parser: Add support for variables and identifiers in function body

This commit adds support for variables and identifiers in the function
body of the parser, stored as a vector.

However, at this point, identifier resolution is not fully implemented,
and we currently accept identifiers without checking if they can be
resolved. This is a known limitation that will be addressed in a future
commit once hash-tables are added to the parser.

Signed-off-by: Carlos Maniero <carlosmaniero@gmail.com>
Reviewed-by: Johnny Richard <johnny@johnnyrichard.com>

6 files changed, 187 insertions, 48 deletions

diff --git a/src/ast.c b/src/ast.c
index b2efc42..645a2dd 100644
--- a/src/ast.c
+++ b/src/ast.c
@@ -67,20 +67,35 @@ ast_node_new()
   return node;
 }
 
+static void
+ast_node_destroy_vector(vector_t *vector)
+{
+  for (size_t i=0; i < vector->size; i++) {
+    ast_node_destroy(vector_at(vector, i));
+  }
+  vector_destroy(vector);
+}
+
 void
 ast_node_destroy(ast_node_t *node)
 {
   switch (node->kind) {
     case AST_FUNCTION_DECLARATION:
-      ast_node_destroy(node->data.function.body);
+      ast_node_destroy_vector(node->data.function.body);
       break;
     case AST_BINARY_OPERATION:
       ast_node_destroy(node->data.binary_operation.left);
       ast_node_destroy(node->data.binary_operation.right);
       break;
+    case AST_VARIABLE_DECLARATION:
+      ast_node_destroy(node->data.variable.value);
+      break;
     case AST_LITERAL:
       break;
     case AST_RETURN_STMT:
+      ast_node_destroy(node->data.return_stmt.argument);
+      break;
+    case AST_IDENTIFIER:
       break;
     case AST_UNKOWN_NODE:
       break;
@@ -103,7 +118,7 @@ ast_node_init_return_stmt(ast_node_t *node, ast_node_t *argument)
 }
 
 void
-ast_node_init_function_declaration(ast_node_t *node, string_view_t name, type_t return_type, ast_node_t* body)
+ast_node_init_function_declaration(ast_node_t *node, string_view_t name, type_t return_type, vector_t *body)
 {
   node->accept_visitor = &ast_node_function_accept_visitor,
   node->kind = AST_FUNCTION_DECLARATION;
@@ -144,3 +159,29 @@ ast_node_init_binary_operation(ast_node_t *node, string_view_t op, ast_node_t *l
       }
   };
 }
+
+void
+ast_node_init_variable_declaration(ast_node_t *node, string_view_t variable_name, type_t type, ast_node_t *value)
+{
+  // FIXME: define the visitor strategy
+  node->kind = AST_VARIABLE_DECLARATION;
+  node->data = (ast_node_data_t) {
+      .variable = {
+        .name = variable_name,
+        .type = type,
+        .value = value
+      }
+  };
+}
+
+void
+ast_node_init_identifier(ast_node_t *node, string_view_t name)
+{
+  // FIXME: define the visitor strategy
+  node->kind = AST_IDENTIFIER;
+  node->data = (ast_node_data_t) {
+      .variable = {
+        .name = name,
+      }
+  };
+}
diff --git a/src/ast.h b/src/ast.h
index 70852b1..5ba21ad 100644
--- a/src/ast.h
+++ b/src/ast.h
@@ -18,6 +18,7 @@
 #define AST_H
 #include <stdint.h>
 #include "string_view.h"
+#include "vector.h"
 
 #define ast_visitor_visit(visitor, node) ast_node_accept_visitor(node, (ast_visitor_t *) visitor);
 
@@ -35,7 +36,7 @@ typedef struct ast_return_stmt_t {
 typedef struct ast_function_declaration_t {
   string_view_t name;
   type_t  return_type;
-  ast_node_t* body;
+  vector_t* body;
 } ast_function_declaration_t;
 
 typedef struct ast_binary_operation_t {
@@ -58,6 +59,16 @@ typedef struct ast_literal_t {
   ast_literal_value_t value;
 } ast_literal_t;
 
