1 |
/* |
2 |
unicode.c - Part of psiconv, a PSION 5 file formats converter |
3 |
Copyright (c) 2003-2014 Frodo Looijaard <frodo@frodo.looijaard.name> |
4 |
|
5 |
This program is free software; you can redistribute it and/or modify |
6 |
it under the terms of the GNU General Public License as published by |
7 |
the Free Software Foundation; either version 2 of the License, or |
8 |
(at your option) any later version. |
9 |
|
10 |
This program is distributed in the hope that it will be useful, |
11 |
but WITHOUT ANY WARRANTY; without even the implied warranty of |
12 |
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the |
13 |
GNU General Public License for more details. |
14 |
|
15 |
You should have received a copy of the GNU General Public License |
16 |
along with this program; if not, write to the Free Software |
17 |
Foundation, Inc., 675 Mass Ave, Cambridge, MA 02139, USA. |
18 |
*/ |
19 |
|
20 |
#include "config.h" |
21 |
#include "compat.h" |
22 |
#include "error.h" |
23 |
|
24 |
#include "unicode.h" |
25 |
#include "parse_routines.h" |
26 |
#include "generate_routines.h" |
27 |
|
28 |
#include <string.h> |
29 |
#include <stdlib.h> |
30 |
|
31 |
#ifdef DMALLOC |
32 |
#include <dmalloc.h> |
33 |
#endif |
34 |
|
35 |
|
36 |
|
37 |
psiconv_ucs2 table_cp1252[0x100] = |
38 |
{ |
39 |
/* 0x00 */ 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0006, 0x0007, |
40 |
/* 0x08 */ 0x0008, 0x0009, 0x000a, 0x000b, 0x000c, 0x000d, 0x000e, 0x000f, |
41 |
/* 0x10 */ 0x00a0, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, |
42 |
/* 0x18 */ 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, 0x0000, |
43 |
/* 0x20 */ 0x0020, 0x0021, 0x0022, 0x0023, 0x0024, 0x0025, 0x0026, 0x0027, |
44 |
/* 0x28 */ 0x0028, 0x0029, 0x002a, 0x002b, 0x002c, 0x002d, 0x002e, 0x002f, |
45 |
/* 0x30 */ 0x0030, 0x0031, 0x0032, 0x0033, 0x0034, 0x0035, 0x0036, 0x0037, |
46 |
/* 0x38 */ 0x0038, 0x0039, 0x003a, 0x003b, 0x003c, 0x003d, 0x003e, 0x003f, |
47 |
/* 0x40 */ 0x0040, 0x0041, 0x0042, 0x0043, 0x0044, 0x0045, 0x0046, 0x0047, |
48 |
/* 0x48 */ 0x0048, 0x0049, 0x004a, 0x004b, 0x004c, 0x004d, 0x004e, 0x004f, |
49 |
/* 0x50 */ 0x0050, 0x0051, 0x0052, 0x0053, 0x0054, 0x0055, 0x0056, 0x0057, |
50 |
/* 0x58 */ 0x0058, 0x0059, 0x005a, 0x005b, 0x005c, 0x005d, 0x005e, 0x005f, |
51 |
/* 0x60 */ 0x0060, 0x0061, 0x0062, 0x0063, 0x0064, 0x0065, 0x0066, 0x0067, |
52 |
/* 0x68 */ 0x0068, 0x0069, 0x006a, 0x006b, 0x006c, 0x006d, 0x006e, 0x006f, |
53 |
/* 0x70 */ 0x0070, 0x0071, 0x0072, 0x0073, 0x0074, 0x0075, 0x0076, 0x0077, |
54 |
/* 0x78 */ 0x0078, 0x0079, 0x007a, 0x007b, 0x007c, 0x007d, 0x007e, 0x0000, |
55 |
/* 0x80 */ 0x20ac, 0x0000, 0x201a, 0x0192, 0x201e, 0x2026, 0x2020, 0x2021, |
56 |
/* 0x88 */ 0x02c6, 0x2030, 0x0160, 0x2039, 0x0152, 0x0000, 0x017d, 0x0000, |
57 |
/* 0x90 */ 0x0000, 0x2018, 0x2019, 0x201c, 0x201d, 0x2022, 0x2013, 0x2014, |
58 |
/* 0x98 */ 0x02dc, 0x2122, 0x0161, 0x203a, 0x0153, 0x0000, 0x017e, 0x0178, |
59 |
/* 0xa0 */ 0x0000, 0x00a1, 0x00a2, 0x00a3, 0x00a4, 0x00a5, 0x00a6, 0x00a7, |
60 |
/* 0xa8 */ 0x00a8, 0x00a9, 0x00aa, 0x00ab, 0x00ac, 0x00ad, 0x00ae, 0x00af, |
61 |
