1 /*
2 * "streamable kanji code filter and converter"
3 * Copyright (c) 1998-2002 HappySize, Inc. All rights reserved.
4 *
5 * LICENSE NOTICES
6 *
7 * This file is part of "streamable kanji code filter and converter",
8 * which is distributed under the terms of GNU Lesser General Public
9 * License (version 2) as published by the Free Software Foundation.
10 *
11 * This software is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 * GNU Lesser General Public License for more details.
15 *
16 * You should have received a copy of the GNU Lesser General Public
17 * License along with "streamable kanji code filter and converter";
18 * if not, write to the Free Software Foundation, Inc., 59 Temple Place,
19 * Suite 330, Boston, MA 02111-1307 USA
20 *
21 * The author of this file:
22 *
23 */
24 /*
25 * The source code included in this files was separated from mbfilter.c
26 * by moriyoshi koizumi <moriyoshi@php.net> on 4 dec 2002.
27 *
28 */
29
30 #ifdef HAVE_CONFIG_H
31 #include "config.h"
32 #endif
33
34 #include "mbfilter.h"
35 #include "mbfilter_utf16.h"
36
37 static const char *mbfl_encoding_utf16_aliases[] = {"utf16", NULL};
38
39 const mbfl_encoding mbfl_encoding_utf16 = {
40 mbfl_no_encoding_utf16,
41 "UTF-16",
42 "UTF-16",
43 (const char *(*)[])&mbfl_encoding_utf16_aliases,
44 NULL,
45 MBFL_ENCTYPE_MWC2BE
46 };
47
48 const mbfl_encoding mbfl_encoding_utf16be = {
49 mbfl_no_encoding_utf16be,
50 "UTF-16BE",
51 "UTF-16BE",
52 NULL,
53 NULL,
54 MBFL_ENCTYPE_MWC2BE
55 };
56
57 const mbfl_encoding mbfl_encoding_utf16le = {
58 mbfl_no_encoding_utf16le,
59 "UTF-16LE",
60 "UTF-16LE",
61 NULL,
62 NULL,
63 MBFL_ENCTYPE_MWC2LE
64 };
65
66 const struct mbfl_convert_vtbl vtbl_utf16_wchar = {
67 mbfl_no_encoding_utf16,
68 mbfl_no_encoding_wchar,
69 mbfl_filt_conv_common_ctor,
70 mbfl_filt_conv_common_dtor,
71 mbfl_filt_conv_utf16_wchar,
72 mbfl_filt_conv_common_flush
73 };
74
75 const struct mbfl_convert_vtbl vtbl_wchar_utf16 = {
76 mbfl_no_encoding_wchar,
77 mbfl_no_encoding_utf16,
78 mbfl_filt_conv_common_ctor,
79 mbfl_filt_conv_common_dtor,
80 mbfl_filt_conv_wchar_utf16be,
81 mbfl_filt_conv_common_flush
82 };
83
84 const struct mbfl_convert_vtbl vtbl_utf16be_wchar = {
85 mbfl_no_encoding_utf16be,
86 mbfl_no_encoding_wchar,
87 mbfl_filt_conv_common_ctor,
88 mbfl_filt_conv_common_dtor,
89 mbfl_filt_conv_utf16be_wchar,
90 mbfl_filt_conv_common_flush
91 };
92
93 const struct mbfl_convert_vtbl vtbl_wchar_utf16be = {
94 mbfl_no_encoding_wchar,
95 mbfl_no_encoding_utf16be,
96 mbfl_filt_conv_common_ctor,
97 mbfl_filt_conv_common_dtor,
98 mbfl_filt_conv_wchar_utf16be,
99 mbfl_filt_conv_common_flush
100 };
101
102 const struct mbfl_convert_vtbl vtbl_utf16le_wchar = {
103 mbfl_no_encoding_utf16le,
104 mbfl_no_encoding_wchar,
105 mbfl_filt_conv_common_ctor,
106 mbfl_filt_conv_common_dtor,
107 mbfl_filt_conv_utf16le_wchar,
108 mbfl_filt_conv_common_flush
