Add KOI8/WIN/ALT support

diff --git a/src/backend/utils/mb/conv.c b/src/backend/utils/mb/conv.c
index 9aac4e9..e2b6ff0 100644 (file)
--- a/src/backend/utils/mb/conv.c
+++ b/src/backend/utils/mb/conv.c
@@ -2,7 +2,7 @@
  * conversion between client encoding and server internal encoding
  * (currently mule internal code (mic) is used)
  * Tatsuo Ishii
- * $Id: conv.c,v 1.5 1999/02/02 18:51:23 momjian Exp $
+ * $Id: conv.c,v 1.6 1999/03/24 07:02:16 ishii Exp $
  */
 #include <stdio.h>
 #include <string.h>
@@ -588,6 +588,262 @@ mic2ascii(unsigned char *mic, unsigned char *p, int len)
        *p = '\0';
 }
 
+/* 
+ * Cyrillic support
+ * currently supported Cyrillic encodings:
+ *
+ * KOI8-R (this is the charset for the mule internal code
+ *             for Cyrillic)
+ * ISO-8859-5
+ * Microsoft's CP1251(windows-1251)
+ * Alternativny Variant (MS-DOS CP866)
+ */
+
+/* koi2mic: KOI8-R to Mule internal code */ 
+static void
+koi2mic(unsigned char *l, unsigned char *p, int len)
+{
+  latin2mic(l, p, len, LC_KOI8_R);
+}
+
+/* mic2koi: Mule internal code to KOI8-R */
+static void
+mic2koi(unsigned char *mic, unsigned char *p, int len)
+{
+  mic2latin(mic, p, len, LC_KOI8_R);
+}
+
+/*
+ * latin2mic_with_table: a generic single byte charset encoding
+ * conversion from a local charset to the mule internal code.
+ * with a encoding conversion table.
+ * the table is ordered according to the local charset,
+ * starting from 128 (0x80). each entry in the table
+ * holds the corresponding code point for the mule internal code.
+ */
+static void
+latin2mic_with_table(
+                    unsigned char *l,  /* local charset string (source) */
+                    unsigned char *p,  /* pointer to store mule internal code
+                                          (destination) */
+                    int len,           /* length of l */
+                    int lc,            /* leading character of p */
+                    unsigned char *tab /* code conversion table */
+                    )
+{
+  unsigned char        c1,c2;
+  
+  while (len-- > 0 && (c1 = *l++)) {
+    if (c1 < 128) {
+      *p++ = c1;
+    } else {
+      c2 = tab[c1 - 128];
+      if (c2) {
+       *p++ = lc;
+       *p++ = c2;
+      } else {
+       *p++ = ' ';     /* cannot convert */
+      }
+    }
+  }
+  *p = '\0';
+}
+
+/*
+ * mic2latin_with_table: a generic single byte charset encoding
+ * conversion from the mule internal code to a local charset
+ * with a encoding conversion table.
+ * the table is ordered according to the second byte of the mule
+ * internal code starting from 128 (0x80). 
+ * each entry in the table
+ * holds the corresponding code point for the local code.
