1 /**************************************************************************** 2 * Copyright (c) 1998,1999,2000 Free Software Foundation, Inc. * 3 * * 4 * Permission is hereby granted, free of charge, to any person obtaining a * 5 * copy of this software and associated documentation files (the * 6 * "Software"), to deal in the Software without restriction, including * 7 * without limitation the rights to use, copy, modify, merge, publish, * 8 * distribute, distribute with modifications, sublicense, and/or sell * 9 * copies of the Software, and to permit persons to whom the Software is * 10 * furnished to do so, subject to the following conditions: * 11 * * 12 * The above copyright notice and this permission notice shall be included * 13 * in all copies or substantial portions of the Software. * 14 * * 15 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS * 16 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF * 17 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. * 18 * IN NO EVENT SHALL THE ABOVE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, * 19 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR * 20 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR * 21 * THE USE OR OTHER DEALINGS IN THE SOFTWARE. * 22 * * 23 * Except as contained in this notice, the name(s) of the above copyright * 24 * holders shall not be used in advertising or otherwise to promote the * 25 * sale, use or other dealings in this Software without prior written * 26 * authorization. * 27 ****************************************************************************/ 28 29 /**************************************************************************** 30 * Author: Zeyd M. Ben-Halim <zmbenhal@netcom.com> 1992,1995 * 31 * and: Eric S. Raymond <esr@snark.thyrsus.com> * 32 ****************************************************************************/ 33 34 /* $FreeBSD: src/contrib/ncurses/ncurses/tinfo/comp_scan.c,v 1.2.2.2 2000/10/12 18:40:51 peter Exp $ */ 35 /* $DragonFly: src/contrib/ncurses/ncurses/tinfo/Attic/comp_scan.c,v 1.2 2003/06/17 04:24:03 dillon Exp $ */ 36 37 /* 38 * comp_scan.c --- Lexical scanner for terminfo compiler. 39 * 40 * _nc_reset_input() 41 * _nc_get_token() 42 * _nc_panic_mode() 43 * int _nc_syntax; 44 * int _nc_curr_line; 45 * long _nc_curr_file_pos; 46 * long _nc_comment_start; 47 * long _nc_comment_end; 48 */ 49 50 #include <curses.priv.h> 51 52 #include <ctype.h> 53 #include <term_entry.h> 54 #include <tic.h> 55 56 MODULE_ID("$Id: comp_scan.c,v 1.47 2000/09/24 01:15:17 tom Exp $") 57 58 /* 59 * Maximum length of string capability we'll accept before raising an error. 60 * Yes, there is a real capability in /etc/termcap this long, an "is". 61 */ 62 #define MAXCAPLEN 600 63 64 #define iswhite(ch) (ch == ' ' || ch == '\t') 65 66 int _nc_syntax = 0; /* termcap or terminfo? */ 67 long _nc_curr_file_pos = 0; /* file offset of current line */ 68 long _nc_comment_start = 0; /* start of comment range before name */ 69 long _nc_comment_end = 0; /* end of comment range before name */ 70 long _nc_start_line = 0; /* start line of current entry */ 71 72 struct token _nc_curr_token = 73 {0, 0, 0}; 74 75 /***************************************************************************** 76 * 77 * Token-grabbing machinery 78 * 79 *****************************************************************************/ 80 81 static bool first_column; /* See 'next_char()' below */ 82 static char