1 /* $NetBSD: gettext.c,v 1.9 2001/02/16 07:20:35 minoura Exp $ */ 2 3 /*- 4 * Copyright (c) 2000, 2001 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 #include <sys/cdefs.h> 30 #if defined(LIBC_SCCS) && !defined(lint) 31 __RCSID("$NetBSD: gettext.c,v 1.9 2001/02/16 07:20:35 minoura Exp $"); 32 #endif /* LIBC_SCCS and not lint */ 33 34 #include <sys/types.h> 35 #include <sys/param.h> 36 #include <sys/stat.h> 37 #include <sys/mman.h> 38 #include <sys/uio.h> 39 40 #include <fcntl.h> 41 #include <stdio.h> 42 #include <stdlib.h> 43 #include <unistd.h> 44 #include <string.h> 45 #if 0 46 #include <util.h> 47 #endif 48 #include <libintl.h> 49 #include <locale.h> 50 #include "libintl_local.h" 51 #include "pathnames.h" 52 53 static const char *lookup_category __P((int)); 54 static const char *split_locale __P((const char *)); 55 static const char *lookup_mofile __P((char *, size_t, const char *, 56 char *, const char *, const char *, struct domainbinding *)); 57 static u_int32_t flip __P((u_int32_t, u_int32_t)); 58 static int validate __P((void *, struct mohandle *)); 59 static int mapit __P((const char *, struct domainbinding *)); 60 static int unmapit __P((struct domainbinding *)); 61 static const char *lookup_hash __P((const char *, struct domainbinding *)); 62 static const char *lookup_bsearch __P((const char *, struct domainbinding *)); 63 static const char *lookup __P((const char *, struct domainbinding *)); 64 65 /* 66 * shortcut functions. the main implementation resides in dcngettext(). 67 */ 68 char * 69 gettext(msgid) 70 const char *msgid; 71 { 72 73 return dcngettext(NULL, msgid, NULL, 1UL, LC_MESSAGES); 74 } 75 76 char * 77 dgettext(domainname, msgid) 78 const char *domainname; 79 const char *msgid; 80 { 81 82 return dcngettext(domainname, msgid, NULL, 1UL, LC_MESSAGES); 83 } 84 85 char * 86 dcgettext(domainname, msgid, category) 87 const char *domainname; 88 const char *msgid; 89 int category; 90 { 91 92 return dcngettext(domainname, msgid, NULL, 1UL, category); 93 } 94 95 char * 96 ngettext(msgid1, msgid2, n) 97 const char *msgid1; 98 const char *msgid2; 99 unsigned long int n; 100 { 101 102 return dcngettext(NULL, msgid1, msgid2, n, LC_MESSAGES); 103 } 104 105 char * 106 dngettext(domainname, msgid1, msgid2, n) 107 const char *domainname; 108 const char *msgid1; 109 const char *msgid2; 110 unsigned long int n; 111 { 112 113 return dcngettext(domainname, msgid1, msgid2, n, LC_MESSAGES); 114 } 115 116 /* 117 * dcngettext() - 118 * lookup internationalized message on database locale/category/domainname 119 * (like ja_JP.eucJP/LC_MESSAGES/domainname). 120 * if n equals to 1, internationalized message will be looked up for msgid1. 121 * otherwise, message will be looked up for msgid2. 122 * if the lookup fails, the function will return msgid1 or msgid2 as is. 123 * 124 * Even though the return type is "char *", caller should not rewrite the 125 * region pointed to by the return value (should be "const char *", but can't 126 * change it for compatibility with other implementations). 