6 +32,10 @@ namespace gcc
         {
             return "Char";
         }
+        else if (is_string_type(type))
+        {
+            return "String";
+        }
         else
         {
             return "<<unknown-type>>";

gcc/elna-generic.cc

@@ -25,7 +25,7 @@ namespace gcc
         if (statement->arguments().size() != 1)
         {
             error_at(get_location(&statement->position()),
-                    "procedure '%s' expects 1 argument, %i given",
+                    "procedure '%s' expects 1 argument, %lu given",
                     statement->name().c_str(), statement->arguments().size());
             return;
         }
@@ -46,6 +46,10 @@ namespace gcc
         {
             format_number = "%c\n";
         }
+        else if (is_string_type(argument_type))
+        {
+            format_number = "%s\n";
+        }
         else
         {
             error_at(get_location(&argument->position()),
@@ -83,6 +87,7 @@ namespace gcc
         tree main_fndecl_type = build_function_type_array(integer_type_node, 2, main_fndecl_type_param);
         this->main_fndecl = build_fn_decl("main", main_fndecl_type);
         tree resdecl = build_decl(UNKNOWN_LOCATION, RESULT_DECL, NULL_TREE, integer_type_node);
+        DECL_CONTEXT(resdecl) = this->main_fndecl;
         DECL_RESULT(this->main_fndecl) = resdecl;
         tree set_result = build2(INIT_EXPR, void_type_node, DECL_RESULT(main_fndecl),
                 build_int_cst_type(integer_type_node, 0));
@@ -139,6 +144,11 @@ namespace gcc
         this->current_expression = build_int_cstu(elna_char_type_node, character->character());
     }
 
+    void generic_visitor::visit(source::string_literal *string)
+    {
+        this->current_expression = build_string_literal(string->string().size() + 1, string->string().c_str());
+    }
+
     void generic_visitor::visit(source::binary_expression *expression)
     {
         expression->lhs().accept(this);
@@ -180,6 +190,8 @@ namespace gcc
                 operator_code = MULT_EXPR;
                 target_type = left_type;
                 break;
+            default:
+                break;
         }
         if (operator_code != ERROR_MARK) // An arithmetic operation.
         {
@@ -219,6 +231,8 @@ namespace gcc
                 operator_code = GE_EXPR;
                 target_type = boolean_type_node;
                 break;
+            default:
+                break;
         }
         gcc_assert(operator_code != ERROR_MARK);
         gcc_assert(target_type != error_mark_node);
@@ -275,6 +289,10 @@ namespace gcc
         {
             declaration_type = elna_char_type_node;
         }
+        else if (declaration->type().base() == "String")
+        {
+            declaration_type = elna_string_type_node;
+        }
         else
         {
             error_at(get_location(&declaration->type().position()),
@@ -286,6 +304,7 @@ namespace gcc
                 get_identifier(declaration->identifier().c_str()), declaration_type);
         auto result = this->symbol_map.insert({ declaration->identifier(), declaration_tree });
 
+        // DECL_CONTEXT(declaration_tree) = this->main_fndecl;
         if (result.second)
         {
             auto declaration_statement = build1_loc(declaration_location, DECL_EXPR,

gcc/elna-tree.cc

@@ -4,8 +4,23 @@
 
 tree elna_global_trees[ELNA_TI_MAX];
 
-void elna_init_ttree(void)
+namespace elna
 {
-    elna_char_type_node = make_unsigned_type(8);
+namespace gcc
-    TYPE_STRING_FLAG(elna_char_type_node) = 1;
+{
+    void init_ttree()
+    {
+        elna_char_type_node = make_unsigned_type(8);
+        elna_string_type_node = build_pointer_type(
+                build_qualified_type(char_type_node, TYPE_QUAL_CONST)); /* const char* */
+        TYPE_STRING_FLAG(elna_char_type_node) = 1;
+    }
+
+    bool is_string_type(tree type)
+    {
+        gcc_assert(TYPE_P(type));
+        return TREE_CODE(type) == POINTER_TYPE
+            && TYPE_MAIN_VARIANT(TREE_TYPE(type)) == char_type_node;
+    }
+}
 }

gcc/elna1.cc

@@ -25,14 +25,14 @@
 
 struct GTY (()) lang_type
 {
-  char dummy;
+    char dummy;
 };
 
