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 file was separated from mbfilter.c
26 * by moriyoshi koizumi <moriyoshi@php.net> on 20 dec 2002.
27 *
28 */
29
30 #include "mbfilter.h"
31 #include "mbfilter_utf32.h"
32
33 static int mbfl_filt_conv_utf32_wchar_flush(mbfl_convert_filter *filter);
34 static size_t mb_utf32_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state);
35 static size_t mb_utf32be_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state);
36 static void mb_wchar_to_utf32be(uint32_t *in, size_t len, mb_convert_buf *buf, bool end);
37 static size_t mb_utf32le_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state);
38 static void mb_wchar_to_utf32le(uint32_t *in, size_t len, mb_convert_buf *buf, bool end);
39
40 static const char *mbfl_encoding_utf32_aliases[] = {"utf32", NULL};
41
42 const mbfl_encoding mbfl_encoding_utf32 = {
43 mbfl_no_encoding_utf32,
44 "UTF-32",
45 "UTF-32",
46 mbfl_encoding_utf32_aliases,
47 NULL,
48 MBFL_ENCTYPE_WCS4,
49 &vtbl_utf32_wchar,
50 &vtbl_wchar_utf32,
51 mb_utf32_to_wchar,
52 mb_wchar_to_utf32be,
53 NULL
54 };
55
56 const mbfl_encoding mbfl_encoding_utf32be = {
57 mbfl_no_encoding_utf32be,
58 "UTF-32BE",
59 "UTF-32BE",
60 NULL,
61 NULL,
62 MBFL_ENCTYPE_WCS4,
63 &vtbl_utf32be_wchar,
64 &vtbl_wchar_utf32be,
65 mb_utf32be_to_wchar,
66 mb_wchar_to_utf32be,
67 NULL
68 };
69
70 const mbfl_encoding mbfl_encoding_utf32le = {
71 mbfl_no_encoding_utf32le,
72 "UTF-32LE",
73 "UTF-32LE",
74 NULL,
75 NULL,
76 MBFL_ENCTYPE_WCS4,
77 &vtbl_utf32le_wchar,
78 &vtbl_wchar_utf32le,
79 mb_utf32le_to_wchar,
80 mb_wchar_to_utf32le,
81 NULL
82 };
83
84 const struct mbfl_convert_vtbl vtbl_utf32_wchar = {
85 mbfl_no_encoding_utf32,
86 mbfl_no_encoding_wchar,
87 mbfl_filt_conv_common_ctor,
88 NULL,
89 mbfl_filt_conv_utf32_wchar,
90 mbfl_filt_conv_utf32_wchar_flush,
91 NULL,
92 };
93
94 const struct mbfl_convert_vtbl vtbl_wchar_utf32 = {
95 mbfl_no_encoding_wchar,
96 mbfl_no_encoding_utf32,
97 mbfl_filt_conv_common_ctor,
98 NULL,
99 mbfl_filt_conv_wchar_utf32be,
100 mbfl_filt_conv_common_flush,
101 NULL,
102 };
103
104 const struct mbfl_convert_vtbl vtbl_utf32be_wchar = {
105 mbfl_no_encoding_utf32be,
106 mbfl_no_encoding_wchar,
107 mbfl_filt_conv_common_ctor,
108 NULL,
109 mbfl_filt_conv_utf32be_wchar,
110 mbfl_filt_conv_utf32_wchar_flush,
111 NULL,
112 };
113
114 const struct mbfl_convert_vtbl vtbl_wchar_utf32be = {
115 mbfl_no_encoding_wchar,
116 mbfl_no_encoding_utf32be,
117 mbfl_filt_conv_common_ctor,
118 NULL,
119 mbfl_filt_conv_wchar_utf32be,
120 mbfl_filt_conv_common_flush,
121 NULL,
122 };
123
124 const struct mbfl_convert_vtbl vtbl_utf32le_wchar = {
125 mbfl_no_encoding_utf32le,
126 mbfl_no_encoding_wchar,
127 mbfl_filt_conv_common_ctor,
128 NULL,
129 mbfl_filt_conv_utf32le_wchar,
130 mbfl_filt_conv_utf32_wchar_flush,
131 NULL,
132 };
133
134 const struct mbfl_convert_vtbl vtbl_wchar_utf32le = {
135 mbfl_no_encoding_wchar,
136 mbfl_no_encoding_utf32le,
137 mbfl_filt_conv_common_ctor,
138 NULL,
139 mbfl_filt_conv_wchar_utf32le,
140 mbfl_filt_conv_common_flush,
141 NULL,
142 };
143
144 #define CK(statement) do { if ((statement) < 0) return (-1); } while (0)
145
emit_char_if_valid(int n,mbfl_convert_filter * filter)146 static int emit_char_if_valid(int n, mbfl_convert_filter *filter)
147 {
148 if (n >= 0 && n < MBFL_WCSPLANE_UTF32MAX && (n < 0xD800 || n > 0xDFFF)) {
149 CK((*filter->output_function)(n, filter->data));
150 } else {
151 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
152 }
153 return 0;
154 }
155
mbfl_filt_conv_utf32_wchar(int c,mbfl_convert_filter * filter)156 int mbfl_filt_conv_utf32_wchar(int c, mbfl_convert_filter *filter)
