1 /* $FreeBSD$ */ 2 /* $NetBSD: citrus_utf7.c,v 1.5 2006/08/23 12:57:24 tnozaki Exp $ */ 3 4 /*- 5 * Copyright (c)2004, 2005 Citrus Project, 6 * All rights reserved. 7 * 8 * Redistribution and use in source and binary forms, with or without 9 * modification, are permitted provided that the following conditions 10 * are met: 11 * 1. Redistributions of source code must retain the above copyright 12 * notice, this list of conditions and the following disclaimer. 13 * 2. Redistributions in binary form must reproduce the above copyright 14 * notice, this list of conditions and the following disclaimer in the 15 * documentation and/or other materials provided with the distribution. 16 * 17 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 18 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 21 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27 * SUCH DAMAGE. 28 * 29 */ 30 31 #include <sys/cdefs.h> 32 33 #include <assert.h> 34 #include <errno.h> 35 #include <limits.h> 36 #include <stdio.h> 37 #include <stdint.h> 38 #include <stdlib.h> 39 #include <string.h> 40 #include <wchar.h> 41 42 #include "citrus_namespace.h" 43 #include "citrus_types.h" 44 #include "citrus_module.h" 45 #include "citrus_stdenc.h" 46 #include "citrus_utf7.h" 47 48 /* ---------------------------------------------------------------------- 49 * private stuffs used by templates 50 */ 51 52 #define EI_MASK UINT16_C(0xff) 53 #define EI_DIRECT UINT16_C(0x100) 54 #define EI_OPTION UINT16_C(0x200) 55 #define EI_SPACE UINT16_C(0x400) 56 57 typedef struct { 58 uint16_t cell[0x80]; 59 } _UTF7EncodingInfo; 60 61 typedef struct { 62 unsigned int 63 mode: 1, /* whether base64 mode */ 64 bits: 4, /* need to hold 0 - 15 */ 65 cache: 22, /* 22 = BASE64_BIT + UTF16_BIT */ 66 surrogate: 1; /* whether surrogate pair or not */ 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 if (psenc->surrogate) { 257 hi = (psenc->cache >> psenc->bits) & UTF16_MAX; 258 if (hi < HISRG_MIN || hi > HISRG_MAX) 259 return (EINVAL); 260 siz = 0; 261 } else { 262 err = _citrus_UTF7_mbtoutf16(ei, &hi, s, n, psenc, &nr); 263 if (nr == (size_t)-1 || nr == (size_t)-2) { 264 *nresult = nr; 265 return (err); 266 } 267 if (err != 0) 268 return (err); 269 n -= nr; 270 siz = nr; 271 if (hi < HISRG_MIN || hi > HISRG_MAX) { 272 u32 = (uint32_t)hi; 273 goto done; 274 } 275 psenc->surrogate = 1; 276 } 277 err = _citrus_UTF7_mbtoutf16(ei, &lo, s, n, psenc, &nr); 278 if (nr == (size_t)-1 || nr == (size_t)-2) { 279 *nresult = nr; 280 return (err); 281 } 282 if (err != 0) 283 return (err); 284 hi -= HISRG_MIN; 285 lo -= LOSRG_MIN; 286 u32 = (hi << 10 | lo) + SRG_BASE; 287 siz += nr; 288 done: 289 if (pwc != NULL) 290 *pwc = (wchar_t)u32; 291 if (u32 == (uint32_t)0) { 292 *nresult = (size_t)0; 293 _citrus_UTF7_init_state(ei, psenc); 294 } else { 295 *nresult = siz; 296 psenc->surrogate = 0; 297 } 298 return (err); 299 } 300 301 static int 302 _citrus_UTF7_utf16tomb(_UTF7EncodingInfo * __restrict ei, 303 char * __restrict s, size_t n __unused, uint16_t u16, 304 _UTF7State * __restrict psenc, size_t * __restrict nresult) 305 { 306 int bits, i; 307 308 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT) 309 return (EINVAL); 310 311 if (ISSAFE(ei, u16)) { 312 if (psenc->mode) { 313 if (psenc->bits > 0) { 314 bits = BASE64_BIT - psenc->bits; 315 i = (psenc->cache << bits) & BASE64_MAX; 316 psenc->ch[psenc->chlen++] = base64[i]; 317 psenc->bits = psenc->cache = 0; 318 } 319 if (u16 == BASE64_OUT || FINDLEN(ei, u16) >= 0) 320 psenc->ch[psenc->chlen++] = BASE64_OUT; 321 psenc->mode = 0; 322 } 323 if (psenc->bits != 0) 324 return (EINVAL); 325 psenc->ch[psenc->chlen++] = (char)u16; 326 if (u16 == BASE64_IN) 327 psenc->ch[psenc->chlen++] = BASE64_OUT; 328 } else { 329 if (!psenc->mode) { 330 if (psenc->bits > 0) 331 return (EINVAL); 332 psenc->ch[psenc->chlen++] = BASE64_IN; 333 psenc->mode = 1; 334 } 335 psenc->cache = (psenc->cache << UTF16_BIT) | u16; 