feat: start work on ast

6 files changed, 104 insertions, 18 deletions

diff --git a/Makefile b/Makefile
index e1140cf..a30b93c 100644
--- a/Makefile
+++ b/Makefile
@@ -1,8 +1,8 @@
 CC      = cc
 LEX     = flex
 CFLAGS  = -std=c17 -Wall -Wextra -Werror -Wpedantic -O2
-SOURCES = lexer.h main.c
-OBJECTS = bin/lexer.o bin/main.o
+SOURCES = ast.c ast.h lexer.h main.c
+OBJECTS = bin/lexer.o bin/ast.o bin/main.o
 SANITIZE = -g -O0 -fsanitize=address,undefined
 
 .PHONY: all debug check format lint clean
@@ -25,6 +25,7 @@ bin/%.o: %.c | bin
 	$(CC) $(CFLAGS) -c -o $@ $<
 
 bin/main.o: lexer.h
+bin/ast.o: ast.h lexer.h
 
 debug: CFLAGS += $(SANITIZE)
 debug: clean sfinae
@@ -39,7 +40,7 @@ format:
 	clang-format -i $(SOURCES)
 
 lint:
-	clang-tidy main.c -- -std=c17
+	clang-tidy ast.c main.c -- -std=c17
 
 clean:
 	rm -rf bin sfinae
diff --git a/ast.c b/ast.c
new file mode 100644
index 0000000..25cecb7
--- /dev/null
+++ b/ast.c
@@ -0,0 +1,68 @@
+#include "ast.h"
+#include "lexer.h"
+
+#include <stdio.h>
+#include <stdlib.h>
+#include <string.h>
+
+static struct Node* node_new(enum NodeKind kind)
+{
+    struct Node* node = calloc(1, sizeof *node);
+    if (node == NULL)
+    {
+        (void)fprintf(stderr, "out of memory\n");
+        exit(1);
+    }
+    node->kind = kind;
+    node->line = yylineno;
+    return node;
+}
+
+struct Node* node_leaf(enum NodeKind kind)
+{
+    return node_new(kind);
+}
+
+struct Node* node_char(char character)
+{
+    struct Node* node = node_new(NODE_CHAR);
+    node->character = character;
+    return node;
+}
+
+struct Node* node_unary(enum NodeKind kind, struct Node* first)
+{
+    struct Node* node = node_new(kind);
+    node->first = first;
+    return node;
+}
+
+struct Node* node_binary(enum NodeKind kind, struct Node* first, struct Node* second)
+{
+    struct Node* node = node_new(kind);
+    node->first = first;
+    node->second = second;
+    return node;
+}
+
+struct Node* node_append(struct Node* list, struct Node* node)
+{
+    struct Node* last = list;
+
+    if (list == NULL)
+        return node;
+    while (last->next != NULL)
+        last = last->next;
+    last->next = node;
+    return list;
+}
+
+void node_free(struct Node* node)
+{
+    if (node == NULL)
+        return;
+    node_free(node->first);
+    node_free(node->second);
+    node_free(node->next);
+    free(node);
+}
diff --git a/ast.h b/ast.h
new file mode 100644
index 0000000..1a6f159
--- /dev/null
+++ b/ast.h
@@ -0,0 +1,27 @@
+#pragma once
+
+enum NodeKind
+{
+    NODE_CHAR,
+    TYPE_DECLTYPE,
+    VALUE_DIFFERENCE,
+    STMT_EMIT
+};
+
+struct Node
+{
+    enum NodeKind kind;
+    int line;
+    char character;
+    struct Node* first;
+    struct Node* second;
+    struct Node* next;
+};
+
+struct Node* node_leaf(enum NodeKind kind);
+struct Node* node_char(char character);
+struct Node* node_unary(enum NodeKind kind, struct Node* first);
+struct Node* node_binary(enum NodeKind kind, struct Node* first, struct Node* second);
+
+struct Node* node_append(struct Node* list, struct Node* node);
+void node_free(struct Node* node);
diff --git a/lexer.h b/lexer.h
index eb7a4d3..4e09ace 100644
--- a/lexer.h
+++ b/lexer.h
@@ -3,8 +3,6 @@
 #include <stdbool.h>
 #include <stdio.h>
 
-// single characters stand for themselves
-// so the named tokens start past them
 enum Token
 {
     TOK_END = 0,
@@ -16,15 +14,9 @@ enum Token
     TOK_CHAR
 };
 
-// provided by flex
-// the input
-// the current line
-// and the scanner itself
-extern FILE* yyin;
-extern int yylineno;
-int yylex(void);
+extern FILE* yyin;   // input
+extern int yylineno; // current line
+int yylex(void);     // scanner
 
-// true once the lexer has refused something
 bool lexer_failed(void);
-// the character inside the most recent TOK_CHAR
 char lexer_character(void);
diff --git a/main.c b/main.c
index 44d9025..634e6d0 100644
--- a/main.c
+++ b/main.c
@@ -11,9 +11,7 @@ static const char* token_name(int token)
         case TOK_COMMIT: return "commit";
         case TOK_DECLTYPE: return "decltype";
         case TOK_DIFFERENCE: return "difference";
-        default:
-            // single characters and character literals are printed by the caller
-            return NULL;
+        default: return NULL;
     }
 }
 
@@ -32,6 +30,7 @@ static int dump_tokens(void)
         else
             (void)printf("%c\n", (char)token);
     }
+
     return lexer_failed() ? 1 : 0;
 }
 
diff --git a/sfinae.l b/sfinae.l
index 8453699..f969a98 100644
--- a/sfinae.l
+++ b/sfinae.l
@@ -20,7 +20,6 @@ static char unescape(char escaped)
     }
 }
 
-// complains, remembers, and hands back TOK_END so the scan stops where it went wrong
 static int fail(const char* message)
 {
     (void)fprintf(stderr, "line %d: %s\n", yylineno, message);