1 /* $NetBSD: charset.c,v 1.5 2023/10/06 05:49:49 simonb Exp $ */ 2 3 /* 4 * Copyright (C) 1984-2023 Mark Nudelman 5 * 6 * You may distribute under the terms of either the GNU General Public 7 * License or the Less License, as specified in the README file. 8 * 9 * For more information, see the README file. 10 */ 11 12 13 /* 14 * Functions to define the character set 15 * and do things specific to the character set. 16 */ 17 18 #include "less.h" 19 #if HAVE_LOCALE 20 #include <locale.h> 21 #include <ctype.h> 22 #include <langinfo.h> 23 #endif 24 25 #include "charset.h" 26 #include "xbuf.h" 27 28 #if MSDOS_COMPILER==WIN32C 29 #define WIN32_LEAN_AND_MEAN 30 #include <windows.h> 31 #endif 32 33 extern int bs_mode; 34 35 public int utf_mode = 0; 36 37 /* 38 * Predefined character sets, 39 * selected by the LESSCHARSET environment variable. 40 */ 41 struct charset { 42 char *name; 43 int *p_flag; 44 char *desc; 45 } charsets[] = { 46 { "ascii", NULL, "8bcccbcc18b95.b" }, 47 { "utf-8", &utf_mode, "8bcccbcc18b95.b126.bb" }, 48 { "iso8859", NULL, "8bcccbcc18b95.33b." }, 49 { "latin3", NULL, "8bcccbcc18b95.33b5.b8.b15.b4.b12.b18.b12.b." }, 50 { "arabic", NULL, "8bcccbcc18b95.33b.3b.7b2.13b.3b.b26.5b19.b" }, 51 { "greek", NULL, "8bcccbcc18b95.33b4.2b4.b3.b35.b44.b" }, 52 { "greek2005", NULL, "8bcccbcc18b95.33b14.b35.b44.b" }, 53 { "hebrew", NULL, "8bcccbcc18b95.33b.b29.32b28.2b2.b" }, 54 { "koi8-r", NULL, "8bcccbcc18b95.b." }, 55 { "KOI8-T", NULL, "8bcccbcc18b95.b8.b6.b8.b.b.5b7.3b4.b4.b3.b.b.3b." }, 56 { "georgianps", NULL, "8bcccbcc18b95.3b11.4b12.2b." }, 57 { "tcvn", NULL, "b..b...bcccbccbbb7.8b95.b48.5b." }, 58 { "TIS-620", NULL, "8bcccbcc18b95.b.4b.11b7.8b." }, 59 { "next", NULL, "8bcccbcc18b95.bb125.bb" }, 60 { "dos", NULL, "8bcccbcc12bc5b95.b." }, 61 { "windows-1251", NULL, "8bcccbcc12bc5b95.b24.b." }, 62 { "windows-1252", NULL, "8bcccbcc12bc5b95.b.b11.b.2b12.b." }, 63 { "windows-1255", NULL, "8bcccbcc12bc5b95.b.b8.b.5b9.b.4b." }, 64 { "ebcdic", NULL, "5bc6bcc7bcc41b.9b7.9b5.b..8b6.10b6.b9.7b9.8b8.17b3.3b9.7b9.8b8.6b10.b.b.b." }, 65 { "IBM-1047", NULL, "4cbcbc3b9cbccbccbb4c6bcc5b3cbbc4bc4bccbc191.b" }, 66 { NULL, NULL, NULL } 67 }; 68 69 /* 70 * Support "locale charmap"/nl_langinfo(CODESET) values, as well as others. 71 */ 72 struct cs_alias { 73 char *name; 74 char *oname; 75 } cs_aliases[] = { 76 { "UTF-8", "utf-8" }, 77 { "utf8", "utf-8" }, 78 { "UTF8", "utf-8" }, 79 { "ANSI_X3.4-1968", "ascii" }, 80 { "US-ASCII", "ascii" }, 81 { "latin1", "iso8859" }, 82 { "ISO-8859-1", "iso8859" }, 83 { "latin9", "iso8859" }, 84 { "ISO-8859-15", "iso8859" }, 85 { "latin2", "iso8859" }, 86 { "ISO-8859-2", "iso8859" }, 87 { "ISO-8859-3", "latin3" }, 88 { "latin4", "iso8859" }, 89 { "ISO-8859-4", "iso8859" }, 90 { "cyrillic", "iso8859" }, 91 { "ISO-8859-5", "iso8859" }, 92 { "ISO-8859-6", "arabic" }, 93 { "ISO-8859-7", "greek" }, 94 { "IBM9005", "greek2005" }, 95 { "ISO-8859-8", "hebrew" }, 96 { "latin5", "iso8859" }, 97 { "ISO-8859-9", "iso8859" }, 98 { "latin6", "iso8859" }, 99 { "ISO-8859-10", "iso8859" }, 100 { "latin7", "iso8859" }, 101 { "ISO-8859-13", "iso8859" }, 102 { "latin8", "iso8859" }, 103 { "ISO-8859-14", "iso8859" }, 104 { "latin10", "iso8859" }, 105 { "ISO-8859-16", "iso8859" }, 106 { "IBM437", "dos" }, 107 { "EBCDIC-US", "ebcdic" }, 108 { "IBM1047", "IBM-1047" }, 109 { "KOI8-R", "koi8-r" }, 110 { "KOI8-U", "koi8-r" }, 111 { "GEORGIAN-PS", "georgianps" }, 112 { "TCVN5712-1", "tcvn" }, 113 { "NEXTSTEP", "next" }, 114 { "windows", "windows-1252" }, /* backward compatibility */ 115 { "CP1251", "windows-1251" }, 116 { "CP1252", "windows-1252" }, 117 { "CP1255", "windows-1255" }, 118 { NULL, NULL } 119 }; 120 121 #define IS_BINARY_CHAR 01 122 #define IS_CONTROL_CHAR 02 123 124 static char chardef[256]; 125 static char *binfmt = NULL; 126 static char *utfbinfmt = NULL; 127 public int binattr = AT_STANDOUT|AT_COLOR_BIN; 128 129 static struct xbuffer user_wide_array; 130 static struct xbuffer user_ubin_array; 131 static struct xbuffer user_compose_array; 132 static struct xbuffer user_prt_array; 133 static struct wchar_range_table user_wide_table; 134 static struct wchar_range_table user_ubin_table; 135 static struct wchar_range_table user_compose_table; 136 static struct wchar_range_table user_prt_table; 137 138 /* 139 * Set a wchar_range_table to the table in an xbuffer. 140 */ 141 static void wchar_range_table_set(struct wchar_range_table *tbl, struct xbuffer *arr) 142 { 143 tbl->table = (struct wchar_range *) arr->data; 144 tbl->count = arr->end / sizeof(struct wchar_range); 145 } 146 147 /* 148 * Skip over a "U" or "U+" prefix before a hex codepoint. 149 */ 150 static char * skip_uprefix(char *s) 151 { 152 if (*s == 'U' || *s == 'u') 153 if (*++s == '+') ++s; 154 return s; 155 } 156 157 /* 158 * Parse a dash-separated range of hex values. 159 */ 160 static void wchar_range_get(char **ss, struct wchar_range *range) 161 { 162 char *s = skip_uprefix(*ss); 163 range->first = lstrtoul(s, &s, 16); 164 if (s[0] == '-') 165 { 166 s = skip_uprefix(&s[1]); 167 range->last = lstrtoul(s, &s, 16); 168 } else 169 { 170 range->last = range->first; 171 } 172 *ss = s; 173 } 174 175 /* 176 * Parse the LESSUTFCHARDEF variable. 