127 * 128 * by default (if domainname == NULL), domainname is taken from the value set 129 * by textdomain(). usually name of the application (like "ls") is used as 130 * domainname. category is usually LC_MESSAGES. 131 * 132 * the code reads in *.mo files generated by GNU gettext. *.mo is a host- 133 * endian encoded file. both endians are supported here, as the files are in 134 * /usr/share/locale! (or we should move those files into /usr/libdata) 135 */ 136 137 static const char * 138 lookup_category(category) 139 int category; 140 { 141 142 switch (category) { 143 case LC_COLLATE: return "LC_COLLATE"; 144 case LC_CTYPE: return "LC_CTYPE"; 145 case LC_MONETARY: return "LC_MONETARY"; 146 case LC_NUMERIC: return "LC_NUMERIC"; 147 case LC_TIME: return "LC_TIME"; 148 case LC_MESSAGES: return "LC_MESSAGES"; 149 } 150 return NULL; 151 } 152 153 /* 154 * XPG syntax: language[_territory[.codeset]][@modifier] 155 * XXX boundary check on "result" is lacking 156 */ 157 static const char * 158 split_locale(lname) 159 const char *lname; 160 { 161 char buf[BUFSIZ], tmp[BUFSIZ]; 162 char *l, *t, *c, *m; 163 static char result[BUFSIZ]; 164 165 memset(result, 0, sizeof(result)); 166 167 if (strlen(lname) + 1 > sizeof(buf)) { 168 fail: 169 return lname; 170 } 171 172 strlcpy(buf, lname, sizeof(buf)); 173 m = strrchr(buf, '@'); 174 if (m) 175 *m++ = '\0'; 176 c = strrchr(buf, '.'); 177 if (c) 178 *c++ = '\0'; 179 t = strrchr(buf, '_'); 180 if (t) 181 *t++ = '\0'; 182 l = buf; 183 if (strlen(l) == 0) 184 goto fail; 185 if (c && !t) 186 goto fail; 187 188 if (m) { 189 if (t) { 190 if (c) { 191 snprintf(tmp, sizeof(tmp), "%s_%s.%s@%s", 192 l, t, c, m); 193 strlcat(result, tmp, sizeof(result)); 194 strlcat(result, ":", sizeof(result)); 195 } 196 snprintf(tmp, sizeof(tmp), "%s_%s@%s", l, t, m); 197 strlcat(result, tmp, sizeof(result)); 198 strlcat(result, ":", sizeof(result)); 199 } 200 snprintf(tmp, sizeof(tmp), "%s@%s", l, m); 201 strlcat(result, tmp, sizeof(result)); 202 strlcat(result, ":", sizeof(result)); 203 } 204 if (t) { 205 if (c) { 206 snprintf(tmp, sizeof(tmp), "%s_%s.%s", l, t, c); 207 strlcat(result, tmp, sizeof(result)); 208 strlcat(result, ":", sizeof(result)); 209 } 210 strlcat(result, tmp, sizeof(result)); 211 strlcat(result, ":", sizeof(result)); 212 } 213 strlcat(result, l, sizeof(result)); 214 215 return result; 216 } 217 218 static const char * 219 lookup_mofile(buf, len, dir, lpath, category, domainname, db) 220 char *buf; 221 size_t len; 222 const char *dir; 223 char *lpath; /* list of locales to be tried */ 224 const char *category; 225 const char *domainname; 226 struct domainbinding *db; 227 { 228 struct stat st; 229 char *p, *q; 230 231 q = lpath; 232 /* CONSTCOND */ 233 while (1) { 234 p = strsep(&q, ":"); 235 if (!p) 236 break; 237 if (!*p) 238 continue; 239 240 /* don't mess with default locales */ 241 if (strcmp(p, "C") == 0 || strcmp(p, "POSIX") == 0) 242 return NULL; 243 244 /* validate pathname */ 245 if (strchr(p, '/') || strchr(category, '/')) 246 continue; 247 #if 1 /*?