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_cn.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_hz.h"
36
37 #include "unicode_table_cp936.h"
38
39 static int mbfl_filt_ident_hz(int c, mbfl_identify_filter *filter);
40
41 const mbfl_encoding mbfl_encoding_hz = {
42 mbfl_no_encoding_hz,
43 "HZ",
44 "HZ-GB-2312",
45 NULL,
46 NULL,
47 MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_SHFTCODE | MBFL_ENCTYPE_GL_UNSAFE
48 };
49
50 const struct mbfl_identify_vtbl vtbl_identify_hz = {
51 mbfl_no_encoding_hz,
52 mbfl_filt_ident_common_ctor,
53 mbfl_filt_ident_common_dtor,
54 mbfl_filt_ident_hz
55 };
56
57 const struct mbfl_convert_vtbl vtbl_hz_wchar = {
58 mbfl_no_encoding_hz,
59 mbfl_no_encoding_wchar,
60 mbfl_filt_conv_common_ctor,
61 mbfl_filt_conv_common_dtor,
62 mbfl_filt_conv_hz_wchar,
63 mbfl_filt_conv_common_flush
64 };
65
66 const struct mbfl_convert_vtbl vtbl_wchar_hz = {
67 mbfl_no_encoding_wchar,
68 mbfl_no_encoding_hz,
69 mbfl_filt_conv_common_ctor,
70 mbfl_filt_conv_common_dtor,
71 mbfl_filt_conv_wchar_hz,
72 mbfl_filt_conv_any_hz_flush
73 };
74
75 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
76
77 /*
78 * HZ => wchar
79 */
80 int
mbfl_filt_conv_hz_wchar(int c,mbfl_convert_filter * filter)81 mbfl_filt_conv_hz_wchar(int c, mbfl_convert_filter *filter)
82 {
83 int c1, s, w;
84
85 switch (filter->status & 0xf) {
86 /* case 0x00: ASCII */
87 /* case 0x10: GB2312 */
88 case 0:
89 if (c == 0x7e) {
90 filter->status += 2;
91 } else if (filter->status == 0x10 && c > 0x20 && c < 0x7f) { /* DBCS first char */
92 filter->cache = c;
93 filter->status += 1;
94 } else if (c >= 0 && c < 0x80) { /* latin, CTLs */
95 CK((*filter->output_function)(c, filter->data));
96 } else {
97 w = c & MBFL_WCSGROUP_MASK;
98 w |= MBFL_WCSGROUP_THROUGH;
99 CK((*filter->output_function)(w, filter->data));
100 }
101 break;
102
103 /* case 0x11: GB2312 second char */
104 case 1:
105 filter->status &= ~0xf;
106 c1 = filter->cache;
107 if (c1 > 0x20 && c1 < 0x7f && c > 0x20 && c < 0x7f) {
108 s = (c1 - 1)*192 + c + 0x40; /* GB2312 */
109 if (s >= 0 && s < cp936_ucs_table_size) {
110 w = cp936_ucs_table[s];
111 } else {
112 w = 0;
113 }
114 if (w <= 0) {
115 w = (c1 << 8) | c;
116 w &= MBFL_WCSPLANE_MASK;
117 w |= MBFL_WCSPLANE_GB2312;
118 }
119 CK((*filter->output_function)(w, filter->data));
120 } else if ((c >= 0 && c < 0x21) || c == 0x7f) { /* CTLs */
121 CK((*filter->output_function)(c, filter->data));
122 } else {
123 w = (c1 << 8) | c;
124 w &= MBFL_WCSGROUP_MASK;
125 w |= MBFL_WCSGROUP_THROUGH;
126 CK((*filter->output_function)(w, filter->data));
127 }
128 break;
129
130 /* '~' */
131 case 2:
132 if (c == 0x7d) { /* '}' */
133 filter->status = 0x0;
134 } else if (c == 0x7b) { /* '{' */
135 filter->status = 0x10;
136 } else if (c == 0x7e) { /* '~' */
137 filter->status = 0x0;
138 CK((*filter->output_function)(0x007e, filter->data));
139 }
140 break;
141
142 default:
143 filter->status = 0;
144 break;
145 }
146
147 return c;
148 }
149
150 /*
151 * wchar => HZ
152 */
153 int
mbfl_filt_conv_wchar_hz(int c,mbfl_convert_filter * filter)154 mbfl_filt_conv_wchar_hz(int c, mbfl_convert_filter *filter)
155 {
156 int s;
157
158 s = 0;
159 if (c >= ucs_a1_cp936_table_min && c < ucs_a1_cp936_table_max) {
160 s = ucs_a1_cp936_table[c - ucs_a1_cp936_table_min];
161 } else if (c >= ucs_a2_cp936_table_min && c < ucs_a2_cp936_table_max) {
