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_utf7imap.h"
36
37 const mbfl_encoding mbfl_encoding_utf7imap = {
38 mbfl_no_encoding_utf7imap,
39 "UTF7-IMAP",
40 NULL,
41 NULL,
42 NULL,
43 MBFL_ENCTYPE_MBCS | MBFL_ENCTYPE_SHFTCODE
44 };
45
46 const struct mbfl_convert_vtbl vtbl_utf7imap_wchar = {
47 mbfl_no_encoding_utf7imap,
48 mbfl_no_encoding_wchar,
49 mbfl_filt_conv_common_ctor,
50 mbfl_filt_conv_common_dtor,
51 mbfl_filt_conv_utf7imap_wchar,
52 mbfl_filt_conv_common_flush };
53
54 const struct mbfl_convert_vtbl vtbl_wchar_utf7imap = {
55 mbfl_no_encoding_wchar,
56 mbfl_no_encoding_utf7imap,
57 mbfl_filt_conv_common_ctor,
58 mbfl_filt_conv_common_dtor,
59 mbfl_filt_conv_wchar_utf7imap,
60 mbfl_filt_conv_wchar_utf7imap_flush };
61
62 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
63
64 /*
65 * UTF7-IMAP => wchar
66 */
mbfl_filt_conv_utf7imap_wchar(int c,mbfl_convert_filter * filter)67 int mbfl_filt_conv_utf7imap_wchar(int c, mbfl_convert_filter *filter)
68 {
69 int s, n;
70
71 n = -1;
72 if (filter->status != 0) { /* Modified Base64 */
73 if (c >= 0x41 && c <= 0x5a) { /* A - Z */
74 n = c - 65;
75 } else if (c >= 0x61 && c <= 0x7a) { /* a - z */
76 n = c - 71;
77 } else if (c >= 0x30 && c <= 0x39) { /* 0 - 9 */
78 n = c + 4;
79 } else if (c == 0x2b) { /* '+' */
80 n = 62;
81 } else if (c == 0x2c) { /* ',' */
82 n = 63;
83 }
84 if (n < 0 || n > 63) {
85 if (c == 0x2d) {
86 if (filter->status == 1) { /* "&-" -> "&" */
87 CK((*filter->output_function)(0x26, filter->data));
88 }
89 } else if (c >= 0 && c < 0x80) { /* ASCII exclude '-' */
90 CK((*filter->output_function)(c, filter->data));
91 } else { /* illegal character */
92 s = c & MBFL_WCSGROUP_MASK;
93 s |= MBFL_WCSGROUP_THROUGH;
94 CK((*filter->output_function)(s, filter->data));
95 }
96 filter->cache = 0;
97 filter->status = 0;
98 return c;
99 }
100 }
101
102 switch (filter->status) {
103 /* directly encoded characters */
104 case 0:
105 if (c == 0x26) { /* '&' shift character */
106 filter->status++;
107 } else if (c >= 0 && c < 0x80) { /* ASCII */
108 CK((*filter->output_function)(c, filter->data));
109 } else { /* illegal character */
110 s = c & MBFL_WCSGROUP_MASK;
111 s |= MBFL_WCSGROUP_THROUGH;
112 CK((*filter->output_function)(s, filter->data));
113 }
114 break;
115
116 /* decode Modified Base64 */
117 case 1:
118 case 2:
119 filter->cache |= n << 10;
120 filter->status = 3;
121 break;
122 case 3:
123 filter->cache |= n << 4;
124 filter->status = 4;
125 break;
126 case 4:
127 s = ((n >> 2) & 0xf) | (filter->cache & 0xffff);
128 n = (n & 0x3) << 14;
129 filter->status = 5;
130 if (s >= 0xd800 && s < 0xdc00) {
131 s = (((s & 0x3ff) << 16) + 0x400000) | n;
132 filter->cache = s;
133 } else if (s >= 0xdc00 && s < 0xe000) {
134 s &= 0x3ff;
135 s |= (filter->cache & 0xfff0000) >> 6;
136 filter->cache = n;
137 if (s >= MBFL_WCSPLANE_SUPMIN && s < MBFL_WCSPLANE_SUPMAX) {
138 CK((*filter->output_function)(s, filter->data));
139 } else { /* illegal character */
140 s &= MBFL_WCSGROUP_MASK;
141 s |= MBFL_WCSGROUP_THROUGH;
142 CK((*filter->output_function)(s, filter->data));
143 }
144 } else {
145 filter->cache = n;
146 CK((*filter->output_function)(s, filter->data));
147 }
148 break;
149
150 case 5:
151 filter->cache |= n << 8;
152 filter->status = 6;
153 break;
154 case 6:
155 filter->cache |= n << 2;
156 filter->status = 7;
157 break;
158 case 7:
159 s = ((n >> 4) & 0x3) | (filter->cache & 0xffff);
160 n = (n & 0xf) << 12;
161 filter->status = 8;
162 if (s >= 0xd800 && s < 0xdc00) {
