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 /* 35 * comp_scan.c --- Lexical scanner for terminfo compiler. 36 * 37 * _nc_reset_input() 38 * _nc_get_token() 39 * _nc_panic_mode() 40 * int _nc_syntax; 41 * int _nc_curr_line; 42 * long _nc_curr_file_pos; 43 * long _nc_comment_start; 44 * long _nc_comment_end; 45 */ 46 47 #include <curses.priv.h> 48 49 #include <ctype.h> 50 #include <term_entry.h> 51 #include <tic.h> 52 53 MODULE_ID("$From: comp_scan.c,v 1.52 2000/12/10 02:55:07 tom Exp $") 54 55 /* 56 * Maximum length of string capability we'll accept before raising an error. 57 * Yes, there is a real capability in /etc/termcap this long, an "is". 58 */ 59 #define MAXCAPLEN 600 60 61 #define iswhite(ch) (ch == ' ' || ch == '\t') 62 63 NCURSES_EXPORT_VAR(int) 64 _nc_syntax = 0; /* termcap or terminfo? */ 65 NCURSES_EXPORT_VAR(long) 66 _nc_curr_file_pos = 0; /* file offset of current line */ 67 NCURSES_EXPORT_VAR(long) 68 _nc_comment_start = 0; /* start of comment range before name */ 69 NCURSES_EXPORT_VAR(long) 70 _nc_comment_end = 0; /* end of comment range before name */ 71 NCURSES_EXPORT_VAR(long) 72 _nc_start_line = 0; /* start line of current entry */ 73 74 NCURSES_EXPORT_VAR(struct token) 75 _nc_curr_token = 76 { 77 0, 0, 0 78 }; 79 80 /***************************************************************************** 81 * 82 * Token-grabbing machinery 83 * 84 *****************************************************************************/ 85 86 static bool first_column; /* See 'next_char()' below */ 87 static char separator; /* capability separator */ 88 static int pushtype; /* type of pushback token */ 89 static char pushname[MAX_NAME_SIZE + 1]; 90 91 #if NCURSES_EXT_FUNCS 92 NCURSES_EXPORT_VAR(bool) _nc_disable_period = FALSE; /* used by tic -a option */ 93 #endif 94 95 static int last_char(void); 96 static int next_char(void); 97 static long stream_pos(void); 98 static bool end_of_stream(void); 99 static void push_back(char c); 100 101 /* Assume we may be looking at a termcap-style continuation */ 102 static inline int 103 eat_escaped_newline(int ch) 104 { 105 if (ch == '\\') 106 while ((ch = next_char()) == '\n' || iswhite(ch)) 107 continue; 108 return ch; 109 } 110 111 /* 112 * int 113 * get_token() 114 * 115 * Scans the input for the next token, storing the specifics in the 116 * global structure 'curr_token' and returning one of the following: 117 * 118 * NAMES A line beginning in column 1. 'name' 119 * will be set to point to everything up to but 120 * not including the first separator on the line. 121 * BOOLEAN An entry consisting of a name followed by 122 * a separator. 'name' will be set to point to 123 * the name of the capability. 124 * NUMBER An entry of the form 125 * name#digits, 126 * 'name' will be set to point to the capability 127 * name and 'valnumber' to the number given. 128 * STRING An entry of the form 129 * name=characters, 130 * 'name' is set to the capability name and 131 * 'valstring' to the string of characters, with 132 * input translations done. 133 * CANCEL An entry of the form 134 * name@, 135 * 'name' is set to the capability name and 136 * 'valnumber' to -1. 