* handle invalid utf8; fixes #211 (hopefully)

lib/mu-str.c

@@ -475,6 +475,16 @@ process_str (const char *str, gboolean xapian_esc, gboolean query_esc)
 	char *norm, *cur;
 
 	norm = g_utf8_normalize (str, -1, G_NORMALIZE_ALL);
+	if (G_UNLIKELY(!norm)) {  /* not valid utf8? */
+		char *u8;
+		u8 = mu_str_utf8ify (str);
+		norm = g_utf8_normalize (u8, -1, G_NORMALIZE_ALL);
+		g_free (u8);
+	}
+
+ 	if (!norm)
+		return NULL;
+
 	gstr = g_string_sized_new (strlen (norm));
 
 	for (cur = norm; cur && *cur; cur = g_utf8_next_char (cur)) {

lib/tests/test-mu-str.c

@@ -221,7 +221,9 @@ test_mu_str_process_term (void)
 		{ "Masha@Аркона.ru", "masha_аркона_ru" },
 		{ "foo:ελληνικά", "foo_ελληνικα" },
 		{ "日本語!!", "日本語__" },
-		{ "￡", "_" }
+		{ "￡", "_" },
+		/* invalid utf8 */
+		{ "Hello\xC3\x2EWorld", "hello__world" }
 	};
 
 	for (i = 0; i != G_N_ELEMENTS(words); ++i) {