Header And Logo
|
00001 /*------------------------------------------------------------------------- 00002 * 00003 * ISO 8859 2-16 <--> UTF8 00004 * 00005 * Portions Copyright (c) 1996-2013, PostgreSQL Global Development Group 00006 * Portions Copyright (c) 1994, Regents of the University of California 00007 * 00008 * IDENTIFICATION 00009 * src/backend/utils/mb/conversion_procs/utf8_and_iso8859/utf8_and_iso8859.c 00010 * 00011 *------------------------------------------------------------------------- 00012 */ 00013 00014 #include "postgres.h" 00015 #include "fmgr.h" 00016 #include "mb/pg_wchar.h" 00017 #include "../../Unicode/iso8859_10_to_utf8.map" 00018 #include "../../Unicode/iso8859_13_to_utf8.map" 00019 #include "../../Unicode/iso8859_14_to_utf8.map" 00020 #include "../../Unicode/iso8859_15_to_utf8.map" 00021 #include "../../Unicode/iso8859_2_to_utf8.map" 00022 #include "../../Unicode/iso8859_3_to_utf8.map" 00023 #include "../../Unicode/iso8859_4_to_utf8.map" 00024 #include "../../Unicode/iso8859_5_to_utf8.map" 00025 #include "../../Unicode/iso8859_6_to_utf8.map" 00026 #include "../../Unicode/iso8859_7_to_utf8.map" 00027 #include "../../Unicode/iso8859_8_to_utf8.map" 00028 #include "../../Unicode/iso8859_9_to_utf8.map" 00029 #include "../../Unicode/utf8_to_iso8859_10.map" 00030 #include "../../Unicode/utf8_to_iso8859_13.map" 00031 #include "../../Unicode/utf8_to_iso8859_14.map" 00032 #include "../../Unicode/utf8_to_iso8859_15.map" 00033 #include "../../Unicode/utf8_to_iso8859_16.map" 00034 #include "../../Unicode/utf8_to_iso8859_2.map" 00035 #include "../../Unicode/utf8_to_iso8859_3.map" 00036 #include "../../Unicode/utf8_to_iso8859_4.map" 00037 #include "../../Unicode/utf8_to_iso8859_5.map" 00038 #include "../../Unicode/utf8_to_iso8859_6.map" 00039 #include "../../Unicode/utf8_to_iso8859_7.map" 00040 #include "../../Unicode/utf8_to_iso8859_8.map" 00041 #include "../../Unicode/utf8_to_iso8859_9.map" 00042 #include "../../Unicode/iso8859_16_to_utf8.map" 00043 00044 PG_MODULE_MAGIC; 00045 00046 PG_FUNCTION_INFO_V1(iso8859_to_utf8); 00047 PG_FUNCTION_INFO_V1(utf8_to_iso8859); 00048 00049 extern Datum iso8859_to_utf8(PG_FUNCTION_ARGS); 00050 extern Datum utf8_to_iso8859(PG_FUNCTION_ARGS); 00051 00052 /* ---------- 00053 * conv_proc( 00054 * INTEGER, -- source encoding id 00055 * INTEGER, -- destination encoding id 00056 * CSTRING, -- source string (null terminated C string) 00057 * CSTRING, -- destination string (null terminated C string) 00058 * INTEGER -- source string length 00059 * ) returns VOID; 00060 * ---------- 00061 */ 00062 00063 typedef struct 00064 { 00065 pg_enc encoding; 00066 pg_local_to_utf *map1; /* to UTF8 map name */ 00067 pg_utf_to_local *map2; /* from UTF8 map name */ 00068 int size1; /* size of map1 */ 00069 int size2; /* size of map2 */ 00070 } pg_conv_map; 00071 00072 static pg_conv_map maps[] = { 00073 {PG_LATIN2, LUmapISO8859_2, ULmapISO8859_2, 00074 sizeof(LUmapISO8859_2) / sizeof(pg_local_to_utf), 00075 sizeof(ULmapISO8859_2) / sizeof(pg_utf_to_local)}, /* ISO-8859-2 Latin 2 */ 00076 {PG_LATIN3, LUmapISO8859_3, ULmapISO8859_3, 00077 sizeof(LUmapISO8859_3) / sizeof(pg_local_to_utf), 00078 sizeof(ULmapISO8859_3) / sizeof(pg_utf_to_local)}, /* ISO-8859-3 Latin 3 */ 00079 {PG_LATIN4, LUmapISO8859_4, ULmapISO8859_4, 00080 sizeof(LUmapISO8859_4) / sizeof(pg_local_to_utf), 00081 sizeof(ULmapISO8859_4) / sizeof(pg_utf_to_local)}, /* ISO-8859-4 Latin 4 */ 00082 {PG_LATIN5, LUmapISO8859_9, ULmapISO8859_9, 00083 sizeof(LUmapISO8859_9) / sizeof(pg_local_to_utf), 00084 sizeof(ULmapISO8859_9) / sizeof(pg_utf_to_local)}, /* ISO-8859-9 Latin 5 */ 00085 {PG_LATIN6, LUmapISO8859_10, ULmapISO8859_10, 00086 sizeof(LUmapISO8859_10) / sizeof(pg_local_to_utf), 00087 sizeof(ULmapISO8859_10) / sizeof(pg_utf_to_local)}, /* ISO-8859-10 Latin 6 */ 00088 {PG_LATIN7, LUmapISO8859_13, ULmapISO8859_13, 00089 sizeof(LUmapISO8859_13) / sizeof(pg_local_to_utf), 00090 sizeof(ULmapISO8859_13) / sizeof(pg_utf_to_local)}, /* ISO-8859-13 Latin 7 */ 00091 {PG_LATIN8, LUmapISO8859_14, ULmapISO8859_14, 00092 sizeof(LUmapISO8859_14) / sizeof(pg_local_to_utf), 00093 sizeof(ULmapISO8859_14) / sizeof(pg_utf_to_local)}, /* ISO-8859-14 Latin 8 */ 00094 {PG_LATIN9, LUmapISO8859_15, ULmapISO8859_15, 00095 sizeof(LUmapISO8859_15) / sizeof(pg_local_to_utf), 00096 sizeof(ULmapISO8859_15) / sizeof(pg_utf_to_local)}, /* ISO-8859-15 Latin 9 */ 00097 {PG_LATIN10, LUmapISO8859_16, ULmapISO8859_16, 00098 sizeof(LUmapISO8859_16) / sizeof(pg_local_to_utf), 00099 sizeof(ULmapISO8859_16) / sizeof(pg_utf_to_local)}, /* ISO-8859-16 Latin 10 */ 00100 {PG_ISO_8859_5, LUmapISO8859_5, ULmapISO8859_5, 00101 sizeof(LUmapISO8859_5) / sizeof(pg_local_to_utf), 00102 sizeof(ULmapISO8859_5) / sizeof(pg_utf_to_local)}, /* ISO-8859-5 */ 00103 {PG_ISO_8859_6, LUmapISO8859_6, ULmapISO8859_6, 00104 sizeof(LUmapISO8859_6) / sizeof(pg_local_to_utf), 00105 sizeof(ULmapISO8859_6) / sizeof(pg_utf_to_local)}, /* ISO-8859-6 */ 00106 {PG_ISO_8859_7, LUmapISO8859_7, ULmapISO8859_7, 00107 sizeof(LUmapISO8859_7) / sizeof(pg_local_to_utf), 00108 sizeof(ULmapISO8859_7) / sizeof(pg_utf_to_local)}, /* ISO-8859-7 */ 00109 {PG_ISO_8859_8, LUmapISO8859_8, ULmapISO8859_8, 00110 sizeof(LUmapISO8859_8) / sizeof(pg_local_to_utf), 00111 sizeof(ULmapISO8859_8) / sizeof(pg_utf_to_local)}, /* ISO-8859-8 */ 00112 }; 00113 00114 Datum 00115 iso8859_to_utf8(PG_FUNCTION_ARGS) 00116 { 00117 int encoding = PG_GETARG_INT32(0); 00118 unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2); 00119 unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3); 00120 int len = PG_GETARG_INT32(4); 00121 int i; 00122 00123 CHECK_ENCODING_CONVERSION_ARGS(-1, PG_UTF8); 00124 00125 for (i = 0; i < sizeof(maps) / sizeof(pg_conv_map); i++) 00126 { 00127 if (encoding == maps[i].encoding) 00128 { 00129 LocalToUtf(src, dest, maps[i].map1, NULL, maps[i].size1, 0, encoding, len); 00130 PG_RETURN_VOID(); 00131 } 00132 } 00133 00134 ereport(ERROR, 00135 (errcode(ERRCODE_INTERNAL_ERROR), 00136 errmsg("unexpected encoding ID %d for ISO 8859 character sets", encoding))); 00137 00138 PG_RETURN_VOID(); 00139 } 00140 00141 Datum 00142 utf8_to_iso8859(PG_FUNCTION_ARGS) 00143 { 00144 int encoding = PG_GETARG_INT32(1); 00145 unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2); 00146 unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3); 00147 int len = PG_GETARG_INT32(4); 00148 int i; 00149 00150 CHECK_ENCODING_CONVERSION_ARGS(PG_UTF8, -1); 00151 00152 for (i = 0; i < sizeof(maps) / sizeof(pg_conv_map); i++) 00153 { 00154 if (encoding == maps[i].encoding) 00155 { 00156 UtfToLocal(src, dest, maps[i].map2, NULL, maps[i].size2, 0, encoding, len); 00157 PG_RETURN_VOID(); 00158 } 00159 } 00160 00161 ereport(ERROR, 00162 (errcode(ERRCODE_INTERNAL_ERROR), 00163 errmsg("unexpected encoding ID %d for ISO 8859 character sets", encoding))); 00164 00165 PG_RETURN_VOID(); 00166 }
1.7.1