1 /* $NetBSD: citrus_utf7.c,v 1.5 2006/08/23 12:57:24 tnozaki Exp $ */ 2 3 /*- 4 * SPDX-License-Identifier: BSD-2-Clause 5 * 6 * Copyright (c)2004, 2005 Citrus Project, 7 * All rights reserved. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 18 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 19 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 20 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 21 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 22 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 23 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 24 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 25 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 26 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 27 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 28 * SUCH DAMAGE. 29 * 30 */ 31 32 #include <sys/cdefs.h> 33 34 #include <assert.h> 35 #include <errno.h> 36 #include <limits.h> 37 #include <stdio.h> 38 #include <stdint.h> 39 #include <stdlib.h> 40 #include <string.h> 41 #include <wchar.h> 42 43 #include "citrus_namespace.h" 44 #include "citrus_types.h" 45 #include "citrus_module.h" 46 #include "citrus_stdenc.h" 47 #include "citrus_utf7.h" 48 49 /* ---------------------------------------------------------------------- 50 * private stuffs used by templates 51 */ 52 53 #define EI_MASK UINT16_C(0xff) 54 #define EI_DIRECT UINT16_C(0x100) 55 #define EI_OPTION UINT16_C(0x200) 56 #define EI_SPACE UINT16_C(0x400) 57 58 typedef struct { 59 uint16_t cell[0x80]; 60 } _UTF7EncodingInfo; 61 62 typedef struct { 63 unsigned int 64 mode: 1, /* whether base64 mode */ 65 bits: 4, /* need to hold 0 - 15 */ 66 cache: 22; /* 22 = BASE64_BIT + UTF16_BIT */ 67 int chlen; 68 char ch[4]; /* BASE64_IN, 3 * 6 = 18, most closed to UTF16_BIT */ 69 } _UTF7State; 70 71 #define _CEI_TO_EI(_cei_) (&(_cei_)->ei) 72 #define _CEI_TO_STATE(_cei_, _func_) (_cei_)->states.s_##_func_ 73 74 #define _FUNCNAME(m) _citrus_UTF7_##m 75 #define _ENCODING_INFO _UTF7EncodingInfo 76 #define _ENCODING_STATE _UTF7State 77 #define _ENCODING_MB_CUR_MAX(_ei_) 4 78 #define _ENCODING_IS_STATE_DEPENDENT 1 79 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_) 0 80 81 static __inline void 82 /*ARGSUSED*/ 83 _citrus_UTF7_init_state(_UTF7EncodingInfo * __restrict ei __unused, 84 _UTF7State * __restrict s) 85 { 86 87 memset((void *)s, 0, sizeof(*s)); 88 } 89 90 #if 0 91 static __inline void 92 /*ARGSUSED*/ 93 _citrus_UTF7_pack_state(_UTF7EncodingInfo * __restrict ei __unused, 94 void *__restrict pspriv, const _UTF7State * __restrict s) 95 { 96 97 memcpy(pspriv, (const void *)s, sizeof(*s)); 98 } 99 100 static __inline void 101 /*ARGSUSED*/ 102 _citrus_UTF7_unpack_state(_UTF7EncodingInfo * __restrict ei __unused, 103 _UTF7State * __restrict s, const void * __restrict pspriv) 104 { 105 106 memcpy((void *)s, pspriv, sizeof(*s)); 107 } 108 #endif 109 110 static const char base64[] = 111 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 112 "abcdefghijklmnopqrstuvwxyz" 113 "0123456789+/"; 114 115 static const char direct[] = 116 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 117 "abcdefghijklmnopqrstuvwxyz" 118 "0123456789'(),-./:?"; 119 120 static const char option[] = "!