Line data Source code
1 : /*-------------------------------------------------------------------------
2 : *
3 : * GB18030 <--> UTF8
4 : *
5 : * Portions Copyright (c) 1996-2025, PostgreSQL Global Development Group
6 : * Portions Copyright (c) 1994, Regents of the University of California
7 : *
8 : * IDENTIFICATION
9 : * src/backend/utils/mb/conversion_procs/utf8_and_gb18030/utf8_and_gb18030.c
10 : *
11 : *-------------------------------------------------------------------------
12 : */
13 :
14 : #include "postgres.h"
15 : #include "fmgr.h"
16 : #include "mb/pg_wchar.h"
17 : #include "../../Unicode/gb18030_to_utf8.map"
18 : #include "../../Unicode/utf8_to_gb18030.map"
19 :
20 12 : PG_MODULE_MAGIC_EXT(
21 : .name = "utf8_and_gb18030",
22 : .version = PG_VERSION
23 : );
24 :
25 12 : PG_FUNCTION_INFO_V1(gb18030_to_utf8);
26 12 : PG_FUNCTION_INFO_V1(utf8_to_gb18030);
27 :
28 : /*
29 : * Convert 4-byte GB18030 characters to and from a linear code space
30 : *
31 : * The first and third bytes can range from 0x81 to 0xfe (126 values),
32 : * while the second and fourth bytes can range from 0x30 to 0x39 (10 values).
33 : */
34 : static inline uint32
35 180 : gb_linear(uint32 gb)
36 : {
37 180 : uint32 b0 = (gb & 0xff000000) >> 24;
38 180 : uint32 b1 = (gb & 0x00ff0000) >> 16;
39 180 : uint32 b2 = (gb & 0x0000ff00) >> 8;
40 180 : uint32 b3 = (gb & 0x000000ff);
41 :
42 180 : return b0 * 12600 + b1 * 1260 + b2 * 10 + b3 -
43 : (0x81 * 12600 + 0x30 * 1260 + 0x81 * 10 + 0x30);
44 : }
45 :
46 : static inline uint32
47 72 : gb_unlinear(uint32 lin)
48 : {
49 72 : uint32 r0 = 0x81 + lin / 12600;
50 72 : uint32 r1 = 0x30 + (lin / 1260) % 10;
51 72 : uint32 r2 = 0x81 + (lin / 10) % 126;
52 72 : uint32 r3 = 0x30 + lin % 10;
53 :
54 72 : return (r0 << 24) | (r1 << 16) | (r2 << 8) | r3;
55 : }
56 :
57 : /*
58 : * Convert word-formatted UTF8 to and from Unicode code points
59 : *
60 : * Probably this should be somewhere else ...
61 : */
62 : static inline uint32
63 54 : unicode_to_utf8word(uint32 c)
64 : {
65 : uint32 word;
66 :
67 54 : if (c <= 0x7F)
68 : {
69 0 : word = c;
70 : }
71 54 : else if (c <= 0x7FF)
72 : {
73 0 : word = (0xC0 | ((c >> 6) & 0x1F)) << 8;
74 0 : word |= 0x80 | (c & 0x3F);
75 : }
76 54 : else if (c <= 0xFFFF)
77 : {
78 54 : word = (0xE0 | ((c >> 12) & 0x0F)) << 16;
79 54 : word |= (0x80 | ((c >> 6) & 0x3F)) << 8;
80 54 : word |= 0x80 | (c & 0x3F);
81 : }
82 : else
83 : {
84 0 : word = (0xF0 | ((c >> 18) & 0x07)) << 24;
85 0 : word |= (0x80 | ((c >> 12) & 0x3F)) << 16;
86 0 : word |= (0x80 | ((c >> 6) & 0x3F)) << 8;
87 0 : word |= 0x80 | (c & 0x3F);
88 : }
89 :
90 54 : return word;
91 : }
92 :
93 : static inline uint32
94 72 : utf8word_to_unicode(uint32 c)
95 : {
96 : uint32 ucs;
97 :
98 72 : if (c <= 0x7F)
99 : {
100 0 : ucs = c;
101 : }
102 72 : else if (c <= 0xFFFF)
103 : {
104 0 : ucs = ((c >> 8) & 0x1F) << 6;
105 0 : ucs |= c & 0x3F;
106 : }
107 72 : else if (c <= 0xFFFFFF)
108 : {
109 72 : ucs = ((c >> 16) & 0x0F) << 12;
110 72 : ucs |= ((c >> 8) & 0x3F) << 6;
111 72 : ucs |= c & 0x3F;
112 : }
113 : else
114 : {
115 0 : ucs = ((c >> 24) & 0x07) << 18;
116 0 : ucs |= ((c >> 16) & 0x3F) << 12;
117 0 : ucs |= ((c >> 8) & 0x3F) << 6;
118 0 : ucs |= c & 0x3F;
119 : }
120 :
121 72 : return ucs;
122 : }
123 :
124 : /*
125 : * Perform mapping of GB18030 ranges to UTF8
126 : *
127 : * General description, and the range we need to convert for U+10000 and up:
128 : * https://htmlpreview.github.io/?https://github.com/unicode-org/icu-data/blob/main/charset/source/gb18030/gb18030.html
129 : *
130 : * Ranges up to U+FFFF:
131 : * https://github.com/unicode-org/icu-data/blob/main/charset/source/gb18030/ranges.txt
132 : *
133 : * All are ranges of 4-byte GB18030 codes.