+typedef struct ast_identifier_t {
+  string_view_t name;
+} ast_identifier_t;
+
+typedef struct ast_variable_declaration_t {
+  string_view_t name;
+  type_t type;
+  ast_node_t* value;
+} ast_variable_declaration_t;
+
 typedef struct ast_visitor_t {
   void (*visit_function)(struct ast_visitor_t *, ast_function_declaration_t *);
   void (*visit_return_stmt)(struct ast_visitor_t *, ast_return_stmt_t *);
@@ -66,18 +77,22 @@ typedef struct ast_visitor_t {
 } ast_visitor_t;
 
 typedef enum {
-  AST_FUNCTION_DECLARATION,
   AST_BINARY_OPERATION,
+  AST_FUNCTION_DECLARATION,
+  AST_IDENTIFIER,
   AST_LITERAL,
   AST_RETURN_STMT,
-  AST_UNKOWN_NODE
+  AST_UNKOWN_NODE,
+  AST_VARIABLE_DECLARATION,
 } ast_node_kind_t;
 
 typedef union {
-  ast_function_declaration_t function;
   ast_binary_operation_t binary_operation;
+  ast_function_declaration_t function;
   ast_literal_t literal;
   ast_return_stmt_t return_stmt;
+  ast_variable_declaration_t variable;
+  ast_identifier_t identifier;
 } ast_node_data_t;
 
 typedef struct ast_node_t {
@@ -91,9 +106,11 @@ void ast_node_accept_visitor(ast_node_t *node, ast_visitor_t *visitor);
 ast_node_t* ast_node_new();
 void ast_node_destroy(ast_node_t *node);
 
-void ast_node_init_function_declaration(ast_node_t *node, string_view_t name, type_t return_type, ast_node_t *body);
-void ast_node_init_return_stmt(ast_node_t *node, ast_node_t *argument);
 void ast_node_init_binary_operation(ast_node_t *node, string_view_t op, ast_node_t *left, ast_node_t *right);
+void ast_node_init_function_declaration(ast_node_t *node, string_view_t name, type_t return_type, vector_t *body);
+void ast_node_init_identifier(ast_node_t *node, string_view_t name);
+void ast_node_init_return_stmt(ast_node_t *node, ast_node_t *argument);
+void ast_node_init_variable_declaration(ast_node_t *node, string_view_t variable_name, type_t type, ast_node_t *value);
 // FIXME: use the naming convention
 void ast_literal_integer_create(ast_node_t* node, uint32_t number);
 
diff --git a/src/gas_assembly_generator.c b/src/gas_assembly_generator.c
index 10a67f3..5572cb7 100644
--- a/src/gas_assembly_generator.c
+++ b/src/gas_assembly_generator.c
@@ -54,7 +54,9 @@ gas_assembly_generator_visit_function(ast_visitor_t *visitor, ast_function_decla
   fprintf(gen->out,".text\n");
   fprintf(gen->out,"_start:\n");
 
-  ast_visitor_visit(visitor, func->body);
+  for (size_t i=0; i < func->body->size; i++) {
+    ast_visitor_visit(visitor, vector_at(func->body, i));
+  }
 }
 
 static void
diff --git a/src/lexer.c b/src/lexer.c
index b641752..e1f0d80 100644
--- a/src/lexer.c
+++ b/src/lexer.c
@@ -154,7 +154,8 @@ lexer_next_token(lexer_t *lexer, token_t *token)
     if (lexer_current_char(lexer) == '+'
         || lexer_current_char(lexer) == '-'
         || lexer_current_char(lexer) == '*'
-        || lexer_current_char(lexer) == '/') {
+        || lexer_current_char(lexer) == '/'
+        || lexer_current_char(lexer) == '=') {
       lexer_define_literal_token_props(lexer, token, TOKEN_OP);
       lexer_drop_char(lexer);
       return;
diff --git a/src/parser.c b/src/parser.c
index a123648..5399cfd 100644
--- a/src/parser.c
+++ b/src/parser.c
@@ -24,6 +24,7 @@
 #include "ast.h"
 #include "lexer.h"
 #include "parser.h"
+#include "vector.h"
 
 void
 parser_init(parser_t *parser, lexer_t *lexer)
@@ -34,32 +35,27 @@ parser_init(parser_t *parser, lexer_t *lexer)
   parser->errors_len = 0;
 }
 
