1 | /* Conversion from and to IBM1364. |
2 | Copyright (C) 2005-2021 Free Software Foundation, Inc. |
3 | This file is part of the GNU C Library. |
4 | Contributed by Masahide Washizawa <washi@jp.ibm.com>, 2005. |
5 | |
6 | The GNU C Library is free software; you can redistribute it and/or |
7 | modify it under the terms of the GNU Lesser General Public |
8 | License as published by the Free Software Foundation; either |
9 | version 2.1 of the License, or (at your option) any later version. |
10 | |
11 | The GNU C Library 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 GNU |
14 | 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 the GNU C Library; if not, see |
18 | <https://www.gnu.org/licenses/>. */ |
19 | |
20 | #include <dlfcn.h> |
21 | #include <stdint.h> |
22 | #include <wchar.h> |
23 | #include <byteswap.h> |
24 | |
25 | #ifndef CHARSET_NAME |
26 | /* This is really the IBM1364 converter, not another module sharing |
27 | the code. */ |
28 | # define "ibm1364.h" |
29 | # define CHARSET_NAME "IBM1364//" |
30 | # define FROM_LOOP from_ibm1364 |
31 | # define TO_LOOP to_ibm1364 |
32 | # define SB_TO_UCS4 __ibm1364sb_to_ucs4 |
33 | # define DB_TO_UCS4_IDX __ibm1364db_to_ucs4_idx |
34 | # define DB_TO_UCS4 __ibm1364db_to_ucs4 |
35 | # define UCS4_TO_SB_IDX __ucs4_to_ibm1364sb_idx |
36 | # define UCS4_TO_SB __ucs4_to_ibm1364sb |
37 | # define UCS4_TO_DB_IDX __ucs4_to_ibm1364db_idx |
38 | # define UCS4_TO_DB __ucs4_to_ibm1364db |
39 | # define UCS_LIMIT 0xffff |
40 | #endif |
41 | |
42 | |
43 | #include DATA_HEADER |
44 | |
45 | /* The shift sequences for this charset (it does not use ESC). */ |
46 | #define SI 0x0F /* Shift In, host code to turn DBCS off. */ |
47 | #define SO 0x0E /* Shift Out, host code to turn DBCS on. */ |
48 | |
49 | /* Definitions used in the body of the `gconv' function. */ |
50 | #define MIN_NEEDED_FROM 1 |
51 | #define MAX_NEEDED_FROM 2 |
52 | #define MIN_NEEDED_TO 4 |
53 | #ifdef HAS_COMBINED |
54 | # define MAX_NEEDED_TO 8 |
55 | #else |
56 | # define MAX_NEEDED_TO 4 |
57 | #endif |
58 | #define ONE_DIRECTION 0 |
59 | #define PREPARE_LOOP \ |
60 | int save_curcs; \ |
61 | int *curcsp = &data->__statep->__count; |
62 | #define , curcsp |
63 | |
64 | /* Definitions of initialization and destructor function. */ |
65 | #define DEFINE_INIT 1 |
66 | #define DEFINE_FINI 1 |
67 | |
68 | |
69 | /* Since this is a stateful encoding we have to provide code which resets |
70 | the output state to the initial state. This has to be done during the |
71 | flushing. */ |
72 | #define EMIT_SHIFT_TO_INIT \ |
73 | if ((data->__statep->__count & ~7) != sb) \ |
74 | { \ |
75 | if (FROM_DIRECTION) \ |
76 | data->__statep->__count &= 7; \ |
77 | else \ |
78 | { \ |
79 | /* We are not in the initial state. To switch back we have \ |
80 | to emit `SI'. */ \ |
81 | if (__glibc_unlikely (outbuf >= outend)) \ |
82 | /* We don't have enough room in the output buffer. */ \ |
83 | status = __GCONV_FULL_OUTPUT; \ |
84 | else \ |
85 | { \ |
86 | /* Write out the shift sequence. */ \ |
87 | *outbuf++ = SI; \ |
88 | data->__statep->__count &= 7; \ |
89 | } \ |
90 | } \ |
91 | } |
92 | |
93 | |
94 | /* Since we might have to reset input pointer we must be able to save |
95 | and retore the state. */ |
96 | #define SAVE_RESET_STATE(Save) \ |
97 | if (Save) \ |
98 | save_curcs = *curcsp; \ |
99 | else \ |
100 | *curcsp = save_curcs |
101 | |
102 | |
103 | /* Current codeset type. */ |
104 | enum |
105 | { |
106 | sb = 0, |
107 | db = 64 |
108 | }; |
109 | |
110 | |
111 | /* Subroutine to write out converted UCS4 from IBM-13XX. */ |
112 | #ifdef HAS_COMBINED |
113 | # define SUB_COMBINED_UCS_FROM_IBM13XX \ |
114 | { \ |
115 | if (res != UCS_LIMIT || ch < __TO_UCS4_COMBINED_MIN \ |
116 | || ch > __TO_UCS4_COMBINED_MAX) \ |
117 | { \ |
118 | put32 (outptr, res); \ |
119 | outptr += 4; \ |
120 | } \ |
121 | else \ |
122 | { \ |
123 | /* This is a combined character. Make sure we have room. */ \ |
124 | if (__glibc_unlikely (outptr + 8 > outend)) \ |
125 | { \ |
126 | result = __GCONV_FULL_OUTPUT; \ |
127 | break; \ |
128 | } \ |
129 | \ |
130 | const struct divide *cmbp \ |
131 | = &DB_TO_UCS4_COMB[ch - __TO_UCS4_COMBINED_MIN]; \ |
132 | assert (cmbp->res1 != 0 && cmbp->res2 != 0); \ |
133 | \ |
134 | put32 (outptr, cmbp->res1); \ |
135 | outptr += 4; \ |
136 | put32 (outptr, cmbp->res2); \ |
137 | outptr += 4; \ |
138 | } \ |
139 | } |
140 | #else |
141 | # define SUB_COMBINED_UCS_FROM_IBM13XX \ |
142 | { \ |
143 | put32 (outptr, res); \ |
144 | outptr += 4; \ |
145 | } |
146 | #endif /* HAS_COMBINED */ |
147 | |
148 | |
149 | /* First, define the conversion function from IBM-13XX to UCS4. */ |
150 | #define MIN_NEEDED_INPUT MIN_NEEDED_FROM |
151 | #define MAX_NEEDED_INPUT MAX_NEEDED_FROM |
152 | #define MIN_NEEDED_OUTPUT MIN_NEEDED_TO |
153 | #define MAX_NEEDED_OUTPUT MAX_NEEDED_TO |
154 | #define LOOPFCT FROM_LOOP |
155 | #define BODY \ |
156 | { \ |
157 | uint32_t ch = *inptr; \ |
158 | \ |
159 | if (__builtin_expect (ch, 0) == SO) \ |
160 | { \ |
161 | /* Shift OUT, change to DBCS converter (redundant escape okay). */ \ |
162 | curcs = db; \ |
163 | ++inptr; \ |
164 | continue; \ |
165 | } \ |
166 | if (__builtin_expect (ch, 0) == SI) \ |
167 | { \ |
168 | /* Shift IN, change to SBCS converter (redundant escape okay). */ \ |
169 | curcs = sb; \ |
170 | ++inptr; \ |
171 | continue; \ |
172 | } \ |
173 | \ |
174 | if (curcs == sb) \ |
175 | { \ |
176 | /* Use the IBM13XX table for single byte. */ \ |
177 | uint32_t res = SB_TO_UCS4[ch]; \ |
178 | if (__builtin_expect (res, L'\1') == L'\0' && ch != '\0') \ |
179 | { \ |
180 | /* This is an illegal character. */ \ |
181 | if (! ignore_errors_p ()) \ |
182 | { \ |
183 | result = __GCONV_ILLEGAL_INPUT; \ |
184 | break; \ |
185 | } \ |
186 | ++*irreversible; \ |
187 | } \ |
188 | else \ |
189 | { \ |
190 | put32 (outptr, res); \ |
191 | outptr += 4; \ |
192 | } \ |
193 | ++inptr; \ |
194 | } \ |
195 | else \ |
196 | { \ |
197 | assert (curcs == db); \ |
198 | \ |
199 | if (__glibc_unlikely (inptr + 1 >= inend)) \ |
200 | { \ |
201 | /* The second character is not available. Store the \ |
202 | intermediate result. */ \ |
203 | result = __GCONV_INCOMPLETE_INPUT; \ |
204 | break; \ |
205 | } \ |
206 | \ |
207 | ch = (ch * 0x100) + inptr[1]; \ |
208 | \ |
209 | /* Use the IBM1364 table for double byte. */ \ |
210 | const struct gap *rp2 = DB_TO_UCS4_IDX; \ |
211 | while (ch > rp2->end) \ |
212 | ++rp2; \ |
213 | \ |
214 | uint32_t res; \ |
215 | if (__builtin_expect (rp2->start == 0xffff, 0) \ |
216 | || __builtin_expect (ch < rp2->start, 0) \ |
217 | || (res = DB_TO_UCS4[ch + rp2->idx], \ |
218 | __builtin_expect (res, L'\1') == L'\0' && ch != '\0')) \ |
219 | { \ |
220 | /* This is an illegal character. */ \ |
221 | if (! ignore_errors_p ()) \ |
222 | { \ |
223 | result = __GCONV_ILLEGAL_INPUT; \ |
224 | break; \ |
225 | } \ |
226 | ++*irreversible; \ |
227 | } \ |
228 | else \ |
229 | { \ |
230 | SUB_COMBINED_UCS_FROM_IBM13XX; \ |
231 | } \ |
232 | inptr += 2; \ |
233 | } \ |
234 | } |
235 | #define LOOP_NEED_FLAGS |
236 | #define , int *curcsp |
237 | #define INIT_PARAMS int curcs = *curcsp & ~7 |
238 | #define UPDATE_PARAMS *curcsp = curcs |
239 | #include <iconv/loop.c> |
240 | |
241 | |
242 | /* Subroutine to convert two UCS4 codes to IBM-13XX. */ |
243 | #ifdef HAS_COMBINED |
244 | # define SUB_COMBINED_UCS_TO_IBM13XX \ |
245 | { \ |
246 | const struct combine *cmbp = UCS4_COMB_TO_DB; \ |
247 | while (cmbp->res1 < ch) \ |
248 | ++cmbp; \ |
249 | /* XXX if last char is beginning of combining store in state */ \ |
250 | if (cmbp->res1 == ch && inptr + 4 < inend) \ |
251 | { \ |
252 | /* See if input is part of a combined character. */ \ |
253 | uint32_t ch_next = get32 (inptr + 4); \ |
254 | while (cmbp->res2 != ch_next) \ |
255 | { \ |
256 | ++cmbp; \ |
257 | if (cmbp->res1 != ch) \ |
258 | goto not_combined; \ |
259 | } \ |
260 | \ |
261 | /* It is a combined character. First make sure we are in \ |
262 | double byte mode. */ \ |
263 | if (curcs == sb) \ |
264 | { \ |
265 | /* We know there is room for at least one byte. */ \ |
266 | *outptr++ = SO; \ |
267 | curcs = db; \ |
268 | } \ |
269 | \ |
270 | if (__glibc_unlikely (outptr + 2 > outend)) \ |
271 | { \ |
272 | result = __GCONV_FULL_OUTPUT; \ |
273 | break; \ |
274 | } \ |
275 | *outptr++ = cmbp->ch[0]; \ |
276 | *outptr++ = cmbp->ch[1]; \ |
277 | inptr += 8; \ |
278 | continue; \ |
279 | \ |
280 | not_combined:; \ |
281 | } \ |
282 | } |
283 | #else |
284 | # define SUB_COMBINED_UCS_TO_IBM13XX |
285 | #endif /* HAS_COMBINED */ |
286 | |
287 | |
288 | /* Next, define the other direction. */ |