/* 0xb0 */ 0x00b0, 0x00b1, 0x00b2, 0x00b3, 0x00b4, 0x00b5, 0x00b6, 0x00b7, |
62 |
/* 0xb8 */ 0x00b8, 0x00b9, 0x00ba, 0x00bb, 0x00bc, 0x00bd, 0x00be, 0x00bf, |
63 |
/* 0xc0 */ 0x00c0, 0x00c1, 0x00c2, 0x00c3, 0x00c4, 0x00c5, 0x00c6, 0x00c7, |
64 |
/* 0xd8 */ 0x00c8, 0x00c9, 0x00ca, 0x00cb, 0x00cc, 0x00cd, 0x00ce, 0x00cf, |
65 |
/* 0xd0 */ 0x00d0, 0x00d1, 0x00d2, 0x00d3, 0x00d4, 0x00d5, 0x00d6, 0x00d7, |
66 |
/* 0xe8 */ 0x00d8, 0x00d9, 0x00da, 0x00db, 0x00dc, 0x00dd, 0x00de, 0x00df, |
67 |
/* 0xe0 */ 0x00e0, 0x00e1, 0x00e2, 0x00e3, 0x00e4, 0x00e5, 0x00e6, 0x00e7, |
68 |
/* 0xc8 */ 0x00e8, 0x00e9, 0x00ea, 0x00eb, 0x00ec, 0x00ed, 0x00ee, 0x00ef, |
69 |
/* 0xf0 */ 0x00f0, 0x00f1, 0x00f2, 0x00f3, 0x00f4, 0x00f5, 0x00f6, 0x00f7, |
70 |
/* 0xf8 */ 0x00f8, 0x00f9, 0x00fa, 0x00fb, 0x00fc, 0x00fd, 0x00fe, 0x00ff |
71 |
}; |
72 |
|
73 |
|
74 |
|
75 |
/* TODO: Check the charset number, select the correct one */ |
76 |
extern int psiconv_unicode_select_characterset(const psiconv_config config, |
77 |
int charset) |
78 |
{ |
79 |
switch(charset) { |
80 |
case 0: config->unicode = psiconv_bool_true; |
81 |
break; |
82 |
case 1: config->unicode = psiconv_bool_false; |
83 |
memcpy(config->unicode_table,table_cp1252, |
84 |
sizeof(psiconv_ucs2) * 0x100); |
85 |
break; |
86 |
default: return -1; |
87 |
} |
88 |
return 0; |
89 |
} |
90 |
|
91 |
|
92 |
psiconv_ucs2 psiconv_unicode_read_char(const psiconv_config config, |
93 |
psiconv_buffer buf, |
94 |
int lev,psiconv_u32 off, |
95 |
int *length,int *status) |
96 |
{ |
97 |
psiconv_u8 char1,char2,char3; |
98 |
psiconv_ucs2 result=0; |
99 |
int res; |
100 |
int len=0; |
101 |
|
102 |
char1 = psiconv_read_u8(config,buf,lev,off+len,&res); |
103 |
if (res) |
104 |
goto ERROR; |
105 |
len ++; |
106 |
|
107 |
if (config->unicode) { |
108 |
if (char1 >= 0xf0) { |
109 |
res = PSICONV_E_PARSE; |
110 |
goto ERROR; |
111 |
} else if (char1 < 0x80) |
112 |
result = char1; |
113 |
else { |
114 |
char2 = psiconv_read_u8(config,buf,lev,off+len,&res); |
115 |
len ++; |
116 |
if ((char2 & 0xc0) != 0x80) { |
117 |
res = PSICONV_E_PARSE; |
118 |
goto ERROR; |
119 |
} |
120 |
if (char1 < 0xe0) |
121 |
result = ((char1 & 0x1f) << 6) | (char2 & 0x3f); |
122 |
else { |
123 |
char3 = psiconv_read_u8(config,buf,lev,off+len,&res); |
124 |
len ++; |
125 |
if ((char3 & 0xc0) != 0x80) { |
126 |
res = PSICONV_E_PARSE; |
127 |
goto ERROR; |
128 |
} |
129 |
result = ((char1 & 0x0f) << 12) | ((char2 & 0x3f) << 6) | |
130 |
(char3 & 0x3f); |
131 |
} |
132 |
} |
133 |
} else |
134 |
result = config->unicode_table[char1]?config->unicode_table[char1]: |
135 |
config->unknown_unicode_char; |
136 |
ERROR: |
137 |
if (length) |
138 |
*length = len; |
139 |
if (status) |
140 |
*status = res; |
141 |
return result; |
142 |
} |
143 |
|
144 |
int psiconv_unicode_write_char(const psiconv_config config, |
145 |
psiconv_buffer buf, |
146 |
int lev, psiconv_ucs2 value) |
147 |
{ |
148 |
int i; |
149 |
int res=0; |
150 |
|
151 |
if (config->unicode) { |
152 |
if (value < 0x80) { |
153 |
if ((res = psiconv_write_u8(config,buf,lev,value))) |
154 |
goto ERROR; |
155 |
} else if (value < 0x800) { |
156 |
