1219019Sgabor/* $FreeBSD$ */ 2219019Sgabor/* $NetBSD: citrus_big5.c,v 1.12 2008/06/14 16:01:07 tnozaki Exp $ */ 3219019Sgabor 4219019Sgabor/*- 5219019Sgabor * Copyright (c)2002, 2006 Citrus Project, 6219019Sgabor * All rights reserved. 7219019Sgabor * 8219019Sgabor * Redistribution and use in source and binary forms, with or without 9219019Sgabor * modification, are permitted provided that the following conditions 10219019Sgabor * are met: 11219019Sgabor * 1. Redistributions of source code must retain the above copyright 12219019Sgabor * notice, this list of conditions and the following disclaimer. 13219019Sgabor * 2. Redistributions in binary form must reproduce the above copyright 14219019Sgabor * notice, this list of conditions and the following disclaimer in the 15219019Sgabor * documentation and/or other materials provided with the distribution. 16219019Sgabor * 17219019Sgabor * THIS SOFTWARE IS PROVIDED BY THE AUTHOR AND CONTRIBUTORS ``AS IS'' AND 18219019Sgabor * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 19219019Sgabor * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 20219019Sgabor * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHOR OR CONTRIBUTORS BE LIABLE 21219019Sgabor * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 22219019Sgabor * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 23219019Sgabor * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 24219019Sgabor * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 25219019Sgabor * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 26219019Sgabor * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 27219019Sgabor * SUCH DAMAGE. 28219019Sgabor */ 29219019Sgabor 30219019Sgabor/*- 31219019Sgabor * Copyright (c) 1993 32219019Sgabor * The Regents of the University of California. All rights reserved. 33219019Sgabor * 34219019Sgabor * This code is derived from software contributed to Berkeley by 35219019Sgabor * Paul Borman at Krystal Technologies. 36219019Sgabor * 37219019Sgabor * Redistribution and use in source and binary forms, with or without 38219019Sgabor * modification, are permitted provided that the following conditions 39219019Sgabor * are met: 40219019Sgabor * 1. Redistributions of source code must retain the above copyright 41219019Sgabor * notice, this list of conditions and the following disclaimer. 42219019Sgabor * 2. Redistributions in binary form must reproduce the above copyright 43219019Sgabor * notice, this list of conditions and the following disclaimer in the 44219019Sgabor * documentation and/or other materials provided with the distribution. 45219019Sgabor * 3. Neither the name of the University nor the names of its contributors 46219019Sgabor * may be used to endorse or promote products derived from this software 47219019Sgabor * without specific prior written permission. 48219019Sgabor * 49219019Sgabor * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 50219019Sgabor * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 51219019Sgabor * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 52219019Sgabor * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 53219019Sgabor * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 54219019Sgabor * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 55219019Sgabor * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 56219019Sgabor * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 57219019Sgabor * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 58219019Sgabor * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 59219019Sgabor * SUCH DAMAGE. 60219019Sgabor */ 61219019Sgabor 62219019Sgabor#include <sys/cdefs.h> 63219019Sgabor#include <sys/queue.h> 64219019Sgabor#include <sys/types.h> 65219019Sgabor 66219019Sgabor#include <assert.h> 67219019Sgabor#include <errno.h> 68219019Sgabor#include <limits.h> 69219019Sgabor#include <stddef.h> 70219019Sgabor#include <stdint.h> 71219019Sgabor#include <stdio.h> 72219019Sgabor#include <stdlib.h> 73219019Sgabor#include <string.h> 74219019Sgabor#include <wchar.h> 75219019Sgabor 76219019Sgabor#include "citrus_namespace.h" 77219019Sgabor#include "citrus_prop.h" 78219019Sgabor#include "citrus_types.h" 79219019Sgabor#include "citrus_bcs.h" 80219019Sgabor#include "citrus_module.h" 81219019Sgabor#include "citrus_stdenc.h" 82219019Sgabor#include "citrus_big5.h" 83219019Sgabor 84219019Sgabor/* ---------------------------------------------------------------------- 85219019Sgabor * private stuffs used by templates 86219019Sgabor */ 87219019Sgabor 88219019Sgabortypedef struct { 89219019Sgabor int chlen; 90219019Sgabor char ch[2]; 91219019Sgabor} _BIG5State; 92219019Sgabor 93219019Sgabortypedef struct _BIG5Exclude { 94219019Sgabor TAILQ_ENTRY(_BIG5Exclude) entry; 95219019Sgabor wint_t end; 96219019Sgabor wint_t start; 97219019Sgabor} _BIG5Exclude; 98219019Sgabor 99219019Sgabortypedef TAILQ_HEAD(_BIG5ExcludeList, _BIG5Exclude) _BIG5ExcludeList; 100219019Sgabor 101219019Sgabortypedef struct { 102219019Sgabor _BIG5ExcludeList excludes; 103219019Sgabor int cell[0x100]; 104219019Sgabor} _BIG5EncodingInfo; 105219019Sgabor 106219019Sgabor#define _CEI_TO_EI(_cei_) (&(_cei_)->ei) 107219019Sgabor#define _CEI_TO_STATE(_cei_, _func_) (_cei_)->states.s_##_func_ 108219019Sgabor 109219019Sgabor#define _FUNCNAME(m) _citrus_BIG5_##m 110219019Sgabor#define _ENCODING_INFO _BIG5EncodingInfo 111219019Sgabor#define _ENCODING_STATE _BIG5State 112219019Sgabor#define _ENCODING_MB_CUR_MAX(_ei_) 2 113219019Sgabor#define _ENCODING_IS_STATE_DEPENDENT 0 114219019Sgabor#define _STATE_NEEDS_EXPLICIT_INIT(_ps_) 0 115219019Sgabor 116219019Sgabor 117219019Sgaborstatic __inline void 118219019Sgabor/*ARGSUSED*/ 119219019Sgabor_citrus_BIG5_init_state(_BIG5EncodingInfo * __restrict ei __unused, 120219019Sgabor _BIG5State * __restrict s) 121219019Sgabor{ 122219019Sgabor 123219019Sgabor memset(s, 0, sizeof(*s)); 124219019Sgabor} 125219019Sgabor 126260264Sdim#if 0 127219019Sgaborstatic __inline void 128219019Sgabor/*ARGSUSED*/ 129219019Sgabor_citrus_BIG5_pack_state(_BIG5EncodingInfo * __restrict ei __unused, 130219019Sgabor void * __restrict pspriv, 131219019Sgabor const _BIG5State * __restrict s) 132219019Sgabor{ 133219019Sgabor 134219019Sgabor memcpy(pspriv, (const void *)s, sizeof(*s)); 135219019Sgabor} 136219019Sgabor 137219019Sgaborstatic __inline void 138219019Sgabor/*ARGSUSED*/ 139219019Sgabor_citrus_BIG5_unpack_state(_BIG5EncodingInfo * __restrict ei __unused, 140219019Sgabor _BIG5State * __restrict s, 141219019Sgabor const void * __restrict pspriv) 142219019Sgabor{ 143219019Sgabor 144219019Sgabor memcpy((void *)s, pspriv, sizeof(*s)); 145219019Sgabor} 146260264Sdim#endif 147219019Sgabor 148219019Sgaborstatic __inline int 149219019Sgabor_citrus_BIG5_check(_BIG5EncodingInfo *ei, unsigned int c) 150219019Sgabor{ 151219019Sgabor 152219019Sgabor return ((ei->cell[c & 0xFF] & 0x1) ? 2 : 1); 153219019Sgabor} 154219019Sgabor 155219019Sgaborstatic __inline int 156219019Sgabor_citrus_BIG5_check2(_BIG5EncodingInfo *ei, unsigned int c) 157219019Sgabor{ 158219019Sgabor 159219019Sgabor return ((ei->cell[c & 0xFF] & 0x2) ? 1 : 0); 160219019Sgabor} 161219019Sgabor 162219019Sgaborstatic __inline int 163219019Sgabor_citrus_BIG5_check_excludes(_BIG5EncodingInfo *ei, wint_t c) 164219019Sgabor{ 165219019Sgabor _BIG5Exclude *exclude; 166219019Sgabor 167219019Sgabor TAILQ_FOREACH(exclude, &ei->excludes, entry) { 168219019Sgabor if (c >= exclude->start && c <= exclude->end) 169219019Sgabor return (EILSEQ); 170219019Sgabor } 171219019Sgabor return (0); 172219019Sgabor} 173219019Sgabor 174219019Sgaborstatic int 175219019Sgabor_citrus_BIG5_fill_rowcol(void ** __restrict ctx, const char * __restrict s, 176219019Sgabor uint64_t start, uint64_t end) 177219019Sgabor{ 178219019Sgabor _BIG5EncodingInfo *ei; 179219019Sgabor uint64_t n; 180219019Sgabor int i; 181219019Sgabor 182219019Sgabor if (start > 0xFF || end > 0xFF) 183219019Sgabor return (EINVAL); 184219019Sgabor ei = (_BIG5EncodingInfo *)ctx; 185219019Sgabor i = strcmp("row", s) ? 1 : 0; 186219019Sgabor i = 1 << i; 187219019Sgabor for (n = start; n <= end; ++n) 188219019Sgabor ei->cell[n & 0xFF] |= i; 189219019Sgabor return (0); 190219019Sgabor} 191219019Sgabor 192219019Sgaborstatic int 193219019Sgabor/*ARGSUSED*/ 194219019Sgabor_citrus_BIG5_fill_excludes(void ** __restrict ctx, 195219019Sgabor const char * __restrict s __unused, uint64_t start, uint64_t end) 196219019Sgabor{ 197219019Sgabor _BIG5EncodingInfo *ei; 198219019Sgabor _BIG5Exclude *exclude; 199219019Sgabor 200219019Sgabor if (start > 0xFFFF || end > 0xFFFF) 201219019Sgabor return (EINVAL); 202219019Sgabor ei = (_BIG5EncodingInfo *)ctx; 203219019Sgabor exclude = TAILQ_LAST(&ei->excludes, _BIG5ExcludeList); 204219019Sgabor if (exclude != NULL && (wint_t)start <= exclude->end) 205219019Sgabor return (EINVAL); 206219019Sgabor exclude = (void *)malloc(sizeof(*exclude)); 207219019Sgabor if (exclude == NULL) 208219019Sgabor return (ENOMEM); 209219019Sgabor exclude->start = (wint_t)start; 210219019Sgabor exclude->end = (wint_t)end; 211219019Sgabor TAILQ_INSERT_TAIL(&ei->excludes, exclude, entry); 212219019Sgabor 213219019Sgabor return (0); 214219019Sgabor} 215219019Sgabor 216219019Sgaborstatic const _citrus_prop_hint_t root_hints[] = { 217219019Sgabor _CITRUS_PROP_HINT_NUM("row", &_citrus_BIG5_fill_rowcol), 218219019Sgabor _CITRUS_PROP_HINT_NUM("col", &_citrus_BIG5_fill_rowcol), 219219019Sgabor _CITRUS_PROP_HINT_NUM("excludes", &_citrus_BIG5_fill_excludes), 220219019Sgabor _CITRUS_PROP_HINT_END 221219019Sgabor}; 222219019Sgabor 223219019Sgaborstatic void 224219019Sgabor/*ARGSUSED*/ 225219019Sgabor_citrus_BIG5_encoding_module_uninit(_BIG5EncodingInfo *ei) 226219019Sgabor{ 227219019Sgabor _BIG5Exclude *exclude; 228219019Sgabor 229219019Sgabor while ((exclude = TAILQ_FIRST(&ei->excludes)) != NULL) { 230219019Sgabor TAILQ_REMOVE(&ei->excludes, exclude, entry); 231219019Sgabor free(exclude); 232219019Sgabor } 233219019Sgabor} 234219019Sgabor 235219019Sgaborstatic int 236219019Sgabor/*ARGSUSED*/ 237219019Sgabor_citrus_BIG5_encoding_module_init(_BIG5EncodingInfo * __restrict ei, 238219019Sgabor const void * __restrict var, size_t lenvar) 239219019Sgabor{ 240219019Sgabor void *ctx = (void *)ei; 241219019Sgabor const char *s; 242219019Sgabor int err; 243219019Sgabor 244219019Sgabor memset((void *)ei, 0, sizeof(*ei)); 245219019Sgabor TAILQ_INIT(&ei->excludes); 246219019Sgabor 247219019Sgabor if (lenvar > 0 && var != NULL) { 248219019Sgabor s = _bcs_skip_ws_len((const char *)var, &lenvar); 249219019Sgabor if (lenvar > 0 && *s != '\0') { 250219019Sgabor err = _citrus_prop_parse_variable( 251219019Sgabor root_hints, (void *)ei, s, lenvar); 252219019Sgabor if (err == 0) 253219019Sgabor return (0); 254219019Sgabor 255219019Sgabor _citrus_BIG5_encoding_module_uninit(ei); 256219019Sgabor memset((void *)ei, 0, sizeof(*ei)); 257219019Sgabor TAILQ_INIT(&ei->excludes); 258219019Sgabor } 259219019Sgabor } 260219019Sgabor 261219019Sgabor /* fallback Big5-1984, for backward compatibility. */ 262219019Sgabor _citrus_BIG5_fill_rowcol((void **)&ctx, "row", 0xA1, 0xFE); 263219019Sgabor _citrus_BIG5_fill_rowcol((void **)&ctx, "col", 0x40, 0x7E); 264219019Sgabor _citrus_BIG5_fill_rowcol((void **)&ctx, "col", 0xA1, 0xFE); 265219019Sgabor 266219019Sgabor return (0); 267219019Sgabor} 268219019Sgabor 269219019Sgaborstatic int 270219019Sgabor/*ARGSUSED*/ 271219019Sgabor_citrus_BIG5_mbrtowc_priv(_BIG5EncodingInfo * __restrict ei, 272219019Sgabor wchar_t * __restrict pwc, 273219019Sgabor char ** __restrict s, size_t n, 274219019Sgabor _BIG5State * __restrict psenc, 275219019Sgabor size_t * __restrict nresult) 276219019Sgabor{ 277219019Sgabor wchar_t wchar; 278219019Sgabor char *s0; 279219019Sgabor int c, chlenbak; 280219019Sgabor 281219019Sgabor s0 = *s; 282219019Sgabor 283219019Sgabor if (s0 == NULL) { 284219019Sgabor _citrus_BIG5_init_state(ei, psenc); 285219019Sgabor *nresult = 0; 286219019Sgabor return (0); 287219019Sgabor } 288219019Sgabor 289219019Sgabor chlenbak = psenc->chlen; 290219019Sgabor 291219019Sgabor /* make sure we have the first byte in the buffer */ 292219019Sgabor switch (psenc->chlen) { 293219019Sgabor case 0: 294219019Sgabor if (n < 1) 295219019Sgabor goto restart; 296219019Sgabor psenc->ch[0] = *s0++; 297219019Sgabor psenc->chlen = 1; 298219019Sgabor n--; 299219019Sgabor break; 300219019Sgabor case 1: 301219019Sgabor break; 302219019Sgabor default: 303219019Sgabor /* illegal state */ 304219019Sgabor goto ilseq; 305219019Sgabor } 306219019Sgabor 307219019Sgabor c = _citrus_BIG5_check(ei, psenc->ch[0] & 0xff); 308219019Sgabor if (c == 0) 309219019Sgabor goto ilseq; 310219019Sgabor while (psenc->chlen < c) { 311219019Sgabor if (n < 1) { 312219019Sgabor goto restart; 313219019Sgabor } 314219019Sgabor psenc->ch[psenc->chlen] = *s0++; 315219019Sgabor psenc->chlen++; 316219019Sgabor n--; 317219019Sgabor } 318219019Sgabor 319219019Sgabor switch (c) { 320219019Sgabor case 1: 321219019Sgabor wchar = psenc->ch[0] & 0xff; 322219019Sgabor break; 323219019Sgabor case 2: 324219019Sgabor if (!_citrus_BIG5_check2(ei, psenc->ch[1] & 0xff)) 325219019Sgabor goto ilseq; 326219019Sgabor wchar = ((psenc->ch[0] & 0xff) << 8) | (psenc->ch[1] & 0xff); 327219019Sgabor break; 328219019Sgabor default: 329219019Sgabor /* illegal state */ 330219019Sgabor goto ilseq; 331219019Sgabor } 332219019Sgabor 333219019Sgabor if (_citrus_BIG5_check_excludes(ei, (wint_t)wchar) != 0) 334219019Sgabor goto ilseq; 335219019Sgabor 336219019Sgabor *s = s0; 337219019Sgabor psenc->chlen = 0; 338219019Sgabor if (pwc) 339219019Sgabor *pwc = wchar; 340219019Sgabor *nresult = wchar ? c - chlenbak : 0; 341219019Sgabor 342219019Sgabor return (0); 343219019Sgabor 344219019Sgaborilseq: 345219019Sgabor psenc->chlen = 0; 346219019Sgabor *nresult = (size_t)-1; 347219019Sgabor return (EILSEQ); 348219019Sgabor 349219019Sgaborrestart: 350219019Sgabor *s = s0; 351219019Sgabor *nresult = (size_t)-2; 352219019Sgabor return (0); 353219019Sgabor} 354219019Sgabor 355219019Sgaborstatic int 356219019Sgabor/*ARGSUSED*/ 357219019Sgabor_citrus_BIG5_wcrtomb_priv(_BIG5EncodingInfo * __restrict ei, 358219019Sgabor char * __restrict s, 359219019Sgabor size_t n, wchar_t wc, _BIG5State * __restrict psenc __unused, 360219019Sgabor size_t * __restrict nresult) 361219019Sgabor{ 362219019Sgabor unsigned char l; 363219019Sgabor int ret; 364219019Sgabor 365219019Sgabor /* check invalid sequence */ 366219019Sgabor if (wc & ~0xffff || 367219019Sgabor _citrus_BIG5_check_excludes(ei, (wint_t)wc) != 0) { 368219019Sgabor ret = EILSEQ; 369219019Sgabor goto err; 370219019Sgabor } 371219019Sgabor 372219019Sgabor if (wc & 0x8000) { 373219019Sgabor if (_citrus_BIG5_check(ei, (wc >> 8) & 0xff) != 2 || 374219019Sgabor !_citrus_BIG5_check2(ei, wc & 0xff)) { 375219019Sgabor ret = EILSEQ; 376219019Sgabor goto err; 377219019Sgabor } 378219019Sgabor l = 2; 379219019Sgabor } else { 380219019Sgabor if (wc & ~0xff || !_citrus_BIG5_check(ei, wc & 0xff)) { 381219019Sgabor ret = EILSEQ; 382219019Sgabor goto err; 383219019Sgabor } 384219019Sgabor l = 1; 385219019Sgabor } 386219019Sgabor 387219019Sgabor if (n < l) { 388219019Sgabor /* bound check failure */ 389219019Sgabor ret = E2BIG; 390219019Sgabor goto err; 391219019Sgabor } 392219019Sgabor 393219019Sgabor if (l == 2) { 394219019Sgabor s[0] = (wc >> 8) & 0xff; 395219019Sgabor s[1] = wc & 0xff; 396219019Sgabor } else 397219019Sgabor s[0] = wc & 0xff; 398219019Sgabor 399219019Sgabor *nresult = l; 400219019Sgabor 401219019Sgabor return (0); 402219019Sgabor 403219019Sgaborerr: 404219019Sgabor *nresult = (size_t)-1; 405219019Sgabor return (ret); 406219019Sgabor} 407219019Sgabor 408219019Sgaborstatic __inline int 409219019Sgabor/*ARGSUSED*/ 410219019Sgabor_citrus_BIG5_stdenc_wctocs(_BIG5EncodingInfo * __restrict ei __unused, 411219019Sgabor _csid_t * __restrict csid, 412219019Sgabor _index_t * __restrict idx, wchar_t wc) 413219019Sgabor{ 414219019Sgabor 415219019Sgabor *csid = (wc < 0x100) ? 0 : 1; 416219019Sgabor *idx = (_index_t)wc; 417219019Sgabor 418219019Sgabor return (0); 419219019Sgabor} 420219019Sgabor 421219019Sgaborstatic __inline int 422219019Sgabor/*ARGSUSED*/ 423219019Sgabor_citrus_BIG5_stdenc_cstowc(_BIG5EncodingInfo * __restrict ei __unused, 424219019Sgabor wchar_t * __restrict wc, 425219019Sgabor _csid_t csid, _index_t idx) 426219019Sgabor{ 427219019Sgabor 428219019Sgabor switch (csid) { 429219019Sgabor case 0: 430219019Sgabor case 1: 431219019Sgabor *wc = (wchar_t)idx; 432219019Sgabor break; 433219019Sgabor default: 434219019Sgabor return (EILSEQ); 435219019Sgabor } 436219019Sgabor 437219019Sgabor return (0); 438219019Sgabor} 439219019Sgabor 440219019Sgaborstatic __inline int 441219019Sgabor/*ARGSUSED*/ 442219019Sgabor_citrus_BIG5_stdenc_get_state_desc_generic(_BIG5EncodingInfo * __restrict ei __unused, 443219019Sgabor _BIG5State * __restrict psenc, 444219019Sgabor int * __restrict rstate) 445219019Sgabor{ 446219019Sgabor 447219019Sgabor *rstate = (psenc->chlen == 0) ? _STDENC_SDGEN_INITIAL : 448219019Sgabor _STDENC_SDGEN_INCOMPLETE_CHAR; 449219019Sgabor return (0); 450219019Sgabor} 451219019Sgabor 452219019Sgabor/* ---------------------------------------------------------------------- 453219019Sgabor * public interface for stdenc 454219019Sgabor */ 455219019Sgabor 456219019Sgabor_CITRUS_STDENC_DECLS(BIG5); 457219019Sgabor_CITRUS_STDENC_DEF_OPS(BIG5); 458219019Sgabor 459219019Sgabor#include "citrus_stdenc_template.h" 460