+ */
+static void
+mic2latin_with_table(
+                    unsigned char *mic,        /* mule internal code (source) */
+                    unsigned char *p,          /* local code (destination) */
+                    int len,                   /* length of p */
+                    int lc,                    /* leading character */
+                    unsigned char *tab         /* code conversion table */
+                    )
+{
+
+  unsigned char        c1,c2;
+
+  while (len-- > 0 && (c1 = *mic++)) {
+    if (c1 < 128) {
+      *p++ = c1;
+    } else if (c1 == lc) {
+      c1 = *mic++;
+      len--;
+      c2 = tab[c1 - 128];
+      if (c2) {
+       *p++ = c2;
+      } else {
+       *p++ = ' ';     /* cannot convert */
+      }
+    } else {
+      *p++ = ' ';      /* bogus character */
+    }
+  }
+  *p = '\0';
+}
+
+/* iso2mic: ISO-8859-5 to Mule internal code */ 
+static void
+iso2mic(unsigned char *l, unsigned char *p, int len)
+{
+  static char iso2koi[] = {
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0xe1, 0xe2, 0xf7, 0xe7, 0xe4, 0xe5, 0xf6, 0xfa, 
+    0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, 0xf0, 
+    0xf2, 0xf3, 0xf4, 0xf5, 0xe6, 0xe8, 0xe3, 0xfe, 
+    0xfb, 0xfd, 0xff, 0xf9, 0xf8, 0xfc, 0xe0, 0xf1, 
+    0xc1, 0xc2, 0xd7, 0xc7, 0xc4, 0xc5, 0xd6, 0xda, 
+    0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, 0xd0, 
+    0xd2, 0xd3, 0xd4, 0xd5, 0xc6, 0xc8, 0xc3, 0xde, 
+    0xdb, 0xdd, 0xdf, 0xd9, 0xd8, 0xdc, 0xc0, 0xd1, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
+  };
+  latin2mic_with_table(l, p, len, LC_KOI8_R, iso2koi);
+}
+
+/* mic2iso: Mule internal code to ISO8859-5 */
+static void
+mic2iso(unsigned char *mic, unsigned char *p, int len)
+{
+  static char koi2iso[] = {
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0xee, 0xd0, 0xd1, 0xe6, 0xd4, 0xd5, 0xe4, 0xd3, 
+    0xe5, 0xd8, 0xd9, 0xda, 0xdb, 0xdc, 0xdd, 0xde, 
+    0xdf, 0xef, 0xe0, 0xe1, 0xe2, 0xe3, 0xd6, 0xd2, 
+    0xec, 0xeb, 0xd7, 0xe8, 0xed, 0xe9, 0xe7, 0xea, 
+    0xce, 0xb0, 0xb1, 0xc6, 0xb4, 0xb5, 0xc4, 0xb3, 
+    0xc5, 0xb8, 0xb9, 0xba, 0xbb, 0xbc, 0xbd, 0xbe, 
+    0xbf, 0xcf, 0xc0, 0xc1, 0xc2, 0xc3, 0xb6, 0xb2, 
+    0xcc, 0xcb, 0xb7, 0xc8, 0xcd, 0xc9, 0xc7, 0xca
+  };
+
+  mic2latin_with_table(mic, p, len, LC_KOI8_R, koi2iso);
+}
+
+/* win2mic: CP1251 to Mule internal code */ 
+static void
+win2mic(unsigned char *l, unsigned char *p, int len)
+{
+  static char win2koi[] = {
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xbd, 0x00, 0x00, 
+    0xb3, 0x00, 0xb4, 0x00, 0x00, 0x00, 0x00, 0xb7, 
+    0x00, 0x00, 0xb6, 0xa6, 0xad, 0x00, 0x00, 0x00, 
+    0xa3, 0x00, 0xa4, 0x00, 0x00, 0x00, 0x00, 0xa7, 
+    0xe1, 0xe2, 0xf7, 0xe7, 0xe4, 0xe5, 0xf6, 0xfa, 
+    0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, 0xf0, 
+    0xf2, 0xf3, 0xf4, 0xf5, 0xe6, 0xe8, 0xe3, 0xfe, 
+    0xfb, 0xfd, 0xff, 0xf9, 0xf8, 0xfc, 0xe0, 0xf1, 
+    0xc1, 0xc2, 0xd7, 0xc7, 0xc4, 0xc5, 0xd6, 0xda, 
+    0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, 0xd0, 
+    0xd2, 0xd3, 0xd4, 0xd5, 0xc6, 0xc8, 0xc3, 0xde, 
+    0xdb, 0xdd, 0xdf, 0xd9, 0xd8, 0xdc, 0xc0, 0xd1
+  };
+  latin2mic_with_table(l, p, len, LC_KOI8_R, win2koi);
+}
+
+/* mic2win: Mule internal code to CP1251 */
+static void
+mic2win(unsigned char *mic, unsigned char *p, int len)
+{
+  static char koi2win[] = {
