1 /**************************************************************************** 2 * Copyright (c) 1998-2012,2013 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 * and: Thomas E. Dickey 1996 on * 33 ****************************************************************************/ 34 35 /* 36 * comp_scan.c --- Lexical scanner for terminfo compiler. 37 * 38 * _nc_reset_input() 39 * _nc_get_token() 40 * _nc_panic_mode() 41 * int _nc_syntax; 42 * int _nc_curr_line; 43 * long _nc_curr_file_pos; 44 * long _nc_comment_start; 45 * long _nc_comment_end; 46 */ 47 48 #include <curses.priv.h> 49 50 #include <ctype.h> 51 #include <tic.h> 52 53 MODULE_ID("$Id: comp_scan.c,v 1.102 2013/11/16 19:57:50 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) _nc_syntax = 0; /* termcap or terminfo? */ 64 NCURSES_EXPORT_VAR (int) _nc_strict_bsd = 1; /* ncurses extended termcap? */ 65 NCURSES_EXPORT_VAR (long) _nc_curr_file_pos = 0; /* file offset of current line */ 66 NCURSES_EXPORT_VAR (long) _nc_comment_start = 0; /* start of comment range before name */ 67 NCURSES_EXPORT_VAR (long) _nc_comment_end = 0; /* end of comment range before name */ 68 NCURSES_EXPORT_VAR (long) _nc_start_line = 0; /* start line of current entry */ 69 70 NCURSES_EXPORT_VAR (struct token) _nc_curr_token = 71 { 72 0, 0, 0 73 }; 74 75 /***************************************************************************** 76 * 77 * Token-grabbing machinery 78 * 79 *****************************************************************************/ 80 81 static bool first_column; /* See 'next_char()' below */ 82 static bool had_newline; 83 static char separator; /* capability separator */ 84 static int pushtype; /* type of pushback token */ 85 static char *pushname; 86 87 #if NCURSES_EXT_FUNCS 88 NCURSES_EXPORT_VAR (bool) _nc_disable_period = FALSE; /* used by tic -a option */ 89 #endif 90 91 /***************************************************************************** 92 * 93 * Character-stream handling 94 * 95 *****************************************************************************/ 96 97 #define LEXBUFSIZ 1024 98 99 static char *bufptr; /* otherwise, the input buffer pointer */ 100 static char *bufstart; /* start of buffer so we can compute offsets */ 101 static FILE *yyin; /* scanner's input file descriptor */ 102 103 /* 104 * _nc_reset_input() 105 * 106 * Resets the input-reading routines. Used on initialization, 107 * or after a seek has been done. Exactly one argument must be 108 * non-null. 109 */ 110 111 NCURSES_EXPORT(void) 112 _nc_reset_input(FILE *fp, char *buf) 113 { 114 pushtype = NO_PUSHBACK; 115 if (pushname != 0) 116 pushname[0] = '\0'; 117 yyin = fp; 118 bufstart = bufptr = buf; 119 _nc_curr_file_pos = 0L; 120 if (fp != 0) 121 _nc_curr_line = 0; 122 _nc_curr_col = 0; 123 } 124 125 /* 126 * int last_char() 127 * 128 * Returns the final nonblank character on the current input buffer 129 */ 130 static int 131 last_char(int from_end) 132 { 133 size_t len = strlen(bufptr); 134 int result = 0; 135 136 while (len--) { 137 if (!isspace(UChar(bufptr[len]))) { 138 if (from_end < (int) len) 139 result = bufptr[(int) len - from_end]; 140 break; 141 } 142 } 143 return result; 144 } 145 146 /* 147 * int next_char() 148 * 149 * Returns the next character in the input stream. Comments and leading 150 * white space are stripped. 