[Minor] Rework utf8 lowercasing

1 files changed, 19 insertions, 27 deletions

diff --git a/src/libutil/str_util.c b/src/libutil/str_util.c
index 1f2c4629f..4ce84fa65 100644
--- a/src/libutil/str_util.c
+++ b/src/libutil/str_util.c
@@ -62,7 +62,7 @@ const guchar lc_map[256] = {
 		0xf8, 0xf9, 0xfa, 0xfb, 0xfc, 0xfd, 0xfe, 0xff
 };
 
-void
+guint
 rspamd_str_lc (gchar *str, guint size)
 {
 	guint leftover = size % 4;
@@ -93,6 +93,7 @@ rspamd_str_lc (gchar *str, guint size)
 		*dest = lc_map[(guchar)str[i]];
 	}
 
+	return size;
 }
 
 gint
@@ -144,42 +145,33 @@ rspamd_lc_cmp (const gchar *s, const gchar *d, gsize l)
  * string to lower case, so some locale peculiarities are simply ignored
  * If the target string is longer than initial one, then we just trim it
  */
-void
+guint
 rspamd_str_lc_utf8 (gchar *str, guint size)
 {
-	const gchar *s = str, *p;
-	gchar *d = str, tst[6];
-	gint remain = size;
-	gint r;
-	gunichar uc;
+	guchar *d = (guchar *)str, tst[6];
+	gint32 i = 0, prev = 0;
+	UChar32 uc;
 
-	while (remain > 0) {
-		p = g_utf8_next_char (s);
+	while (i < size) {
+		prev = i;
 
-		if (p - s > remain) {
-			break;
-		}
+		U8_NEXT ((guint8*)str, i, size, uc);
+		uc = u_tolower (uc);
 
-		uc = g_utf8_get_char (s);
-		uc = g_unichar_tolower (uc);
+		gint32 olen = 0;
+		U8_APPEND_UNSAFE (tst, olen, uc);
 
-		if (remain >= 6) {
-			r = g_unichar_to_utf8 (uc, d);
+		if (olen <= (i - prev)) {
+			memcpy (d, tst, olen);
+			d += olen;
 		}
 		else {
-			/* We must be cautious here to avoid broken unicode being append */
-			r = g_unichar_to_utf8 (uc, tst);
-			if (r > remain) {
-				break;
-			}
-			else {
-				memcpy (d, tst, r);
-			}
+			/* Lowercasing has increased the length, so we need to ignore it */
+			d += i - prev;
 		}
-		remain -= r;
-		s = p;
-		d += r;
 	}
+
+	return d - (guchar *)str;
 }
 
 gboolean