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0xb8, 0xba, 0x00, 0xb3, 0xbf, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xb4, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0xa8, 0xaa, 0x00, 0xb2, 0xaf, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xa5, 0x00, 0x00, 
+    0xfe, 0xe0, 0xe1, 0xf6, 0xe4, 0xe5, 0xf4, 0xe3, 
+    0xf5, 0xe8, 0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 
+    0xef, 0xff, 0xf0, 0xf1, 0xf2, 0xf3, 0xe6, 0xe2, 
+    0xfc, 0xfb, 0xe7, 0xf8, 0xfd, 0xf9, 0xf7, 0xfa, 
+    0xde, 0xc0, 0xc1, 0xd6, 0xc4, 0xc5, 0xd4, 0xc3, 
+    0xd5, 0xc8, 0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 
+    0xcf, 0xdf, 0xd0, 0xd1, 0xd2, 0xd3, 0xc6, 0xc2, 
+    0xdc, 0xdb, 0xc7, 0xd8, 0xdd, 0xd9, 0xd7, 0xda
+  };
+  mic2latin_with_table(mic, p, len, LC_KOI8_R, koi2win);
+}
+
+/* alt2mic: CP866 to Mule internal code */ 
+static void
+alt2mic(unsigned char *l, unsigned char *p, int len)
+{
+  static char alt2koi[] = {
+    0xe1, 0xe2, 0xf7, 0xe7, 0xe4, 0xe5, 0xf6, 0xfa, 
+    0xe9, 0xea, 0xeb, 0xec, 0xed, 0xee, 0xef, 0xf0, 
+    0xf2, 0xf3, 0xf4, 0xf5, 0xe6, 0xe8, 0xe3, 0xfe, 
+    0xfb, 0xfd, 0xff, 0xf9, 0xf8, 0xfc, 0xe0, 0xf1, 
+    0xc1, 0xc2, 0xd7, 0xc7, 0xc4, 0xc5, 0xd6, 0xda, 
+    0xc9, 0xca, 0xcb, 0xcc, 0xcd, 0xce, 0xcf, 0xd0, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xbd, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0xd2, 0xd3, 0xd4, 0xd5, 0xc6, 0xc8, 0xc3, 0xde, 
+    0xdb, 0xdd, 0xdf, 0xd9, 0xd8, 0xdc, 0xc0, 0xd1, 
+    0xb3, 0xa3, 0xb4, 0xa4, 0xb7, 0xa7, 0x00, 0x00, 
+    0xb6, 0xa6, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00
+  };
+  latin2mic_with_table(l, p, len, LC_KOI8_R, alt2koi);
+}
+
+/* mic2alt: Mule internal code to CP866 */
+static void
+mic2alt(unsigned char *mic, unsigned char *p, int len)
+{
+  static char koi2alt[] = {
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0xf1, 0xf3, 0x00, 0xf9, 0xf5, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xad, 0x00, 0x00, 
+    0x00, 0x00, 0x00, 0xf0, 0xf2, 0x00, 0xf8, 0xf4, 
+    0x00, 0x00, 0x00, 0x00, 0x00, 0xbd, 0x00, 0x00, 
+    0xee, 0xa0, 0xa1, 0xe6, 0xa4, 0xa5, 0xe4, 0xa3, 
+    0xe5, 0xa8, 0xa9, 0xaa, 0xab, 0xac, 0xad, 0xae, 
+    0xaf, 0xef, 0xe0, 0xe1, 0xe2, 0xe3, 0xa6, 0xa2, 
+    0xec, 0xeb, 0xa7, 0xe8, 0xed, 0xe9, 0xe7, 0xea, 
+    0x9e, 0x80, 0x81, 0x96, 0x84, 0x85, 0x94, 0x83, 
+    0x95, 0x88, 0x89, 0x8a, 0x8b, 0x8c, 0x8d, 0x8e, 
+    0x8f, 0x9f, 0x90, 0x91, 0x92, 0x93, 0x86, 0x82, 
+    0x9c, 0x9b, 0x87, 0x98, 0x9d, 0x99, 0x97, 0x9a
+  };
+  mic2latin_with_table(mic, p, len, LC_KOI8_R, koi2alt);
+}
+
+/*
+ * end of Cyrillic support
+ */
+
 pg_encoding_conv_tbl pg_conv_tbl[] = {
        {SQL_ASCII, "SQL_ASCII", 0, ascii2mic, mic2ascii},      /* SQL/ACII */
        {EUC_JP, "EUC_JP", 0, euc_jp2mic, mic2euc_jp},          /* EUC_JP */
@@ -600,7 +856,10 @@ pg_encoding_conv_tbl pg_conv_tbl[] = {
        {LATIN2, "LATIN2", 0, latin22mic, mic2latin2},          /* ISO 8859 Latin 2 */
        {LATIN3, "LATIN3", 0, latin32mic, mic2latin3},          /* ISO 8859 Latin 3 */
        {LATIN4, "LATIN4", 0, latin42mic, mic2latin4},          /* ISO 8859 Latin 4 */
-       {LATIN5, "LATIN5", 0, latin52mic, mic2latin5},          /* ISO 8859 Latin 5 */
+       {LATIN5, "LATIN5", 0, iso2mic, mic2iso},                /* ISO 8859 Latin 5 */
+       {KOI8, "KOI8", 0, koi2mic, mic2koi},    /* KOI8-R */
+       {WIN, "WIN", 0, win2mic, mic2win},      /* CP1251 */
+       {ALT, "ALT", 0, alt2mic, mic2alt},      /* CP866 */
        {SJIS, "SJIS", 1, sjis2mic, mic2sjis},          /* SJIS */