137 * EOF The end of the file has been reached. 138 * 139 * A `separator' is either a comma or a semicolon, depending on whether 140 * we are in termcap or terminfo mode. 141 * 142 */ 143 144 NCURSES_EXPORT(int) 145 _nc_get_token(void) 146 { 147 static const char terminfo_punct[] = "@%&*!#"; 148 long number; 149 int type; 150 int ch; 151 char *numchk; 152 char numbuf[80]; 153 unsigned found; 154 static char buffer[MAX_ENTRY_SIZE]; 155 char *ptr; 156 int dot_flag = FALSE; 157 long token_start; 158 159 if (pushtype != NO_PUSHBACK) { 160 int retval = pushtype; 161 162 _nc_set_type(pushname); 163 DEBUG(3, ("pushed-back token: `%s', class %d", 164 _nc_curr_token.tk_name, pushtype)); 165 166 pushtype = NO_PUSHBACK; 167 pushname[0] = '\0'; 168 169 /* currtok wasn't altered by _nc_push_token() */ 170 return (retval); 171 } 172 173 if (end_of_stream()) 174 return (EOF); 175 176 start_token: 177 token_start = stream_pos(); 178 while ((ch = next_char()) == '\n' || iswhite(ch)) 179 continue; 180 181 ch = eat_escaped_newline(ch); 182 183 if (ch == EOF) 184 type = EOF; 185 else { 186 /* if this is a termcap entry, skip a leading separator */ 187 if (separator == ':' && ch == ':') 188 ch = next_char(); 189 190 if (ch == '.' 191 #if NCURSES_EXT_FUNCS 192 && !_nc_disable_period 193 #endif 194 ) { 195 dot_flag = TRUE; 196 DEBUG(8, ("dot-flag set")); 197 198 while ((ch = next_char()) == '.' || iswhite(ch)) 199 continue; 200 } 201 202 if (ch == EOF) { 203 type = EOF; 204 goto end_of_token; 205 } 206 207 /* have to make some punctuation chars legal for terminfo */ 208 if (!isalnum(ch) 209 #if NCURSES_EXT_FUNCS 210 && !(ch == '.' && _nc_disable_period) 211 #endif 212 && !strchr(terminfo_punct, (char) ch)) { 213 _nc_warning("Illegal character (expected alphanumeric or %s) - %s", 214 terminfo_punct, unctrl((chtype) ch)); 215 _nc_panic_mode(separator); 216 goto start_token; 217 } 218 219 ptr = buffer; 220 *(ptr++) = ch; 221 222 if (first_column) { 223 char *desc; 224 225 _nc_comment_start = token_start; 226 _nc_comment_end = _nc_curr_file_pos; 227 _nc_start_line = _nc_curr_line; 228 229 _nc_syntax = ERR; 230 while ((ch = next_char()) != '\n') { 231 if (ch == EOF) 232 _nc_err_abort("premature EOF"); 233 else if (ch == ':' && last_char() != ',') { 234 _nc_syntax = SYN_TERMCAP; 235 separator = ':'; 236 break; 237 } else if (ch == ',') { 238 _nc_syntax = SYN_TERMINFO; 239 separator = ','; 240 /* 241 * Fall-through here is not an accident. 242 * The idea is that if we see a comma, we 243 * figure this is terminfo unless we 244 * subsequently run into a colon -- but 245 * we don't stop looking for that colon until 246 * hitting a newline. This allows commas to 247 * be embedded in description fields of 248 * either syntax. 249 */ 250 /* FALLTHRU */ 251 } else 252 ch = eat_escaped_newline(ch); 253 254 *ptr++ = ch; 255 } 256 ptr[0] = '\0'; 257 if (_nc_syntax == ERR) { 258 /* 259 * Grrr...what we ought to do here is barf, 260 * complaining that the entry is malformed. 261 * But because a couple of name fields in the 262 * 8.2 termcap file end with |\, we just have 263 * to assume it's termcap syntax. 