From 66c0a2ed96c0f40b7618234f9040f03b16dea116 Mon Sep 17 00:00:00 2001
From: Rui Ueyama <ruiu@cs.stanford.edu>
Date: Sun, 26 Apr 2020 10:11:02 +0900
Subject: [PATCH] Concatenate adjacent string literals

---
 preprocess.c  | 27 +++++++++++++++++++++++++++
 test/string.c |  6 ++++++
 test/test.h   |  1 +
 3 files changed, 34 insertions(+)

diff --git a/preprocess.c b/preprocess.c
index bfaaa4a5..d47fe2a4 100644
--- a/preprocess.c
+++ b/preprocess.c
@@ -919,6 +919,32 @@ static void init_macros(void) {
   add_builtin("__LINE__", line_macro);
 }
 
+// Concatenate adjacent string literals into a single string literal
+// as per the C spec.
+static void join_adjacent_string_literals(Token *tok1) {
+  while (tok1->kind != TK_EOF) {
+    Token *tok2 = tok1->next;
+
+    if (tok1->kind != TK_STR || tok2->kind != TK_STR) {
+      tok1 = tok1->next;
+      continue;
+    }
+
+    Token *t = copy_token(tok1);
+    t->ty = array_of(tok1->ty->base, tok1->ty->array_len + tok2->ty->array_len - 1);
+    t->str = calloc(1, t->ty->size);
+    t->next = tok2->next;
+
+    int i = 0;
+    for (int j = 0; j < tok1->ty->size - tok1->ty->base->size; i++, j++)
+      t->str[i] = tok1->str[j];
+    for (int j = 0; j < tok2->ty->size; i++, j++)
+      t->str[i] = tok2->str[j];
+
+    *tok1 = *t;
+  }
+}
+
 // Entry point function of the preprocessor.
 Token *preprocess(Token *tok) {
   init_macros();
@@ -926,5 +952,6 @@ Token *preprocess(Token *tok) {
   if (cond_incl)
     error_tok(cond_incl->tok, "unterminated conditional directive");
   convert_keywords(tok);
+  join_adjacent_string_literals(tok);
   return tok;
 }
diff --git a/test/string.c b/test/string.c
index 303e72b2..25f1988a 100644
--- a/test/string.c
+++ b/test/string.c
@@ -35,6 +35,12 @@ int main() {
   ASSERT(0, "\x00"[0]);
   ASSERT(119, "\x77"[0]);
 
+  ASSERT(7, sizeof("abc" "def"));
+  ASSERT(9, sizeof("abc" "d" "efgh"));
+  ASSERT(0, strcmp("abc" "d" "\nefgh", "abcd\nefgh"));
+  ASSERT(0, !strcmp("abc" "d", "abcd\nefgh"));
+  ASSERT(0, strcmp("\x9" "0", "\t0"));
+
   printf("OK\n");
   return 0;
 }
diff --git a/test/test.h b/test/test.h
index 3d5f5a94..b2eb5d7f 100644
--- a/test/test.h
+++ b/test/test.h
@@ -4,5 +4,6 @@ int assert(int expected, int actual, char *code);
 int printf(char *fmt, ...);
 int sprintf(char *buf, char *fmt, ...);
 int strcmp(char *p, char *q);
+int strncmp(char *p, char *q, long n);
 int memcmp(char *p, char *q, long n);
 void exit(int n);