289 | #define MIN_NEEDED_INPUT MIN_NEEDED_TO |
290 | #define MAX_NEEDED_INPUT MAX_NEEDED_TO |
291 | #define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM |
292 | #define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM |
293 | #define LOOPFCT TO_LOOP |
294 | #define BODY \ |
295 | { \ |
296 | uint32_t ch = get32 (inptr); \ |
297 | \ |
298 | if (__glibc_unlikely (ch >= UCS_LIMIT)) \ |
299 | { \ |
300 | UNICODE_TAG_HANDLER (ch, 4); \ |
301 | \ |
302 | if (! ignore_errors_p ()) \ |
303 | { \ |
304 | result = __GCONV_ILLEGAL_INPUT; \ |
305 | break; \ |
306 | } \ |
307 | ++*irreversible; \ |
308 | inptr += 4; \ |
309 | continue; \ |
310 | } \ |
311 | \ |
312 | SUB_COMBINED_UCS_TO_IBM13XX; \ |
313 | \ |
314 | const struct gap *rp1 = UCS4_TO_SB_IDX; \ |
315 | while (ch > rp1->end) \ |
316 | ++rp1; \ |
317 | \ |
318 | /* Use the UCS4 table for single byte. */ \ |
319 | const char *cp; \ |
320 | if (__builtin_expect (ch < rp1->start, 0) \ |
321 | || (cp = UCS4_TO_SB[ch + rp1->idx], \ |
322 | __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \ |
323 | { \ |
324 | /* Use the UCS4 table for double byte. */ \ |
325 | const struct gap *rp2 = UCS4_TO_DB_IDX; \ |
326 | while (ch > rp2->end) \ |
327 | ++rp2; \ |
328 | \ |
329 | if (__builtin_expect (ch < rp2->start, 0) \ |
330 | || (cp = UCS4_TO_DB[ch + rp2->idx], \ |
331 | __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \ |
332 | { \ |
333 | /* This is an illegal character. */ \ |
334 | if (! ignore_errors_p ()) \ |
335 | { \ |
336 | result = __GCONV_ILLEGAL_INPUT; \ |
337 | break; \ |
338 | } \ |
339 | ++*irreversible; \ |
340 | } \ |
341 | else \ |
342 | { \ |
343 | if (curcs == sb) \ |
344 | { \ |
345 | /* We know there is room for at least one byte. */ \ |
346 | *outptr++ = SO; \ |
347 | curcs = db; \ |
348 | } \ |
349 | \ |
350 | if (__glibc_unlikely (outptr + 2 > outend)) \ |
351 | { \ |
352 | result = __GCONV_FULL_OUTPUT; \ |
353 | break; \ |
354 | } \ |
355 | *outptr++ = cp[0]; \ |
356 | *outptr++ = cp[1]; \ |
357 | } \ |
358 | } \ |
359 | else \ |
360 | { \ |
361 | if (__glibc_unlikely (curcs == db)) \ |
362 | { \ |
363 | /* We know there is room for at least one byte. */ \ |
364 | *outptr++ = SI; \ |
365 | curcs = sb; \ |
366 | \ |
367 | if (__glibc_unlikely (outptr >= outend)) \ |
368 | { \ |
369 | result = __GCONV_FULL_OUTPUT; \ |
370 | break; \ |
371 | } \ |
372 | } \ |
373 | \ |
374 | *outptr++ = cp[0]; \ |
375 | } \ |
376 | \ |
377 | /* Now that we wrote the output increment the input pointer. */ \ |
378 | inptr += 4; \ |
379 | } |
380 | #define LOOP_NEED_FLAGS |
381 | #define , int *curcsp |
382 | #define INIT_PARAMS int curcs = *curcsp & ~7 |
383 | #define REINIT_PARAMS curcs = *curcsp & ~7 |
384 | #define UPDATE_PARAMS *curcsp = curcs |
385 | #include <iconv/loop.c> |
386 | |
387 | /* Now define the toplevel functions. */ |
388 | #include <iconv/skeleton.c> |
389 | |