177 */ 178 static void ichardef_utf(char *s) 179 { 180 xbuf_init(&user_wide_array); 181 xbuf_init(&user_ubin_array); 182 xbuf_init(&user_compose_array); 183 xbuf_init(&user_prt_array); 184 185 if (s != NULL) 186 { 187 while (s[0] != '\0') 188 { 189 struct wchar_range range; 190 wchar_range_get(&s, &range); 191 if (range.last == 0) 192 { 193 error("invalid hex number(s) in LESSUTFCHARDEF", NULL_PARG); 194 quit(QUIT_ERROR); 195 } 196 if (*s++ != ':') 197 { 198 error("missing colon in LESSUTFCHARDEF", NULL_PARG); 199 quit(QUIT_ERROR); 200 } 201 switch (*s++) 202 { 203 case 'b': 204 xbuf_add_data(&user_ubin_array, (unsigned char *) &range, sizeof(range)); 205 break; 206 case 'c': 207 xbuf_add_data(&user_compose_array, (unsigned char *) &range, sizeof(range)); 208 break; 209 case 'w': 210 xbuf_add_data(&user_wide_array, (unsigned char *) &range, sizeof(range)); 211 xbuf_add_data(&user_prt_array, (unsigned char *) &range, sizeof(range)); 212 break; 213 case 'p': case '.': 214 xbuf_add_data(&user_prt_array, (unsigned char *) &range, sizeof(range)); 215 break; 216 case '\0': 217 s--; 218 break; 219 default: 220 /* Ignore unknown character attribute. */ 221 break; 222 } 223 if (s[0] == ',') ++s; 224 } 225 } 226 wchar_range_table_set(&user_wide_table, &user_wide_array); 227 wchar_range_table_set(&user_ubin_table, &user_ubin_array); 228 wchar_range_table_set(&user_compose_table, &user_compose_array); 229 wchar_range_table_set(&user_prt_table, &user_prt_array); 230 } 231 232 /* 233 * Define a charset, given a description string. 234 * The string consists of 256 letters, 235 * one for each character in the charset. 236 * If the string is shorter than 256 letters, missing letters 237 * are taken to be identical to the last one. 238 * A decimal number followed by a letter is taken to be a 239 * repetition of the letter. 240 * 241 * Each letter is one of: 242 * . normal character 243 * b binary character 244 * c control character 245 */ 246 static void ichardef(char *s) 247 { 248 char *cp; 249 int n; 250 char v; 251 252 n = 0; 253 v = 0; 254 cp = chardef; 255 while (*s != '\0') 256 { 257 switch (*s++) 258 { 259 case '.': 260 v = 0; 261 break; 262 case 'c': 263 v = IS_CONTROL_CHAR; 264 break; 265 case 'b': 266 v = IS_BINARY_CHAR|IS_CONTROL_CHAR; 267 break; 268 269 case '0': case '1': case '2': case '3': case '4': 270 case '5': case '6': case '7': case '8': case '9': 271 if (ckd_mul(&n, n, 10) || ckd_add(&n, n, s[-1] - '0')) 272 goto invalid_chardef; 273 continue; 274 275 default: 276 invalid_chardef: 277 error("invalid chardef", NULL_PARG); 278 quit(QUIT_ERROR); 279 /*NOTREACHED*/ 280 } 281 282 do 283 { 284 if (cp >= chardef + sizeof(chardef)) 285 { 286 error("chardef longer than 256", NULL_PARG); 287 quit(QUIT_ERROR); 288 /*NOTREACHED*/ 289 } 290 *cp++ = v; 291 } while (--n > 0); 292 n = 0; 293 } 294 295 while (cp < chardef + sizeof(chardef)) 296 *cp++ = v; 297 } 298 299 /* 300 * Define a charset, given a charset name. 301 * The valid charset names are listed in the "charsets" array. 