xf.li | bdd93d5 | 2023-05-12 07:10:14 -0700 | [diff] [blame] | 1 | /* Conversion from and to IBM1364. |
| 2 | Copyright (C) 2005-2016 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 | <http://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 DATA_HEADER "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 EXTRA_LOOP_ARGS , 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 LOOPFCT FROM_LOOP |
| 154 | #define BODY \ |
| 155 | { \ |
| 156 | uint32_t ch = *inptr; \ |
| 157 | \ |
| 158 | if (__builtin_expect (ch, 0) == SO) \ |
| 159 | { \ |
| 160 | /* Shift OUT, change to DBCS converter. */ \ |
| 161 | if (curcs == db) \ |
| 162 | { \ |
| 163 | result = __GCONV_ILLEGAL_INPUT; \ |
| 164 | break; \ |
| 165 | } \ |
| 166 | curcs = db; \ |
| 167 | ++inptr; \ |
| 168 | continue; \ |
| 169 | } \ |
| 170 | if (__builtin_expect (ch, 0) == SI) \ |
| 171 | { \ |
| 172 | /* Shift IN, change to SBCS converter. */ \ |
| 173 | if (curcs == sb) \ |
| 174 | { \ |
| 175 | result = __GCONV_ILLEGAL_INPUT; \ |
| 176 | break; \ |
| 177 | } \ |
| 178 | curcs = sb; \ |
| 179 | ++inptr; \ |
| 180 | continue; \ |
| 181 | } \ |
| 182 | \ |
| 183 | if (curcs == sb) \ |
| 184 | { \ |
| 185 | /* Use the IBM13XX table for single byte. */ \ |
| 186 | uint32_t res = SB_TO_UCS4[ch]; \ |
| 187 | if (__builtin_expect (res, L'\1') == L'\0' && ch != '\0') \ |
| 188 | { \ |
| 189 | /* This is an illegal character. */ \ |
| 190 | if (! ignore_errors_p ()) \ |
| 191 | { \ |
| 192 | result = __GCONV_ILLEGAL_INPUT; \ |
| 193 | break; \ |
| 194 | } \ |
| 195 | ++*irreversible; \ |
| 196 | } \ |
| 197 | else \ |
| 198 | { \ |
| 199 | put32 (outptr, res); \ |
| 200 | outptr += 4; \ |
| 201 | } \ |
| 202 | ++inptr; \ |
| 203 | } \ |
| 204 | else \ |
| 205 | { \ |
| 206 | assert (curcs == db); \ |
| 207 | \ |
| 208 | if (__glibc_unlikely (inptr + 1 >= inend)) \ |
| 209 | { \ |
| 210 | /* The second character is not available. Store the \ |
| 211 | intermediate result. */ \ |
| 212 | result = __GCONV_INCOMPLETE_INPUT; \ |
| 213 | break; \ |
| 214 | } \ |
| 215 | \ |
| 216 | ch = (ch * 0x100) + inptr[1]; \ |
| 217 | \ |
| 218 | /* Use the IBM1364 table for double byte. */ \ |
| 219 | const struct gap *rp2 = DB_TO_UCS4_IDX; \ |
| 220 | while (ch > rp2->end) \ |
| 221 | ++rp2; \ |
| 222 | \ |
| 223 | uint32_t res; \ |
| 224 | if (__builtin_expect (rp2->start == 0xffff, 0) \ |
| 225 | || __builtin_expect (ch < rp2->start, 0) \ |
| 226 | || (res = DB_TO_UCS4[ch + rp2->idx], \ |
| 227 | __builtin_expect (res, L'\1') == L'\0' && ch != '\0')) \ |
| 228 | { \ |
| 229 | /* This is an illegal character. */ \ |
| 230 | if (! ignore_errors_p ()) \ |
| 231 | { \ |
| 232 | result = __GCONV_ILLEGAL_INPUT; \ |
| 233 | break; \ |
| 234 | } \ |
| 235 | ++*irreversible; \ |
| 236 | } \ |
| 237 | else \ |
| 238 | { \ |
| 239 | SUB_COMBINED_UCS_FROM_IBM13XX; \ |
| 240 | } \ |
| 241 | inptr += 2; \ |
| 242 | } \ |
| 243 | } |
| 244 | #define LOOP_NEED_FLAGS |
| 245 | #define EXTRA_LOOP_DECLS , int *curcsp |
| 246 | #define INIT_PARAMS int curcs = *curcsp & ~7 |
| 247 | #define UPDATE_PARAMS *curcsp = curcs |
| 248 | #include <iconv/loop.c> |
| 249 | |
| 250 | |
| 251 | /* Subroutine to convert two UCS4 codes to IBM-13XX. */ |
| 252 | #ifdef HAS_COMBINED |
| 253 | # define SUB_COMBINED_UCS_TO_IBM13XX \ |
| 254 | { \ |
| 255 | const struct combine *cmbp = UCS4_COMB_TO_DB; \ |
| 256 | while (cmbp->res1 < ch) \ |
| 257 | ++cmbp; \ |
| 258 | /* XXX if last char is beginning of combining store in state */ \ |
| 259 | if (cmbp->res1 == ch && inptr + 4 < inend) \ |
| 260 | { \ |
| 261 | /* See if input is part of a combined character. */ \ |
| 262 | uint32_t ch_next = get32 (inptr + 4); \ |
| 263 | while (cmbp->res2 != ch_next) \ |
| 264 | { \ |
| 265 | ++cmbp; \ |
| 266 | if (cmbp->res1 != ch) \ |
| 267 | goto not_combined; \ |
