1 /* $NetBSD: citrus_utf7.c,v 1.9 2026/06/30 23:18:22 riastradh Exp $ */ 2 3 /*- 4 * Copyright (c)2004, 2005 Citrus Project, 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26 * SUCH DAMAGE. 27 * 28 */ 29 30 #include <sys/cdefs.h> 31 #if defined(LIB_SCCS) && !defined(lint) 32 __RCSID("$NetBSD: citrus_utf7.c,v 1.9 2026/06/30 23:18:22 riastradh Exp $"); 33 #endif /* LIB_SCCS and not lint */ 34 35 #include <assert.h> 36 #include <errno.h> 37 #include <string.h> 38 #include <stdio.h> 39 #include <stdint.h> 40 #include <stdlib.h> 41 #include <limits.h> 42 #include <wchar.h> 43 44 #include "citrus_namespace.h" 45 #include "citrus_types.h" 46 #include "citrus_module.h" 47 #include "citrus_ctype.h" 48 #include "citrus_stdenc.h" 49 #include "citrus_utf7.h" 50 51 /* ---------------------------------------------------------------------- 52 * private stuffs used by templates 53 */ 54 55 typedef struct { 56 uint16_t cell[0x80]; 57 #define EI_MASK UINT16_C(0xff) 58 #define EI_DIRECT UINT16_C(0x100) 59 #define EI_OPTION UINT16_C(0x200) 60 #define EI_SPACE UINT16_C(0x400) 61 } _UTF7EncodingInfo; 62 63 typedef struct { 64 unsigned int 65 mode: 1, /* whether base64 mode */ 66 bits: 4, /* need to hold 0 - 15 */ 67 cache: 22, /* 22 = BASE64_BIT + UTF16_BIT */ 68 surrogate: 1; /* whether surrogate pair or not */ 69 unsigned char chlen; 70 char ch[4]; /* BASE64_IN, 3 * 6 = 18, most closed to UTF16_BIT */ 71 } _UTF7State; 72 73 __CTASSERT(__arraycount(((_UTF7State *)0)->ch) <= CHAR_MAX); 74 75 typedef struct { 76 _UTF7EncodingInfo ei; 77 struct { 78 /* for future multi-locale facility */ 79 _UTF7State s_mblen; 80 _UTF7State s_mbrlen; 81 _UTF7State s_mbrtowc; 82 _UTF7State s_mbtowc; 83 _UTF7State s_mbsrtowcs; 84 _UTF7State s_mbsnrtowcs; 85 _UTF7State s_wcrtomb; 86 _UTF7State s_wcsrtombs; 87 _UTF7State s_wcsnrtombs; 88 _UTF7State s_wctomb; 89 } states; 90 } _UTF7CTypeInfo; 91 92 #define _CEI_TO_EI(_cei_) (&(_cei_)->ei) 93 #define _CEI_TO_STATE(_cei_, _func_) (_cei_)->states.s_##_func_ 94 95 #define _FUNCNAME(m) _citrus_UTF7_##m 96 #define _ENCODING_INFO _UTF7EncodingInfo 97 #define _CTYPE_INFO _UTF7CTypeInfo 98 #define _ENCODING_STATE _UTF7State 99 #define _ENCODING_MB_CUR_MAX(_ei_) 4 100 #define _ENCODING_IS_STATE_DEPENDENT 1 101 #define _STATE_NEEDS_EXPLICIT_INIT(_ps_) 0 102 103 static __inline void 104 /*ARGSUSED*/ 105 _citrus_UTF7_init_state(_UTF7EncodingInfo * __restrict ei, 106 _UTF7State * __restrict s) 107 { 108 /* ei appears to be unused */ 109 _DIAGASSERT(s != NULL); 110 111 memset((void *)s, 0, sizeof(*s)); 112 } 113 114 static __inline void 115 /*ARGSUSED*/ 116 _citrus_UTF7_pack_state(_UTF7EncodingInfo * __restrict ei, 117 void *__restrict pspriv, const _UTF7State * __restrict s) 118 { 119 /* ei seem to be unused */ 120 _DIAGASSERT(pspriv != NULL); 121 _DIAGASSERT(s != NULL); 122 123 memcpy(pspriv, (const void *)s, sizeof(*s)); 124 } 125 126 static __inline void 127 /*ARGSUSED*/ 128 _citrus_UTF7_unpack_state(_UTF7EncodingInfo * __restrict ei, 129 _UTF7State * __restrict s, const void * __restrict pspriv) 130 { 131 /* ei seem to be unused */ 132 _DIAGASSERT(s != NULL); 133 _DIAGASSERT(pspriv != NULL); 134 135 memcpy((void *)s, pspriv, sizeof(*s)); 136 } 137 138 static const char base64[] = 139 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 140 "abcdefghijklmnopqrstuvwxyz" 141 "0123456789+/"; 142 143 static const char direct[] = 144 "ABCDEFGHIJKLMNOPQRSTUVWXYZ" 145 "abcdefghijklmnopqrstuvwxyz" 146 "0123456789(),-./:?"; 147 148 static const char option[] = "!\"#$%&';<=>@[]^_`{|}"; 149 static const char spaces[] = " \t\r\n"; 150 151 #define BASE64_BIT 6 152 #define UTF16_BIT 16 153 154 #define BASE64_MAX 0x3f 155 #define UTF16_MAX UINT16_C(0xffff) 156 #define UTF32_MAX UINT32_C(0x10ffff) 157 158 #define BASE64_IN '+' 159 #define BASE64_OUT '-' 160 161 #define SHIFT7BIT(c) ((c) >> 7) 162 #define ISSPECIAL(c) ((c) == '\0' || (c) == BASE64_IN) 163 164 #define FINDLEN(ei, c) \ 165 (SHIFT7BIT((c)) ? -1 : (((ei)->cell[(c)] & EI_MASK) - 1)) 166 167 #define ISDIRECT(ei, c) (!SHIFT7BIT((c)) && (ISSPECIAL((c)) || \ 168 ei->cell[(c)] & (EI_DIRECT | EI_OPTION | EI_SPACE))) 169 170 #define ISSAFE(ei, c) (!SHIFT7BIT((c)) && (ISSPECIAL((c)) || \ 171 (c < 0x80 && ei->cell[(c)] & (EI_DIRECT | EI_SPACE)))) 172 173 /* surrogate pair */ 174 #define SRG_BASE UINT32_C(0x10000) 175 #define HISRG_MIN UINT16_C(0xd800) 176 #define HISRG_MAX UINT16_C(0xdbff) 177 #define LOSRG_MIN UINT16_C(0xdc00) 178 #define LOSRG_MAX UINT16_C(0xdfff) 179 180 static int 181 _citrus_UTF7_mbtoutf16(_UTF7EncodingInfo * __restrict ei, 182 uint16_t * __restrict u16, const char ** __restrict s, size_t n, 183 _UTF7State * __restrict psenc, size_t * __restrict nresult) 184 { 185 _UTF7State sv; 186 const char *s0; 187 int i, done, len; 188 189 _DIAGASSERT(ei != NULL); 190 _DIAGASSERT(s != NULL && *s != NULL); 191 _DIAGASSERT(psenc != NULL); 192 193 s0 = *s; 194 sv = *psenc; 195 196 for (i = 0, done = 0; done == 0; i++) { 197 _DIAGASSERT(i < __arraycount(psenc->ch)); 198 _DIAGASSERT(i <= psenc->chlen); 199 _DIAGASSERT(psenc->chlen <= __arraycount(psenc->ch)); 200 if (i == psenc->chlen) { 201 if (n-- < 1) { 202 *nresult = (size_t)-2; 203 *s = s0; 204 sv.chlen = psenc->chlen; 205 *psenc = sv; 206 return 0; 207 } 208 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 209 psenc->ch[psenc->chlen++] = *s0++; 210 } 211 if (SHIFT7BIT((int)psenc->ch[i])) 212 goto ilseq; 213 if (!psenc->mode) { 214 if (psenc->bits > 0 || psenc->cache > 0) 215 return EINVAL; 216 if (psenc->ch[i] == BASE64_IN) { 217 psenc->mode = 1; 218 } else { 219 if (!ISDIRECT(ei, (int)psenc->ch[i])) 220 goto ilseq; 221 *u16 = (uint16_t)psenc->ch[i]; 222 done = 1; 223 continue; 224 } 225 } else { 226 if (psenc->ch[i] == BASE64_OUT && psenc->cache == 0) { 227 psenc->mode = 0; 228 *u16 = (uint16_t)BASE64_IN; 229 done = 1; 230 continue; 231 } 232 len = FINDLEN(ei, (int)psenc->ch[i]); 233 if (len < 0) { 234 if (psenc->bits >= BASE64_BIT) 235 return EINVAL; 236 psenc->mode = 0; 237 psenc->bits = psenc->cache = 0; 238 if (psenc->ch[i] != BASE64_OUT) { 239 if (!ISDIRECT(ei, (int)psenc->ch[i])) 240 goto ilseq; 241 *u16 = (uint16_t)psenc->ch[i]; 242 done = 1; 243 } 244 } else { 245 psenc->cache = 246 (psenc->cache << BASE64_BIT) | len; 247 switch (psenc->bits) { 248 case 0: case 2: case 4: case 6: case 8: 249 psenc->bits += BASE64_BIT; 250 break; 251 case 10: case 12: case 14: 252 psenc->bits -= (UTF16_BIT - BASE64_BIT); 253 *u16 = (psenc->cache >> psenc->bits) 254 & UTF16_MAX; 255 done = 1; 256 break; 257 default: 258 return EINVAL; 259 } 260 } 261 } 262 } 263 264 if (psenc->chlen > i) 265 return EINVAL; 266 psenc->chlen = 0; 267 *nresult = (size_t)((*u16 == 0) ? 0 : s0 - *s); 268 *s = s0; 269 270 return 0; 271 272 ilseq: 273 *nresult = (size_t)-1; 274 return EILSEQ; 275 } 276 277 static int 278 _citrus_UTF7_mbrtowc_priv(_UTF7EncodingInfo * __restrict ei, 279 wchar_t * __restrict pwc, const char ** __restrict s, size_t n, 280 _UTF7State * __restrict psenc, size_t * __restrict nresult) 281 { 282 const char *s0; 283 uint32_t u32; 284 uint16_t hi, lo; 285 size_t siz, nr; 286 int err; 287 288 _DIAGASSERT(ei != NULL); 289 /* pwc may be null */ 290 _DIAGASSERT(s != NULL); 291 _DIAGASSERT(psenc != NULL); 292 293 if (*s == NULL) { 294 _citrus_UTF7_init_state(ei, psenc); 295 *nresult = (size_t)_ENCODING_IS_STATE_DEPENDENT; 296 return 0; 297 } 298 s0 = *s; 299 if (psenc->surrogate) { 300 hi = (psenc->cache >> 2) & UTF16_MAX; 301 if (hi < HISRG_MIN || hi > HISRG_MAX) 302 return EINVAL; 303 siz = 0; 304 } else { 305 err = _citrus_UTF7_mbtoutf16(ei, &hi, &s0, n, psenc, &nr); 306 if (nr == (size_t)-1 || nr == (size_t)-2) { 307 *nresult = nr; 308 return err; 309 } 310 if (err != 0) 311 return err; 312 n -= nr; 313 siz = nr; 314 if (hi < HISRG_MIN || hi > HISRG_MAX) { 315 u32 = (uint32_t)hi; 316 goto done; 317 } 318 psenc->surrogate = 1; 319 } 320 err = _citrus_UTF7_mbtoutf16(ei, &lo, &s0, n, psenc, &nr); 321 if (nr == (size_t)-1 || nr == (size_t)-2) { 322 *nresult = nr; 323 return err; 324 } 325 if (err != 0) 326 return err; 327 hi -= HISRG_MIN; 328 lo -= LOSRG_MIN; 329 u32 = (hi << 10 | lo) + SRG_BASE; 330 siz += nr; 331 done: 332 *s = s0; 333 if (pwc != NULL) 334 *pwc = (wchar_t)u32; 335 if (u32 == (uint32_t)0) { 336 *nresult = (size_t)0; 337 _citrus_UTF7_init_state(ei, psenc); 338 } else { 339 *nresult = siz; 340 psenc->surrogate = 0; 341 } 342 return err; 343 } 344 345 static int 346 _citrus_UTF7_utf16tomb(_UTF7EncodingInfo * __restrict ei, 347 char * __restrict s, size_t n, uint16_t u16, 348 _UTF7State * __restrict psenc, size_t * __restrict nresult) 349 { 350 int bits, i; 351 352 _DIAGASSERT(ei != NULL); 353 _DIAGASSERT(psenc != NULL); 354 355 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT) 356 return EINVAL; 357 358 if (ISSAFE(ei, u16)) { 359 if (psenc->mode) { 360 if (psenc->bits > 0) { 361 bits = BASE64_BIT - psenc->bits; 362 i = (psenc->cache << bits) & BASE64_MAX; 363 _DIAGASSERT(psenc->chlen < 364 __arraycount(psenc->ch)); 