-static bool
-expected_token(token_t *token, parser_t *parser, token_kind_t kind)
+static void
+parser_error_push_unexpected_kind(parser_t *parser, token_t *token, token_kind_t expected)
 {
-  lexer_next_token(parser->lexer, token);
-  parser_error_t error;
+  parser_error_t* error = &parser->errors[parser->errors_len++];
+  error->token = *token;
 
   if (token->kind == TOKEN_EOF) {
-    error.token = *token;
-    sprintf(
-        error.message,
-        "expected '%s' but got end of file",
-        token_kind_to_str(kind)
-    );
-
-    parser->errors[parser->errors_len++] = error;
-    return false;
+    sprintf(error->message, "expected '%s' but got end of file", token_kind_to_str(expected));
+    return;
   }
 
+  sprintf(error->message, "expected '%s' but got '%s'", token_kind_to_str(expected), token_kind_to_str(token->kind));
+}
+
+static bool
+expected_token(token_t *token, parser_t *parser, token_kind_t kind)
+{
+  lexer_next_token(parser->lexer, token);
+
   if (token->kind != kind) {
-    error.token = *token;
-    sprintf(
-        error.message,
-        "expected '%s' but got '%s'",
-        token_kind_to_str(kind), token_kind_to_str(token->kind)
-    );
-    parser->errors[parser->errors_len++] = error;
+    parser_error_push_unexpected_kind(parser, token, kind);
     return false;
   }
 
@@ -120,6 +116,10 @@ parser_parse_factor(parser_t *parser, ast_node_t *node)
     parser_parse_expression(parser, node);
     if (!drop_expected_token(parser, TOKEN_CPAREN)) return false;
     return true;
+  } else if (token.kind == TOKEN_NAME) {
+    /// FIXME: Check if the identifier is defined
+    ast_node_init_identifier(node, token.value);
+    return true;
   }
 
   // FIXME: Extract this erros logic to a function
@@ -196,33 +196,112 @@ parser_parse_expression(parser_t *parser, ast_node_t *node)
 bool
 parser_parse_return_stmt(parser_t *parser, ast_node_t *node)
 {
-  token_t return_keyword_token;
+  ast_node_t *argument_token = ast_node_new();
+  if (!parser_parse_expression(parser, argument_token)) return false;
 
-  if (!drop_expected_token(parser, TOKEN_OCURLY)) return false;
+  if (!drop_expected_token(parser, TOKEN_SEMICOLON)) return false;
+
+  ast_node_init_return_stmt(node, argument_token);
+  return true;
+}
+
+bool
+parser_parse_variable_definition(parser_t *parser, string_view_t variable_name, ast_node_t *node)
+{
+  if (!drop_expected_token(parser, TOKEN_COLON)) return false;
+
+  type_t type;
+
+  // FIXME: change the parameters order
+  if (!parser_parse_type(&type, parser)) return false;
 
-  if(!expected_token(&return_keyword_token, parser, TOKEN_NAME)) return false;
+  token_t equal_token;
 
