From c09bd088c1125f2e3f96ddbc245a98e4f82772ab Mon Sep 17 00:00:00 2001
From: hyung-hwan <hyunghwan.chung@gmail.com>
Date: Fri, 7 Jan 2022 02:04:28 +0000
Subject: [PATCH] limited the utf8 ranges

---
 hio/lib/utf8.c | 14 +++++++++++++-
 hio/t/t-002.c  | 19 +++++++++++++++++++
 2 files changed, 32 insertions(+), 1 deletion(-)

diff --git a/hio/lib/utf8.c b/hio/lib/utf8.c
index 2d21ec5..8e97c2a 100644
--- a/hio/lib/utf8.c
+++ b/hio/lib/utf8.c
@@ -24,6 +24,8 @@
 
 #include "hio-prv.h"
 
+/*#define RETAIN_RFC2279 1*/
+
 /*
  * from RFC 2279 UTF-8, a transformation format of ISO 10646
  *
@@ -34,6 +36,12 @@
  * 4:4 00010000-001FFFFF  11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
  * inv 00200000-03FFFFFF  111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
  * inv 04000000-7FFFFFFF  1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx
+ *
+ * RFC3629 limits the ranges like this:
+ * 1:2 00000000-0000007F  0xxxxxxx
+ * 2:2 00000080-000007FF  110xxxxx 10xxxxxx
+ * 3:2 00000800-0000FFFF  1110xxxx 10xxxxxx 10xxxxxx
+ * 4:4 00010000-0010FFFF  11110xxx 10xxxxxx 10xxxxxx 10xxxxxx
  */
 
 struct __utf8_t
@@ -53,9 +61,13 @@ static __utf8_t utf8_table[] =
 	{0x00000000ul, 0x0000007Ful, 0x00, 0x80, 0x7F, 1},
 	{0x00000080ul, 0x000007FFul, 0xC0, 0xE0, 0x1F, 2},
 	{0x00000800ul, 0x0000FFFFul, 0xE0, 0xF0, 0x0F, 3},
+#if defined(RETAIN_RFC2279)
 	{0x00010000ul, 0x001FFFFFul, 0xF0, 0xF8, 0x07, 4},
 	{0x00200000ul, 0x03FFFFFFul, 0xF8, 0xFC, 0x03, 5},
 	{0x04000000ul, 0x7FFFFFFFul, 0xFC, 0xFE, 0x01, 6}
+#else
+	{0x00010000ul, 0x0010FFFFul, 0xF0, 0xF8, 0x07, 4}
+#endif
 };
 
 static HIO_INLINE __utf8_t* get_utf8_slot (hio_uch_t uc)
@@ -79,7 +91,7 @@ static HIO_INLINE __utf8_t* get_utf8_slot (hio_uch_t uc)
 
 hio_oow_t hio_uc_to_utf8 (hio_uch_t uc, hio_bch_t* utf8, hio_oow_t size)
 {
-	__utf8_t* cur = get_utf8_slot (uc);
+	__utf8_t* cur = get_utf8_slot(uc);
 
 	if (cur == HIO_NULL) return 0; /* illegal character */
 
diff --git a/hio/t/t-002.c b/hio/t/t-002.c
index 484a3ac..08f657d 100644
--- a/hio/t/t-002.c
+++ b/hio/t/t-002.c
@@ -50,6 +50,25 @@ int main ()
 		T_ASSERT1 (v == 0 && *endptr == '\0' && is_sober == 1, "integer in E notation");
 	}
 
+	{
+		hio_bch_t tmp[10];
+		hio_oow_t x;
+		hio_uch_t uc;
+
+		x = hio_uc_to_utf8(0x2665, tmp, HIO_COUNTOF(tmp));
+		T_ASSERT1 (x == 3 && (hio_uint8_t)tmp[0] == 0xE2 && (hio_uint8_t)tmp[1] == 0x99 && (hio_uint8_t)tmp[2] == 0xA5, "unicode to utf8 conversion");
+
+		x = hio_utf8_to_uc(tmp, x, &uc);
+		T_ASSERT1 (x == 3 && uc == 0x2665, "utf8 to unicode conversion");
+
+	#if (HIO_SIZEOF_UCH_T > 2)
+		x = hio_uc_to_utf8(0x1F3E9, tmp, HIO_COUNTOF(tmp));
+		T_ASSERT1 (x == 4 && (hio_uint8_t)tmp[0] == 0xF0 && (hio_uint8_t)tmp[1] == 0x9F && (hio_uint8_t)tmp[2] == 0x8F && (hio_uint8_t)tmp[3] == 0xA9, "unicode to utf8 conversion");
+
+		x = hio_utf8_to_uc(tmp, x, &uc);
+		T_ASSERT1 (x == 4 && uc == 0x1F3E9, "utf8 to unicode conversion");
+	#endif
+	}
 	return 0;
 
 oops: