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 static __inline void 91 /*ARGSUSED*/ 92 _citrus_UTF7_pack_state(_UTF7EncodingInfo * __restrict ei __unused, 93 void *__restrict pspriv, const _UTF7State * __restrict s) 94 { 95 96 memcpy(pspriv, (const void *)s, sizeof(*s)); 97 } 98 99 static __inline void 100 /*ARGSUSED*/ 101 _citrus_UTF7_unpack_state(_UTF7EncodingInfo * __restrict ei __unused, 102 _UTF7State * __restrict s, const void * __restrict pspriv) 103 { 104 105 memcpy((void *)s, pspriv, sizeof(*s)); 106 } 107 108 static const char base64[] = 109 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 110 "abcdefghijklmnopqrstuvwxyz" 111 "0123456789+/"; 112 113 static const char direct[] = 114 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 115 "abcdefghijklmnopqrstuvwxyz" 116 "0123456789(),-./:?"; 117 118 static const char option[] = "!\"#$%&';<=>@[]^_`{|}"; 119 static const char spaces[] = " \t\r\n"; 120 121 #define BASE64_BIT 6 122 #define UTF16_BIT 16 123 124 #define BASE64_MAX 0x3f 125 #define UTF16_MAX UINT16_C(0xffff) 126 #define UTF32_MAX UINT32_C(0x10ffff) 127 128 #define BASE64_IN '+' 129 #define BASE64_OUT '-' 130 131 #define SHIFT7BIT(c) ((c) >> 7) 132 #define ISSPECIAL(c) ((c) == '\0' || (c) == BASE64_IN) 133 134 #define FINDLEN(ei, c) \ 135 (SHIFT7BIT((c)) ? -1 : (((ei)->cell[(c)] & EI_MASK) - 1)) 136 137 #define ISDIRECT(ei, c) (!SHIFT7BIT((c)) && (ISSPECIAL((c)) || \ 138 ei->cell[(c)] & (EI_DIRECT | EI_OPTION | EI_SPACE))) 139 140 #define ISSAFE(ei, c) (!SHIFT7BIT((c)) && (ISSPECIAL((c)) || \ 141 (c < 0x80 && ei->cell[(c)] & (EI_DIRECT | EI_SPACE)))) 142 143 /* surrogate pair */ 144 #define SRG_BASE UINT32_C(0x10000) 145 #define HISRG_MIN UINT16_C(0xd800) 146 #define HISRG_MAX UINT16_C(0xdbff) 147 #define LOSRG_MIN UINT16_C(0xdc00) 148 #define LOSRG_MAX UINT16_C(0xdfff) 149 150 static int 151 _citrus_UTF7_mbtoutf16(_UTF7EncodingInfo * __restrict ei, 152 uint16_t * __restrict u16, const char ** __restrict s, size_t n, 153 _UTF7State * __restrict psenc, size_t * __restrict nresult) 154 { 155 _UTF7State sv; 156 const char *s0; 157 int done, i, len; 158 159 s0 = *s; 160 sv = *psenc; 161 162 for (i = 0, done = 0; done == 0; i++) { 163 if (i == psenc->chlen) { 164 if (n-- < 1) { 165 *nresult = (size_t)-2; 166 *s = s0; 167 sv.chlen = psenc->chlen; 168 *psenc = sv; 169 return (0); 170 } 171 psenc->ch[psenc->chlen++] = *s0++; 172 } 173 if (SHIFT7BIT((int)psenc->ch[i])) 174 goto ilseq; 175 if (!psenc->mode) { 176 if (psenc->bits > 0 || psenc->cache > 0) 177 return (EINVAL); 178 if (psenc->ch[i] == BASE64_IN) 179 psenc->mode = 1; 180 else { 181 if (!ISDIRECT(ei, (int)psenc->ch[i])) 182 goto ilseq; 183 *u16 = (uint16_t)psenc->ch[i]; 184 done = 1; 185 continue; 186 } 187 } else { 188 if (psenc->ch[i] == BASE64_OUT && psenc->cache == 0) { 189 psenc->mode = 0; 190 *u16 = (uint16_t)BASE64_IN; 191 done = 1; 192 continue; 193 } 194 len = FINDLEN(ei, (int)psenc->ch[i]); 195 if (len < 0) { 196 if (psenc->bits >= BASE64_BIT) 197 return (EINVAL); 198 psenc->mode = 0; 199 psenc->bits = psenc->cache = 0; 200 if (psenc->ch[i] != BASE64_OUT) { 201 if (!ISDIRECT(ei, (int)psenc->ch[i])) 202 goto ilseq; 203 *u16 = (uint16_t)psenc->ch[i]; 204 done = 1; 205 } 206 } else { 207 psenc->cache = 208 (psenc->cache << BASE64_BIT) | len; 209 switch (psenc->bits) { 210 case 0: case 2: case 4: case 6: case 8: 211 psenc->bits += BASE64_BIT; 212 break; 213 case 10: case 12: case 14: 214 psenc->bits -= (UTF16_BIT - BASE64_BIT); 215 *u16 = (psenc->cache >> psenc->bits) & 216 UTF16_MAX; 217 done = 1; 218 break; 219 default: 220 return (EINVAL); 221 } 222 } 223 } 224 } 225 226 if (psenc->chlen > i) 227 return (EINVAL); 228 psenc->chlen = 0; 229 *nresult = (size_t)((*u16 == 0) ? 0 : s0 - *s); 230 *s = s0; 231 232 return (0); 233 234 ilseq: 235 *nresult = (size_t)-1; 236 return (EILSEQ); 237 } 238 239 static int 240 _citrus_UTF7_mbrtowc_priv(_UTF7EncodingInfo * __restrict ei, 241 wchar_t * __restrict pwc, const char ** __restrict s, size_t n, 242 _UTF7State * __restrict psenc, size_t * __restrict nresult) 243 { 244 const char *s0; 245 uint32_t u32; 246 uint16_t hi, lo; 247 size_t nr, siz; 248 int err; 249 250 if (*s == NULL) { 251 _citrus_UTF7_init_state(ei, psenc); 252 *nresult = (size_t)_ENCODING_IS_STATE_DEPENDENT; 253 return (0); 254 } 255 s0 = *s; 256 if (psenc->surrogate) { 257 hi = (psenc->cache >> 2) & 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, &s0, 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, &s0, 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 *s = s0; 290 if (pwc != NULL) 291 *pwc = (wchar_t)u32; 292 if (u32 == (uint32_t)0) { 293 *nresult = (size_t)0; 294 _citrus_UTF7_init_state(ei, psenc); 295 } else { 296 *nresult = siz; 297 psenc->surrogate = 0; 298 } 299 return (err); 300 } 301 302 static int 303 _citrus_UTF7_utf16tomb(_UTF7EncodingInfo * __restrict ei, 304 char * __restrict s, size_t n __unused, uint16_t u16, 305 _UTF7State * __restrict psenc, size_t * __restrict nresult) 306 { 307 int bits, i; 308 309 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT) 310 return (EINVAL); 311 312 if (ISSAFE(ei, u16)) { 313 if (psenc->mode) { 314 if (psenc->bits > 0) { 315 bits = BASE64_BIT - psenc->bits; 316 i = (psenc->cache << bits) & BASE64_MAX; 317 psenc->ch[psenc->chlen++] = base64[i]; 318 psenc->bits = psenc->cache = 0; 319 } 320 if (u16 == BASE64_OUT || FINDLEN(ei, u16) >= 0) 321 psenc->ch[psenc->chlen++] = BASE64_OUT; 322 psenc->mode = 0; 323 } 324 if (psenc->bits != 0) 325 return (EINVAL); 326 psenc->ch[psenc->chlen++] = (char)u16; 327 if (u16 == BASE64_IN) 328 psenc->ch[psenc->chlen++] = BASE64_OUT; 329 } else { 330 if (!psenc->mode) { 331 if (psenc->bits > 0) 332 return (EINVAL); 333 psenc->ch[psenc->chlen++] = BASE64_IN; 334 psenc->mode = 1; 335 } 336 psenc->cache = (psenc->cache << UTF16_BIT) | u16; 337 bits = UTF16_BIT + psenc->bits; 338 psenc->bits = bits % BASE64_BIT; 339 while ((bits -= BASE64_BIT) >= 0) { 340 i = (psenc->cache >> bits) & BASE64_MAX; 341 psenc->ch[psenc->chlen++] = base64[i]; 342 } 343 } 344 memcpy(s, psenc->ch, psenc->chlen); 345 *nresult = psenc->chlen; 346 psenc->chlen = 0; 347 348 return (0); 349 } 350 351 static int 352 _citrus_UTF7_wcrtomb_priv(_UTF7EncodingInfo * __restrict ei, 353 char * __restrict s, size_t n, wchar_t wchar, 354 _UTF7State * __restrict psenc, size_t * __restrict nresult) 355 { 356 uint32_t u32; 357 uint16_t u16[2]; 358 int err, i, len; 359 size_t nr, siz; 360 361 u32 = (uint32_t)wchar; 362 if (u32 <= UTF16_MAX) { 363 