109 };
110
111 const struct mbfl_convert_vtbl vtbl_wchar_utf16le = {
112 mbfl_no_encoding_wchar,
113 mbfl_no_encoding_utf16le,
114 mbfl_filt_conv_common_ctor,
115 mbfl_filt_conv_common_dtor,
116 mbfl_filt_conv_wchar_utf16le,
117 mbfl_filt_conv_common_flush
118 };
119
120 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
121
122 /*
123 * UTF-16 => wchar
124 */
mbfl_filt_conv_utf16_wchar(int c,mbfl_convert_filter * filter)125 int mbfl_filt_conv_utf16_wchar(int c, mbfl_convert_filter *filter)
126 {
127 int n, endian;
128
129 endian = filter->status & 0xff00;
130 switch (filter->status & 0x0f) {
131 case 0:
132 if (endian) {
133 n = c & 0xff;
134 } else {
135 n = (c & 0xff) << 8;
136 }
137 filter->cache |= n;
138 filter->status++;
139 break;
140 default:
141 if (endian) {
142 n = (c & 0xff) << 8;
143 } else {
144 n = c & 0xff;
145 }
146 n |= filter->cache & 0xffff;
147 filter->status &= ~0x0f;
148 if (n >= 0xd800 && n < 0xdc00) {
149 filter->cache = ((n & 0x3ff) << 16) + 0x400000;
150 } else if (n >= 0xdc00 && n < 0xe000) {
151 n &= 0x3ff;
152 n |= (filter->cache & 0xfff0000) >> 6;
153 filter->cache = 0;
154 if (n >= MBFL_WCSPLANE_SUPMIN && n < MBFL_WCSPLANE_SUPMAX) {
155 CK((*filter->output_function)(n, filter->data));
156 } else { /* illegal character */
157 n &= MBFL_WCSGROUP_MASK;
158 n |= MBFL_WCSGROUP_THROUGH;
159 CK((*filter->output_function)(n, filter->data));
160 }
161 } else {
162 int is_first = filter->status & 0x10;
163 filter->cache = 0;
164 filter->status |= 0x10;
165 if (!is_first) {
166 if (n == 0xfffe) {
167 if (endian) {
168 filter->status &= ~0x100; /* big-endian */
169 } else {
170 filter->status |= 0x100; /* little-endian */
171 }
172 break;
173 } else if (n == 0xfeff) {
174 break;
175 }
176 }
177 CK((*filter->output_function)(n, filter->data));
178 }
179 break;
180 }
181
182 return c;
183 }
184
185 /*
186 * UTF-16BE => wchar
187 */
mbfl_filt_conv_utf16be_wchar(int c,mbfl_convert_filter * filter)188 int mbfl_filt_conv_utf16be_wchar(int c, mbfl_convert_filter *filter)
189 {
190 int n;
191
192 switch (filter->status) {
193 case 0:
194 filter->status = 1;
195 n = (c & 0xff) << 8;
196 filter->cache |= n;
197 break;
198 default:
199 filter->status = 0;
200 n = (filter->cache & 0xff00) | (c & 0xff);
201 if (n >= 0xd800 && n < 0xdc00) {
202 filter->cache = ((n & 0x3ff) << 16) + 0x400000;
203 } else if (n >= 0xdc00 && n < 0xe000) {
204 n &= 0x3ff;
205 n |= (filter->cache & 0xfff0000) >> 6;
206 filter->cache = 0;
207 if (n >= MBFL_WCSPLANE_SUPMIN && n < MBFL_WCSPLANE_SUPMAX) {
208 CK((*filter->output_function)(n, filter->data));
209 } else { /* illegal character */
210 n &= MBFL_WCSGROUP_MASK;
211 n |= MBFL_WCSGROUP_THROUGH;
212 CK((*filter->output_function)(n, filter->data));
213 }
214 } else {
215 filter->cache = 0;
216 CK((*filter->output_function)(n, filter->data));
217 }
218 break;
219 }
220
221 return c;
222 }
223