-  if (!string_view_eq(return_keyword_token.value, string_view_from_str("return"))) {
+  if (!expected_token(&equal_token, parser, TOKEN_OP)) return false;
+  
+  if (!string_view_eq(equal_token.value, string_view_from_str("="))) {
     parser_error_t error;
-    error.token = return_keyword_token;
+    error.token = equal_token;
 
     sprintf(
         error.message,
-        "expected 'return' keyword but got '"SVFMT"'",
-        SVARG(&return_keyword_token.value)
+        "expected '=' but got "SVFMT,
+        SVARG(&equal_token.value)
     );
 
     parser->errors[parser->errors_len++] = error;
     return false;
   }
 
-  ast_node_t *argument_token = ast_node_new();
-  if (!parser_parse_expression(parser, argument_token)) return false;
+  ast_node_t *expression = ast_node_new();
 
-  if (!drop_expected_token(parser, TOKEN_SEMICOLON)) return false;
-  if (!drop_expected_token(parser, TOKEN_CCURLY)) return false;
+  if (!parser_parse_expression(parser, expression) || !drop_expected_token(parser, TOKEN_SEMICOLON)) {
+    ast_node_destroy(expression);
+    return false;
+  }
+
+  ast_node_init_variable_declaration(node, variable_name, type, expression);
 
-  ast_node_init_return_stmt(node, argument_token);
+  return true;
+}
+
+bool
+parser_parse_block_declarations(parser_t *parser, vector_t *body)
+{
+  token_t current_token;
+  lexer_next_token(parser->lexer, &current_token);
+
+  while (current_token.kind != TOKEN_CCURLY && current_token.kind != TOKEN_EOF) {
+    if (current_token.kind != TOKEN_NAME) {
+      parser_error_push_unexpected_kind(parser, &current_token, TOKEN_NAME);
+      return false;
+    }
+
+    if (string_view_eq(current_token.value, string_view_from_str("return"))) {
+      ast_node_t *return_node = ast_node_new();
+      bool parsed_return = parser_parse_return_stmt(parser, return_node);
+
+      if (!parsed_return) {
+        ast_node_destroy(return_node);
+        return false;
+      }
+
+      vector_push_back(body, return_node);
+    }
+    else {
+      ast_node_t *variable_node = ast_node_new();
+      bool parsed_variable = parser_parse_variable_definition(parser, current_token.value, variable_node);
+
+      if (!parsed_variable) {
+        ast_node_destroy(variable_node);
+        return false;
+      }
+
+      vector_push_back(body, variable_node);
+    }
+
+    lexer_next_token(parser->lexer, &current_token);
+  }
+
+  if (current_token.kind != TOKEN_CCURLY) {
+    parser_error_push_unexpected_kind(parser, &current_token, TOKEN_CCURLY);
+    return false;
+  }
+
+  ast_node_t *latest_node = vector_at(body, body->size - 1);
+
+  if (latest_node->kind != AST_RETURN_STMT) {
+    parser_error_t error;
+    error.token = current_token;
+
+    sprintf(
+        error.message,
+        "expected 'return' keyword."
+    );
+
+    parser->errors[parser->errors_len++] = error;
+    return false;
+  }
   return true;
 }
 
@@ -245,19 +324,17 @@ parser_parse_function_declaration(parser_t *parser, ast_node_t *node)
     return false;
   }
 
-  ast_node_t *return_node = ast_node_new();
+  if (!drop_expected_token(parser, TOKEN_OCURLY)) return false;
 
-  bool parsed_return = parser_parse_return_stmt(parser, return_node);
+  vector_t* body = vector_new();
 
-  if (!parsed_return) {
-    return false;
-  }
+  if (!parser_parse_block_declarations(parser, body)) return false;
 
   ast_node_init_function_declaration(
     node,
     func_name_token.value,
     return_type,
-    return_node
+    body
   );
 
   return true;
diff --git a/src/parser.h b/src/parser.h
index 00d1aae..4b6b5a9 100644
--- a/src/parser.h
+++ b/src/parser.h
@@ -29,6 +29,7 @@ typedef struct parser_error_t {
 typedef struct parser_t {
   lexer_t *lexer;
   int errors_len;
+  // FIXME: replace with vector
   parser_error_t errors[1];
 } parser_t;