151 * 152 * The global state variable 'firstcolumn' is set TRUE if the character 153 * returned is from the first column of the input line. 154 * 155 * The global variable _nc_curr_line is incremented for each new line. 156 * The global variable _nc_curr_file_pos is set to the file offset of the 157 * beginning of each line. 158 */ 159 160 static int 161 next_char(void) 162 { 163 static char *result; 164 static size_t allocated; 165 int the_char; 166 167 if (!yyin) { 168 if (result != 0) { 169 FreeAndNull(result); 170 FreeAndNull(pushname); 171 allocated = 0; 172 } 173 /* 174 * An string with an embedded null will truncate the input. This is 175 * intentional (we don't read binary files here). 176 */ 177 if (bufptr == 0 || *bufptr == '\0') 178 return (EOF); 179 if (*bufptr == '\n') { 180 _nc_curr_line++; 181 _nc_curr_col = 0; 182 } else if (*bufptr == '\t') { 183 _nc_curr_col = (_nc_curr_col | 7); 184 } 185 } else if (!bufptr || !*bufptr) { 186 /* 187 * In theory this could be recoded to do its I/O one character at a 188 * time, saving the buffer space. In practice, this turns out to be 189 * quite hard to get completely right. Try it and see. If you 190 * succeed, don't forget to hack push_back() correspondingly. 191 */ 192 size_t used; 193 size_t len; 194 195 do { 196 bufstart = 0; 197 used = 0; 198 do { 199 if (used + (LEXBUFSIZ / 4) >= allocated) { 200 allocated += (allocated + LEXBUFSIZ); 201 result = typeRealloc(char, allocated, result); 202 if (result == 0) 203 return (EOF); 204 if (bufstart) 205 bufstart = result; 206 } 207 if (used == 0) 208 _nc_curr_file_pos = ftell(yyin); 209 210 if (fgets(result + used, (int) (allocated - used), yyin) != 0) { 211 bufstart = result; 212 if (used == 0) { 213 if (_nc_curr_line == 0 214 && IS_TIC_MAGIC(result)) { 215 _nc_err_abort("This is a compiled terminal description, not a source"); 216 } 217 _nc_curr_line++; 218 _nc_curr_col = 0; 219 } 220 } else { 221 if (used != 0) 222 _nc_STRCAT(result, "\n", allocated); 223 } 224 if ((bufptr = bufstart) != 0) { 225 used = strlen(bufptr); 226 while (iswhite(*bufptr)) { 227 if (*bufptr == '\t') { 228 _nc_curr_col = (_nc_curr_col | 7) + 1; 229 } else { 230 _nc_curr_col++; 231 } 232 bufptr++; 233 } 234 235 /* 236 * Treat a trailing <cr><lf> the same as a <newline> so we 237 * can read files on OS/2, etc. 238 */ 239 if ((len = strlen(bufptr)) > 1) { 240 if (bufptr[len - 1] == '\n' 241 && bufptr[len - 2] == '\r') { 242 len--; 243 bufptr[len - 1] = '\n'; 244 bufptr[len] = '\0'; 245 } 246 } 247 } else { 248 return (EOF); 249 } 250 } while (bufptr[len - 1] != '\n'); /* complete a line */ 251 } while (result[0] == '#'); /* ignore comments */ 252 } else if (*bufptr == '\t') { 253 _nc_curr_col = (_nc_curr_col | 7); 254 } 255 256 first_column = (bufptr == bufstart); 257 if (first_column) 258 had_newline = FALSE; 259 260 _nc_curr_col++; 261 the_char = *bufptr++; 262 return UChar(the_char); 263 } 264 265 static void 266 push_back(int c) 267 /* push a character back onto the input stream */ 268 { 269 if (bufptr == bufstart) 270 _nc_syserr_abort("Can't backspace off beginning of line"); 271 *--bufptr = (char) c; 272 _nc_curr_col--; 273 } 274 275 static long 276 stream_pos(void) 277 /* return our current character position in the input stream */ 278 { 279 return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0)); 280 } 281 282 static bool 283 end_of_stream(void) 284 /* are we at end of input? */ 285 { 286 