302 */ 303 static int icharset(char *name, int no_error) 304 { 305 struct charset *p; 306 struct cs_alias *a; 307 308 if (name == NULL || *name == '\0') 309 return (0); 310 311 /* First see if the name is an alias. */ 312 for (a = cs_aliases; a->name != NULL; a++) 313 { 314 if (strcmp(name, a->name) == 0) 315 { 316 name = a->oname; 317 break; 318 } 319 } 320 321 for (p = charsets; p->name != NULL; p++) 322 { 323 if (strcmp(name, p->name) == 0) 324 { 325 ichardef(p->desc); 326 if (p->p_flag != NULL) 327 { 328 #if MSDOS_COMPILER==WIN32C 329 *(p->p_flag) = 1 + (GetConsoleOutputCP() != CP_UTF8); 330 #else 331 *(p->p_flag) = 1; 332 #endif 333 } 334 return (1); 335 } 336 } 337 338 if (!no_error) { 339 error("invalid charset name", NULL_PARG); 340 quit(QUIT_ERROR); 341 } 342 return (0); 343 } 344 345 #if HAVE_LOCALE 346 /* 347 * Define a charset, given a locale name. 348 */ 349 static void ilocale(void) 350 { 351 int c; 352 353 for (c = 0; c < (int) sizeof(chardef); c++) 354 { 355 if (isprint(c)) 356 chardef[c] = 0; 357 else if (iscntrl(c)) 358 chardef[c] = IS_CONTROL_CHAR; 359 else 360 chardef[c] = IS_BINARY_CHAR|IS_CONTROL_CHAR; 361 } 362 } 363 #endif 364 365 /* 366 * Define the printing format for control (or binary utf) chars. 367 */ 368 public void setfmt(char *s, char **fmtvarptr, int *attrptr, char *default_fmt, int for_printf) 369 { 370 if (s && utf_mode) 371 { 372 /* It would be too hard to account for width otherwise. */ 373 char constant *t = s; 374 while (*t) 375 { 376 if (*t < ' ' || *t > '~') 377 { 378 s = default_fmt; 379 goto attr; 380 } 381 t++; 382 } 383 } 384 385 if (s == NULL || *s == '\0') 386 s = default_fmt; 387 else if (for_printf && 388 ((*s == '*' && (s[1] == '\0' || s[2] == '\0' || strchr(s + 2, 'n'))) || 389 (*s != '*' && strchr(s, 'n')))) 390 /* %n is evil */ 391 s = default_fmt; 392 393 /* 394 * Select the attributes if it starts with "*". 395 */ 396 attr: 397 if (*s == '*' && s[1] != '\0') 398 { 399 switch (s[1]) 400 { 401 case 'd': *attrptr = AT_BOLD; break; 402 case 'k': *attrptr = AT_BLINK; break; 403 case 's': *attrptr = AT_STANDOUT; break; 404 case 'u': *attrptr = AT_UNDERLINE; break; 405 default: *attrptr = AT_NORMAL; break; 406 } 407 s += 2; 408 } 409 *fmtvarptr = s; 410 } 411 412 /* 413 * 414 */ 415 static void set_charset(void) 416 { 417 char *s; 418 419 #if MSDOS_COMPILER==WIN32C 420 /* 421 * If the Windows console is using UTF-8, we'll use it too. 422 */ 423 if (GetConsoleOutputCP() == CP_UTF8) 424 if (icharset("utf-8", 1)) 425 return; 426 #endif 427 428 ichardef_utf(lgetenv("LESSUTFCHARDEF")); 429 430 /* 431 * See if environment variable LESSCHARSET is defined. 432 */ 433 s = lgetenv("LESSCHARSET"); 434 if (icharset(s, 0)) 435 return; 436 437 /* 438 * LESSCHARSET is not defined: try LESSCHARDEF. 