*/ 248 if (strchr(domainname, '/')) 249 continue; 250 #endif 251 252 snprintf(buf, len, "%s/%s/%s/%s.mo", dir, p, 253 category, domainname); 254 if (stat(buf, &st) < 0) 255 continue; 256 if ((st.st_mode & S_IFMT) != S_IFREG) 257 continue; 258 259 if (mapit(buf, db) == 0) 260 return buf; 261 } 262 263 return NULL; 264 } 265 266 static u_int32_t 267 flip(v, magic) 268 u_int32_t v; 269 u_int32_t magic; 270 { 271 272 if (magic == MO_MAGIC) 273 return v; 274 else if (magic == MO_MAGIC_SWAPPED) { 275 v = ((v >> 24) & 0xff) | ((v >> 8) & 0xff00) | 276 ((v << 8) & 0xff0000) | ((v << 24) & 0xff000000); 277 return v; 278 } else { 279 abort(); 280 /*NOTREACHED*/ 281 } 282 } 283 284 static int 285 validate(arg, mohandle) 286 void *arg; 287 struct mohandle *mohandle; 288 { 289 char *p; 290 291 p = (char *)arg; 292 if (p < (char *)mohandle->addr || 293 p > (char *)mohandle->addr + mohandle->len) 294 return 0; 295 else 296 return 1; 297 } 298 299 int 300 mapit(path, db) 301 const char *path; 302 struct domainbinding *db; 303 { 304 int fd; 305 struct stat st; 306 char *base; 307 u_int32_t magic, revision; 308 struct moentry *otable, *ttable; 309 struct moentry_h *p; 310 struct mo *mo; 311 size_t l; 312 int i; 313 char *v; 314 struct mohandle *mohandle = &db->mohandle; 315 316 if (mohandle->addr && mohandle->addr != MAP_FAILED && 317 mohandle->mo.mo_magic) 318 return 0; /*already opened*/ 319 320 unmapit(db); 321 322 #if 0 323 if (secure_path(path) != 0) 324 goto fail; 325 #endif 326 if (stat(path, &st) < 0) 327 goto fail; 328 if ((st.st_mode & S_IFMT) != S_IFREG || st.st_size > GETTEXT_MMAP_MAX) 329 goto fail; 330 fd = open(path, O_RDONLY); 331 if (fd < 0) 332 goto fail; 333 if (read(fd, &magic, sizeof(magic)) != sizeof(magic) || 334 (magic != MO_MAGIC && magic != MO_MAGIC_SWAPPED)) { 335 close(fd); 336 goto fail; 337 } 338 if (read(fd, &revision, sizeof(revision)) != sizeof(revision) || 339 flip(revision, magic) != MO_REVISION) { 340 close(fd); 341 goto fail; 342 } 343 mohandle->addr = mmap(NULL, (size_t)st.st_size, PROT_READ, 344 MAP_FILE | MAP_SHARED, fd, (off_t)0); 345 if (!mohandle->addr || mohandle->addr == MAP_FAILED) { 346 close(fd); 347 goto fail; 348 } 349 close(fd); 350 mohandle->len = (size_t)st.st_size; 351 352 base = mohandle->addr; 353 mo = (struct mo *)mohandle->addr; 354 355 /* flip endian. do not flip magic number! */ 356 mohandle->mo.mo_magic = mo->mo_magic; 357 mohandle->mo.mo_revision = flip(mo->mo_revision, magic); 358 mohandle->mo.mo_nstring = flip(mo->mo_nstring, magic); 359 360 /* validate otable/ttable */ 361 otable = (struct moentry *)(base + flip(mo->mo_otable, magic)); 362 ttable = (struct moentry *)(base + flip(mo->mo_ttable, magic)); 363 if (!validate(otable, mohandle) || 364 !validate(&otable[mohandle->mo.mo_nstring], mohandle)) { 365 unmapit(db); 366 goto fail; 367 } 368 if (!validate(ttable, mohandle) || 369 !validate(&ttable[mohandle->mo.mo_nstring], mohandle)) { 370 unmapit(db); 371 goto fail; 372 } 373 374 /* allocate [ot]table, and convert to normal pointer representation. */ 375 l = sizeof(struct moentry_h) * mohandle->mo.mo_nstring; 376 mohandle->mo.mo_otable = (struct moentry_h *)malloc(l); 377 if (!mohandle->mo.mo_otable) { 378 unmapit(db); 379 goto fail; 380 } 381 mohandle->mo.mo_ttable = (struct moentry_h *)malloc(l); 382 if (!mohandle->mo.mo_ttable) { 383 unmapit(db); 