365 psenc->ch[psenc->chlen++] = base64[i]; 366 psenc->bits = psenc->cache = 0; 367 } 368 if (u16 == BASE64_OUT || FINDLEN(ei, u16) >= 0) { 369 _DIAGASSERT(psenc->chlen < 370 __arraycount(psenc->ch)); 371 psenc->ch[psenc->chlen++] = BASE64_OUT; 372 } 373 psenc->mode = 0; 374 } 375 if (psenc->bits != 0) 376 return EINVAL; 377 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 378 psenc->ch[psenc->chlen++] = (char)u16; 379 if (u16 == BASE64_IN) { 380 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 381 psenc->ch[psenc->chlen++] = BASE64_OUT; 382 } 383 } else { 384 if (!psenc->mode) { 385 if (psenc->bits > 0) 386 return EINVAL; 387 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 388 psenc->ch[psenc->chlen++] = BASE64_IN; 389 psenc->mode = 1; 390 } 391 psenc->cache = (psenc->cache << UTF16_BIT) | u16; 392 bits = UTF16_BIT + psenc->bits; 393 psenc->bits = bits % BASE64_BIT; 394 while ((bits -= BASE64_BIT) >= 0) { 395 i = (psenc->cache >> bits) & BASE64_MAX; 396 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 397 psenc->ch[psenc->chlen++] = base64[i]; 398 } 399 } 400 if (n < (size_t)psenc->chlen) 401 return E2BIG; 402 _DIAGASSERT(psenc->chlen <= sizeof(psenc->ch)); 403 memcpy(s, psenc->ch, psenc->chlen); 404 *nresult = psenc->chlen; 405 psenc->chlen = 0; 406 407 return 0; 408 } 409 410 static int 411 _citrus_UTF7_wcrtomb_priv(_UTF7EncodingInfo * __restrict ei, 412 char * __restrict s, size_t n, wchar_t wchar, 413 _UTF7State * __restrict psenc, size_t * __restrict nresult) 414 { 415 uint32_t u32; 416 uint16_t u16[2]; 417 int err, len, i; 418 size_t siz, nr; 419 char buf[8], *bufp = buf; 420 _UTF7State psenc_save = *psenc; 421 422 _DIAGASSERT(ei != NULL); 423 _DIAGASSERT(s != NULL); 424 _DIAGASSERT(psenc != NULL); 425 _DIAGASSERT(nresult != NULL); 426 427 u32 = (uint32_t)wchar; 428 if (u32 <= UTF16_MAX) { 429 u16[0] = (uint16_t)u32; 430 len = 1; 431 } else if (u32 <= UTF32_MAX) { 432 u32 -= SRG_BASE; 433 u16[0] = (u32 >> 10) + HISRG_MIN; 434 u16[1] = ((uint16_t)(u32 & UINT32_C(0x3ff))) + LOSRG_MIN; 435 len = 2; 436 } else { 437 *nresult = (size_t)-1; 438 return EILSEQ; 439 } 440 siz = 0; 441 if (n > sizeof(buf)) /* paranoia */ 442 n = sizeof(buf); 443 for (i = 0; i < len; ++i) { 444 err = _citrus_UTF7_utf16tomb(ei, bufp, n, u16[i], psenc, &nr); 445 if (err != 0) { 446 *nresult = (size_t)-1; 447 *psenc = psenc_save; /* restore state */ 448 return err; 449 } 450 bufp += nr; 451 n -= nr; 452 siz += nr; 453 } 454 _DIAGASSERT(siz <= sizeof(buf)); 455 memcpy(s, buf, siz); 456 *nresult = siz; 457 458 return 0; 459 } 460 461 static int 462 /* ARGSUSED */ 463 _citrus_UTF7_put_state_reset(_UTF7EncodingInfo * __restrict ei, 464 char * __restrict s, size_t n, _UTF7State * __restrict psenc, 465 size_t * __restrict nresult) 466 { 467 int bits, pos; 468 469 _DIAGASSERT(ei != NULL); 470 _DIAGASSERT(s != NULL); 471 _DIAGASSERT(psenc != NULL); 472 _DIAGASSERT(nresult != NULL); 473 474 if (psenc->chlen != 0 || psenc->bits > BASE64_BIT || psenc->surrogate) 475 return EINVAL; 476 477 if (psenc->mode) { 478 if (psenc->bits > 0) { 479 if (n-- < 1) 480 return E2BIG; 481 bits = BASE64_BIT - psenc->bits; 482 pos = (psenc->cache << bits) & BASE64_MAX; 483 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 484 psenc->ch[psenc->chlen++] = base64[pos]; 485 _DIAGASSERT(psenc->chlen < __arraycount(psenc->ch)); 486 psenc->ch[psenc->chlen++] = BASE64_OUT; 487 psenc->bits = psenc->cache = 0; 488 } 489 psenc->mode = 0; 490 } 491 if (psenc->bits != 0) 492 return EINVAL; 493 if (n-- < 1) 494 return E2BIG; 495 496 _DIAGASSERT(n >= psenc->chlen); 497 *nresult = (size_t)psenc->chlen; 498 _DIAGASSERT(psenc->chlen <= sizeof(psenc->ch)); 499 if (psenc->chlen > 0) { 500 memcpy(s, psenc->ch, psenc->chlen); 501 psenc->chlen = 0; 502 } 503 504 return 0; 505 } 506 507 static __inline int 508 /*ARGSUSED*/ 509 _citrus_UTF7_stdenc_wctocs(_UTF7EncodingInfo * __restrict ei, 510 _csid_t * __restrict csid, 511 _index_t * __restrict idx, wchar_t wc) 512 { 513 /* ei seem to be unused */ 514 _DIAGASSERT(csid != NULL); 515 _DIAGASSERT(idx != NULL); 516 517 *csid = 0; 518 *idx = (_index_t)wc; 519 520 return 0; 521 } 522 523 static __inline int 524 /*ARGSUSED*/ 525 _citrus_UTF7_stdenc_cstowc(_UTF7EncodingInfo * __restrict ei, 526 wchar_t * __restrict wc, 527 _csid_t csid, _index_t idx) 528 { 529 /* ei seem to be unused */ 530 _DIAGASSERT(wc != NULL); 531 532 if (csid != 0) 533 return EILSEQ; 534 *wc = (wchar_t)idx; 535 536 return 0; 537 } 538 539 static __inline int 540 /*ARGSUSED*/ 541 _citrus_UTF7_stdenc_get_state_desc_generic(_UTF7EncodingInfo * __restrict ei, 542 _UTF7State * __restrict psenc, 543 int * __restrict rstate) 544 { 545 546 if (psenc->chlen == 0) 547 *rstate = _STDENC_SDGEN_INITIAL; 548 else 549 *rstate = _STDENC_SDGEN_INCOMPLETE_CHAR; 550 551 return 0; 552 } 553 554 static void 555 /*ARGSUSED*/ 556 _citrus_UTF7_encoding_module_uninit(_UTF7EncodingInfo *ei) 557 { 558 /* ei seems to be unused */ 559 } 560 561 static int 562 /*ARGSUSED*/ 563 _citrus_UTF7_encoding_module_init(_UTF7EncodingInfo * __restrict ei, 564 const void * __restrict var, size_t lenvar) 565 { 566 const char *s; 567 568 _DIAGASSERT(ei != NULL); 569 /* var may be null */ 570 571 memset(ei, 0, sizeof(*ei)); 572 573 #define FILL(str, flag) \ 574 do { \ 575 for (s = str; *s != '\0'; s++) \ 576 ei->cell[*s & 0x7f] |= flag; \ 577 } while (0) 578 579 FILL(base64, (s - base64) + 1); 580 FILL(direct, EI_DIRECT); 581 FILL(option, EI_OPTION); 582 FILL(spaces, EI_SPACE); 583 584 return 0; 585 } 586 587 /* ---------------------------------------------------------------------- 588 * public interface for ctype 589 */ 590 591 _CITRUS_CTYPE_DECLS(UTF7); 592 _CITRUS_CTYPE_DEF_OPS(UTF7); 593 594 #include "citrus_ctype_template.h" 595 596 /* ---------------------------------------------------------------------- 597 * public interface for stdenc 598 */ 599 600 _CITRUS_STDENC_DECLS(UTF7); 601 _CITRUS_STDENC_DEF_OPS(UTF7); 602 603 #include "citrus_stdenc_template.h" 604