157 {
158 if (filter->status < 3) {
159 filter->cache = (filter->cache << 8) | (c & 0xFF);
160 filter->status++;
161 } else {
162 int n = ((unsigned int)filter->cache << 8) | (c & 0xFF);
163 filter->cache = filter->status = 0;
164
165 if (n == 0xFFFE0000) {
166 /* Found a little-endian byte order mark */
167 filter->filter_function = mbfl_filt_conv_utf32le_wchar;
168 } else {
169 filter->filter_function = mbfl_filt_conv_utf32be_wchar;
170 if (n != 0xFEFF) {
171 CK(emit_char_if_valid(n, filter));
172 }
173 }
174 }
175
176 return 0;
177 }
178
mbfl_filt_conv_utf32be_wchar(int c,mbfl_convert_filter * filter)179 int mbfl_filt_conv_utf32be_wchar(int c, mbfl_convert_filter *filter)
180 {
181 if (filter->status < 3) {
182 filter->cache = (filter->cache << 8) | (c & 0xFF);
183 filter->status++;
184 } else {
185 int n = ((unsigned int)filter->cache << 8) | (c & 0xFF);
186 filter->cache = filter->status = 0;
187 CK(emit_char_if_valid(n, filter));
188 }
189 return 0;
190 }
191
mbfl_filt_conv_wchar_utf32be(int c,mbfl_convert_filter * filter)192 int mbfl_filt_conv_wchar_utf32be(int c, mbfl_convert_filter *filter)
193 {
194 if (c >= 0 && c < MBFL_WCSPLANE_UTF32MAX) {
195 CK((*filter->output_function)((c >> 24) & 0xff, filter->data));
196 CK((*filter->output_function)((c >> 16) & 0xff, filter->data));
197 CK((*filter->output_function)((c >> 8) & 0xff, filter->data));
198 CK((*filter->output_function)(c & 0xff, filter->data));
199 } else {
200 CK(mbfl_filt_conv_illegal_output(c, filter));
201 }
202
203 return 0;
204 }
205
mbfl_filt_conv_utf32le_wchar(int c,mbfl_convert_filter * filter)206 int mbfl_filt_conv_utf32le_wchar(int c, mbfl_convert_filter *filter)
207 {
208 if (filter->status < 3) {
209 filter->cache |= ((c & 0xFFU) << (8 * filter->status));
210 filter->status++;
211 } else {
212 int n = ((c & 0xFFU) << 24) | filter->cache;
213 filter->cache = filter->status = 0;
214 CK(emit_char_if_valid(n, filter));
215 }
216 return 0;
217 }
218
mbfl_filt_conv_wchar_utf32le(int c,mbfl_convert_filter * filter)219 int mbfl_filt_conv_wchar_utf32le(int c, mbfl_convert_filter *filter)
220 {
221 if (c >= 0 && c < MBFL_WCSPLANE_UTF32MAX) {
222 CK((*filter->output_function)(c & 0xff, filter->data));
223 CK((*filter->output_function)((c >> 8) & 0xff, filter->data));
224 CK((*filter->output_function)((c >> 16) & 0xff, filter->data));
225 CK((*filter->output_function)((c >> 24) & 0xff, filter->data));
226 } else {
227 CK(mbfl_filt_conv_illegal_output(c, filter));
228 }
229
230 return 0;
231 }
232
mbfl_filt_conv_utf32_wchar_flush(mbfl_convert_filter * filter)233 static int mbfl_filt_conv_utf32_wchar_flush(mbfl_convert_filter *filter)
234 {
235 if (filter->status) {
236 /* Input string was truncated */
237 CK((*filter->output_function)(MBFL_BAD_INPUT, filter->data));
238 }
239 filter->cache = filter->status = 0;
240
241 if (filter->flush_function) {
242 (*filter->flush_function)(filter->data);
243 }
244
245 return 0;
246 }
247
248 #define DETECTED_BE 1
249 #define DETECTED_LE 2
250
mb_utf32_to_wchar(unsigned char ** in,size_t * in_len,uint32_t * buf,size_t bufsize,unsigned int * state)251 static size_t mb_utf32_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state)
252 {
253 if (*state == DETECTED_BE) {
254 return mb_utf32be_to_wchar(in, in_len, buf, bufsize, NULL);
255 } else if (*state == DETECTED_LE) {
256 return mb_utf32le_to_wchar(in, in_len, buf, bufsize, NULL);
257 } else if (*in_len >= 4) {
258 unsigned char *p = *in;
259 uint32_t c1 = *p++;
260 uint32_t c2 = *p++;
261 uint32_t c3 = *p++;
262 uint32_t c4 = *p++;
263 uint32_t w = (c1 << 24) | (c2 << 16) | (c3 << 8) | c4;
264
265 if (w == 0xFFFE0000) {
266 /* Little-endian BOM */
267 *in = p;
268 *in_len -= 4;
269 *state = DETECTED_LE;
270 return mb_utf32le_to_wchar(in, in_len, buf, bufsize, NULL);
271 } else if (w == 0xFEFF) {
272 /* Big-endian BOM; don't send it to output */
273 *in = p;
274 *in_len -= 4;
275 }
276 }
277
278 *state = DETECTED_BE;
279 return mb_utf32be_to_wchar(in, in_len, buf, bufsize, NULL);
280 }
281
mb_utf32be_to_wchar(unsigned char ** in,size_t * in_len,uint32_t * buf,size_t bufsize,unsigned int * state)282 static size_t mb_utf32be_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state)
283 {
284 unsigned char *p = *in, *e = p + (*in_len & ~3);
285 uint32_t *out = buf, *limit = buf + bufsize;
286
287 while (p < e && out < limit) {
288 uint32_t c1 = *p++;
289 uint32_t c2 = *p++;
290 uint32_t c3 = *p++;
291 uint32_t c4 = *p++;
292 uint32_t w = (c1 << 24) | (c2 << 16) | (c3 << 8) | c4;
293
294 if (w < MBFL_WCSPLANE_UTF32MAX && (w < 0xD800 || w > 0xDFFF)) {
295 *out++ = w;
296 } else {
297 *out++ = MBFL_BAD_INPUT;
298 }
299 }
300
301 if (p == e && (*in_len & 0x3) && out < limit) {
302 /* There are 1-3 trailing bytes, which shouldn't be there */
303 *out++ = MBFL_BAD_INPUT;
304 p = *in + *in_len;
305 }
306
307 *in_len -= (p - *in);
308 *in = p;
309 return out - buf;
310 }
311
mb_wchar_to_utf32be(uint32_t * in,size_t len,mb_convert_buf * buf,bool end)312 static void mb_wchar_to_utf32be(uint32_t *in, size_t len, mb_convert_buf *buf, bool end)
313 {
314 unsigned char *out, *limit;
315 MB_CONVERT_BUF_LOAD(buf, out, limit);
316 MB_CONVERT_BUF_ENSURE(buf, out, limit, len * 4);
317
318 while (len--) {
319 uint32_t w = *in++;
320 if (w < MBFL_WCSPLANE_UTF32MAX) {
321 out = mb_convert_buf_add4(out, (w >> 24) & 0xFF, (w >> 16) & 0xFF, (w >> 8) & 0xFF, w & 0xFF);
322 } else {
323 MB_CONVERT_ERROR(buf, out, limit, w, mb_wchar_to_utf32be);
324 MB_CONVERT_BUF_ENSURE(buf, out, limit, len * 4);
325 }
326 }
327
328 MB_CONVERT_BUF_STORE(buf, out, limit);
329 }
330
mb_utf32le_to_wchar(unsigned char ** in,size_t * in_len,uint32_t * buf,size_t bufsize,unsigned int * state)331 static size_t mb_utf32le_to_wchar(unsigned char **in, size_t *in_len, uint32_t *buf, size_t bufsize, unsigned int *state)
332 {
333 unsigned char *p = *in, *e = p + (*in_len & ~3);
334 uint32_t *out = buf, *limit = buf + bufsize;
335
336 while (p < e && out < limit) {
337 uint32_t c1 = *p++;
338 uint32_t c2 = *p++;
339 uint32_t c3 = *p++;
340 uint32_t c4 = *p++;
341 uint32_t w = (c4 << 24) | (c3 << 16) | (c2 << 8) | c1;
342
343 if (w < MBFL_WCSPLANE_UTF32MAX && (w < 0xD800 || w > 0xDFFF)) {
344 *out++ = w;
345 } else {
346 *out++ = MBFL_BAD_INPUT;
347 }
348 }
349
350 if (p == e && (*in_len & 0x3) && out < limit) {
351 /* There are 1-3 trailing bytes, which shouldn't be there */
352 *out++ = MBFL_BAD_INPUT;
353 p = *in + *in_len;
354 }
355
356 *in_len -= (p - *in);
357 *in = p;
358 return out - buf;
359 }
360
mb_wchar_to_utf32le(uint32_t * in,size_t len,mb_convert_buf * buf,bool end)361 static void mb_wchar_to_utf32le(uint32_t *in, size_t len, mb_convert_buf *buf, bool end)
362 {
363 unsigned char *out, *limit;
364 MB_CONVERT_BUF_LOAD(buf, out, limit);
365 MB_CONVERT_BUF_ENSURE(buf, out, limit, len * 4);
366
367 while (len--) {
368 uint32_t w = *in++;
369 if (w < MBFL_WCSPLANE_UTF32MAX) {
370 out = mb_convert_buf_add4(out, w & 0xFF, (w >> 8) & 0xFF, (w >> 16) & 0xFF, (w >> 24) & 0xFF);
371 } else {
372 MB_CONVERT_ERROR(buf, out, limit, w, mb_wchar_to_utf32le);
373 MB_CONVERT_BUF_ENSURE(buf, out, limit, len * 4);
374 }
375 }
376
377 MB_CONVERT_BUF_STORE(buf, out, limit);
378 }
379