separator; /* capability separator */ 83 static int pushtype; /* type of pushback token */ 84 static char pushname[MAX_NAME_SIZE + 1]; 85 86 #if NCURSES_EXT_FUNCS 87 bool _nc_disable_period = FALSE; /* used by tic -a option */ 88 #endif 89 90 static int last_char(void); 91 static int next_char(void); 92 static long stream_pos(void); 93 static bool end_of_stream(void); 94 static void push_back(char c); 95 96 /* Assume we may be looking at a termcap-style continuation */ 97 static inline int 98 eat_escaped_newline(int ch) 99 { 100 if (ch == '\\') 101 while ((ch = next_char()) == '\n' || iswhite(ch)) 102 continue; 103 return ch; 104 } 105 106 /* 107 * int 108 * get_token() 109 * 110 * Scans the input for the next token, storing the specifics in the 111 * global structure 'curr_token' and returning one of the following: 112 * 113 * NAMES A line beginning in column 1. 'name' 114 * will be set to point to everything up to but 115 * not including the first separator on the line. 116 * BOOLEAN An entry consisting of a name followed by 117 * a separator. 'name' will be set to point to 118 * the name of the capability. 119 * NUMBER An entry of the form 120 * name#digits, 121 * 'name' will be set to point to the capability 122 * name and 'valnumber' to the number given. 123 * STRING An entry of the form 124 * name=characters, 125 * 'name' is set to the capability name and 126 * 'valstring' to the string of characters, with 127 * input translations done. 128 * CANCEL An entry of the form 129 * name@, 130 * 'name' is set to the capability name and 131 * 'valnumber' to -1. 132 * EOF The end of the file has been reached. 133 * 134 * A `separator' is either a comma or a semicolon, depending on whether 135 * we are in termcap or terminfo mode. 136 * 137 */ 138 139 int 140 _nc_get_token(void) 141 { 142 static const char terminfo_punct[] = "@%&*!#"; 143 long number; 144 int type; 145 int ch; 146 char *numchk; 147 char numbuf[80]; 148 unsigned found; 149 static char buffer[MAX_ENTRY_SIZE]; 150 char *ptr; 151 int dot_flag = FALSE; 152 long token_start; 153 154 if (pushtype != NO_PUSHBACK) { 155 int retval = pushtype; 156 157 _nc_set_type(pushname); 158 DEBUG(3, ("pushed-back token: `%s', class %d", 159 _nc_curr_token.tk_name, pushtype)); 160 161 pushtype = NO_PUSHBACK; 162 pushname[0] = '\0'; 163 164 /* currtok wasn't altered by _nc_push_token() */ 165 return (retval); 166 } 167 168 if (end_of_stream()) 169 return (EOF); 170 171 start_token: 172 token_start = stream_pos(); 173 while ((ch = next_char()) == '\n' || iswhite(ch)) 174 continue; 175 176 ch = eat_escaped_newline(ch); 177 178 if (ch == EOF) 179 type = EOF; 180 else { 181 /* if this is a termcap entry, skip a leading separator */ 182 if (separator == ':' && ch == ':') 183 ch = next_char(); 184 185 if (ch == '.' 186 #if NCURSES_EXT_FUNCS 187 && !_nc_disable_period 188 #endif 189 ) { 190 dot_flag = TRUE; 191 DEBUG(8, ("dot-flag set")); 192 193 while ((ch = next_char()) == '.' || iswhite(ch)) 194 continue; 195 } 196 197 if (ch == EOF) { 198 type = EOF; 199 goto end_of_token; 200 } 201 202 /* have to make some punctuation chars legal for terminfo */ 203 if (!isalnum(ch) 204 #if NCURSES_EXT_FUNCS 205 && !