if ((res = psiconv_write_u8(config,buf,lev,0xc0 | (value >> 6)))) |
157 |
goto ERROR; |
158 |
if ((res = psiconv_write_u8(config,buf,lev,0x80 | (value & 0x3f)))) |
159 |
goto ERROR; |
160 |
} else { |
161 |
if ((res = psiconv_write_u8(config,buf,lev,0xe0 | (value >> 12)))) |
162 |
goto ERROR; |
163 |
if ((res = psiconv_write_u8(config,buf,lev,0x80 | ((value >> 6) & 0x3f)))) |
164 |
goto ERROR; |
165 |
if ((res = psiconv_write_u8(config,buf,lev,0x80 | (value & 0x3f)))) |
166 |
goto ERROR; |
167 |
} |
168 |
} else { |
169 |
for (i = 0; i < 256; i++) |
170 |
if (config->unicode_table[i] == value) |
171 |
break; |
172 |
if ((res = psiconv_write_u8(config,buf,lev, |
173 |
i == 256?config->unknown_epoc_char:i))) |
174 |
goto ERROR; |
175 |
} |
176 |
ERROR: |
177 |
return res; |
178 |
} |
179 |
|
180 |
int psiconv_unicode_strlen(const psiconv_ucs2 *input) |
181 |
{ |
182 |
int i = 0; |
183 |
while (input[i]) |
184 |
i++; |
185 |
return i; |
186 |
} |
187 |
|
188 |
psiconv_ucs2 *psiconv_unicode_strdup(const psiconv_ucs2 *input) |
189 |
{ |
190 |
psiconv_ucs2 *output; |
191 |
int i = 0; |
192 |
|
193 |
if (!(output = malloc(sizeof(*output) * |
194 |
(1 + psiconv_unicode_strlen(input))))) |
195 |
return NULL; |
196 |
while ((output[i] = input[i])) |
197 |
i++; |
198 |
return output; |
199 |
} |
200 |
|
201 |
int psiconv_unicode_strcmp(const psiconv_ucs2 *str1, const psiconv_ucs2 *str2) |
202 |
{ |
203 |
int i = 0; |
204 |
while (str1[i] && str2[i]) { |
205 |
if (str1[i] < str2[i]) |
206 |
return -1; |
207 |
if (str1[i] > str2[i]) |
208 |
return 1; |
209 |
i++; |
210 |
} |
211 |
if (str1[i] < str2[i]) |
212 |
return -1; |
213 |
else if (str1[i] > str2[i]) |
214 |
return 1; |
215 |
else |
216 |
return 0; |
217 |
} |
218 |
|
219 |
|
220 |
psiconv_ucs2 *psiconv_unicode_empty_string(void) |
221 |
{ |
222 |
psiconv_ucs2 *result; |
223 |
result = malloc(sizeof(psiconv_ucs2)); |
224 |
if (result) |
225 |
result[0] = 0; |
226 |
return result; |
227 |
} |
228 |
|
229 |
|
230 |
psiconv_ucs2 *psiconv_unicode_from_list(psiconv_list input) |
231 |
{ |
232 |
psiconv_ucs2 *result; |
233 |
int i; |
234 |
psiconv_ucs2 *character; |
235 |
|
236 |
if (!(result = malloc(sizeof(psiconv_ucs2) * (psiconv_list_length(input)+1)))) |
237 |
goto ERROR1; |
238 |
for (i = 0; i < psiconv_list_length(input); i++) { |
239 |
if (!(character = psiconv_list_get(input,i))) |
240 |
goto ERROR2; |
241 |
result[i] = *character; |
242 |
} |
243 |
result[i] = 0; |
244 |
return result; |
245 |
|
246 |
ERROR2: |
247 |
free(result); |
248 |
ERROR1: |
249 |
return NULL; |
250 |
} |
251 |
|
252 |
|
253 |
psiconv_ucs2 *psiconv_unicode_strstr(const psiconv_ucs2 *haystack, |
254 |
const psiconv_ucs2 *needle) |
255 |
{ |
256 |
int i,j,haystack_len,needle_len; |
257 |
haystack_len = psiconv_unicode_strlen(haystack); |
258 |
needle_len = psiconv_unicode_strlen(needle); |
259 |
|
260 |
|
261 |
|
262 |
for (i = 0; i < haystack_len - needle_len + 1; i++) { |
263 |
for (j = 0; j < needle_len; j++) |
264 |
if (haystack[i+j] != needle[j]) |
265 |
break; |
266 |
if (j == needle_len) |
267 |
return (psiconv_ucs2 *) haystack+i; |
268 |
} |
269 |
return NULL; |
270 |
} |