| 268 | } \ |
| 269 | \ |
| 270 | /* It is a combined character. First make sure we are in \ |
| 271 | double byte mode. */ \ |
| 272 | if (curcs == sb) \ |
| 273 | { \ |
| 274 | /* We know there is room for at least one byte. */ \ |
| 275 | *outptr++ = SO; \ |
| 276 | curcs = db; \ |
| 277 | } \ |
| 278 | \ |
| 279 | if (__glibc_unlikely (outptr + 2 > outend)) \ |
| 280 | { \ |
| 281 | result = __GCONV_FULL_OUTPUT; \ |
| 282 | break; \ |
| 283 | } \ |
| 284 | *outptr++ = cmbp->ch[0]; \ |
| 285 | *outptr++ = cmbp->ch[1]; \ |
| 286 | inptr += 8; \ |
| 287 | continue; \ |
| 288 | \ |
| 289 | not_combined:; \ |
| 290 | } \ |
| 291 | } |
| 292 | #else |
| 293 | # define SUB_COMBINED_UCS_TO_IBM13XX |
| 294 | #endif /* HAS_COMBINED */ |
| 295 | |
| 296 | |
| 297 | /* Next, define the other direction. */ |
| 298 | #define MIN_NEEDED_INPUT MIN_NEEDED_TO |
| 299 | #define MIN_NEEDED_OUTPUT MIN_NEEDED_FROM |
| 300 | #define MAX_NEEDED_OUTPUT MAX_NEEDED_FROM |
| 301 | #define LOOPFCT TO_LOOP |
| 302 | #define BODY \ |
| 303 | { \ |
| 304 | uint32_t ch = get32 (inptr); \ |
| 305 | \ |
| 306 | if (__glibc_unlikely (ch >= UCS_LIMIT)) \ |
| 307 | { \ |
| 308 | UNICODE_TAG_HANDLER (ch, 4); \ |
| 309 | \ |
| 310 | if (! ignore_errors_p ()) \ |
| 311 | { \ |
| 312 | result = __GCONV_ILLEGAL_INPUT; \ |
| 313 | break; \ |
| 314 | } \ |
| 315 | ++*irreversible; \ |
| 316 | inptr += 4; \ |
| 317 | continue; \ |
| 318 | } \ |
| 319 | \ |
| 320 | SUB_COMBINED_UCS_TO_IBM13XX; \ |
| 321 | \ |
| 322 | const struct gap *rp1 = UCS4_TO_SB_IDX; \ |
| 323 | while (ch > rp1->end) \ |
| 324 | ++rp1; \ |
| 325 | \ |
| 326 | /* Use the UCS4 table for single byte. */ \ |
| 327 | const char *cp; \ |
| 328 | if (__builtin_expect (ch < rp1->start, 0) \ |
| 329 | || (cp = UCS4_TO_SB[ch + rp1->idx], \ |
| 330 | __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \ |
| 331 | { \ |
| 332 | /* Use the UCS4 table for double byte. */ \ |
| 333 | const struct gap *rp2 = UCS4_TO_DB_IDX; \ |
| 334 | while (ch > rp2->end) \ |
| 335 | ++rp2; \ |
| 336 | \ |
| 337 | if (__builtin_expect (ch < rp2->start, 0) \ |
| 338 | || (cp = UCS4_TO_DB[ch + rp2->idx], \ |
| 339 | __builtin_expect (cp[0], L'\1') == L'\0' && ch != '\0')) \ |
| 340 | { \ |
| 341 | /* This is an illegal character. */ \ |
| 342 | if (! ignore_errors_p ()) \ |
| 343 | { \ |
| 344 | result = __GCONV_ILLEGAL_INPUT; \ |
| 345 | break; \ |
| 346 | } \ |
| 347 | ++*irreversible; \ |
| 348 | } \ |
| 349 | else \ |
| 350 | { \ |
| 351 | if (curcs == sb) \ |
| 352 | { \ |
| 353 | /* We know there is room for at least one byte. */ \ |
| 354 | *outptr++ = SO; \ |
| 355 | curcs = db; \ |
| 356 | } \ |
| 357 | \ |
| 358 | if (__glibc_unlikely (outptr + 2 > outend)) \ |
| 359 | { \ |
| 360 | result = __GCONV_FULL_OUTPUT; \ |
| 361 | break; \ |
| 362 | } \ |
| 363 | *outptr++ = cp[0]; \ |
| 364 | *outptr++ = cp[1]; \ |
| 365 | } \ |
| 366 | } \ |
| 367 | else \ |
| 368 | { \ |
| 369 | if (__glibc_unlikely (curcs == db)) \ |
| 370 | { \ |
| 371 | /* We know there is room for at least one byte. */ \ |
| 372 | *outptr++ = SI; \ |
| 373 | curcs = sb; \ |
| 374 | \ |
| 375 | if (__glibc_unlikely (outptr >= outend)) \ |
| 376 | { \ |
| 377 | result = __GCONV_FULL_OUTPUT; \ |
| 378 | break; \ |
| 379 | } \ |
| 380 | } \ |
| 381 | \ |
| 382 | *outptr++ = cp[0]; \ |
| 383 | } \ |
| 384 | \ |
| 385 | /* Now that we wrote the output increment the input pointer. */ \ |
| 386 | inptr += 4; \ |
| 387 | } |
| 388 | #define LOOP_NEED_FLAGS |
| 389 | #define EXTRA_LOOP_DECLS , int *curcsp |
| 390 | #define INIT_PARAMS int curcs = *curcsp & ~7 |
| 391 | #define REINIT_PARAMS curcs = *curcsp & ~7 |
| 392 | #define UPDATE_PARAMS *curcsp = curcs |
| 393 | #include <iconv/loop.c> |
| 394 | |
| 395 | /* Now define the toplevel functions. */ |
| 396 | #include <iconv/skeleton.c> |