 /* Language-dependent contents of a decl.  */
 
 struct GTY (()) lang_decl
 {
-  char dummy;
+    char dummy;
 };
 
 /* Language-dependent contents of an identifier.  This must include a
@@ -40,7 +40,7 @@ struct GTY (()) lang_decl
 
 struct GTY (()) lang_identifier
 {
-  struct tree_identifier common;
+    struct tree_identifier common;
 };
 
 /* The resulting tree type.  */
@@ -50,15 +50,15 @@ union GTY ((desc ("TREE_CODE (&%h.generic) == IDENTIFIER_NODE"),
 			"TS_COMMON) ? ((union lang_tree_node *) TREE_CHAIN "
 			"(&%h.generic)) : NULL"))) lang_tree_node
 {
-  union tree_node GTY ((tag ("0"), desc ("tree_node_structure (&%h)"))) generic;
+    union tree_node GTY ((tag ("0"), desc ("tree_node_structure (&%h)"))) generic;
-  struct lang_identifier GTY ((tag ("1"))) identifier;
+    struct lang_identifier GTY ((tag ("1"))) identifier;
 };
 
 /* We don't use language_function.  */
 
 struct GTY (()) language_function
 {
-  int dummy;
+    int dummy;
 };
 
 /* Language hooks.  */
@@ -66,9 +66,8 @@ struct GTY (()) language_function
 static bool elna_langhook_init(void)
 {
     build_common_tree_nodes(false);
-    elna_init_ttree();
+    elna::gcc::init_ttree();
 
-    /* I don't know why this has to be done explicitly.  */
     void_list_node = build_tree_list(NULL_TREE, void_type_node);
 
     build_common_builtin_nodes();
@@ -82,7 +81,7 @@ static void elna_parse_file(const char *filename)
 
     if (!file)
     {
-      fatal_error(UNKNOWN_LOCATION, "cannot open filename %s: %m", filename);
+        fatal_error(UNKNOWN_LOCATION, "cannot open filename %s: %m", filename);
     }
 
     elna::source::driver driver{ filename };
@@ -90,7 +89,7 @@ static void elna_parse_file(const char *filename)
     yy::parser parser(lexer, driver);
 
     linemap_add(line_table, LC_ENTER, 0, filename, 1);
-    if (auto result = parser())
+    if (parser())
     {
         for (const auto& error : driver.errors())
         {
@@ -110,56 +109,77 @@ static void elna_parse_file(const char *filename)
 
 static void elna_langhook_parse_file(void)
 {
-    for (int i = 0; i < num_in_fnames; i++)
+    for (unsigned int i = 0; i < num_in_fnames; i++)
     {
-        elna_parse_file (in_fnames[i]);
+        elna_parse_file(in_fnames[i]);
     }
 }
 
