Improve documentation about MULE encoding.

tglsfdc · tglsfdc · commit 09022de1f557 · 2012-07-04T00:29:57.000-04:00
This commit improves the comments in pg_wchar.h and creates #define symbols
for some formerly hard-coded values.  No substantive code changes.

Tatsuo Ishii and Tom Lane
diff --git a/src/backend/utils/mb/conversion_procs/euc_tw_and_big5/euc_tw_and_big5.c b/src/backend/utils/mb/conversion_procs/euc_tw_and_big5/euc_tw_and_big5.c
@@ -168,7 +168,8 @@ euc_tw2mic(const unsigned char *euc, unsigned char *p, int len)
 					*p++ = LC_CNS11643_2;
 				else
 				{
-					*p++ = 0x9d;	/* LCPRV2 */
+					/* other planes are MULE private charsets */
+					*p++ = LCPRV2_B;
 					*p++ = c1 - 0xa3 + LC_CNS11643_3;
 				}
 				*p++ = euc[2];
@@ -235,9 +236,9 @@ mic2euc_tw(const unsigned char *mic, unsigned char *p, int len)
 			*p++ = mic[1];
 			*p++ = mic[2];
 		}
-		else if (c1 == 0x9d &&
+		else if (c1 == LCPRV2_B &&
 				 mic[1] >= LC_CNS11643_3 && mic[1] <= LC_CNS11643_7)
-		{						/* LCPRV2? */
+		{
 			*p++ = SS2;
 			*p++ = mic[1] - LC_CNS11643_3 + 0xa3;
 			*p++ = mic[2];
@@ -286,10 +287,9 @@ big52mic(const unsigned char *big5, unsigned char *p, int len)
 		cnsBuf = BIG5toCNS(big5buf, &lc);
 		if (lc != 0)
 		{
+			/* Planes 3 and 4 are MULE private charsets */
 			if (lc == LC_CNS11643_3 || lc == LC_CNS11643_4)
-			{
-				*p++ = 0x9d;	/* LCPRV2 */
-			}
+				*p++ = LCPRV2_B;
 			*p++ = lc;			/* Plane No. */
 			*p++ = (cnsBuf >> 8) & 0x00ff;
 			*p++ = cnsBuf & 0x00ff;
@@ -332,10 +332,9 @@ mic2big5(const unsigned char *mic, unsigned char *p, int len)
 		if (l < 0)
 			report_invalid_encoding(PG_MULE_INTERNAL,
 									(const char *) mic, len);
-		/* 0x9d means LCPRV2 */
-		if (c1 == LC_CNS11643_1 || c1 == LC_CNS11643_2 || c1 == 0x9d)
+		if (c1 == LC_CNS11643_1 || c1 == LC_CNS11643_2 || c1 == LCPRV2_B)
 		{
-			if (c1 == 0x9d)
+			if (c1 == LCPRV2_B)
 			{
 				c1 = mic[1];	/* get plane no. */
 				cnsBuf = (mic[2] << 8) | mic[3];
diff --git a/src/backend/utils/mb/wchar.c b/src/backend/utils/mb/wchar.c
@@ -742,6 +742,12 @@ pg_mule_dsplen(const unsigned char *s)
 {
 	int			len;
 
+	/*
+	 * Note: it's not really appropriate to assume that all multibyte charsets
+	 * are double-wide on screen.  But this seems an okay approximation for
+	 * the MULE charsets we currently support.
+	 */
+
 	if (IS_LC1(*s))
 		len = 1;
 	else if (IS_LCPRV1(*s))
diff --git a/src/include/mb/pg_wchar.h b/src/include/mb/pg_wchar.h
@@ -36,36 +36,60 @@ typedef unsigned int pg_wchar;
 #define ISSJISHEAD(c) (((c) >= 0x81 && (c) <= 0x9f) || ((c) >= 0xe0 && (c) <= 0xfc))
 #define ISSJISTAIL(c) (((c) >= 0x40 && (c) <= 0x7e) || ((c) >= 0x80 && (c) <= 0xfc))
 