u16[0] = (uint16_t)u32; 364 len = 1; 365 } else if (u32 <= UTF32_MAX) { 366 u32 -= SRG_BASE; 367 u16[0] = (u32 >> 10) + HISRG_MIN; 368 u16[1] = ((uint16_t)(u32 & UINT32_C(0x3ff))) + LOSRG_MIN; 369 len = 2; 370 } else { 371 *nresult = (size_t)-1; 372 return (EILSEQ); 373 } 374 siz = 0; 375 for (i = 0; i < len; ++i) { 376 err = _citrus_UTF7_utf16tomb(ei, s, n, u16[i], psenc, &nr); 377 if (err != 0) 378 return (err); /* XXX: state has been modified */ 379 s += nr; 380 n -= nr; 381 siz += nr; 382 } 383 *nresult = siz; 384 385 return (0); 386 } 387 388 static int 389 /* ARGSUSED */ 390 _citrus_UTF7_put_state_reset(_UTF7EncodingInfo * __restrict ei __unused, 391 char * __restrict s, size_t n, _UTF7State * __restrict psenc, 392 size_t * __restrict nresult) 393 { 394 int bits, pos; 395 396 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT || psenc->surrogate) 397 return (EINVAL); 398 399 if (psenc->mode) { 400 if (psenc->bits > 0) { 401 if (n-- < 1) 402 return (E2BIG); 403 bits = BASE64_BIT - psenc->bits; 404 pos = (psenc->cache << bits) & BASE64_MAX; 405 psenc->ch[psenc->chlen++] = base64[pos]; 406 psenc->ch[psenc->chlen++] = BASE64_OUT; 407 psenc->bits = psenc->cache = 0; 408 } 409 psenc->mode = 0; 410 } 411 if (psenc->bits != 0) 412 return (EINVAL); 413 if (n-- < 1) 414 return (E2BIG); 415 416 *nresult = (size_t)psenc->chlen; 417 if (psenc->chlen > 0) { 418 memcpy(s, psenc->ch, psenc->chlen); 419 psenc->chlen = 0; 420 } 421 422 return (0); 423 } 424 425 static __inline int 426 /*ARGSUSED*/ 427 _citrus_UTF7_stdenc_wctocs(_UTF7EncodingInfo * __restrict ei __unused, 428 _csid_t * __restrict csid, _index_t * __restrict idx, wchar_t wc) 429 { 430 431 *csid = 0; 432 *idx = (_index_t)wc; 433 434 return (0); 435 } 436 437 static __inline int 438 /*ARGSUSED*/ 439 _citrus_UTF7_stdenc_cstowc(_UTF7EncodingInfo * __restrict ei __unused, 440 wchar_t * __restrict wc, _csid_t csid, _index_t idx) 441 { 442 443 if (csid != 0) 444 return (EILSEQ); 445 *wc = (wchar_t)idx; 446 447 return (0); 448 } 449 450 static __inline int 451 /*ARGSUSED*/ 452 _citrus_UTF7_stdenc_get_state_desc_generic(_UTF7EncodingInfo * __restrict ei __unused, 453 _UTF7State * __restrict psenc, int * __restrict rstate) 454 { 455 456 *rstate = (psenc->chlen == 0) ? _STDENC_SDGEN_INITIAL : 457 _STDENC_SDGEN_INCOMPLETE_CHAR; 458 return (0); 459 } 460 461 static void 462 /*ARGSUSED*/ 463 _citrus_UTF7_encoding_module_uninit(_UTF7EncodingInfo *ei __unused) 464 { 465 466 /* ei seems to be unused */ 467 } 468 469 static int 470 /*ARGSUSED*/ 471 _citrus_UTF7_encoding_module_init(_UTF7EncodingInfo * __restrict ei, 472 const void * __restrict var __unused, size_t lenvar __unused) 473 { 474 const char *s; 475 476 memset(ei, 0, sizeof(*ei)); 477 478 #define FILL(str, flag) \ 479 do { \ 480 for (s = str; *s != '\0'; s++) \ 481 ei->cell[*s & 0x7f] |= flag; \ 482 } while (/*CONSTCOND*/0) 483 484 FILL(base64, (s - base64) + 1); 485 FILL(direct, EI_DIRECT); 486 FILL(option, EI_OPTION); 487 FILL(spaces, EI_SPACE); 488 489 return (0); 490 } 491 492 /* ---------------------------------------------------------------------- 493 * public interface for stdenc 494 */ 495 496 _CITRUS_STDENC_DECLS(UTF7); 497 _CITRUS_STDENC_DEF_OPS(UTF7); 498 499 #include "citrus_stdenc_template.h" 500