439 */ 440 s = lgetenv("LESSCHARDEF"); 441 if (!isnullenv(s)) 442 { 443 ichardef(s); 444 return; 445 } 446 447 #if HAVE_LOCALE 448 #ifdef CODESET 449 /* 450 * Try using the codeset name as the charset name. 451 */ 452 s = nl_langinfo(CODESET); 453 if (icharset(s, 1)) 454 return; 455 #endif 456 #endif 457 458 #if HAVE_STRSTR 459 /* 460 * Check whether LC_ALL, LC_CTYPE or LANG look like UTF-8 is used. 461 */ 462 if ((s = lgetenv("LC_ALL")) != NULL || 463 (s = lgetenv("LC_CTYPE")) != NULL || 464 (s = lgetenv("LANG")) != NULL) 465 { 466 if ( strstr(s, "UTF-8") != NULL || strstr(s, "utf-8") != NULL 467 || strstr(s, "UTF8") != NULL || strstr(s, "utf8") != NULL) 468 if (icharset("utf-8", 1)) 469 return; 470 } 471 #endif 472 473 #if HAVE_LOCALE 474 /* 475 * Get character definitions from locale functions, 476 * rather than from predefined charset entry. 477 */ 478 ilocale(); 479 #else 480 #if MSDOS_COMPILER 481 /* 482 * Default to "dos". 483 */ 484 (void) icharset("dos", 1); 485 #else 486 /* 487 * Default to "latin1". 488 */ 489 (void) icharset("latin1", 1); 490 #endif 491 #endif 492 } 493 494 /* 495 * Initialize charset data structures. 496 */ 497 public void init_charset(void) 498 { 499 char *s; 500 501 #if HAVE_LOCALE 502 setlocale(LC_ALL, ""); 503 #endif 504 505 set_charset(); 506 507 s = lgetenv("LESSBINFMT"); 508 setfmt(s, &binfmt, &binattr, "*s<%02X>", TRUE); 509 510 s = lgetenv("LESSUTFBINFMT"); 511 setfmt(s, &utfbinfmt, &binattr, "<U+%04lX>", TRUE); 512 } 513 514 /* 515 * Is a given character a "binary" character? 516 */ 517 public int binary_char(LWCHAR c) 518 { 519 if (utf_mode) 520 return (is_ubin_char(c)); 521 c &= 0377; 522 return (chardef[c] & IS_BINARY_CHAR); 523 } 524 525 /* 526 * Is a given character a "control" character? 527 */ 528 public int control_char(LWCHAR c) 529 { 530 c &= 0377; 531 return (chardef[c] & IS_CONTROL_CHAR); 532 } 533 534 /* 535 * Return the printable form of a character. 536 * For example, in the "ascii" charset '\3' is printed as "^C". 537 */ 538 public char * prchar(LWCHAR c) 539 { 540 /* {{ This buffer can be overrun if LESSBINFMT is a long string. }} */ 541 static char buf[MAX_PRCHAR_LEN+1]; 542 543 c &= 0377; 544 if ((c < 128 || !utf_mode) && !control_char(c)) 545 SNPRINTF1(buf, sizeof(buf), "%c", (int) c); 546 else if (c == ESC) 547 strcpy(buf, "ESC"); 548 #if IS_EBCDIC_HOST 549 else if (!binary_char(c) && c < 64) 550 SNPRINTF1(buf, sizeof(buf), "^%c", 551 /* 552 * This array roughly inverts CONTROL() #defined in less.h, 553 * and should be kept in sync with CONTROL() and IBM-1047. 554 */ 555 "@ABC.I.?...KLMNO" 556 "PQRS.JH.XY.." 557 "\\]^_" 558 "......W[.....EFG" 559 "..V....D....TU.Z"[c]); 560 #else 561 else if (c < 128 && !control_char(c ^ 0100)) 562 SNPRINTF1(buf, sizeof(buf), "^%c", (int) (c ^ 0100)); 563 #endif 564 else 565 SNPRINTF1(buf, sizeof(buf), binfmt, c); 566 return (buf); 567 } 568 569 /* 570 * Return the printable form of a UTF-8 character. 