(ch == '.' && _nc_disable_period) 206 #endif 207 && !strchr(terminfo_punct, (char) ch)) { 208 _nc_warning("Illegal character (expected alphanumeric or %s) - %s", 209 terminfo_punct, unctrl(ch)); 210 _nc_panic_mode(separator); 211 goto start_token; 212 } 213 214 ptr = buffer; 215 *(ptr++) = ch; 216 217 if (first_column) { 218 char *desc; 219 220 _nc_comment_start = token_start; 221 _nc_comment_end = _nc_curr_file_pos; 222 _nc_start_line = _nc_curr_line; 223 224 _nc_syntax = ERR; 225 while ((ch = next_char()) != '\n') { 226 if (ch == EOF) 227 _nc_err_abort("premature EOF"); 228 else if (ch == ':' && last_char() != ',') { 229 _nc_syntax = SYN_TERMCAP; 230 separator = ':'; 231 break; 232 } else if (ch == ',') { 233 _nc_syntax = SYN_TERMINFO; 234 separator = ','; 235 /* 236 * Fall-through here is not an accident. 237 * The idea is that if we see a comma, we 238 * figure this is terminfo unless we 239 * subsequently run into a colon -- but 240 * we don't stop looking for that colon until 241 * hitting a newline. This allows commas to 242 * be embedded in description fields of 243 * either syntax. 244 */ 245 /* FALLTHRU */ 246 } else 247 ch = eat_escaped_newline(ch); 248 249 *ptr++ = ch; 250 } 251 ptr[0] = '\0'; 252 if (_nc_syntax == ERR) { 253 /* 254 * Grrr...what we ought to do here is barf, 255 * complaining that the entry is malformed. 256 * But because a couple of name fields in the 257 * 8.2 termcap file end with |\, we just have 258 * to assume it's termcap syntax. 259 */ 260 _nc_syntax = SYN_TERMCAP; 261 separator = ':'; 262 } else if (_nc_syntax == SYN_TERMINFO) { 263 /* throw away trailing /, *$/ */ 264 for (--ptr; iswhite(*ptr) || *ptr == ','; ptr--) 265 continue; 266 ptr[1] = '\0'; 267 } 268 269 /* 270 * This is the soonest we have the terminal name 271 * fetched. Set up for following warning messages. 272 */ 273 ptr = strchr(buffer, '|'); 274 if (ptr == (char *) NULL) 275 ptr = buffer + strlen(buffer); 276 ch = *ptr; 277 *ptr = '\0'; 278 _nc_set_type(buffer); 279 *ptr = ch; 280 281 /* 282 * Compute the boundary between the aliases and the 283 * description field for syntax-checking purposes. 284 */ 285 desc = strrchr(buffer, '|'); 286 if (desc) { 287 if (*desc == '\0') 288 _nc_warning("empty longname field"); 289 #ifndef FREEBSD_NATIVE 290 else if (strchr(desc, ' ') == (char *) NULL) 291 _nc_warning("older tic versions may treat the description field as an alias"); 292 #endif 293 } 294 if (!desc) 295 desc = buffer + strlen(buffer); 296 297 /* 298 * Whitespace in a name field other than the long name 299 * can confuse rdist and some termcap tools. Slashes 300 * are a no-no. Other special characters can be 301 * dangerous due to shell expansion. 302 */ 303 for (ptr = buffer; ptr < desc; ptr++) { 304 if (isspace(*ptr)) { 305 _nc_warning("whitespace in name or alias field"); 306 break; 307 } else if (*ptr == '/') { 308 _nc_warning("slashes aren't allowed in names or aliases"); 309 break; 310 } else if (strchr("$[]!*?", *ptr)) { 311 _nc_warning("dubious character `%c' in name or alias field", *ptr); 312 break; 313 } 314 } 315 316 ptr = buffer; 317 318 _nc_curr_token.tk_name = buffer; 319 type = NAMES; 320 } else { 321 while ((ch = next_char()) != EOF) { 322 if (!isalnum(ch)) { 323 if (_nc_syntax == SYN_TERMINFO) { 324 if (ch != '_') 325 break; 326 } else { /* allow ';' for "k;" */ 327 if (ch != ';') 328 break; 329 } 330 } 331 *(ptr++) = ch; 332 } 333 334 *ptr++ = '\0'; 335 switch (ch) { 336 case ',': 337 case ':': 338 if (ch != separator) 339 _nc_err_abort("Separator inconsistent with syntax"); 340 _nc_curr_token.tk_name = buffer; 341 type = BOOLEAN; 342 break; 343 case '@': 344 if ((ch = next_char()) != separator) 345 _nc_warning("Missing separator after `%s', have %s", 346 buffer, unctrl(ch)); 347 _nc_curr_token.tk_name = buffer; 348 type = CANCEL; 349 break; 350 351 case '#': 352 found = 0; 353 while (isalnum(ch = next_char())) { 354 numbuf[found++] = ch; 355 if (found >= sizeof(numbuf) - 1) 356 break; 357 } 358 numbuf[found] = '\0'; 359 number = strtol(numbuf, &numchk, 0); 360 if (numchk == numbuf) 361 _nc_warning("no value given for `%s'", buffer); 362 if ((*numchk != '\0') || (ch != separator)) 363 _nc_warning("Missing separator"); 364 _nc_curr_token.tk_name = buffer; 365 _nc_curr_token.tk_valnumber = number; 366 type = NUMBER; 367 break; 368 369 case '=': 370 ch = _nc_trans_string(ptr, buffer + sizeof(buffer)); 371 if (ch != separator) 372 _nc_warning("Missing separator"); 373 _nc_curr_token.tk_name = buffer; 374 _nc_curr_token.tk_valstring = ptr; 375 type = STRING; 376 break; 377 378 case EOF: 379 type = EOF; 380 break; 381 default: 382 /* just to get rid of the compiler warning */ 383 type = UNDEF; 384 _nc_warning("Illegal character - %s", unctrl(ch)); 385 } 386 } /* end else (first_column == FALSE) */ 387 } /* end else (ch != EOF) */ 388 389 end_of_token: 390 391 #ifdef TRACE 392 if (dot_flag == TRUE) 393 DEBUG(8, ("Commented out ")); 394 395 if (_nc_tracing >= DEBUG_LEVEL(7)) { 396 switch (type) { 397 case BOOLEAN: 398 _tracef("Token: Boolean; name='%s'", 399 _nc_curr_token.tk_name); 400 break; 401 402 case NUMBER: 403 _tracef("Token: Number; name='%s', value=%d", 404 _nc_curr_token.tk_name, 405 _nc_curr_token.tk_valnumber); 406 break; 407 408 case STRING: 409 _tracef("Token: String; name='%s', value=%s", 410 _nc_curr_token.tk_name, 411 _nc_visbuf(_nc_curr_token.tk_valstring)); 412 break; 413 414 case CANCEL: 415 _tracef("Token: Cancel; name='%s'", 416 _nc_curr_token.tk_name); 417 break; 418 419 case NAMES: 420 421 _tracef("Token: Names; value='%s'", 422 _nc_curr_token.tk_name); 423 break; 424 425 case EOF: 426 _tracef("Token: End of file"); 427 break; 428 429 default: 430 _nc_warning("Bad token type"); 431 } 432 } 433 #endif 434 435 if (dot_flag == TRUE) /* if commented out, use the next one */ 436 type = _nc_get_token(); 437 438 DEBUG(3, ("token: `%s', class %d", _nc_curr_token.tk_name, type)); 439 440 return (type); 441 } 442 443 /* 444 * char 445 * trans_string(ptr) 446 * 447 * Reads characters using next_char() until encountering a separator, nl, 448 * or end-of-file. The returned value is the character which caused 449 * reading to stop. The following translations are done on the input: 450 * 451 * ^X goes to ctrl-X (i.e. X & 037) 452 * {\E,\n,\r,\b,\t,\f} go to 453 * {ESCAPE,newline,carriage-return,backspace,tab,formfeed} 454 * {\^,\\} go to {carat,backslash} 455 * \ddd (for ddd = up to three octal digits) goes to the character ddd 456 * 457 * \e == \E 458 * \0 == \200 459 * 460 */ 461 462 char 463 _nc_trans_string(char *ptr, char *last) 464 { 465 int count = 0; 466 int number; 467 int i, c; 468 chtype ch, last_ch = '\0'; 469 bool ignored = FALSE; 470 bool long_warning = FALSE; 471 472 while ((ch = c = next_char()) != (chtype) separator && c != EOF) { 473 if (ptr == (last - 1)) 474 break; 475 if ((_nc_syntax == SYN_TERMCAP) && c == '\n') 476 break; 477 if (ch == '^' && last_ch != '%') { 478 ch = c = next_char(); 479 if (c == EOF) 480 _nc_err_abort("Premature EOF"); 481 482 if (!