From 4c7acd92f0b3666c726252f9fc93e37153193f69 Mon Sep 17 00:00:00 2001
From: Alex Kotov <kotovalexarian@gmail.com>
Date: Sun, 7 May 2023 15:01:02 +0400
Subject: [PATCH] Escape in strings

---
 src/builtins.c   | 13 ++++++++++++-
 src/lexer.c      | 10 ++++++++++
 src/lexer.h      |  1 +
 tests/arcana.scm | 36 ++++++++++++++++++++++--------------
 4 files changed, 45 insertions(+), 15 deletions(-)

diff --git a/src/builtins.c b/src/builtins.c
index c59019f..9a81d72 100644
--- a/src/builtins.c
+++ b/src/builtins.c
@@ -655,7 +655,18 @@ struct Object *func_display(
         printf("%li", object->number.i64);
         break;
     case TYPE_STRING:
-        printf("\"%s\"", object->s);
+        putchar('"');
+        for (char *str = object->s; *str; ++str) {
+            const char chr = *str;
+            if (chr == '\\') {
+                printf("\\\\");
+            } else if (chr == '\"') {
+                printf("\\\"");
+            } else {
+                putchar(chr);
+            }
+        }
+        putchar('"');
         break;
     }
 
diff --git a/src/lexer.c b/src/lexer.c
index efb257c..642fdb4 100644
--- a/src/lexer.c
+++ b/src/lexer.c
@@ -354,6 +354,8 @@ void Lexer_lex(const Lexer self, const char chr)
         if (chr == '"') {
             token_finish(self);
             token_start(self, STATE_STRING_END);
+        } else if (chr == '\\') {
+            self->state = STATE_STRING_ESCAPE;
         } else {
             token_putc(self, chr);
         }
@@ -389,5 +391,13 @@ void Lexer_lex(const Lexer self, const char chr)
             assert(0);
         }
         break;
+    case STATE_STRING_ESCAPE:
+        if (chr == '\\' || chr == '"') {
+            token_putc(self, chr);
+            self->state = STATE_STRING_INSIDE;
+        } else {
+            assert(0);
+        }
+        break;
     }
 }
diff --git a/src/lexer.h b/src/lexer.h
index 452981d..be0187a 100644
--- a/src/lexer.h
+++ b/src/lexer.h
@@ -31,6 +31,7 @@ enum Lexer_State {
     STATE_STRING_START,
     STATE_STRING_INSIDE,
     STATE_STRING_END,
+    STATE_STRING_ESCAPE,
 };
 
 typedef struct Lexer {
diff --git a/tests/arcana.scm b/tests/arcana.scm
index 2565e46..5b3738f 100644
--- a/tests/arcana.scm
+++ b/tests/arcana.scm
@@ -20,22 +20,30 @@
   (assert-equal
     (arcana/tokenize "#false")
     (list (cons 'TOKEN_TAG "false")))
-  ;(assert-equal
-  ;  (arcana/tokenize "\"\"")
-  ;  (list (cons 'TOKEN_STRING "\"\"")))
-  ;(assert-equal
-  ;  (arcana/tokenize "\"qwe\"")
-  ;  (list (cons 'TOKEN_STRING "\"qwe\"")))
+  (assert-equal
+    (arcana/tokenize "\"\"")
+    (list (cons 'TOKEN_STRING "")))
+  (assert-equal
+    (arcana/tokenize "\"qwe\"")
+    (list (cons 'TOKEN_STRING "qwe")))
+  (assert-equal
+    (arcana/tokenize "(displayln \"qwe\")")
+    (list
+      (cons 'TOKEN_ROUND_OPEN  "(")
+      (cons 'TOKEN_IDENT       "displayln")
+      (cons 'TOKEN_STRING      "qwe")
+      (cons 'TOKEN_ROUND_CLOSE ")")))
+
   (assert-equal
     (arcana/tokenize "(displayln (list 1))")
-      (list
-        (cons 'TOKEN_ROUND_OPEN  "(")
-        (cons 'TOKEN_IDENT       "displayln")
-        (cons 'TOKEN_ROUND_OPEN  "(")
-        (cons 'TOKEN_IDENT       "list")
-        (cons 'TOKEN_NUM         "1")
-        (cons 'TOKEN_ROUND_CLOSE ")")
-        (cons 'TOKEN_ROUND_CLOSE ")")))
+    (list
+      (cons 'TOKEN_ROUND_OPEN  "(")
+      (cons 'TOKEN_IDENT       "displayln")
+      (cons 'TOKEN_ROUND_OPEN  "(")
+      (cons 'TOKEN_IDENT       "list")
+      (cons 'TOKEN_NUM         "1")
+      (cons 'TOKEN_ROUND_CLOSE ")")
+      (cons 'TOKEN_ROUND_CLOSE ")")))
 
   ;;; arcana/typeof ;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;;
   (assert-equal 'null      (arcana/typeof '()))