571 */ 572 public char * prutfchar(LWCHAR ch) 573 { 574 static char buf[MAX_PRCHAR_LEN+1]; 575 576 if (ch == ESC) 577 strcpy(buf, "ESC"); 578 else if (ch < 128 && control_char(ch)) 579 { 580 if (!control_char(ch ^ 0100)) 581 SNPRINTF1(buf, sizeof(buf), "^%c", ((char) ch) ^ 0100); 582 else 583 SNPRINTF1(buf, sizeof(buf), binfmt, (char) ch); 584 } else if (is_ubin_char(ch)) 585 { 586 SNPRINTF1(buf, sizeof(buf), utfbinfmt, ch); 587 } else 588 { 589 char *p = buf; 590 if (ch >= 0x80000000) 591 ch = 0xFFFD; /* REPLACEMENT CHARACTER */ 592 put_wchar(&p, ch); 593 *p = '\0'; 594 } 595 return (buf); 596 } 597 598 /* 599 * Get the length of a UTF-8 character in bytes. 600 */ 601 public int utf_len(int ch) 602 { 603 if ((ch & 0x80) == 0) 604 return 1; 605 if ((ch & 0xE0) == 0xC0) 606 return 2; 607 if ((ch & 0xF0) == 0xE0) 608 return 3; 609 if ((ch & 0xF8) == 0xF0) 610 return 4; 611 if ((ch & 0xFC) == 0xF8) 612 return 5; 613 if ((ch & 0xFE) == 0xFC) 614 return 6; 615 /* Invalid UTF-8 encoding. */ 616 return 1; 617 } 618 619 /* 620 * Does the parameter point to the lead byte of a well-formed UTF-8 character? 621 */ 622 public int is_utf8_well_formed(char *ss, int slen) 623 { 624 int i; 625 int len; 626 unsigned char *s = (unsigned char *) ss; 627 628 if (IS_UTF8_INVALID(s[0])) 629 return (0); 630 631 len = utf_len(s[0]); 632 if (len > slen) 633 return (0); 634 if (len == 1) 635 return (1); 636 if (len == 2) 637 { 638 if (s[0] < 0xC2) 639 return (0); 640 } else 641 { 642 unsigned char mask; 643 mask = (~((1 << (8-len)) - 1)) & 0xFF; 644 if (s[0] == mask && (s[1] & mask) == 0x80) 645 return (0); 646 } 647 648 for (i = 1; i < len; i++) 649 if (!IS_UTF8_TRAIL(s[i])) 650 return (0); 651 return (1); 652 } 653 654 /* 655 * Skip bytes until a UTF-8 lead byte (11xxxxxx) or ASCII byte (0xxxxxxx) is found. 656 */ 657 public void utf_skip_to_lead(char **pp, char *limit) 658 { 659 do { 660 ++(*pp); 661 } while (*pp < limit && !IS_UTF8_LEAD((*pp)[0] & 0377) && !IS_ASCII_OCTET((*pp)[0])); 662 } 663 664 665 /* 666 * Get the value of a UTF-8 character. 667 */ 668 public LWCHAR get_wchar(constant char *p) 669 { 670 switch (utf_len(p[0])) 671 { 672 case 1: 673 default: 674 /* 0xxxxxxx */ 675 return (LWCHAR) 676 (p[0] & 0xFF); 677 case 2: 678 /* 110xxxxx 10xxxxxx */ 679 return (LWCHAR) ( 680 ((p[0] & 0x1F) << 6) | 681 (p[1] & 0x3F)); 682 case 3: 683 /* 1110xxxx 10xxxxxx 10xxxxxx */ 684 return (LWCHAR) ( 685 ((p[0] & 0x0F) << 12) | 686 ((p[1] & 0x3F) << 6) | 687 (p[2] & 0x3F)); 688 case 4: 689 /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */ 690 return (LWCHAR) ( 691 ((p[0] & 0x07) << 18) | 692 ((p[1] & 0x3F) << 12) | 693 ((p[2] & 0x3F) << 6) | 694 (p[3] & 0x3F)); 695 case 5: 696 /* 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */ 697 return (LWCHAR) ( 698 ((p[0] & 0x03) << 24) | 699 ((p[1] & 0x3F) << 18) | 700 ((p[2] & 0x3F) << 12) | 701 ((p[3] & 0x3F) << 6) | 702 (p[4] & 0x3F)); 703 case 6: 704 /* 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */ 705 return (LWCHAR) ( 706 ((p[0] & 0x01) << 30) | 707 ((p[1] & 0x3F) << 24) | 708 ((p[2] & 0x3F) << 18) | 709 ((p[3] & 0x3F) << 12) | 710 ((p[4] & 0x3F) << 6) | 711 (p[5] & 0x3F)); 712 } 713 } 714 715 /* 716 * Store a character into a UTF-8 string. 