224 /*
225 * wchar => UTF-16BE
226 */
mbfl_filt_conv_wchar_utf16be(int c,mbfl_convert_filter * filter)227 int mbfl_filt_conv_wchar_utf16be(int c, mbfl_convert_filter *filter)
228 {
229 int n;
230
231 if (c >= 0 && c < MBFL_WCSPLANE_UCS2MAX) {
232 CK((*filter->output_function)((c >> 8) & 0xff, filter->data));
233 CK((*filter->output_function)(c & 0xff, filter->data));
234 } else if (c >= MBFL_WCSPLANE_SUPMIN && c < MBFL_WCSPLANE_SUPMAX) {
235 n = ((c >> 10) - 0x40) | 0xd800;
236 CK((*filter->output_function)((n >> 8) & 0xff, filter->data));
237 CK((*filter->output_function)(n & 0xff, filter->data));
238 n = (c & 0x3ff) | 0xdc00;
239 CK((*filter->output_function)((n >> 8) & 0xff, filter->data));
240 CK((*filter->output_function)(n & 0xff, filter->data));
241 } else {
242 if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
243 CK(mbfl_filt_conv_illegal_output(c, filter));
244 }
245 }
246
247 return c;
248 }
249
250 /*
251 * UTF-16LE => wchar
252 */
mbfl_filt_conv_utf16le_wchar(int c,mbfl_convert_filter * filter)253 int mbfl_filt_conv_utf16le_wchar(int c, mbfl_convert_filter *filter)
254 {
255 int n;
256
257 switch (filter->status) {
258 case 0:
259 filter->status = 1;
260 n = c & 0xff;
261 filter->cache |= n;
262 break;
263 default:
264 filter->status = 0;
265 n = (filter->cache & 0xff) | ((c & 0xff) << 8);
266 if (n >= 0xd800 && n < 0xdc00) {
267 filter->cache = ((n & 0x3ff) << 16) + 0x400000;
268 } else if (n >= 0xdc00 && n < 0xe000) {
269 n &= 0x3ff;
270 n |= (filter->cache & 0xfff0000) >> 6;
271 filter->cache = 0;
272 if (n >= MBFL_WCSPLANE_SUPMIN && n < MBFL_WCSPLANE_SUPMAX) {
273 CK((*filter->output_function)(n, filter->data));
274 } else { /* illegal character */
275 n &= MBFL_WCSGROUP_MASK;
276 n |= MBFL_WCSGROUP_THROUGH;
277 CK((*filter->output_function)(n, filter->data));
278 }
279 } else {
280 filter->cache = 0;
281 CK((*filter->output_function)(n, filter->data));
282 }
283 break;
284 }
285
286 return c;
287 }
288
289 /*
290 * wchar => UTF-16LE
291 */
mbfl_filt_conv_wchar_utf16le(int c,mbfl_convert_filter * filter)292 int mbfl_filt_conv_wchar_utf16le(int c, mbfl_convert_filter *filter)
293 {
294 int n;
295
296 if (c >= 0 && c < MBFL_WCSPLANE_UCS2MAX) {
297 CK((*filter->output_function)(c & 0xff, filter->data));
298 CK((*filter->output_function)((c >> 8) & 0xff, filter->data));
299 } else if (c >= MBFL_WCSPLANE_SUPMIN && c < MBFL_WCSPLANE_SUPMAX) {
300 n = ((c >> 10) - 0x40) | 0xd800;
301 CK((*filter->output_function)(n & 0xff, filter->data));
302 CK((*filter->output_function)((n >> 8) & 0xff, filter->data));
303 n = (c & 0x3ff) | 0xdc00;
304 CK((*filter->output_function)(n & 0xff, filter->data));
305 CK((*filter->output_function)((n >> 8) & 0xff, filter->data));
306 } else {
307 if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
308 CK(mbfl_filt_conv_illegal_output(c, filter));
309 }
310 }
311
312 return c;
313 }
314