162 s = ucs_a2_cp936_table[c - ucs_a2_cp936_table_min];
163 } else if (c >= ucs_a3_cp936_table_min && c < ucs_a3_cp936_table_max) {
164 s = ucs_a3_cp936_table[c - ucs_a3_cp936_table_min];
165 } else if (c >= ucs_i_cp936_table_min && c < ucs_i_cp936_table_max) {
166 s = ucs_i_cp936_table[c - ucs_i_cp936_table_min];
167 } else if (c >= ucs_hff_cp936_table_min && c < ucs_hff_cp936_table_max) {
168 if (c == 0xff04) {
169 s = 0xa1e7;
170 } else if (c == 0xff5e) {
171 s = 0xa1ab;
172 } else if (c >= 0xff01 && c <= 0xff5d) {
173 s = c - 0xff01 + 0xa3a1;
174 } else if (c >= 0xffe0 && c <= 0xffe5) {
175 s = ucs_hff_s_cp936_table[c-0xffe0];
176 }
177 }
178 if (s & 0x8000) {
179 s -= 0x8080;
180 }
181
182 if (s <= 0) {
183 if (c == 0) {
184 s = 0;
185 } else if (s <= 0) {
186 s = -1;
187 }
188 } else if ((s >= 0x80 && s < 0x2121) || (s > 0x8080)) {
189 s = -1;
190 }
191 if (s >= 0) {
192 if (s < 0x80) { /* ASCII */
193 if ((filter->status & 0xff00) != 0) {
194 CK((*filter->output_function)(0x7e, filter->data)); /* '~' */
195 CK((*filter->output_function)(0x7d, filter->data)); /* '}' */
196 }
197 filter->status = 0;
198 if (s == 0x7e){
199 CK((*filter->output_function)(0x7e, filter->data));
200 }
201 CK((*filter->output_function)(s, filter->data));
202 } else { /* GB 2312-80 */
203 if ((filter->status & 0xff00) != 0x200) {
204 CK((*filter->output_function)(0x7e, filter->data)); /* '~' */
205 CK((*filter->output_function)(0x7b, filter->data)); /* '{' */
206 }
207 filter->status = 0x200;
208 CK((*filter->output_function)((s >> 8) & 0x7f, filter->data));
209 CK((*filter->output_function)(s & 0x7f, filter->data));
210 }
211 } else {
212 if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
213 CK(mbfl_filt_conv_illegal_output(c, filter));
214 }
215 }
216
217 return c;
218 }
219
220 int
mbfl_filt_conv_any_hz_flush(mbfl_convert_filter * filter)221 mbfl_filt_conv_any_hz_flush(mbfl_convert_filter *filter)
222 {
223 /* back to latin */
224 if ((filter->status & 0xff00) != 0) {
225 CK((*filter->output_function)(0x7e, filter->data)); /* ~ */
226 CK((*filter->output_function)(0x7d, filter->data)); /* '{' */
227 }
228 filter->status &= 0xff;
229 return 0;
230 }
231
mbfl_filt_ident_hz(int c,mbfl_identify_filter * filter)232 static int mbfl_filt_ident_hz(int c, mbfl_identify_filter *filter)
233 {
234 switch (filter->status & 0xf) {
235 /* case 0x00: ASCII */
236 /* case 0x10: GB2312 */
237 case 0:
238 if (c == 0x7e) {
239 filter->status += 2;
240 } else if (filter->status == 0x10 && c > 0x20 && c < 0x7f) { /* DBCS first char */
241 filter->status += 1;
242 } else if (c >= 0 && c < 0x80) { /* latin, CTLs */
243 ;
244 } else {
245 filter->flag = 1; /* bad */
246 }
247 break;
248
249 /* case 0x11: GB2312 second char */
250 case 1:
251 filter->status &= ~0xf;
252 if (c < 0x21 || c > 0x7e) { /* bad */
253 filter->flag = 1;
254 }
255 break;
256
257 case 2:
258 if (c == 0x7d) { /* '}' */
259 filter->status = 0;
260 } else if (c == 0x7b) { /* '{' */
261 filter->status = 0x10;
262 } else if (c == 0x7e) { /* '~' */
263 filter->status = 0;
264 } else {
265 filter->flag = 1; /* bad */
266 filter->status &= ~0xf;
267 }
268 break;
269
270 default:
271 filter->status = 0;
272 break;
273 }
274
275 return c;
276 }
277
278
279 /*
280 * Local variables:
281 * tab-width: 4
282 * c-basic-offset: 4
283 * End:
284 */
285