(is7bits(ch) && isprint(ch))) { 483 _nc_warning("Illegal ^ character - %s", unctrl(ch)); 484 } 485 if (ch == '?') { 486 *(ptr++) = '\177'; 487 if (_nc_tracing) 488 _nc_warning("Allow ^? as synonym for \\177"); 489 } else { 490 if ((ch &= 037) == 0) 491 ch = 128; 492 *(ptr++) = (char) (ch); 493 } 494 } else if (ch == '\\') { 495 ch = c = next_char(); 496 if (c == EOF) 497 _nc_err_abort("Premature EOF"); 498 499 if (ch >= '0' && ch <= '7') { 500 number = ch - '0'; 501 for (i = 0; i < 2; i++) { 502 ch = c = next_char(); 503 if (c == EOF) 504 _nc_err_abort("Premature EOF"); 505 506 if (c < '0' || c > '7') { 507 if (isdigit(c)) { 508 _nc_warning("Non-octal digit `%c' in \\ sequence", c); 509 /* allow the digit; it'll do less harm */ 510 } else { 511 push_back((char) c); 512 break; 513 } 514 } 515 516 number = number * 8 + c - '0'; 517 } 518 519 if (number == 0) 520 number = 0200; 521 *(ptr++) = (char) number; 522 } else { 523 switch (c) { 524 case 'E': 525 case 'e': 526 *(ptr++) = '\033'; 527 break; 528 529 case 'a': 530 *(ptr++) = '\007'; 531 break; 532 533 case 'l': 534 case 'n': 535 *(ptr++) = '\n'; 536 break; 537 538 case 'r': 539 *(ptr++) = '\r'; 540 break; 541 542 case 'b': 543 *(ptr++) = '\010'; 544 break; 545 546 case 's': 547 *(ptr++) = ' '; 548 break; 549 550 case 'f': 551 *(ptr++) = '\014'; 552 break; 553 554 case 't': 555 *(ptr++) = '\t'; 556 break; 557 558 case '\\': 559 *(ptr++) = '\\'; 560 break; 561 562 case '^': 563 *(ptr++) = '^'; 564 break; 565 566 case ',': 567 *(ptr++) = ','; 568 break; 569 570 case ':': 571 *(ptr++) = ':'; 572 break; 573 574 case '\n': 575 continue; 576 577 default: 578 _nc_warning("Illegal character %s in \\ sequence", 579 unctrl(ch)); 580 *(ptr++) = (char) ch; 581 } /* endswitch (ch) */ 582 } /* endelse (ch < '0' || ch > '7') */ 583 } 584 /* end else if (ch == '\\') */ 585 else if (ch == '\n' && (_nc_syntax == SYN_TERMINFO)) { 586 /* newlines embedded in a terminfo string are ignored */ 587 ignored = TRUE; 588 } else { 589 *(ptr++) = (char) ch; 590 } 591 592 if (!ignored) { 593 last_ch = ch; 594 count++; 595 } 596 ignored = FALSE; 597 598 if (count > MAXCAPLEN && !long_warning) { 599 _nc_warning("Very long string found. Missing separator?"); 600 long_warning = TRUE; 601 } 602 } /* end while */ 603 604 *ptr = '\0'; 605 606 return (ch); 607 } 608 609 /* 610 * _nc_push_token() 611 * 612 * Push a token of given type so that it will be reread by the next 613 * get_token() call. 614 */ 615 616 void 617 _nc_push_token(int tokclass) 618 { 619 /* 620 * This implementation is kind of bogus, it will fail if we ever do 621 * more than one pushback at a time between get_token() calls. It 622 * relies on the fact that curr_tok is static storage that nothing 623 * but get_token() touches. 624 */ 625 pushtype = tokclass; 626 _nc_get_type(pushname); 627 628 DEBUG(3, ("pushing token: `%s', class %d", 629 _nc_curr_token.tk_name, pushtype)); 630 } 631 632 /* 633 * Panic mode error recovery - skip everything until a "ch" is found. 634 */ 635 void 636 _nc_panic_mode(char ch) 637 { 638 int c; 639 640 for (;;) { 641 c = next_char(); 642 if (c == ch) 643 return; 644 if (c == EOF) 645 return; 646 } 647 } 648 649 /***************************************************************************** 650 * 651 * Character-stream handling 652 * 653 *****************************************************************************/ 654 655 #define LEXBUFSIZ 1024 656 657 static char *bufptr; /* otherwise, the input buffer pointer */ 658 static char *bufstart; /* start of buffer so we can compute offsets */ 659 static FILE *yyin; /* scanner's input