From dea6dce494e9c403f1f9e2178cd8c7d0aad79738 Mon Sep 17 00:00:00 2001
From: kdx <kikoodx@paranoici.org>
Date: Tue, 13 Jun 2023 04:25:19 +0200
Subject: escape sequences

---
 samples/helloworld.golem |  3 +--
 src/main.c               | 56 +++++++++++++++++++++++++++++++++++++-----------
 testing.sh               |  1 -
 3 files changed, 44 insertions(+), 16 deletions(-)

diff --git a/samples/helloworld.golem b/samples/helloworld.golem
index 37698c0..57333b5 100644
--- a/samples/helloworld.golem
+++ b/samples/helloworld.golem
@@ -5,8 +5,7 @@ define iterations = 9;
 main() {
 	local i;
 
-	strcpy(str, "Hello, World!");
-	strappend(str, '\n');
+	strcpy(str, "Hello, World!\n");
 
 	i = 0;
 	loop {
diff --git a/src/main.c b/src/main.c
index 01ade54..a76d098 100644
--- a/src/main.c
+++ b/src/main.c
@@ -96,6 +96,33 @@ is_punct(const char *p)
 	return (strchr("+-/*()<>,;{}[]=&^|", p[0]) != NULL);
 }
 
+static int
+escaped_char(char *s, char **skipped)
+{
+	if (*s == '\\') {
+		int rv = -1;
+		switch (s[1]) {
+		case 'a': rv = '\a'; break;
+		case 'b': rv = '\b'; break;
+		case 'f': rv = '\f'; break;
+		case 'n': rv = '\n'; break;
+		case 'r': rv = '\r'; break;
+		case 't': rv = '\t'; break;
+		case 'v': rv = '\v'; break;
+		case '\\':rv = '\\'; break;
+		case '\'':rv = '\''; break;
+		case '"': rv = '"';  break;
+		default: break;
+		}
+		if (rv >= 0) {
+			*skipped = s + 2;
+			return rv;
+		}
+	}
+	*skipped = s + 1;
+	return *s;
+}
+
 static Token *
 tokenize(char *p)
 {
@@ -139,17 +166,14 @@ tokenize(char *p)
 			continue;
 		}
 
-		if (*p == '\'' && p[1] != '\0' && p[2] == '\'') {
-			cur = cur->next = new_token(TOK_NUM, p, p + 3);
-			cur->val = (unsigned char)p[1];
-			p += 3;
-			continue;
-		}
-
-		if (strncmp(p, "'\\n'", 4) == 0) {
-			cur = cur->next = new_token(TOK_NUM, p, p + 4);
-			cur->val = '\n';
-			p += 4;
+		if (*p == '\'' && p[1] != '\'') {
+			char *q = p;
+			const int c = escaped_char(p + 1, &p);
+			if (*p != '\'')
+				error("unclosed single quotes", 0);
+			p += 1;
+			cur = cur->next = new_token(TOK_NUM, q, p);
+			cur->val = (unsigned char)c;
 			continue;
 		}
 
@@ -1141,8 +1165,14 @@ codegen(Node *node)
 
 	for (int i = 0; i < strings_size; i++) {
 		printf("@__str_%x\n", strings_size - i - 1);
-		for (int k = 1; k < strings->lhs->len - 1; k++)
-			printf("\t%04x\n", (unsigned)strings->lhs->loc[k]);
+		for (int k = 1; k < strings->lhs->len - 1;) {
+			char *q = strings->lhs->loc + k;
+			const int c = escaped_char(q, &q);
+			if (c == '"') /* TODO: handle this */
+				error("can't escape double quotes", 0);
+			printf("\t%04x\n", (unsigned)c);
+			k = q - strings->lhs->loc;
+		}
 		printf("\t0000\n");
 		strings = strings->next;
 	}
diff --git a/testing.sh b/testing.sh
index 3539b20..7d35ec0 100755
--- a/testing.sh
+++ b/testing.sh
@@ -55,5 +55,4 @@ test "$1" "main() return;"
 test "$1" "main() { local a; a = 5; dbg a; a++; dbg a; }"
 test "$1" "main() { local a; a = 5; dbg a; a--; dbg a; }"
 test "$1" "main() { wrt [\"(\"]; wrt '\n'; }"
-exit
 rm -f build/tmp.*
-- 
cgit v1.2.3