384 goto fail; 385 } 386 p = mohandle->mo.mo_otable; 387 for (i = 0; i < mohandle->mo.mo_nstring; i++) { 388 p[i].len = flip(otable[i].len, magic); 389 p[i].off = base + flip(otable[i].off, magic); 390 391 if (!validate(p[i].off, mohandle) || 392 !validate(p[i].off + p[i].len + 1, mohandle)) { 393 unmapit(db); 394 goto fail; 395 } 396 } 397 p = mohandle->mo.mo_ttable; 398 for (i = 0; i < mohandle->mo.mo_nstring; i++) { 399 p[i].len = flip(ttable[i].len, magic); 400 p[i].off = base + flip(ttable[i].off, magic); 401 402 if (!validate(p[i].off, mohandle) || 403 !validate(p[i].off + p[i].len + 1, mohandle)) { 404 unmapit(db); 405 goto fail; 406 } 407 } 408 409 /* grab MIME-header and charset field */ 410 mohandle->mo.mo_header = lookup("", db); 411 if (mohandle->mo.mo_header) 412 v = strstr(mohandle->mo.mo_header, "charset="); 413 else 414 v = NULL; 415 if (v) { 416 mohandle->mo.mo_charset = strdup(v + 8); 417 if (!mohandle->mo.mo_charset) 418 goto fail; 419 v = strchr(mohandle->mo.mo_charset, '\n'); 420 if (v) 421 *v = '\0'; 422 } 423 424 /* 425 * XXX check charset, reject it if we are unable to support the charset 426 * with the current locale. 427 * for example, if we are using euc-jp locale and we are looking at 428 * *.mo file encoded by euc-kr (charset=euc-kr), we should reject 429 * the *.mo file as we cannot support it. 430 */ 431 432 return 0; 433 434 fail: 435 return -1; 436 } 437 438 static int 439 unmapit(db) 440 struct domainbinding *db; 441 { 442 struct mohandle *mohandle = &db->mohandle; 443 444 /* unmap if there's already mapped region */ 445 if (mohandle->addr && mohandle->addr != MAP_FAILED) 446 munmap(mohandle->addr, mohandle->len); 447 mohandle->addr = NULL; 448 if (mohandle->mo.mo_otable) 449 free(mohandle->mo.mo_otable); 450 if (mohandle->mo.mo_ttable) 451 free(mohandle->mo.mo_ttable); 452 if (mohandle->mo.mo_charset) 453 free(mohandle->mo.mo_charset); 454 memset(&mohandle->mo, 0, sizeof(mohandle->mo)); 455 return 0; 456 } 457 458 /* ARGSUSED */ 459 static const char * 460 lookup_hash(msgid, db) 461 const char *msgid; 462 struct domainbinding *db; 463 { 464 465 /* 466 * XXX should try a hashed lookup here, but to do so, we need to 467 * look inside the GPL'ed *.c and re-implement... 468 */ 469 return NULL; 470 } 471 472 static const char * 473 lookup_bsearch(msgid, db) 474 const char *msgid; 475 struct domainbinding *db; 476 { 477 int top, bottom, middle, omiddle; 478 int n; 479 struct mohandle *mohandle = &db->mohandle; 480 481 top = 0; 482 bottom = mohandle->mo.mo_nstring; 483 omiddle = -1; 484 /* CONSTCOND */ 485 while (1) { 486 if (top > bottom) 487 break; 488 middle = (top + bottom) / 2; 489 /* avoid possible infinite loop, when the data is not sorted */ 490 if (omiddle == middle) 491 break; 492 if (middle < 0 || middle >= mohandle->mo.mo_nstring) 493 break; 494 495 n = strcmp(msgid, mohandle->mo.mo_otable[middle].off); 496 if (n == 0) 497 return (const char *)mohandle->mo.mo_ttable[middle].off; 498 else if (n < 0) 499 bottom = middle; 500 else 501 top = middle; 502 omiddle = middle; 503 } 504 505 return NULL; 506 } 507 508 static const char * 509 lookup(msgid, db) 