Add a tokenizer to allow space characters between tokens

main.c

@@ -1,33 +1,122 @@
+#include <ctype.h>
+#include <stdarg.h>
+#include <stdbool.h>
 #include <stdio.h>
 #include <stdlib.h>
+#include <string.h>
 
-int main(int argc, char **argv) {
-  if (argc != 2) {
-    fprintf(stderr, "%s: invalid number of arguments\n", argv[0]);
-    return 1;
-  }
+typedef enum {
+  TK_PUNCT, // Punctuators
+  TK_NUM,   // Numeric literals
+  TK_EOF,   // End-of-file markers
+} TokenKind;
 
-  char *p = argv[1];
+// Token type
+typedef struct Token Token;
+struct Token {
+  TokenKind kind; // Token kind
+  Token *next;    // Next token
+  int val;        // If kind is TK_NUM, its value
+  char *loc;      // Token location
+  int len;        // Token length
+};
 
-  printf("  .globl main\n");
-  printf("main:\n");
-  printf("  mov $%ld, %%rax\n", strtol(p, &p, 10));
+// Reports an error and exit.
+static void error(char *fmt, ...) {
+  va_list ap;
+  va_start(ap, fmt);
+  vfprintf(stderr, fmt, ap);
+  fprintf(stderr, "\n");
+  exit(1);
+}
+
+// Consumes the current token if it matches `s`.
+static bool equal(Token *tok, char *op) {
+  return memcmp(tok->loc, op, tok->len) == 0 && op[tok->len] == '\0';
+}
+
+// Ensure that the current token is `s`.
+static Token *skip(Token *tok, char *s) {
+  if (!equal(tok, s))
+    error("expected '%s'", s);
+  return tok->next;
+}
+
+// Ensure that the current token is TK_NUM.
+static int get_number(Token *tok) {
+  if (tok->kind != TK_NUM)
+    error("expected a number");
+  return tok->val;
+}
+
+// Create a new token.
+static Token *new_token(TokenKind kind, char *start, char *end) {
+  Token *tok = calloc(1, sizeof(Token));
+  tok->kind = kind;
+  tok->loc = start;
+  tok->len = end - start;
+  return tok;
+}
+
+// Tokenize `p` and returns new tokens.
+static Token *tokenize(char *p) {
+  Token head = {};
+  Token *cur = &head;
 
   while (*p) {
-    if (*p == '+') {
+    // Skip whitespace characters.
+    if (isspace(*p)) {
       p++;
-      printf("  add $%ld, %%rax\n", strtol(p, &p, 10));
       continue;
     }
 
-    if (*p == '-') {
+    // Numeric literal
+    if (isdigit(*p)) {
+      cur = cur->next = new_token(TK_NUM, p, p);
+      char *q = p;
+      cur->val = strtoul(p, &p, 10);
+      cur->len = p - q;
+      continue;
+    }
+
+    // Punctuator
+    if (*p == '+' || *p == '-') {
+      cur = cur->next = new_token(TK_PUNCT, p, p + 1);
       p++;
-      printf("  sub $%ld, %%rax\n", strtol(p, &p, 10));
       continue;
     }
 
-    fprintf(stderr, "unexpected character: '%c'\n", *p);
-    return 1;
+    error("invalid token");
+  }
+
+  cur = cur->next = new_token(TK_EOF, p, p);
+  return head.next;
+}
+
+int main(int argc, char **argv) {
+  if (argc != 2)
+    error("%s: invalid number of arguments", argv[0]);
+
+  Token *tok = tokenize(argv[1]);
+
+  printf("  .globl main\n");
+  printf("main:\n");
+
+  // The first token must be a number
+  printf("  mov $%d, %%rax\n", get_number(tok));
+  tok = tok->next;
+
+  // ... followed by either `+ <number>` or `- <number>`.
+  while (tok->kind != TK_EOF) {
+    if (equal(tok, "+")) {
+      printf("  add $%d, %%rax\n", get_number(tok->next));
+      tok = tok->next->next;
+      continue;
+    }
+
+    tok = skip(tok, "-");
+    printf("  sub $%d, %%rax\n", get_number(tok));
+    tok = tok->next;
   }
 
   printf("  ret\n");

test.sh

@@ -19,5 +19,6 @@ assert() {
 assert 0 0
 assert 42 42
 assert 21 '5+20-4'
+assert 41 ' 12 + 34 - 5 '
 
 echo OK