717 */ 718 public void put_wchar(char **pp, LWCHAR ch) 719 { 720 if (!utf_mode || ch < 0x80) 721 { 722 /* 0xxxxxxx */ 723 *(*pp)++ = (char) ch; 724 } else if (ch < 0x800) 725 { 726 /* 110xxxxx 10xxxxxx */ 727 *(*pp)++ = (char) (0xC0 | ((ch >> 6) & 0x1F)); 728 *(*pp)++ = (char) (0x80 | (ch & 0x3F)); 729 } else if (ch < 0x10000) 730 { 731 /* 1110xxxx 10xxxxxx 10xxxxxx */ 732 *(*pp)++ = (char) (0xE0 | ((ch >> 12) & 0x0F)); 733 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F)); 734 *(*pp)++ = (char) (0x80 | (ch & 0x3F)); 735 } else if (ch < 0x200000) 736 { 737 /* 11110xxx 10xxxxxx 10xxxxxx 10xxxxxx */ 738 *(*pp)++ = (char) (0xF0 | ((ch >> 18) & 0x07)); 739 *(*pp)++ = (char) (0x80 | ((ch >> 12) & 0x3F)); 740 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F)); 741 *(*pp)++ = (char) (0x80 | (ch & 0x3F)); 742 } else if (ch < 0x4000000) 743 { 744 /* 111110xx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */ 745 *(*pp)++ = (char) (0xF0 | ((ch >> 24) & 0x03)); 746 *(*pp)++ = (char) (0x80 | ((ch >> 18) & 0x3F)); 747 *(*pp)++ = (char) (0x80 | ((ch >> 12) & 0x3F)); 748 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F)); 749 *(*pp)++ = (char) (0x80 | (ch & 0x3F)); 750 } else 751 { 752 /* 1111110x 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx 10xxxxxx */ 753 *(*pp)++ = (char) (0xF0 | ((ch >> 30) & 0x01)); 754 *(*pp)++ = (char) (0x80 | ((ch >> 24) & 0x3F)); 755 *(*pp)++ = (char) (0x80 | ((ch >> 18) & 0x3F)); 756 *(*pp)++ = (char) (0x80 | ((ch >> 12) & 0x3F)); 757 *(*pp)++ = (char) (0x80 | ((ch >> 6) & 0x3F)); 758 *(*pp)++ = (char) (0x80 | (ch & 0x3F)); 759 } 760 } 761 762 /* 763 * Step forward or backward one character in a string. 764 */ 765 public LWCHAR step_char(char **pp, signed int dir, constant char *limit) 766 { 767 LWCHAR ch; 768 int len; 769 char *p = *pp; 770 771 if (!utf_mode) 772 { 773 /* It's easy if chars are one byte. */ 774 if (dir > 0) 775 ch = (LWCHAR) (unsigned char) ((p < limit) ? *p++ : 0); 776 else 777 ch = (LWCHAR) (unsigned char) ((p > limit) ? *--p : 0); 778 } else if (dir > 0) 779 { 780 len = utf_len(*p); 781 if (p + len > limit) 782 { 783 ch = 0; 784 p = (char *) limit; 785 } else 786 { 787 ch = get_wchar(p); 788 p += len; 789 } 790 } else 791 { 792 while (p > limit && IS_UTF8_TRAIL(p[-1])) 793 p--; 794 if (p > limit) 795 ch = get_wchar(--p); 796 else 797 ch = 0; 798 } 799 *pp = p; 800 return ch; 801 } 802 803 /* 804 * Unicode characters data 805 * Actual data is in the generated *.uni files. 