510 const char *msgid; 511 struct domainbinding *db; 512 { 513 const char *v; 514 515 v = lookup_hash(msgid, db); 516 if (v) 517 return v; 518 519 return lookup_bsearch(msgid, db); 520 } 521 522 char * 523 dcngettext(domainname, msgid1, msgid2, n, category) 524 const char *domainname; 525 const char *msgid1; 526 const char *msgid2; 527 unsigned long int n; 528 int category; 529 { 530 const char *msgid; 531 char path[PATH_MAX]; 532 static char lpath[PATH_MAX]; 533 static char olpath[PATH_MAX]; 534 const char *locale; 535 const char *language; 536 const char *cname = NULL; 537 const char *v; 538 static char *ocname = NULL; 539 static char *odomainname = NULL; 540 struct domainbinding *db; 541 542 msgid = (n == 1) ? msgid1 : msgid2; 543 if (msgid == NULL) 544 return NULL; 545 546 if (!domainname) 547 domainname = __current_domainname; 548 cname = lookup_category(category); 549 if (!domainname || !cname) 550 goto fail; 551 552 language = getenv("LANGUAGE"); 553 locale = setlocale(LC_MESSAGES, NULL); /*XXX*/ 554 if (locale) 555 locale = split_locale(locale); 556 if (language && locale) { 557 if (strlen(language) + strlen(locale) + 2 > sizeof(lpath)) 558 goto fail; 559 snprintf(lpath, sizeof(lpath), "%s:%s", language, locale); 560 } else if (language) { 561 if (strlen(language) + 1 > sizeof(lpath)) 562 goto fail; 563 strlcpy(lpath, language, sizeof(lpath)); 564 } else if (locale) { 565 if (strlen(locale) + 1 > sizeof(lpath)) 566 goto fail; 567 strlcpy(lpath, locale, sizeof(lpath)); 568 } else 569 goto fail; 570 571 for (db = __bindings; db; db = db->next) 572 if (strcmp(db->domainname, domainname) == 0) 573 break; 574 if (!db) { 575 if (!bindtextdomain(domainname, _PATH_TEXTDOMAIN)) 576 goto fail; 577 db = __bindings; 578 } 579 580 /* don't bother looking it up if the values are the same */ 581 if (odomainname && strcmp(domainname, odomainname) == 0 && 582 ocname && strcmp(cname, ocname) == 0 && strcmp(lpath, olpath) == 0 && 583 db->mohandle.mo.mo_magic) 584 goto found; 585 586 /* try to find appropriate file, from $LANGUAGE */ 587 if (lookup_mofile(path, sizeof(path), db->path, lpath, cname, 588 domainname, db) == NULL) 589 goto fail; 590 591 if (odomainname) 592 free(odomainname); 593 if (ocname) 594 free(ocname); 595 odomainname = strdup(domainname); 596 ocname = strdup(cname); 597 if (!odomainname || !ocname) { 598 if (odomainname) 599 free(odomainname); 600 if (ocname) 601 free(ocname); 602 odomainname = ocname = NULL; 603 goto fail; 604 } 605 606 strlcpy(olpath, lpath, sizeof(olpath)); 607 608 found: 609 v = lookup(msgid, db); 610 if (v) { 611 /* 612 * XXX call iconv() here, if translated text is encoded 613 * differently from currently-selected encoding (locale). 614 * look at Content-type header in *.mo file, in string obtained 615 * by gettext(""). 616 */ 617 618 /* 619 * Given the amount of printf-format security issues, it may 620 * be a good idea to validate if the original msgid and the 621 * translated message format string carry the same printf-like 622 * format identifiers. 623 */ 624 625 msgid = v; 626 } 627 628 fail: 629 /* LINTED const cast */ 630 return (char *)msgid; 631 } 632