-/*
- * Leading byte types or leading prefix byte for MULE internal code.
- * See http://www.xemacs.org for more details.	(there is a doc titled
- * "XEmacs Internals Manual", "MULE Character Sets and Encodings"
- * section.)
- */
-/*
- * Is a leading byte for "official" single byte encodings?
- */
-#define IS_LC1(c)	((unsigned char)(c) >= 0x81 && (unsigned char)(c) <= 0x8d)
-/*
- * Is a prefix byte for "private" single byte encodings?
- */
-#define IS_LCPRV1(c)	((unsigned char)(c) == 0x9a || (unsigned char)(c) == 0x9b)
-/*
- * Is a leading byte for "official" multibyte encodings?
- */
-#define IS_LC2(c)	((unsigned char)(c) >= 0x90 && (unsigned char)(c) <= 0x99)
-/*
- * Is a prefix byte for "private" multibyte encodings?
- */
-#define IS_LCPRV2(c)	((unsigned char)(c) == 0x9c || (unsigned char)(c) == 0x9d)
-
 /*----------------------------------------------------
- * leading characters
+ * MULE Internal Encoding (MIC)
+ *
+ * This encoding follows the design used within XEmacs; it is meant to
+ * subsume many externally-defined character sets.  Each character includes
+ * identification of the character set it belongs to, so the encoding is
+ * general but somewhat bulky.
+ *
+ * Currently PostgreSQL supports 5 types of MULE character sets:
+ *
+ * 1) 1-byte ASCII characters.  Each byte is below 0x80.
+ *
+ * 2) "Official" single byte charsets such as ISO-8859-1 (Latin1).
+ *    Each MULE character consists of 2 bytes: LC1 + C1, where LC1 is
+ *    an identifier for the charset (in the range 0x81 to 0x8d) and C1
+ *    is the character code (in the range 0xa0 to 0xff).
+ *
+ * 3) "Private" single byte charsets such as SISHENG.  Each MULE
+ *    character consists of 3 bytes: LCPRV1 + LC12 + C1, where LCPRV1
+ *    is a private-charset flag, LC12 is an identifier for the charset,
+ *    and C1 is the character code (in the range 0xa0 to 0xff).
+ *    LCPRV1 is either 0x9a (if LC12 is in the range 0xa0 to 0xdf)
+ *    or 0x9b (if LC12 is in the range 0xe0 to 0xef).
+ *
+ * 4) "Official" multibyte charsets such as JIS X0208.  Each MULE
+ *    character consists of 3 bytes: LC2 + C1 + C2, where LC2 is
+ *    an identifier for the charset (in the range 0x90 to 0x99) and C1
+ *    and C2 form the character code (each in the range 0xa0 to 0xff).
+ *
+ * 5) "Private" multibyte charsets such as CNS 11643-1992 Plane 3.
+ *    Each MULE character consists of 4 bytes: LCPRV2 + LC22 + C1 + C2,
+ *    where LCPRV2 is a private-charset flag, LC22 is an identifier for
+ *    the charset, and C1 and C2 form the character code (each in the range
+ *    0xa0 to 0xff).  LCPRV2 is either 0x9c (if LC22 is in the range 0xf0
+ *    to 0xf4) or 0x9d (if LC22 is in the range 0xf5 to 0xfe).
+ *
+ * "Official" encodings are those that have been assigned code numbers by
+ * the XEmacs project; "private" encodings have Postgres-specific charset
+ * identifiers.
+ *
+ * See the "XEmacs Internals Manual", available at http://www.xemacs.org,
+ * for more details.  Note that for historical reasons, Postgres'
+ * private-charset flag values do not match what XEmacs says they should be,
+ * so this isn't really exactly MULE (not that private charsets would be
+ * interoperable anyway).
  *----------------------------------------------------
  */
 
 /*
- * Official single byte encodings (0x81-0x8e)
+ * Charset identifiers (also called "leading bytes" in the MULE documentation)
+ */
+
+/*
+ * Charset IDs for official single byte encodings (0x81-0x8e)
  */
 #define LC_ISO8859_1		0x81	/* ISO8859 Latin 1 */
 #define LC_ISO8859_2		0x82	/* ISO8859 Latin 2 */
@@ -79,21 +103,19 @@ typedef unsigned int pg_wchar;
 #define LC_JISX0201R		0x8a	/* Japanese 1 byte Roman */
 /* Note that 0x8b seems to be unused as of Emacs 20.7.
  * However, there might be a chance that 0x8b could be used
- * in later version of Emacs.
+ * in later versions of Emacs.
  */
 #define LC_KOI8_R			0x8b	/* Cyrillic KOI8-R */
-#define LC_KOI8_U			0x8b	/* Cyrillic KOI8-U */
 #define LC_ISO8859_5		0x8c	/* ISO8859 Cyrillic */
 #define LC_ISO8859_9		0x8d	/* ISO8859 Latin 5 (not supported yet) */
 /* #define FREE				0x8e	free (unused) */
+/* #define CONTROL_1		0x8f	control characters (unused) */
 
-/*
- * Unused
- */
-#define CONTROL_1			0x8f	/* control characters (unused) */
+/* Is a leading byte for "official" single byte encodings? */
+#define IS_LC1(c)	((unsigned char)(c) >= 0x81 && (unsigned char)(c) <= 0x8d)
 
 /*
- * Official multibyte byte encodings (0x90-0x99)
+ * Charset IDs for official multibyte encodings (0x90-0x99)
  * 0x9a-0x9d are free. 0x9e and 0x9f are reserved.
  */
 #define LC_JISX0208_1978	0x90	/* Japanese Kanji, old JIS (not supported) */
@@ -108,45 +130,70 @@ typedef unsigned int pg_wchar;
 #define LC_BIG5_1			0x98	/* Plane 1 Chinese traditional (not supported) */
 #define LC_BIG5_2			0x99	/* Plane 1 Chinese traditional (not supported) */
 