file descriptor */ 660 661 /* 662 * _nc_reset_input() 663 * 664 * Resets the input-reading routines. Used on initialization, 665 * or after a seek has been done. Exactly one argument must be 666 * non-null. 667 */ 668 669 void 670 _nc_reset_input(FILE * fp, char *buf) 671 { 672 pushtype = NO_PUSHBACK; 673 pushname[0] = '\0'; 674 yyin = fp; 675 bufstart = bufptr = buf; 676 _nc_curr_file_pos = 0L; 677 if (fp != 0) 678 _nc_curr_line = 0; 679 _nc_curr_col = 0; 680 } 681 682 /* 683 * int last_char() 684 * 685 * Returns the final nonblank character on the current input buffer 686 */ 687 static int 688 last_char(void) 689 { 690 size_t len = strlen(bufptr); 691 while (len--) { 692 if (!isspace(bufptr[len])) 693 return bufptr[len]; 694 } 695 return 0; 696 } 697 698 /* 699 * int next_char() 700 * 701 * Returns the next character in the input stream. Comments and leading 702 * white space are stripped. 703 * 704 * The global state variable 'firstcolumn' is set TRUE if the character 705 * returned is from the first column of the input line. 706 * 707 * The global variable _nc_curr_line is incremented for each new line. 708 * The global variable _nc_curr_file_pos is set to the file offset of the 709 * beginning of each line. 710 */ 711 712 static int 713 next_char(void) 714 { 715 if (!yyin) { 716 if (*bufptr == '\0') 717 return (EOF); 718 if (*bufptr == '\n') { 719 _nc_curr_line++; 720 _nc_curr_col = 0; 721 } 722 } else if (!bufptr || !*bufptr) { 723 /* 724 * In theory this could be recoded to do its I/O one 725 * character at a time, saving the buffer space. In 726 * practice, this turns out to be quite hard to get 727 * completely right. Try it and see. If you succeed, 728 * don't forget to hack push_back() correspondingly. 729 */ 730 static char line[LEXBUFSIZ]; 731 size_t len; 732 733 do { 734 _nc_curr_file_pos = ftell(yyin); 735 736 if ((bufstart = fgets(line, LEXBUFSIZ, yyin)) != NULL) { 737 _nc_curr_line++; 738 _nc_curr_col = 0; 739 } 740 bufptr = bufstart; 741 } while 742 (bufstart != NULL && line[0] == '#'); 743 744 if (bufstart == NULL || *bufstart == 0) 745 return (EOF); 746 747 while (iswhite(*bufptr)) 748 bufptr++; 749 750 /* 751 * Treat a trailing <cr><lf> the same as a <newline> so we can read 752 * files on OS/2, etc. 753 */ 754 if ((len = strlen(bufptr)) > 1) { 755 if (bufptr[len - 1] == '\n' 756 && bufptr[len - 2] == '\r') { 757 len--; 758 bufptr[len - 1] = '\n'; 759 bufptr[len] = '\0'; 760 } 761 } 762 763 /* 764 * If we don't have a trailing newline, it's because the line is simply 765 * too long. Give up. (FIXME: We could instead reallocate the line 766 * buffer and allow arbitrary-length lines). 767 */ 768 if (len == 0 || (bufptr[len - 1] != '\n')) 769 return (EOF); 770 } 771 772 first_column = (bufptr == bufstart); 773 774 _nc_curr_col++; 775 return (*bufptr++); 776 } 777 778 static void 779 push_back(char c) 780 /* push a character back onto the input stream */ 781 { 782 if (bufptr == bufstart) 783 _nc_syserr_abort("Can't backspace off beginning of line"); 784 *--bufptr = c; 785 } 786 787 static long 788 stream_pos(void) 789 /* return our current character position in the input stream */ 790 { 791 return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0)); 792 } 793 794 static bool 795 end_of_stream(void) 796 /* are we at end of input? */ 797 { 798 return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0')) 799 ? TRUE : FALSE); 800 } 801 802 /* comp_scan.c ends here */ 803