264 */ 265 _nc_syntax = SYN_TERMCAP; 266 separator = ':'; 267 } else if (_nc_syntax == SYN_TERMINFO) { 268 /* throw away trailing /, *$/ */ 269 for (--ptr; iswhite(*ptr) || *ptr == ','; ptr--) 270 continue; 271 ptr[1] = '\0'; 272 } 273 274 /* 275 * This is the soonest we have the terminal name 276 * fetched. Set up for following warning messages. 277 */ 278 ptr = strchr(buffer, '|'); 279 if (ptr == (char *) NULL) 280 ptr = buffer + strlen(buffer); 281 ch = *ptr; 282 *ptr = '\0'; 283 _nc_set_type(buffer); 284 *ptr = ch; 285 286 /* 287 * Compute the boundary between the aliases and the 288 * description field for syntax-checking purposes. 289 */ 290 desc = strrchr(buffer, '|'); 291 if (desc) { 292 if (*desc == '\0') 293 _nc_warning("empty longname field"); 294 else if (strchr(desc, ' ') == (char *) NULL) 295 _nc_warning("older tic versions may treat the description field as an alias"); 296 } 297 if (!desc) 298 desc = buffer + strlen(buffer); 299 300 /* 301 * Whitespace in a name field other than the long name 302 * can confuse rdist and some termcap tools. Slashes 303 * are a no-no. Other special characters can be 304 * dangerous due to shell expansion. 305 */ 306 for (ptr = buffer; ptr < desc; ptr++) { 307 if (isspace(CharOf(*ptr))) { 308 _nc_warning("whitespace in name or alias field"); 309 break; 310 } else if (*ptr == '/') { 311 _nc_warning("slashes aren't allowed in names or aliases"); 312 break; 313 } else if (strchr("$[]!*?", *ptr)) { 314 _nc_warning("dubious character `%c' in name or alias field", *ptr); 315 break; 316 } 317 } 318 319 ptr = buffer; 320 321 _nc_curr_token.tk_name = buffer; 322 type = NAMES; 323 } else { 324 while ((ch = next_char()) != EOF) { 325 if (!isalnum(ch)) { 326 if (_nc_syntax == SYN_TERMINFO) { 327 if (ch != '_') 328 break; 329 } else { /* allow ';' for "k;" */ 330 if (ch != ';') 331 break; 332 } 333 } 334 *(ptr++) = ch; 335 } 336 337 *ptr++ = '\0'; 338 switch (ch) { 339 case ',': 340 case ':': 341 if (ch != separator) 342 _nc_err_abort("Separator inconsistent with syntax"); 343 _nc_curr_token.tk_name = buffer; 344 type = BOOLEAN; 345 break; 346 case '@': 347 if ((ch = next_char()) != separator) 348 _nc_warning("Missing separator after `%s', have %s", 349 buffer, unctrl((chtype) ch)); 350 _nc_curr_token.tk_name = buffer; 351 type = CANCEL; 352 break; 353 354 case '#': 355 found = 0; 356 while (isalnum(ch = next_char())) { 357 numbuf[found++] = ch; 358 if (found >= sizeof(numbuf) - 1) 359 break; 360 } 361 numbuf[found] = '\0'; 362 number = strtol(numbuf, &numchk, 0); 363 if (numchk == numbuf) 364 _nc_warning("no value given for `%s'", buffer); 365 if ((*numchk != '\0') || (ch != separator)) 366 _nc_warning("Missing separator"); 367 _nc_curr_token.tk_name = buffer; 368 _nc_curr_token.tk_valnumber = number; 369 type = NUMBER; 370 break; 371 372 case '=': 373 ch = _nc_trans_string(ptr, buffer + sizeof(buffer)); 374 if (ch != separator) 375 _nc_warning("Missing separator"); 376 _nc_curr_token.tk_name = buffer; 377 _nc_curr_token.tk_valstring = ptr; 378 type = STRING; 379 break; 380 381 case EOF: 382 type = EOF; 383 break; 384 default: 385 /* just to get rid of the compiler warning */ 386 type = UNDEF; 387 _nc_warning("Illegal character - %s", unctrl((chtype) ch)); 388 } 389 } /* end else (first_column == FALSE) */ 390 } /* end else (ch != EOF) */ 391 392 end_of_token: 393 394 #ifdef TRACE 395 if (dot_flag == TRUE) 396 DEBUG(8, ("Commented out ")); 397 398 if (_nc_tracing >= DEBUG_LEVEL(7)) { 399 switch (type) { 400 case BOOLEAN: 401 _tracef("Token: Boolean; name='%s'", 402 _nc_curr_token.tk_name); 403 break; 404 405 case NUMBER: 406 _tracef("Token: Number; name='%s', value=%d", 407 _nc_curr_token.tk_name, 408 _nc_curr_token.tk_valnumber); 409 break; 410 411 case STRING: 412 _tracef("Token: String; name='%s', value=%s", 413 _nc_curr_token.tk_name, 414 _nc_visbuf(_nc_curr_token.tk_valstring)); 415 break; 416 417 case CANCEL: 418 _tracef("Token: Cancel; name='%s'", 419 _nc_curr_token.tk_name); 420 break; 421 422 case NAMES: 423 424 _tracef("Token: Names; value='%s'", 425 _nc_curr_token.tk_name); 426 break; 427 428 case EOF: 429 _tracef("Token: End of file"); 430 break; 431 432 default: 433 _nc_warning("Bad token type"); 434 } 435 } 436 #endif 437 438 if (dot_flag == TRUE) /* if commented out, use the next one */ 439 type = _nc_get_token(); 440 441 DEBUG(3, ("token: `%s', class %d", _nc_curr_token.tk_name, type)); 442 443 return (type); 444 } 445 446 /* 447 * char 448 * trans_string(ptr) 449 * 450 * Reads characters using next_char() until encountering a separator, nl, 451 * or end-of-file. The returned value is the character which caused 452 * reading to stop. The following translations are done on the input: 453 * 454 * ^X goes to ctrl-X (i.e. X & 037) 455 * {\E,\n,\r,\b,\t,\f} go to 456 * {ESCAPE,newline,carriage-return,backspace,tab,formfeed} 457 * {\^,\\} go to {carat,backslash} 458 * \ddd (for ddd = up to three octal digits) goes to the character ddd 459 * 460 * \e == \E 461 * \0 == \200 462 * 463 */ 464 465 NCURSES_EXPORT(char) 466 _nc_trans_string(char *ptr, char *last) 467 { 468 int count = 0; 469 int number = 0; 470 int i, c; 471 chtype ch, last_ch = '\0'; 472 bool ignored = FALSE; 473 bool long_warning = FALSE; 474 475 while ((ch = c = next_char()) != (chtype) separator && c != EOF) { 476 if (ptr == (last - 1)) 477 break; 478 if ((_nc_syntax == SYN_TERMCAP) && c == '\n') 479 break; 480 if (ch == '^' && last_ch != '%') { 481 ch = c = next_char(); 482 if (c == EOF) 483 _nc_err_abort("Premature EOF"); 484 485 if (!(is7bits(ch) && isprint(ch))) { 486 _nc_warning("Illegal ^ character - %s", unctrl(ch)); 487 } 488 if (ch == '?') { 489 *(ptr++) = '\177'; 490 if (_nc_tracing) 491 _nc_warning("Allow ^? as synonym for \\177"); 492 } else { 493 if ((ch &= 037) == 0) 494 ch = 128; 495 *(ptr++) = (char) (ch); 496 } 497 } else if (ch == '\\') { 498 ch = c = next_char(); 499 if (c == EOF) 500 _nc_err_abort("Premature EOF"); 501 502 if (ch >= '0' && ch <= '7') { 503 number = ch - '0'; 504 for (i = 0; i < 2; i++) { 505 ch = c = next_char(); 506 if (c == EOF) 507 _nc_err_abort("Premature EOF"); 508 509 if (c < '0' || c > '7') { 510 if (isdigit(c)) { 511 _nc_warning("Non-octal digit `%c' in \\ sequence", c); 512 /* allow the digit; it'll do less harm */ 513 } else { 514 push_back((char) c); 515 break; 516 } 517 } 