-static tree
+static tree elna_langhook_type_for_mode(enum machine_mode mode, int unsignedp)
-elna_langhook_type_for_mode (enum machine_mode mode, int unsignedp)
 {
-  if (mode == TYPE_MODE (float_type_node))
+    if (mode == TYPE_MODE(float_type_node))
-    return float_type_node;
-
-  if (mode == TYPE_MODE (double_type_node))
-    return double_type_node;
-
-  if (mode == TYPE_MODE (intQI_type_node))
-    return unsignedp ? unsigned_intQI_type_node : intQI_type_node;
-  if (mode == TYPE_MODE (intHI_type_node))
-    return unsignedp ? unsigned_intHI_type_node : intHI_type_node;
-  if (mode == TYPE_MODE (intSI_type_node))
-    return unsignedp ? unsigned_intSI_type_node : intSI_type_node;
-  if (mode == TYPE_MODE (intDI_type_node))
-    return unsignedp ? unsigned_intDI_type_node : intDI_type_node;
-  if (mode == TYPE_MODE (intTI_type_node))
-    return unsignedp ? unsigned_intTI_type_node : intTI_type_node;
-
-  if (mode == TYPE_MODE (integer_type_node))
-    return unsignedp ? unsigned_type_node : integer_type_node;
-
-  if (mode == TYPE_MODE (long_integer_type_node))
-    return unsignedp ? long_unsigned_type_node : long_integer_type_node;
-
-  if (mode == TYPE_MODE (long_long_integer_type_node))
-    return unsignedp ? long_long_unsigned_type_node
-		     : long_long_integer_type_node;
-
-  if (COMPLEX_MODE_P (mode))
     {
-      if (mode == TYPE_MODE (complex_float_type_node))
+        return float_type_node;
-	return complex_float_type_node;
-      if (mode == TYPE_MODE (complex_double_type_node))
-	return complex_double_type_node;
-      if (mode == TYPE_MODE (complex_long_double_type_node))
-	return complex_long_double_type_node;
-      if (mode == TYPE_MODE (complex_integer_type_node) && !unsignedp)
-	return complex_integer_type_node;
     }
-
+    else if (mode == TYPE_MODE(double_type_node))
-  /* gcc_unreachable */
+    {
-  return NULL;
+        return double_type_node;
+    }
+    if (mode == TYPE_MODE(intQI_type_node))
+    {
+        return unsignedp ? unsigned_intQI_type_node : intQI_type_node;
+    }
+    else if (mode == TYPE_MODE(intHI_type_node))
+    {
+        return unsignedp ? unsigned_intHI_type_node : intHI_type_node;
+    }
+    else if (mode == TYPE_MODE(intSI_type_node))
+    {
+        return unsignedp ? unsigned_intSI_type_node : intSI_type_node;
+    }
+    else if (mode == TYPE_MODE(intDI_type_node))
+    {
+        return unsignedp ? unsigned_intDI_type_node : intDI_type_node;
+    }
+    else if (mode == TYPE_MODE(intTI_type_node))
+    {
+        return unsignedp ? unsigned_intTI_type_node : intTI_type_node;
+    }
+    else if (mode == TYPE_MODE(integer_type_node))
+    {
+        return unsignedp ? unsigned_type_node : integer_type_node;
+    }
+    else if (mode == TYPE_MODE(long_integer_type_node))
+    {
+        return unsignedp ? long_unsigned_type_node : long_integer_type_node;
+    }
+    else if (mode == TYPE_MODE(long_long_integer_type_node))
+    {
+        return unsignedp
+            ? long_long_unsigned_type_node
+            : long_long_integer_type_node;
+    }
+    if (COMPLEX_MODE_P(mode))
+    {
+        if (mode == TYPE_MODE(complex_float_type_node))
+        {
+            return complex_float_type_node;
+        }
+        if (mode == TYPE_MODE(complex_double_type_node))
+        {
+            return complex_double_type_node;
+        }
+        if (mode == TYPE_MODE(complex_long_double_type_node))
+        {
+            return complex_long_double_type_node;
+        }
+        if (mode == TYPE_MODE(complex_integer_type_node) && !unsignedp)
+        {
+            return complex_integer_type_node;
+        }
+    }
+    /* gcc_unreachable */
+    return nullptr;
 }
 
 static tree elna_langhook_type_for_size(unsigned int bits ATTRIBUTE_UNUSED,

include/elna/gcc/elna-generic.h

@@ -31,6 +31,7 @@ namespace gcc
         void visit(source::number_literal<double> *literal) override;
         void visit(source::number_literal<bool> *boolean) override;
         void visit(source::char_literal *character) override;
+        void visit(source::string_literal *string) override;
         void visit(source::binary_expression *expression) override;
         void visit(source::constant_definition *definition) override;
         void visit(source::declaration *declaration) override;

include/elna/gcc/elna-tree.h

@@ -7,12 +7,21 @@
 
 enum elna_tree_index
 {
-  ELNA_TI_CHAR_TYPE,
+    ELNA_TI_CHAR_TYPE,
-  ELNA_TI_MAX
+    ELNA_TI_STRING_TYPE,
+    ELNA_TI_MAX
 };
 
 extern GTY(()) tree elna_global_trees[ELNA_TI_MAX];
 