336 bits = UTF16_BIT + psenc->bits; 337 psenc->bits = bits % BASE64_BIT; 338 while ((bits -= BASE64_BIT) >= 0) { 339 i = (psenc->cache >> bits) & BASE64_MAX; 340 psenc->ch[psenc->chlen++] = base64[i]; 341 } 342 } 343 memcpy(s, psenc->ch, psenc->chlen); 344 *nresult = psenc->chlen; 345 psenc->chlen = 0; 346 347 return (0); 348 } 349 350 static int 351 _citrus_UTF7_wcrtomb_priv(_UTF7EncodingInfo * __restrict ei, 352 char * __restrict s, size_t n, wchar_t wchar, 353 _UTF7State * __restrict psenc, size_t * __restrict nresult) 354 { 355 uint32_t u32; 356 uint16_t u16[2]; 357 int err, i, len; 358 size_t nr, siz; 359 360 u32 = (uint32_t)wchar; 361 if (u32 <= UTF16_MAX) { 362 u16[0] = (uint16_t)u32; 363 len = 1; 364 } else if (u32 <= UTF32_MAX) { 365 u32 -= SRG_BASE; 366 u16[0] = (u32 >> 10) + HISRG_MIN; 367 u16[1] = ((uint16_t)(u32 & UINT32_C(0x3ff))) + LOSRG_MIN; 368 len = 2; 369 } else { 370 *nresult = (size_t)-1; 371 return (EILSEQ); 372 } 373 siz = 0; 374 for (i = 0; i < len; ++i) { 375 err = _citrus_UTF7_utf16tomb(ei, s, n, u16[i], psenc, &nr); 376 if (err != 0) 377 return (err); /* XXX: state has been modified */ 378 s += nr; 379 n -= nr; 380 siz += nr; 381 } 382 *nresult = siz; 383 384 return (0); 385 } 386 387 static int 388 /* ARGSUSED */ 389 _citrus_UTF7_put_state_reset(_UTF7EncodingInfo * __restrict ei __unused, 390 char * __restrict s, size_t n, _UTF7State * __restrict psenc, 391 size_t * __restrict nresult) 392 { 393 int bits, pos; 394 395 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT || psenc->surrogate) 396 return (EINVAL); 397 398 if (psenc->mode) { 399 if (psenc->bits > 0) { 400 if (n-- < 1) 401 return (E2BIG); 402 bits = BASE64_BIT - psenc->bits; 403 pos = (psenc->cache << bits) & BASE64_MAX; 404 psenc->ch[psenc->chlen++] = base64[pos]; 405 psenc->ch[psenc->chlen++] = BASE64_OUT; 406 psenc->bits = psenc->cache = 0; 407 } 408 psenc->mode = 0; 409 } 410 if (psenc->bits != 0) 411 return (EINVAL); 412 if (n-- < 1) 413 return (E2BIG); 414 415 *nresult = (size_t)psenc->chlen; 416 if (psenc->chlen > 0) { 417 memcpy(s, psenc->ch, psenc->chlen); 418 psenc->chlen = 0; 419 } 420 421 return (0); 422 } 423 424 static __inline int 425 /*ARGSUSED*/ 426 _citrus_UTF7_stdenc_wctocs(_UTF7EncodingInfo * __restrict ei __unused, 427 _csid_t * __restrict csid, _index_t * __restrict idx, wchar_t wc) 428 { 429 430 *csid = 0; 431 *idx = (_index_t)wc; 432 433 return (0); 434 } 435 436 static __inline int 437 /*ARGSUSED*/ 438 _citrus_UTF7_stdenc_cstowc(_UTF7EncodingInfo * __restrict ei __unused, 439 wchar_t * __restrict wc, _csid_t csid, _index_t idx) 440 { 441 442 if (csid != 0) 443 return (EILSEQ); 444 *wc = (wchar_t)idx; 445 446 return (0); 447 } 448 449 static __inline int 450 /*ARGSUSED*/ 451 _citrus_UTF7_stdenc_get_state_desc_generic(_UTF7EncodingInfo * __restrict ei __unused, 452 _UTF7State * __restrict psenc, int * __restrict rstate) 453 { 454 455 *rstate = (psenc->chlen == 0) ? _STDENC_SDGEN_INITIAL : 456 _STDENC_SDGEN_INCOMPLETE_CHAR; 457 return (0); 458 } 459 460 static void 461 /*ARGSUSED*/ 462 _citrus_UTF7_encoding_module_uninit(_UTF7EncodingInfo *ei __unused) 463 { 464 465 /* ei seems to be unused */ 466 } 467 468 static int 469 /*ARGSUSED*/ 470 _citrus_UTF7_encoding_module_init(_UTF7EncodingInfo * __restrict ei, 471 const void * __restrict var __unused, size_t lenvar __unused) 472 { 473 const char *s; 474 475 memset(ei, 0, sizeof(*ei)); 476 477 #define FILL(str, flag) \ 478 do { \ 479 for (s = str; *s != '\0'; s++) \ 480 ei->cell[*s & 0x7f] |= flag; \ 481 } while (/*CONSTCOND*/0) 482 483 FILL(base64, (s - base64) + 1); 484 FILL(direct, EI_DIRECT); 485 FILL(option, EI_OPTION); 486 FILL(spaces, EI_SPACE); 487 488 return (0); 489 } 490 491 /* ---------------------------------------------------------------------- 492 * public interface for stdenc 493 */ 494 495 _CITRUS_STDENC_DECLS(UTF7); 496 _CITRUS_STDENC_DEF_OPS(UTF7); 497 498 #include "citrus_stdenc_template.h" 499