518 519 number = number * 8 + c - '0'; 520 } 521 522 if (number == 0) 523 number = 0200; 524 *(ptr++) = (char) number; 525 } else { 526 switch (c) { 527 case 'E': 528 case 'e': 529 *(ptr++) = '\033'; 530 break; 531 532 case 'a': 533 *(ptr++) = '\007'; 534 break; 535 536 case 'l': 537 case 'n': 538 *(ptr++) = '\n'; 539 break; 540 541 case 'r': 542 *(ptr++) = '\r'; 543 break; 544 545 case 'b': 546 *(ptr++) = '\010'; 547 break; 548 549 case 's': 550 *(ptr++) = ' '; 551 break; 552 553 case 'f': 554 *(ptr++) = '\014'; 555 break; 556 557 case 't': 558 *(ptr++) = '\t'; 559 break; 560 561 case '\\': 562 *(ptr++) = '\\'; 563 break; 564 565 case '^': 566 *(ptr++) = '^'; 567 break; 568 569 case ',': 570 *(ptr++) = ','; 571 break; 572 573 case ':': 574 *(ptr++) = ':'; 575 break; 576 577 case '\n': 578 continue; 579 580 default: 581 _nc_warning("Illegal character %s in \\ sequence", 582 unctrl(ch)); 583 *(ptr++) = (char) ch; 584 } /* endswitch (ch) */ 585 } /* endelse (ch < '0' || ch > '7') */ 586 } 587 /* end else if (ch == '\\') */ 588 else if (ch == '\n' && (_nc_syntax == SYN_TERMINFO)) { 589 /* newlines embedded in a terminfo string are ignored */ 590 ignored = TRUE; 591 } else { 592 *(ptr++) = (char) ch; 593 } 594 595 if (!ignored) { 596 last_ch = ch; 597 count++; 598 } 599 ignored = FALSE; 600 601 if (count > MAXCAPLEN && !long_warning) { 602 _nc_warning("Very long string found. Missing separator?"); 603 long_warning = TRUE; 604 } 605 } /* end while */ 606 607 *ptr = '\0'; 608 609 return (ch); 610 } 611 612 /* 613 * _nc_push_token() 614 * 615 * Push a token of given type so that it will be reread by the next 616 * get_token() call. 617 */ 618 619 NCURSES_EXPORT(void) 620 _nc_push_token(int tokclass) 621 { 622 /* 623 * This implementation is kind of bogus, it will fail if we ever do 624 * more than one pushback at a time between get_token() calls. It 625 * relies on the fact that curr_tok is static storage that nothing 626 * but get_token() touches. 627 */ 628 pushtype = tokclass; 629 _nc_get_type(pushname); 630 631 DEBUG(3, ("pushing token: `%s', class %d", 632 _nc_curr_token.tk_name, pushtype)); 633 } 634 635 /* 636 * Panic mode error recovery - skip everything until a "ch" is found. 637 */ 638 NCURSES_EXPORT(void) 639 _nc_panic_mode(char ch) 640 { 641 int c; 642 643 for (;;) { 644 c = next_char(); 645 if (c == ch) 646 return; 647 if (c == EOF) 648 return; 649 } 650 } 651 652 /***************************************************************************** 653 * 654 * Character-stream handling 655 * 656 *****************************************************************************/ 657 658 #define LEXBUFSIZ 1024 659 660 static char *bufptr; /* otherwise, the input buffer pointer */ 661 static char *bufstart; /* start of buffer so we can compute offsets */ 662 static FILE *yyin; /* scanner's input file descriptor */ 663 664 /* 665 * _nc_reset_input() 666 * 667 * Resets the input-reading routines. Used on initialization, 668 * or after a seek has been done. Exactly one argument must be 669 * non-null. 670 */ 671 672 NCURSES_EXPORT(void) 673 _nc_reset_input(FILE * fp, char *buf) 674 { 675 pushtype = NO_PUSHBACK; 