134 : */
135 : static uint32
136 90 : conv_18030_to_utf8(uint32 code)
137 : {
138 : #define conv18030(minunicode, mincode, maxcode) \
139 : if (code >= mincode && code <= maxcode) \
140 : return unicode_to_utf8word(gb_linear(code) - gb_linear(mincode) + minunicode)
141 :
142 90 : conv18030(0x0452, 0x8130D330, 0x8136A531);
143 90 : conv18030(0x2643, 0x8137A839, 0x8138FD38);
144 90 : conv18030(0x361B, 0x8230A633, 0x8230F237);
145 90 : conv18030(0x3CE1, 0x8231D438, 0x8232AF32);
146 90 : conv18030(0x4160, 0x8232C937, 0x8232F837);
147 90 : conv18030(0x44D7, 0x8233A339, 0x8233C931);
148 90 : conv18030(0x478E, 0x8233E838, 0x82349638);
149 90 : conv18030(0x49B8, 0x8234A131, 0x8234E733);
150 90 : conv18030(0x9FA6, 0x82358F33, 0x8336C738);
151 90 : conv18030(0xE865, 0x8336D030, 0x84308534);
152 90 : conv18030(0xFA2A, 0x84309C38, 0x84318537);
153 36 : conv18030(0xFFE6, 0x8431A234, 0x8431A439);
154 36 : conv18030(0x10000, 0x90308130, 0xE3329A35);
155 : /* No mapping exists */
156 36 : return 0;
157 : }
158 :
159 : /*
160 : * Perform mapping of UTF8 ranges to GB18030
161 : */
162 : static uint32
163 72 : conv_utf8_to_18030(uint32 code)
164 : {
165 72 : uint32 ucs = utf8word_to_unicode(code);
166 :
167 : #define convutf8(minunicode, maxunicode, mincode) \
168 : if (ucs >= minunicode && ucs <= maxunicode) \
169 : return gb_unlinear(ucs - minunicode + gb_linear(mincode))
170 :
171 72 : convutf8(0x0452, 0x200F, 0x8130D330);
172 72 : convutf8(0x2643, 0x2E80, 0x8137A839);
173 72 : convutf8(0x361B, 0x3917, 0x8230A633);
174 72 : convutf8(0x3CE1, 0x4055, 0x8231D438);
175 72 : convutf8(0x4160, 0x4336, 0x8232C937);
176 72 : convutf8(0x44D7, 0x464B, 0x8233A339);
177 72 : convutf8(0x478E, 0x4946, 0x8233E838);
178 72 : convutf8(0x49B8, 0x4C76, 0x8234A131);
179 72 : convutf8(0x9FA6, 0xD7FF, 0x82358F33);
180 18 : convutf8(0xE865, 0xF92B, 0x8336D030);
181 18 : convutf8(0xFA2A, 0xFE2F, 0x84309C38);
182 0 : convutf8(0xFFE6, 0xFFFF, 0x8431A234);
183 0 : convutf8(0x10000, 0x10FFFF, 0x90308130);
184 : /* No mapping exists */
185 0 : return 0;
186 : }
187 :
188 : /* ----------
189 : * conv_proc(
190 : * INTEGER, -- source encoding id
191 : * INTEGER, -- destination encoding id
192 : * CSTRING, -- source string (null terminated C string)
193 : * CSTRING, -- destination string (null terminated C string)
194 : * INTEGER, -- source string length
195 : * BOOL -- if true, don't throw an error if conversion fails
196 : * ) returns INTEGER;
197 : *
198 : * Returns the number of bytes successfully converted.
199 : * ----------
200 : */
201 : Datum
202 276 : gb18030_to_utf8(PG_FUNCTION_ARGS)
203 : {
204 276 : unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
205 276 : unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
206 276 : int len = PG_GETARG_INT32(4);
207 276 : bool noError = PG_GETARG_BOOL(5);
208 : int converted;
209 :
210 276 : CHECK_ENCODING_CONVERSION_ARGS(PG_GB18030, PG_UTF8);
211 :
212 276 : converted = LocalToUtf(src, len, dest,
213 : &gb18030_to_unicode_tree,
214 : NULL, 0,
215 : conv_18030_to_utf8,
216 : PG_GB18030,
217 : noError);
218 :
219 168 : PG_RETURN_INT32(converted);
220 : }
221 :
222 : Datum
223 330 : utf8_to_gb18030(PG_FUNCTION_ARGS)
224 : {
225 330 : unsigned char *src = (unsigned char *) PG_GETARG_CSTRING(2);
226 330 : unsigned char *dest = (unsigned char *) PG_GETARG_CSTRING(3);
227 330 : int len = PG_GETARG_INT32(4);
228 330 : bool noError = PG_GETARG_BOOL(5);
229 : int converted;
230 :
231 330 : CHECK_ENCODING_CONVERSION_ARGS(PG_UTF8, PG_GB18030);
232 :
233 330 : converted = UtfToLocal(src, len, dest,
234 : &gb18030_from_unicode_tree,
235 : NULL, 0,
236 : conv_utf8_to_18030,
237 : PG_GB18030,
238 : noError);
239 :
240 240 : PG_RETURN_INT32(converted);
241 : }
|