163 s = (((s & 0x3ff) << 16) + 0x400000) | n;
164 filter->cache = s;
165 } else if (s >= 0xdc00 && s < 0xe000) {
166 s &= 0x3ff;
167 s |= (filter->cache & 0xfff0000) >> 6;
168 filter->cache = n;
169 if (s >= MBFL_WCSPLANE_SUPMIN && s < MBFL_WCSPLANE_SUPMAX) {
170 CK((*filter->output_function)(s, filter->data));
171 } else { /* illegal character */
172 s &= MBFL_WCSGROUP_MASK;
173 s |= MBFL_WCSGROUP_THROUGH;
174 CK((*filter->output_function)(s, filter->data));
175 }
176 } else {
177 filter->cache = n;
178 CK((*filter->output_function)(s, filter->data));
179 }
180 break;
181
182 case 8:
183 filter->cache |= n << 6;
184 filter->status = 9;
185 break;
186 case 9:
187 s = n | (filter->cache & 0xffff);
188 filter->status = 2;
189 if (s >= 0xd800 && s < 0xdc00) {
190 s = (((s & 0x3ff) << 16) + 0x400000);
191 filter->cache = s;
192 } else if (s >= 0xdc00 && s < 0xe000) {
193 s &= 0x3ff;
194 s |= (filter->cache & 0xfff0000) >> 6;
195 filter->cache = 0;
196 if (s >= MBFL_WCSPLANE_SUPMIN && s < MBFL_WCSPLANE_SUPMAX) {
197 CK((*filter->output_function)(s, filter->data));
198 } else { /* illegal character */
199 s &= MBFL_WCSGROUP_MASK;
200 s |= MBFL_WCSGROUP_THROUGH;
201 CK((*filter->output_function)(s, filter->data));
202 }
203 } else {
204 filter->cache = 0;
205 CK((*filter->output_function)(s, filter->data));
206 }
207 break;
208
209 default:
210 filter->status = 0;
211 break;
212 }
213
214 return c;
215 }
216
217 static const unsigned char mbfl_utf7imap_base64_table[] =
218 {
219 /* 'A', 'B', 'C', 'D', 'E', 'F', 'G', 'H', 'I', 'J', 'K', 'L', 'M', */
220 0x41,0x42,0x43,0x44,0x45,0x46,0x47,0x48,0x49,0x4a,0x4b,0x4c,0x4d,
221 /* 'N', 'O', 'P', 'Q', 'R', 'S', 'T', 'U', 'V', 'W', 'X', 'Y', 'Z', */
222 0x4e,0x4f,0x50,0x51,0x52,0x53,0x54,0x55,0x56,0x57,0x58,0x59,0x5a,
223 /* 'a', 'b', 'c', 'd', 'e', 'f', 'g', 'h', 'i', 'j', 'k', 'l', 'm', */
224 0x61,0x62,0x63,0x64,0x65,0x66,0x67,0x68,0x69,0x6a,0x6b,0x6c,0x6d,
225 /* 'n', 'o', 'p', 'q', 'r', 's', 't', 'u', 'v', 'w', 'x', 'y', 'z', */
226 0x6e,0x6f,0x70,0x71,0x72,0x73,0x74,0x75,0x76,0x77,0x78,0x79,0x7a,
227 /* '0', '1', '2', '3', '4', '5', '6', '7', '8', '9', '+', ',', '\0' */
228 0x30,0x31,0x32,0x33,0x34,0x35,0x36,0x37,0x38,0x39,0x2b,0x2c,0x00
229 };
230
231 /*
232 * wchar => UTF7-IMAP
233 */
mbfl_filt_conv_wchar_utf7imap(int c,mbfl_convert_filter * filter)234 int mbfl_filt_conv_wchar_utf7imap(int c, mbfl_convert_filter *filter)
235 {
236 int n, s;
237
238 n = 0;
239 if (c == 0x26) {
240 n = 1;
241 } else if ((c >= 0x20 && c <= 0x7e) || c == 0) {
242 n = 2;
243 } else if (c >= 0 && c < MBFL_WCSPLANE_UCS2MAX) {
244 ;
245 } else if (c >= MBFL_WCSPLANE_SUPMIN && c < MBFL_WCSPLANE_SUPMAX) {
246 s = ((c >> 10) - 0x40) | 0xd800;
247 CK((*filter->filter_function)(s, filter));
248 s = (c & 0x3ff) | 0xdc00;
249 CK((*filter->filter_function)(s, filter));
250 return c;
251 } else {
252 if (filter->illegal_mode != MBFL_OUTPUTFILTER_ILLEGAL_MODE_NONE) {
253 CK(mbfl_filt_conv_illegal_output(c, filter));
254 }
255 return c;
256 }
257
258 switch (filter->status) {
259 case 0:
260 if (n != 0) { /* directly encode characters */
261 CK((*filter->output_function)(c, filter->data));
262 if (n == 1) {
263 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
264 }
265 } else { /* Modified Base64 */
266 CK((*filter->output_function)(0x26, filter->data)); /* '&' */
267 filter->status = 1;
268 filter->cache = c;
269 }
270 break;
271
272 /* encode Modified Base64 */
273 case 1:
274 s = filter->cache;