676 pushname[0] = '\0'; 677 yyin = fp; 678 bufstart = bufptr = buf; 679 _nc_curr_file_pos = 0L; 680 if (fp != 0) 681 _nc_curr_line = 0; 682 _nc_curr_col = 0; 683 } 684 685 /* 686 * int last_char() 687 * 688 * Returns the final nonblank character on the current input buffer 689 */ 690 static int 691 last_char(void) 692 { 693 size_t len = strlen(bufptr); 694 while (len--) { 695 if (!isspace(CharOf(bufptr[len]))) 696 return bufptr[len]; 697 } 698 return 0; 699 } 700 701 /* 702 * int next_char() 703 * 704 * Returns the next character in the input stream. Comments and leading 705 * white space are stripped. 706 * 707 * The global state variable 'firstcolumn' is set TRUE if the character 708 * returned is from the first column of the input line. 709 * 710 * The global variable _nc_curr_line is incremented for each new line. 711 * The global variable _nc_curr_file_pos is set to the file offset of the 712 * beginning of each line. 713 */ 714 715 static int 716 next_char(void) 717 { 718 if (!yyin) { 719 if (*bufptr == '\0') 720 return (EOF); 721 if (*bufptr == '\n') { 722 _nc_curr_line++; 723 _nc_curr_col = 0; 724 } 725 } else if (!bufptr || !*bufptr) { 726 /* 727 * In theory this could be recoded to do its I/O one 728 * character at a time, saving the buffer space. In 729 * practice, this turns out to be quite hard to get 730 * completely right. Try it and see. If you succeed, 731 * don't forget to hack push_back() correspondingly. 732 */ 733 static char line[LEXBUFSIZ]; 734 size_t len; 735 736 do { 737 _nc_curr_file_pos = ftell(yyin); 738 739 if ((bufstart = fgets(line, LEXBUFSIZ, yyin)) != NULL) { 740 _nc_curr_line++; 741 _nc_curr_col = 0; 742 } 743 bufptr = bufstart; 744 } while 745 (bufstart != NULL && line[0] == '#'); 746 747 if (bufstart == NULL || *bufstart == 0) 748 return (EOF); 749 750 while (iswhite(*bufptr)) 751 bufptr++; 752 753 /* 754 * Treat a trailing <cr><lf> the same as a <newline> so we can read 755 * files on OS/2, etc. 756 */ 757 if ((len = strlen(bufptr)) > 1) { 758 if (bufptr[len - 1] == '\n' 759 && bufptr[len - 2] == '\r') { 760 len--; 761 bufptr[len - 1] = '\n'; 762 bufptr[len] = '\0'; 763 } 764 } 765 766 /* 767 * If we don't have a trailing newline, it's because the line is simply 768 * too long. Give up. (FIXME: We could instead reallocate the line 769 * buffer and allow arbitrary-length lines). 770 */ 771 if (len == 0 || (bufptr[len - 1] != '\n')) 772 return (EOF); 773 } 774 775 first_column = (bufptr == bufstart); 776 777 _nc_curr_col++; 778 return (*bufptr++); 779 } 780 781 static void 782 push_back(char c) 783 /* push a character back onto the input stream */ 784 { 785 if (bufptr == bufstart) 786 _nc_syserr_abort("Can't backspace off beginning of line"); 787 *--bufptr = c; 788 } 789 790 static long 791 stream_pos(void) 792 /* return our current character position in the input stream */ 793 { 794 return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0)); 795 } 796 797 static bool 798 end_of_stream(void) 799 /* are we at end of input? */ 800 { 801 return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0')) 802 ? TRUE : FALSE); 803 } 804 805 /* comp_scan.c ends here */ 806