        {BIG5, "BIG5", 1, big52mic, mic2big5},          /* Big5 */
        {-1, "", 0, 0, 0}                       /* end mark */

diff --git a/src/backend/utils/mb/wchar.c b/src/backend/utils/mb/wchar.c
index 2a1141f..78f22c1 100644 (file)
--- a/src/backend/utils/mb/wchar.c
+++ b/src/backend/utils/mb/wchar.c
@@ -1,7 +1,7 @@
 /*
  * conversion functions between pg_wchar and multi-byte streams.
  * Tatsuo Ishii
- * $Id: wchar.c,v 1.5 1999/02/02 18:51:23 momjian Exp $
+ * $Id: wchar.c,v 1.6 1999/03/24 07:02:17 ishii Exp $
  */
 
 #include "mb/pg_wchar.h"
@@ -416,40 +416,40 @@ pg_big5_mblen(const unsigned char *s)
 }
 
 pg_wchar_tbl pg_wchar_table[] = {
-       {pg_ascii2wchar_with_len, pg_ascii_mblen},
-       {pg_eucjp2wchar_with_len, pg_eucjp_mblen},
-       {pg_euccn2wchar_with_len, pg_euccn_mblen},
-       {pg_euckr2wchar_with_len, pg_euckr_mblen},
-       {pg_euctw2wchar_with_len, pg_euctw_mblen},
-       {pg_utf2wchar_with_len, pg_utf_mblen},
-       {pg_mule2wchar_with_len, pg_mule_mblen},
-       {pg_latin12wchar_with_len, pg_latin1_mblen},
-       {pg_latin12wchar_with_len, pg_latin1_mblen},
-       {pg_latin12wchar_with_len, pg_latin1_mblen},
-       {pg_latin12wchar_with_len, pg_latin1_mblen},
-       {pg_latin12wchar_with_len, pg_latin1_mblen},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, 0},
-       {0, pg_sjis_mblen},
-       {0, pg_big5_mblen}
+       {pg_ascii2wchar_with_len, pg_ascii_mblen},      /* 0 */
+       {pg_eucjp2wchar_with_len, pg_eucjp_mblen},      /* 1 */
+       {pg_euccn2wchar_with_len, pg_euccn_mblen},      /* 2 */
+       {pg_euckr2wchar_with_len, pg_euckr_mblen},      /* 3 */
+       {pg_euctw2wchar_with_len, pg_euctw_mblen},      /* 4 */
+       {pg_utf2wchar_with_len, pg_utf_mblen},          /* 5 */
+       {pg_mule2wchar_with_len, pg_mule_mblen},        /* 6 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 7 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 8 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 9 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 10 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 11 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 12 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 13 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 14 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 15 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 16 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 17 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 18 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 19 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 20 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 21 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 22 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 23 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 24 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 25 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 26 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 27 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 28 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 29 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 30 */
+       {pg_latin12wchar_with_len, pg_latin1_mblen},    /* 31 */
+       {0, pg_sjis_mblen},                             /* 32 */
+       {0, pg_big5_mblen}                              /* 33 */
 };
 
 /* returns the byte length of a word for mule internal code */