275 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 10) & 0x3f], filter->data));
276 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 4) & 0x3f], filter->data));
277 if (n != 0) {
278 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s << 2) & 0x3c], filter->data));
279 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
280 CK((*filter->output_function)(c, filter->data));
281 if (n == 1) {
282 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
283 }
284 filter->status = 0;
285 } else {
286 filter->status = 2;
287 filter->cache = ((s & 0xf) << 16) | c;
288 }
289 break;
290
291 case 2:
292 s = filter->cache;
293 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 14) & 0x3f], filter->data));
294 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 8) & 0x3f], filter->data));
295 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 2) & 0x3f], filter->data));
296 if (n != 0) {
297 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s << 4) & 0x30], filter->data));
298 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
299 CK((*filter->output_function)(c, filter->data));
300 if (n == 1) {
301 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
302 }
303 filter->status = 0;
304 } else {
305 filter->status = 3;
306 filter->cache = ((s & 0x3) << 16) | c;
307 }
308 break;
309
310 case 3:
311 s = filter->cache;
312 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 12) & 0x3f], filter->data));
313 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(s >> 6) & 0x3f], filter->data));
314 CK((*filter->output_function)(mbfl_utf7imap_base64_table[s & 0x3f], filter->data));
315 if (n != 0) {
316 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
317 CK((*filter->output_function)(c, filter->data));
318 if (n == 1) {
319 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
320 }
321 filter->status = 0;
322 } else {
323 filter->status = 1;
324 filter->cache = c;
325 }
326 break;
327
328 default:
329 filter->status = 0;
330 break;
331 }
332
333 return c;
334
335 }
336
mbfl_filt_conv_wchar_utf7imap_flush(mbfl_convert_filter * filter)337 int mbfl_filt_conv_wchar_utf7imap_flush(mbfl_convert_filter *filter)
338 {
339 int status, cache;
340
341 status = filter->status;
342 cache = filter->cache;
343 filter->status = 0;
344 filter->cache = 0;
345 /* flush fragments */
346 switch (status) {
347 case 1:
348 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 10) & 0x3f], filter->data));
349 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 4) & 0x3f], filter->data));
350 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache << 2) & 0x3c], filter->data));
351 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
352 break;
353
354 case 2:
355 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 14) & 0x3f], filter->data));
356 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 8) & 0x3f], filter->data));
357 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 2) & 0x3f], filter->data));
358 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache << 4) & 0x30], filter->data));
359 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
360 break;
361
362 case 3:
363 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 12) & 0x3f], filter->data));
364 CK((*filter->output_function)(mbfl_utf7imap_base64_table[(cache >> 6) & 0x3f], filter->data));
365 CK((*filter->output_function)(mbfl_utf7imap_base64_table[cache & 0x3f], filter->data));
366 CK((*filter->output_function)(0x2d, filter->data)); /* '-' */
367 break;
368 }
369 return 0;
370 }
371
372
373