\"#$%&*;<=>@[]^_`{|}"; 121 static const char spaces[] = " \t\r\n"; 122 123 #define BASE64_BIT 6 124 #define UTF16_BIT 16 125 126 #define BASE64_MAX 0x3f 127 #define UTF16_MAX UINT16_C(0xffff) 128 #define UTF32_MAX UINT32_C(0x10ffff) 129 130 #define BASE64_IN '+' 131 #define BASE64_OUT '-' 132 133 #define SHIFT7BIT(c) ((c) >> 7) 134 #define ISSPECIAL(c) ((c) == '\0' || (c) == BASE64_IN) 135 136 #define FINDLEN(ei, c) \ 137 (SHIFT7BIT((c)) ? -1 : (((ei)->cell[(c)] & EI_MASK) - 1)) 138 139 #define ISDIRECT(ei, c) (!SHIFT7BIT((c)) && (ISSPECIAL((c)) || \ 140 ei->cell[(c)] & (EI_DIRECT | EI_OPTION | EI_SPACE))) 141 142 #define ISSAFE(ei, c) (!SHIFT7BIT((c)) && (ISSPECIAL((c)) || \ 143 (c < 0x80 && ei->cell[(c)] & (EI_DIRECT | EI_SPACE)))) 144 145 /* surrogate pair */ 146 #define SRG_BASE UINT32_C(0x10000) 147 #define HISRG_MIN UINT16_C(0xd800) 148 #define HISRG_MAX UINT16_C(0xdbff) 149 #define LOSRG_MIN UINT16_C(0xdc00) 150 #define LOSRG_MAX UINT16_C(0xdfff) 151 152 static int 153 _citrus_UTF7_mbtoutf16(_UTF7EncodingInfo * __restrict ei, 154 uint16_t * __restrict u16, char ** __restrict s, size_t n, 155 _UTF7State * __restrict psenc, size_t * __restrict nresult) 156 { 157 char *s0; 158 int done, i, len; 159 160 *nresult = 0; 161 s0 = *s; 162 163 for (i = 0, done = 0; done == 0; i++) { 164 if (i == psenc->chlen) { 165 if (n-- < 1) { 166 *nresult = (size_t)-2; 167 *s = s0; 168 return (0); 169 } 170 psenc->ch[psenc->chlen++] = *s0++; 171 } 172 if (SHIFT7BIT((int)psenc->ch[i])) 173 goto ilseq; 174 if (!psenc->mode) { 175 if (psenc->bits > 0 || psenc->cache > 0) 176 return (EINVAL); 177 if (psenc->ch[i] == BASE64_IN) 178 psenc->mode = 1; 179 else { 180 if (!ISDIRECT(ei, (int)psenc->ch[i])) 181 goto ilseq; 182 *u16 = (uint16_t)psenc->ch[i]; 183 done = 1; 184 continue; 185 } 186 } else { 187 if (psenc->ch[i] == BASE64_OUT && psenc->cache == 0) { 188 psenc->mode = 0; 189 *u16 = (uint16_t)BASE64_IN; 190 done = 1; 191 continue; 192 } 193 len = FINDLEN(ei, (int)psenc->ch[i]); 194 if (len < 0) { 195 if (psenc->bits >= BASE64_BIT) 196 return (EINVAL); 197 psenc->mode = 0; 198 psenc->bits = psenc->cache = 0; 199 if (psenc->ch[i] != BASE64_OUT) { 200 if (!ISDIRECT(ei, (int)psenc->ch[i])) 201 goto ilseq; 202 *u16 = (uint16_t)psenc->ch[i]; 203 done = 1; 204 } else { 205 psenc->chlen--; 206 i--; 207 } 208 } else { 209 psenc->cache = 210 (psenc->cache << BASE64_BIT) | len; 211 switch (psenc->bits) { 212 case 0: case 2: case 4: case 6: case 8: 213 psenc->bits += BASE64_BIT; 214 break; 215 case 10: case 12: case 14: 216 psenc->bits -= (UTF16_BIT - BASE64_BIT); 217 *u16 = (psenc->cache >> psenc->bits) & 218 UTF16_MAX; 219 done = 1; 220 break; 221 default: 222 return (EINVAL); 223 } 224 } 225 } 226 } 227 228 if (psenc->chlen > i) 229 return (EINVAL); 230 psenc->chlen = 0; 231 *nresult = (size_t)((*u16 == 0) ? 