 #define elna_char_type_node elna_global_trees[ELNA_TI_CHAR_TYPE]
+#define elna_string_type_node elna_global_trees[ELNA_TI_STRING_TYPE]
 
-void elna_init_ttree(void);
+namespace elna
+{
+namespace gcc
+{
+    void init_ttree();
+    bool is_string_type(tree type);
+}
+}

include/elna/source/ast.h

@@ -5,6 +5,8 @@
 
 #include <cstdint>
 #include <memory>
+#include <string>
+#include <vector>
 #include "elna/source/result.h"
 #include "elna/source/types.h"
 
@@ -49,6 +51,7 @@ namespace source
     template<typename T>
     class number_literal;
     class char_literal;
+    class string_literal;
 
     /**
      * Interface for AST visitors.
@@ -73,6 +76,7 @@ namespace source
         virtual void visit(number_literal<double> *) = 0;
         virtual void visit(number_literal<bool> *) = 0;
         virtual void visit(char_literal *) = 0;
+        virtual void visit(string_literal *) = 0;
     };
 
     /**
@@ -80,7 +84,7 @@ namespace source
      */
     struct empty_visitor : parser_visitor
     {
-        virtual void visit(declaration *declaration) override;
+        virtual void visit(declaration *) override;
         virtual void visit(constant_definition *definition) override;
         virtual void visit(procedure_definition *definition) override;
         virtual void visit(call_statement *statement) override;
@@ -92,12 +96,13 @@ namespace source
         virtual void visit(program *program) override;
         virtual void visit(binary_expression *expression) override;
         virtual void visit(unary_expression *expression) override;
-        virtual void visit(type_expression *variable) override;
+        virtual void visit(type_expression *) override;
-        virtual void visit(variable_expression *variable) override;
+        virtual void visit(variable_expression *) override;
-        virtual void visit(number_literal<std::int32_t> *number) override;
+        virtual void visit(number_literal<std::int32_t> *) override;
-        virtual void visit(number_literal<double> *number) override;
+        virtual void visit(number_literal<double> *) override;
-        virtual void visit(number_literal<bool> *boolean) override;
+        virtual void visit(number_literal<bool> *) override;
-        virtual void visit(char_literal *character) override;
+        virtual void visit(char_literal *) override;
+        virtual void visit(string_literal *) override;
     };
 
     /**
@@ -408,9 +413,9 @@ namespace source
 
     class block : public node
     {
-        std::unique_ptr<statement> m_body;
         std::vector<std::unique_ptr<definition>> m_definitions;
         std::vector<std::unique_ptr<declaration>> m_declarations;
+        std::unique_ptr<statement> m_body;
 
     public:
         block(const struct position position, std::vector<std::unique_ptr<definition>>&& definitions,
@@ -465,6 +470,17 @@ namespace source
         unsigned char character() const noexcept;
     };
 
+    class string_literal : public expression
+    {
+        std::string m_string;
+
+    public:
+        string_literal(const struct position position, const std::string& value);
+        virtual void accept(parser_visitor *visitor) override;
+
+        const std::string& string() const noexcept;
+    };
+
     class variable_expression : public expression
     {
         std::string m_name;

source/ast.cc

@@ -7,7 +7,7 @@ namespace elna
 {
 namespace source
 {
-    void empty_visitor::visit(declaration *declaration)
+    void empty_visitor::visit(declaration *)
     {
     }
 
@@ -87,27 +87,31 @@ namespace source
         expression->operand().accept(this);
     }
 