+/* Is a leading byte for "official" multibyte encodings? */
+#define IS_LC2(c)	((unsigned char)(c) >= 0x90 && (unsigned char)(c) <= 0x99)
+
 /*
- * Private single byte encodings (0xa0-0xef)
+ * Postgres-specific prefix bytes for "private" single byte encodings
+ * (According to the MULE docs, we should be using 0x9e for this)
  */
-#define LC_SISHENG			0xa0/* Chinese SiSheng characters for
-								 * PinYin/ZhuYin (not supported) */
-#define LC_IPA				0xa1/* IPA (International Phonetic Association)
-								 * (not supported) */
-#define LC_VISCII_LOWER		0xa2/* Vietnamese VISCII1.1 lower-case (not
-								 * supported) */
-#define LC_VISCII_UPPER		0xa3/* Vietnamese VISCII1.1 upper-case (not
-								 * supported) */
+#define LCPRV1_A		0x9a
+#define LCPRV1_B		0x9b
+#define IS_LCPRV1(c)	((unsigned char)(c) == LCPRV1_A || (unsigned char)(c) == LCPRV1_B)
+
+/*
+ * Postgres-specific prefix bytes for "private" multibyte encodings
+ * (According to the MULE docs, we should be using 0x9f for this)
+ */
+#define LCPRV2_A		0x9c
+#define LCPRV2_B		0x9d
+#define IS_LCPRV2(c)	((unsigned char)(c) == LCPRV2_A || (unsigned char)(c) == LCPRV2_B)
+
+/*
+ * Charset IDs for private single byte encodings (0xa0-0xef)
+ */
+#define LC_SISHENG			0xa0	/* Chinese SiSheng characters for
+									 * PinYin/ZhuYin (not supported) */
+#define LC_IPA				0xa1	/* IPA (International Phonetic Association)
+									 * (not supported) */
+#define LC_VISCII_LOWER		0xa2	/* Vietnamese VISCII1.1 lower-case (not
+									 * supported) */
+#define LC_VISCII_UPPER		0xa3	/* Vietnamese VISCII1.1 upper-case (not
+									 * supported) */
 #define LC_ARABIC_DIGIT		0xa4	/* Arabic digit (not supported) */
 #define LC_ARABIC_1_COLUMN	0xa5	/* Arabic 1-column (not supported) */
 #define LC_ASCII_RIGHT_TO_LEFT	0xa6	/* ASCII (left half of ISO8859-1) with
 										 * right-to-left direction (not
 										 * supported) */
-#define LC_LAO				0xa7/* Lao characters (ISO10646 0E80..0EDF) (not
-								 * supported) */
+#define LC_LAO				0xa7	/* Lao characters (ISO10646 0E80..0EDF)
+									 * (not supported) */
 #define LC_ARABIC_2_COLUMN	0xa8	/* Arabic 1-column (not supported) */
 
 /*
- * Private multibyte encodings (0xf0-0xff)
+ * Charset IDs for private multibyte encodings (0xf0-0xff)
  */
-#define LC_INDIAN_1_COLUMN	0xf0/* Indian charset for 1-column width glypps
-								 * (not supported) */
-#define LC_TIBETAN_1_COLUMN 0xf1	/* Tibetan 1 column glyph (not supported) */
+#define LC_INDIAN_1_COLUMN	0xf0	/* Indian charset for 1-column width glyphs
+									 * (not supported) */
+#define LC_TIBETAN_1_COLUMN 0xf1	/* Tibetan 1-column width glyphs
+									 * (not supported) */
 #define LC_ETHIOPIC			0xf5	/* Ethiopic characters (not supported) */
 #define LC_CNS11643_3		0xf6	/* CNS 11643-1992 Plane 3 */
 #define LC_CNS11643_4		0xf7	/* CNS 11643-1992 Plane 4 */
 #define LC_CNS11643_5		0xf8	/* CNS 11643-1992 Plane 5 */
 #define LC_CNS11643_6		0xf9	/* CNS 11643-1992 Plane 6 */
 #define LC_CNS11643_7		0xfa	/* CNS 11643-1992 Plane 7 */
-#define LC_INDIAN_2_COLUMN	0xfb/* Indian charset for 2-column width glypps
-								 * (not supported) */
+#define LC_INDIAN_2_COLUMN	0xfb	/* Indian charset for 2-column width glyphs
+									 * (not supported) */
 #define LC_TIBETAN			0xfc	/* Tibetan (not supported) */
 /* #define FREE				0xfd	free (unused) */
 /* #define FREE				0xfe	free (unused) */
 /* #define FREE				0xff	free (unused) */
 
+/*----------------------------------------------------
+ * end of MULE stuff
+ *----------------------------------------------------
+ */
+
 /*
  * PostgreSQL encoding identifiers
  *