806 */ 807 808 #define DECLARE_RANGE_TABLE_START(name) \ 809 static struct wchar_range name##_array[] = { 810 #define DECLARE_RANGE_TABLE_END(name) \ 811 }; struct wchar_range_table name##_table = { name##_array, sizeof(name##_array)/sizeof(*name##_array) }; 812 813 DECLARE_RANGE_TABLE_START(compose) 814 #include "compose.uni" 815 DECLARE_RANGE_TABLE_END(compose) 816 817 DECLARE_RANGE_TABLE_START(ubin) 818 #include "ubin.uni" 819 DECLARE_RANGE_TABLE_END(ubin) 820 821 DECLARE_RANGE_TABLE_START(wide) 822 #include "wide.uni" 823 DECLARE_RANGE_TABLE_END(wide) 824 825 DECLARE_RANGE_TABLE_START(fmt) 826 #include "fmt.uni" 827 DECLARE_RANGE_TABLE_END(fmt) 828 829 /* comb_table is special pairs, not ranges. */ 830 static struct wchar_range comb_table[] = { 831 {0x0644,0x0622}, {0x0644,0x0623}, {0x0644,0x0625}, {0x0644,0x0627}, 832 }; 833 834 835 static int is_in_table(LWCHAR ch, struct wchar_range_table *table) 836 { 837 int hi; 838 int lo; 839 840 /* Binary search in the table. */ 841 if (table->table == NULL || table->count == 0 || ch < table->table[0].first) 842 return 0; 843 lo = 0; 844 hi = table->count - 1; 845 while (lo <= hi) 846 { 847 int mid = (lo + hi) / 2; 848 if (ch > table->table[mid].last) 849 lo = mid + 1; 850 else if (ch < table->table[mid].first) 851 hi = mid - 1; 852 else 853 return 1; 854 } 855 return 0; 856 } 857 858 /* 859 * Is a character a UTF-8 composing character? 860 * If a composing character follows any char, the two combine into one glyph. 861 */ 862 public int is_composing_char(LWCHAR ch) 863 { 864 if (is_in_table(ch, &user_prt_table)) return 0; 865 return is_in_table(ch, &user_compose_table) || 866 is_in_table(ch, &compose_table) || 867 (bs_mode != BS_CONTROL && is_in_table(ch, &fmt_table)); 868 } 869 870 /* 871 * Should this UTF-8 character be treated as binary? 872 */ 873 public int is_ubin_char(LWCHAR ch) 874 { 875 if (is_in_table(ch, &user_prt_table)) return 0; 876 return is_in_table(ch, &user_ubin_table) || 877 is_in_table(ch, &ubin_table) || 878 (bs_mode == BS_CONTROL && is_in_table(ch, &fmt_table)); 879 } 880 881 /* 882 * Is this a double width UTF-8 character? 883 */ 884 public int is_wide_char(LWCHAR ch) 885 { 886 return is_in_table(ch, &user_wide_table) || 887 is_in_table(ch, &wide_table); 888 } 889 890 /* 891 * Is a character a UTF-8 combining character? 892 * A combining char acts like an ordinary char, but if it follows 893 * a specific char (not any char), the two combine into one glyph. 894 */ 895 public int is_combining_char(LWCHAR ch1, LWCHAR ch2) 896 { 897 /* The table is small; use linear search. */ 898 int i; 899 for (i = 0; i < sizeof(comb_table)/sizeof(*comb_table); i++) 900 { 901 if (ch1 == comb_table[i].first && 902 ch2 == comb_table[i].last) 903 return 1; 904 } 905 return 0; 906 } 907 908