-    void empty_visitor::visit(type_expression *variable)
+    void empty_visitor::visit(type_expression *)
     {
     }
 
-    void empty_visitor::visit(variable_expression *variable)
+    void empty_visitor::visit(variable_expression *)
     {
     }
 
-    void empty_visitor::visit(number_literal<std::int32_t> *number)
+    void empty_visitor::visit(number_literal<std::int32_t> *)
     {
     }
 
-    void empty_visitor::visit(number_literal<double> *number)
+    void empty_visitor::visit(number_literal<double> *)
     {
     }
 
-    void empty_visitor::visit(number_literal<bool> *character)
+    void empty_visitor::visit(number_literal<bool> *)
     {
     }
 
-    void empty_visitor::visit(char_literal *character)
+    void empty_visitor::visit(char_literal *)
+    {
+    }
+
+    void empty_visitor::visit(string_literal *)
     {
     }
 
@@ -313,6 +317,21 @@ namespace source
         return m_character;
     }
 
+    string_literal::string_literal(const struct position position, const std::string& value)
+        : expression(position), m_string(value)
+    {
+    }
+
+    void string_literal::accept(parser_visitor *visitor)
+    {
+        visitor->visit(this);
+    }
+
+    const std::string& string_literal::string() const noexcept
+    {
+        return m_string;
+    }
+
     variable_expression::variable_expression(const struct position position, const std::string& name)
         : expression(position), m_name(name)
     {
@@ -550,6 +569,7 @@ namespace source
             case binary_operator::greater_equal:
                 return ">=";
         }
+        __builtin_unreachable();
     };
 }
 }

source/lexer.ll

@@ -80,6 +80,10 @@ false                   {
                             return yy::parser::make_CHARACTER(
 									std::string(yytext, 1, strlen(yytext) - 2), this->location);
                         }
+\"[^\"]*\"              {
+                            return yy::parser::make_STRING(
+									std::string(yytext, 1, strlen(yytext) - 2), this->location);
+                        }
 \(                      {
                             return yy::parser::make_LEFT_PAREN(this->location);
                         }

source/parser.yy

@@ -60,6 +60,7 @@
 %token <std::int32_t> INTEGER "integer"
 %token <float> FLOAT "float"
 %token <std::string> CHARACTER "character"
+%token <std::string> STRING "string"
 %token <bool> BOOLEAN
 %token IF WHILE DO
 %token CONST VAR PROCEDURE
@@ -76,6 +77,7 @@
 %type <std::unique_ptr<elna::source::number_literal<double>>> float_literal;
 %type <std::unique_ptr<elna::source::number_literal<bool>>> boolean_literal;
 %type <std::unique_ptr<elna::source::char_literal>> character_literal;
+%type <std::unique_ptr<elna::source::string_literal>> string_literal;
 %type <std::unique_ptr<elna::source::constant_definition>> constant_definition;
 %type <std::vector<std::unique_ptr<elna::source::constant_definition>>> constant_definition_part constant_definitions;
 %type <std::unique_ptr<elna::source::declaration>> variable_declaration;
@@ -153,7 +155,11 @@ float_literal: FLOAT
         };
 character_literal: CHARACTER
         {
-            $$ = std::make_unique<elna::source::char_literal>(elna::source::make_position(@1), $1[0]); 
+            $$ = std::make_unique<elna::source::char_literal>(elna::source::make_position(@1), $1.at(0)); 
+        };
+string_literal: STRING
+        {
+            $$ = std::make_unique<elna::source::string_literal>(elna::source::make_position(@1), $1); 
         };
 boolean_literal: BOOLEAN
         {
@@ -194,6 +200,7 @@ pointer:
     | float_literal { $$ = std::move($1); }
     | boolean_literal { $$ = std::move($1); }
     | character_literal { $$ = std::move($1); }
+    | string_literal { $$ = std::move($1); }
     | variable_expression { $$ = std::move($1); }
     | LEFT_PAREN expression RIGHT_PAREN { $$ = std::move($2); }
 summand: