Add cheot-ga-keut composed make-up sequence support in EUC-KR codec.

Lib/test/cjkencodings_test.py

@@ -376,13 +376,20 @@ teststring = {
 "\xcc\xc7\xce\x2c\x20\xb1\xd7\xb8\xae\xb0\xed\x20\xc0\xce\xc5\xcd"
 "\xc7\xc1\xb8\xae\xc6\xc3\x0a\xc8\xaf\xb0\xe6\xc0\xba\x20\xc6\xc4"
 "\xc0\xcc\xbd\xe3\xc0\xbb\x20\xbd\xba\xc5\xa9\xb8\xb3\xc6\xc3\xb0"
-"\xfa\x20\xbf\xa9\xb7\xc1\x20\xba\xd0\xbe\xdf\xbf\xa1\xbc\xad\xbf"
+"\xfa\x20\xbf\xa9\xb7\xaf\x20\xba\xd0\xbe\xdf\xbf\xa1\xbc\xad\xbf"
 "\xcd\x20\xb4\xeb\xba\xce\xba\xd0\xc0\xc7\x20\xc7\xc3\xb7\xa7\xc6"
 "\xfb\xbf\xa1\xbc\xad\xc0\xc7\x20\xba\xfc\xb8\xa5\x0a\xbe\xd6\xc7"
 "\xc3\xb8\xae\xc4\xc9\xc0\xcc\xbc\xc7\x20\xb0\xb3\xb9\xdf\xc0\xbb"
 "\x20\xc7\xd2\x20\xbc\xf6\x20\xc0\xd6\xb4\xc2\x20\xc0\xcc\xbb\xf3"
 "\xc0\xfb\xc0\xce\x20\xbe\xf0\xbe\xee\xb7\xce\x20\xb8\xb8\xb5\xe9"
-"\xbe\xee\xc1\xdd\xb4\xcf\xb4\xd9\x2e\x0a\x0a",
+"\xbe\xee\xc1\xdd\xb4\xcf\xb4\xd9\x2e\x0a\x0a\xa1\xd9\xc3\xb9\xb0"
+"\xa1\xb3\xa1\x3a\x20\xb3\xaf\xbe\xc6\xb6\xf3\x20\xa4\xd4\xa4\xb6"
+"\xa4\xd0\xa4\xd4\xa4\xd4\xa4\xb6\xa4\xd0\xa4\xd4\xbe\xb1\x7e\x20"
+"\xa4\xd4\xa4\xa4\xa4\xd2\xa4\xb7\xc5\xad\x21\x20\xa4\xd4\xa4\xa8"
+"\xa4\xd1\xa4\xb7\xb1\xdd\xbe\xf8\xc0\xcc\x20\xc0\xfc\xa4\xd4\xa4"
+"\xbe\xa4\xc8\xa4\xb2\xb4\xcf\xb4\xd9\x2e\x20\xa4\xd4\xa4\xb2\xa4"
+"\xce\xa4\xaa\x2e\x20\xb1\xd7\xb7\xb1\xb0\xc5\x20\xa4\xd4\xa4\xb7"
+"\xa4\xd1\xa4\xb4\xb4\xd9\x2e\x0a",
 "\xe2\x97\x8e\x20\xed\x8c\x8c\xec\x9d\xb4\xec\x8d\xac\x28\x50\x79"
 "\x74\x68\x6f\x6e\x29\xec\x9d\x80\x20\xeb\xb0\xb0\xec\x9a\xb0\xea"
 "\xb8\xb0\x20\xec\x89\xbd\xea\xb3\xa0\x2c\x20\xea\xb0\x95\xeb\xa0"
@@ -404,7 +411,7 @@ teststring = {
 "\xec\x9d\xb8\xed\x84\xb0\xed\x94\x84\xeb\xa6\xac\xed\x8c\x85\x0a"
 "\xed\x99\x98\xea\xb2\xbd\xec\x9d\x80\x20\xed\x8c\x8c\xec\x9d\xb4"
 "\xec\x8d\xac\xec\x9d\x84\x20\xec\x8a\xa4\xed\x81\xac\xeb\xa6\xbd"
-"\xed\x8c\x85\xea\xb3\xbc\x20\xec\x97\xac\xeb\xa0\xa4\x20\xeb\xb6"
+"\xed\x8c\x85\xea\xb3\xbc\x20\xec\x97\xac\xeb\x9f\xac\x20\xeb\xb6"
 "\x84\xec\x95\xbc\xec\x97\x90\xec\x84\x9c\xec\x99\x80\x20\xeb\x8c"
 "\x80\xeb\xb6\x80\xeb\xb6\x84\xec\x9d\x98\x20\xed\x94\x8c\xeb\x9e"
 "\xab\xed\x8f\xbc\xec\x97\x90\xec\x84\x9c\xec\x9d\x98\x20\xeb\xb9"
@@ -413,7 +420,13 @@ teststring = {
 "\x84\x20\xed\x95\xa0\x20\xec\x88\x98\x20\xec\x9e\x88\xeb\x8a\x94"
 "\x20\xec\x9d\xb4\xec\x83\x81\xec\xa0\x81\xec\x9d\xb8\x20\xec\x96"
 "\xb8\xec\x96\xb4\xeb\xa1\x9c\x20\xeb\xa7\x8c\xeb\x93\xa4\xec\x96"
-"\xb4\xec\xa4\x8d\xeb\x8b\x88\xeb\x8b\xa4\x2e\x0a\x0a"),
+"\xb4\xec\xa4\x8d\xeb\x8b\x88\xeb\x8b\xa4\x2e\x0a\x0a\xe2\x98\x86"
+"\xec\xb2\xab\xea\xb0\x80\xeb\x81\x9d\x3a\x20\xeb\x82\xa0\xec\x95"
+"\x84\xeb\x9d\xbc\x20\xec\x93\x94\xec\x93\x94\xec\x93\xa9\x7e\x20"
+"\xeb\x8b\x81\xed\x81\xbc\x21\x20\xeb\x9c\xbd\xea\xb8\x88\xec\x97"
+"\x86\xec\x9d\xb4\x20\xec\xa0\x84\xed\x99\xa5\xeb\x8b\x88\xeb\x8b"
+"\xa4\x2e\x20\xeb\xb7\x81\x2e\x20\xea\xb7\xb8\xeb\x9f\xb0\xea\xb1"
+"\xb0\x20\xec\x9d\x8e\xeb\x8b\xa4\x2e\x0a"),
 'gb18030': (
 "\x50\x79\x74\x68\x6f\x6e\xa3\xa8\xc5\xc9\xc9\xad\xa3\xa9\xd3\xef"
 "\xd1\xd4\xca\xc7\xd2\xbb\xd6\xd6\xb9\xa6\xc4\xdc\xc7\xbf\xb4\xf3"

Lib/test/test_codecencodings_kr.py

@@ -30,6 +30,24 @@ class Test_EUCKR(test_multibytecodec_support.TestBase, unittest.TestCase):
         ("abc\x80\x80\xc1\xc4", "replace", u"abc\ufffd\uc894"),
         ("abc\x80\x80\xc1\xc4\xc8", "replace", u"abc\ufffd\uc894\ufffd"),
         ("abc\x80\x80\xc1\xc4", "ignore",  u"abc\uc894"),
+
+        # composed make-up sequence errors
+        ("\xa4\xd4", "strict", None),
+        ("\xa4\xd4\xa4", "strict", None),
+        ("\xa4\xd4\xa4\xb6", "strict", None),
+        ("\xa4\xd4\xa4\xb6\xa4", "strict", None),
+        ("\xa4\xd4\xa4\xb6\xa4\xd0", "strict", None),
+        ("\xa4\xd4\xa4\xb6\xa4\xd0\xa4", "strict", None),
+        ("\xa4\xd4\xa4\xb6\xa4\xd0\xa4\xd4", "strict", u"\uc4d4"),
+        ("\xa4\xd4\xa4\xb6\xa4\xd0\xa4\xd4x", "strict", u"\uc4d4x"),
+        ("a\xa4\xd4\xa4\xb6\xa4", "replace", u"a\ufffd"),
+        ("\xa4\xd4\xa3\xb6\xa4\xd0\xa4\xd4", "strict", None),
+        ("\xa4\xd4\xa4\xb6\xa3\xd0\xa4\xd4", "strict", None),
+        ("\xa4\xd4\xa4\xb6\xa4\xd0\xa3\xd4", "strict", None),
+        ("\xa4\xd4\xa4\xff\xa4\xd0\xa4\xd4", "replace", u"\ufffd"),
+        ("\xa4\xd4\xa4\xb6\xa4\xff\xa4\xd4", "replace", u"\ufffd"),
+        ("\xa4\xd4\xa4\xb6\xa4\xd0\xa4\xff", "replace", u"\ufffd"),
+        ("\xc1\xc4", "strict", u"\uc894"),
     )
 
 class Test_JOHAB(test_multibytecodec_support.TestBase, unittest.TestCase):

Lib/test/test_codecmaps_kr.py

@@ -20,6 +20,10 @@ class TestEUCKRMap(test_multibytecodec_support.TestBase_Mapping,
     encoding = 'euc_kr'
     mapfileurl = 'http://people.freebsd.org/~perky/i18n/EUC-KR.TXT'
 
+    # A4D4 HANGUL FILLER indicates the begin of 8-bytes make-up sequence.
+    pass_enctest = [('\xa4\xd4', u'\u3164')]
+    pass_dectest = [('\xa4\xd4', u'\u3164')]
+
 
 class TestJOHABMap(test_multibytecodec_support.TestBase_Mapping,
                    unittest.TestCase):

Misc/NEWS

@@ -240,6 +240,9 @@ Core and builtins
 Library
 -------
 
+- EUC-KR codec now handles the cheot-ga-keut composed make-up hangul
+  syllables.
+
 - GB18030 codec now can encode additional two-byte characters that
   are missing in GBK.
 

Modules/cjkcodecs/_codecs_kr.c

@@ -11,6 +11,26 @@
  * EUC-KR codec
  */
 
+#define EUCKR_JAMO_FIRSTBYTE	0xA4
+#define EUCKR_JAMO_FILLER	0xD4
+
+static const unsigned char u2cgk_choseong[19] = {
+	0xa1, 0xa2, 0xa4, 0xa7, 0xa8, 0xa9, 0xb1, 0xb2,
+	0xb3, 0xb5, 0xb6, 0xb7, 0xb8, 0xb9, 0xba, 0xbb,
+	0xbc, 0xbd, 0xbe
+};
+static const unsigned char u2cgk_jungseong[21] = {
+	0xbf, 0xc0, 0xc1, 0xc2, 0xc3, 0xc4, 0xc5, 0xc6,
+	0xc7, 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce,
+	0xcf, 0xd0, 0xd1, 0xd2, 0xd3
+};
+static const unsigned char u2cgk_jongseong[28] = {
+	0xd4, 0xa1, 0xa2, 0xa3, 0xa4, 0xa5, 0xa6, 0xa7,
+	0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 0xaf, 0xb0,
+	0xb1, 0xb2, 0xb4, 0xb5, 0xb6, 0xb7, 0xb8, 0xba,
+	0xbb, 0xbc, 0xbd, 0xbe
+};
+
 ENCODER(euc_kr)
 {
 	while (inleft > 0) {
@@ -28,17 +48,57 @@ ENCODER(euc_kr)
 		TRYMAP_ENC(cp949, code, c);
 		else return 1;
 
-		if (code & 0x8000) /* MSB set: CP949 */
-			return 1;
-
+		if ((code & 0x8000) == 0) {
+			/* KS X 1001 coded character */
 			OUT1((code >> 8) | 0x80)
 			OUT2((code & 0xFF) | 0x80)
 			NEXT(1, 2)
 		}
+		else {	/* Mapping is found in CP949 extension,
+			 * but we encode it in KS X 1001:1998 Annex 3,
+			 * make-up sequence for EUC-KR. */
+
+			REQUIRE_OUTBUF(8)
+
+			/* syllable composition precedence */
+			OUT1(EUCKR_JAMO_FIRSTBYTE)
+			OUT2(EUCKR_JAMO_FILLER)
+
+			/* All codepoints in CP949 extension are in unicode
+			 * Hangul Syllable area. */
+			assert(0xac00 <= c && c <= 0xd7a3);
+			c -= 0xac00;
+
+			OUT3(EUCKR_JAMO_FIRSTBYTE)
+			OUT4(u2cgk_choseong[c / 588])
+			NEXT_OUT(4)
+
+			OUT1(EUCKR_JAMO_FIRSTBYTE)
+			OUT2(u2cgk_jungseong[(c / 28) % 21])
+			OUT3(EUCKR_JAMO_FIRSTBYTE)
+			OUT4(u2cgk_jongseong[c % 28])
+			NEXT(1, 4)
+		}
+	}
 
 	return 0;
 }
 
+#define NONE	127
+
+static const unsigned char cgk2u_choseong[] = { /* [A1, BE] */
+	   0,    1, NONE,    2, NONE, NONE,    3,    4,
+	   5, NONE, NONE, NONE, NONE, NONE, NONE, NONE,
+	   6,    7,    8, NONE,    9,   10,   11,   12,
+	  13,   14,   15,   16,   17,   18
+};
+static const unsigned char cgk2u_jongseong[] = { /* [A1, BE] */
+	   1,    2,    3,    4,    5,    6,    7, NONE,
+	   8,    9,   10,   11,   12,   13,   14,   15,
+	  16,   17, NONE,   18,   19,   20,   21,   22,
+	NONE,   23,   24,   25,   26,   27
+};
+
 DECODER(euc_kr)
 {
 	while (inleft > 0) {
@@ -54,13 +114,50 @@ DECODER(euc_kr)
 
 		REQUIRE_INBUF(2)
 
-		TRYMAP_DEC(ksx1001, **outbuf, c ^ 0x80, IN2 ^ 0x80) {
+		if (c == EUCKR_JAMO_FIRSTBYTE &&
+		    IN2 == EUCKR_JAMO_FILLER) {
+			/* KS X 1001:1998 Annex 3 make-up sequence */
+			DBCHAR cho, jung, jong;
+
+			REQUIRE_INBUF(8)
+			if ((*inbuf)[2] != EUCKR_JAMO_FIRSTBYTE ||
+			    (*inbuf)[4] != EUCKR_JAMO_FIRSTBYTE ||
+			    (*inbuf)[6] != EUCKR_JAMO_FIRSTBYTE)
+				return 8;
+
+			c = (*inbuf)[3];
+			if (0xa1 <= c && c <= 0xbe)
+				cho = cgk2u_choseong[c - 0xa1];
+			else
+				cho = NONE;
+
+			c = (*inbuf)[5];
+			jung = (0xbf <= c && c <= 0xd3) ? c - 0xbf : NONE;
+
+			c = (*inbuf)[7];
+			if (c == EUCKR_JAMO_FILLER)
+				jong = 0;
+			else if (0xa1 <= c && c <= 0xbe)
+				jong = cgk2u_jongseong[c - 0xa1];
+			else
+				jong = NONE;
+
+			if (cho == NONE || jung == NONE || jong == NONE)
+				return 8;
+
+			OUT1(0xac00 + cho*588 + jung*28 + jong);
+			NEXT(8, 1)
+		}
+		else TRYMAP_DEC(ksx1001, **outbuf, c ^ 0x80, IN2 ^ 0x80) {
 			NEXT(2, 1)
-		} else return 2;
+		}
+		else
+			return 2;
 	}
 
 	return 0;
 }
+#undef NONE
 
 
 /*