0 : s0 - *s); 232 *s = s0; 233 234 return (0); 235 236 ilseq: 237 *nresult = (size_t)-1; 238 return (EILSEQ); 239 } 240 241 static int 242 _citrus_UTF7_mbrtowc_priv(_UTF7EncodingInfo * __restrict ei, 243 wchar_t * __restrict pwc, char ** __restrict s, size_t n, 244 _UTF7State * __restrict psenc, size_t * __restrict nresult) 245 { 246 uint32_t u32; 247 uint16_t hi, lo; 248 size_t nr, siz; 249 int err; 250 251 if (*s == NULL) { 252 _citrus_UTF7_init_state(ei, psenc); 253 *nresult = (size_t)_ENCODING_IS_STATE_DEPENDENT; 254 return (0); 255 } 256 err = _citrus_UTF7_mbtoutf16(ei, &hi, s, n, psenc, &nr); 257 if (nr == (size_t)-1 || nr == (size_t)-2) { 258 *nresult = nr; 259 return (err); 260 } 261 if (err != 0) 262 return (err); 263 n -= nr; 264 siz = nr; 265 if (hi < HISRG_MIN || hi > HISRG_MAX) { 266 u32 = (uint32_t)hi; 267 goto done; 268 } 269 err = _citrus_UTF7_mbtoutf16(ei, &lo, s, n, psenc, &nr); 270 if (nr == (size_t)-1 || nr == (size_t)-2) { 271 psenc->chlen = 1; /* make get_state_desc return incomplete */ 272 *nresult = nr; 273 return (err); 274 } 275 if (err != 0) 276 return (err); 277 if (lo < LOSRG_MIN || lo > LOSRG_MAX) { 278 *nresult = (size_t)-1; 279 return (EILSEQ); 280 } 281 hi -= HISRG_MIN; 282 lo -= LOSRG_MIN; 283 u32 = (hi << 10 | lo) + SRG_BASE; 284 siz += nr; 285 done: 286 if (pwc != NULL) 287 *pwc = (wchar_t)u32; 288 if (u32 == (uint32_t)0) { 289 *nresult = (size_t)0; 290 _citrus_UTF7_init_state(ei, psenc); 291 } else { 292 *nresult = siz; 293 } 294 return (err); 295 } 296 297 static int 298 _citrus_UTF7_utf16tomb(_UTF7EncodingInfo * __restrict ei, 299 char * __restrict s, size_t n __unused, uint16_t u16, 300 _UTF7State * __restrict psenc, size_t * __restrict nresult) 301 { 302 int bits, i; 303 304 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT) 305 return (EINVAL); 306 307 if (ISSAFE(ei, u16)) { 308 if (psenc->mode) { 309 if (psenc->bits > 0) { 310 bits = BASE64_BIT - psenc->bits; 311 i = (psenc->cache << bits) & BASE64_MAX; 312 psenc->ch[psenc->chlen++] = base64[i]; 313 psenc->bits = psenc->cache = 0; 314 } 315 if (u16 == BASE64_OUT || FINDLEN(ei, u16) >= 0) 316 psenc->ch[psenc->chlen++] = BASE64_OUT; 317 psenc->mode = 0; 318 } 319 if (psenc->bits != 0) 320 return (EINVAL); 321 psenc->ch[psenc->chlen++] = (char)u16; 322 if (u16 == BASE64_IN) 323 psenc->ch[psenc->chlen++] = BASE64_OUT; 324 } else { 325 if (!psenc->mode) { 326 if (psenc->bits > 0) 327 return (EINVAL); 328 psenc->ch[psenc->chlen++] = BASE64_IN; 329 psenc->mode = 1; 330 } 331 psenc->cache = (psenc->cache << UTF16_BIT) | u16; 332 bits = UTF16_BIT + psenc->bits; 333 psenc->bits = bits % BASE64_BIT; 334 while ((bits -= BASE64_BIT) >= 0) { 335 i = (psenc->cache >> bits) & BASE64_MAX; 336 psenc->ch[psenc->chlen++] = base64[i]; 337 } 338 } 339 memcpy(s, psenc->ch, psenc->chlen); 340 *nresult = psenc->chlen; 341 psenc->chlen = 0; 342 343 return (0); 344 } 345 346 static int 347 _citrus_UTF7_wcrtomb_priv(_UTF7EncodingInfo * __restrict ei, 348 char * __restrict s, size_t n, wchar_t wchar, 349 _UTF7State * __restrict psenc, size_t * __restrict nresult) 350 { 351 uint32_t u32; 352 uint16_t u16[2]; 353 int err, i, len; 354 size_t nr, siz; 355 356 u32 = (uint32_t)wchar; 357 if (u32 <= UTF16_MAX) { 358 u16[0] = (uint16_t)u32; 359 len = 1; 360 } else if (u32 <= UTF32_MAX) { 361 u32 -= SRG_BASE; 362 u16[0] = (u32 >> 10) + HISRG_MIN; 363 u16[1] = ((uint16_t)(u32 & UINT32_C(0x3ff))) + LOSRG_MIN; 364 len = 2; 365 } else { 366 *nresult = (size_t)-1; 367 return (EILSEQ); 368 } 369 siz = 0; 370 for (i = 0; i < len; ++i) { 371 err = _citrus_UTF7_utf16tomb(ei, s, n, u16[i], psenc, &nr); 372 if (err != 0) 373 return (err); /* XXX: state has been modified */ 374 s += nr; 375 n -= nr; 376 siz += nr; 377 } 378 *nresult = siz; 379 380 return (0); 381 } 382 383 static int 384 /* ARGSUSED */ 385 _citrus_UTF7_put_state_reset(_UTF7EncodingInfo * __restrict ei __unused, 386 char * __restrict s, size_t n, _UTF7State * __restrict psenc, 387 size_t * __restrict nresult) 388 { 389 int bits, pos; 390 391 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT) 392 return (EINVAL); 393 394 if (psenc->mode) { 395 if (psenc->bits > 0) { 396 if (n-- < 1) 397 return (E2BIG); 398 bits = BASE64_BIT - psenc->bits; 399 pos = (psenc->cache << bits) & BASE64_MAX; 400 psenc->ch[psenc->chlen++] = base64[pos]; 401 psenc->ch[psenc->chlen++] = BASE64_OUT; 402 psenc->bits = psenc->cache = 0; 403 } 404 psenc->mode = 0; 405 } 406 if (psenc->bits != 0) 407 return (EINVAL); 408 if (n-- < 1) 409 return (E2BIG); 410 411 *nresult = (size_t)psenc->chlen; 412 if (psenc->chlen > 0) { 413 memcpy(s, psenc->ch, psenc->chlen); 414 psenc->chlen = 0; 415 } 416 417 return (0); 418 } 419 420 static __inline int 421 /*ARGSUSED*/ 422 _citrus_UTF7_stdenc_wctocs(_UTF7EncodingInfo * __restrict ei __unused, 423 _csid_t * __restrict csid, _index_t * __restrict idx, wchar_t wc) 424 { 425 426 *csid = 0; 427 *idx = (_index_t)wc; 428 429 return (0); 430 } 431 432 static __inline int 433 /*ARGSUSED*/ 434 _citrus_UTF7_stdenc_cstowc(_UTF7EncodingInfo * __restrict ei __unused, 435 wchar_t * __restrict wc, _csid_t csid, _index_t idx) 436 { 437 438 if (csid != 0) 439 return (EILSEQ); 440 *wc = (wchar_t)idx; 441 442 return (0); 443 } 444 445 static __inline int 446 /*ARGSUSED*/ 447 _citrus_UTF7_stdenc_get_state_desc_generic(_UTF7EncodingInfo * __restrict ei __unused, 448 _UTF7State * __restrict psenc, int * __restrict rstate) 449 { 450 451 *rstate = (psenc->chlen == 0) ? _STDENC_SDGEN_INITIAL : 452 _STDENC_SDGEN_INCOMPLETE_CHAR; 453 return (0); 454 } 455 456 static void 457 /*ARGSUSED*/ 458 _citrus_UTF7_encoding_module_uninit(_UTF7EncodingInfo *ei __unused) 459 { 460 461 /* ei seems to be unused */ 462 } 463 464 static int 465 /*ARGSUSED*/ 466 _citrus_UTF7_encoding_module_init(_UTF7EncodingInfo * __restrict ei, 467 const void * __restrict var __unused, size_t lenvar __unused) 468 { 469 const char *s; 470 471 memset(ei, 0, sizeof(*ei)); 472 473 #define FILL(str, flag) \ 474 do { \ 475 for (s = str; *s != '\0'; s++) \ 476 ei->cell[*s & 0x7f] |= flag; \ 477 } while (/*CONSTCOND*/0) 478 479 FILL(base64, (s - base64) + 1); 480 FILL(direct, EI_DIRECT); 481 FILL(option, EI_OPTION); 482 FILL(spaces, EI_SPACE); 483 484 return (0); 485 } 486 487 /* ---------------------------------------------------------------------- 488 * public interface for stdenc 489 */ 490 491 _CITRUS_STDENC_DECLS(UTF7); 492 _CITRUS_STDENC_DEF_OPS(UTF7); 493 494 #include "citrus_stdenc_template.h" 495