return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0')) 287 ? TRUE : FALSE); 288 } 289 290 /* Assume we may be looking at a termcap-style continuation */ 291 static NCURSES_INLINE int 292 eat_escaped_newline(int ch) 293 { 294 if (ch == '\\') 295 while ((ch = next_char()) == '\n' || iswhite(ch)) 296 continue; 297 return ch; 298 } 299 300 #define TOK_BUF_SIZE MAX_ENTRY_SIZE 301 302 #define OkToAdd() \ 303 ((tok_ptr - tok_buf) < (TOK_BUF_SIZE - 2)) 304 305 #define AddCh(ch) \ 306 *tok_ptr++ = (char) ch; \ 307 *tok_ptr = '\0' 308 309 static char *tok_buf; 310 311 /* 312 * int 313 * get_token() 314 * 315 * Scans the input for the next token, storing the specifics in the 316 * global structure 'curr_token' and returning one of the following: 317 * 318 * NAMES A line beginning in column 1. 'name' 319 * will be set to point to everything up to but 320 * not including the first separator on the line. 321 * BOOLEAN An entry consisting of a name followed by 322 * a separator. 'name' will be set to point to 323 * the name of the capability. 324 * NUMBER An entry of the form 325 * name#digits, 326 * 'name' will be set to point to the capability 327 * name and 'valnumber' to the number given. 328 * STRING An entry of the form 329 * name=characters, 330 * 'name' is set to the capability name and 331 * 'valstring' to the string of characters, with 332 * input translations done. 333 * CANCEL An entry of the form 334 * name@, 335 * 'name' is set to the capability name and 336 * 'valnumber' to -1. 337 * EOF The end of the file has been reached. 338 * 339 * A `separator' is either a comma or a semicolon, depending on whether 340 * we are in termcap or terminfo mode. 341 * 342 */ 343 344 NCURSES_EXPORT(int) 345 _nc_get_token(bool silent) 346 { 347 static const char terminfo_punct[] = "@%&*!#"; 348 349 char *after_name; /* after primary name */ 350 char *after_list; /* after primary and alias list */ 351 char *numchk; 352 char *tok_ptr; 353 char *s; 354 char numbuf[80]; 355 int ch, c0, c1; 356 int dot_flag = FALSE; 357 int type; 358 long number; 359 long token_start; 360 unsigned found; 361 #ifdef TRACE 362 int old_line; 363 int old_col; 364 #endif 365 366 if (pushtype != NO_PUSHBACK) { 367 int retval = pushtype; 368 369 _nc_set_type(pushname != 0 ? pushname : ""); 370 DEBUG(3, ("pushed-back token: `%s', class %d", 371 _nc_curr_token.tk_name, pushtype)); 372 373 pushtype = NO_PUSHBACK; 374 if (pushname != 0) 375 pushname[0] = '\0'; 376 377 /* currtok wasn't altered by _nc_push_token() */ 378 return (retval); 379 } 380 381 if (end_of_stream()) { 382 yyin = 0; 383 (void) next_char(); /* frees its allocated memory */ 384 if (tok_buf != 0) { 385 if (_nc_curr_token.tk_name == tok_buf) 386 _nc_curr_token.tk_name = 0; 387 } 388 return (EOF); 389 } 390 391 start_token: 392 token_start = stream_pos(); 393 while ((ch = next_char()) == '\n' || iswhite(ch)) { 394 if (ch == '\n') 395 had_newline = TRUE; 396 continue; 397 } 398 399 ch = eat_escaped_newline(ch); 400 _nc_curr_token.tk_valstring = 0; 401 402 #ifdef TRACE 403 old_line = _nc_curr_line; 404 old_col = _nc_curr_col; 405 #endif 406 if (ch == EOF) 407 type = EOF; 408 else { 409 /* if this is a termcap entry, skip a leading separator */ 410 if (separator == ':' && ch == ':') 411 ch = next_char(); 412 413 if (ch == '.' 414 #if NCURSES_EXT_FUNCS 415 && !_nc_disable_period 416 #endif 417 ) { 418 dot_flag = TRUE; 419 DEBUG(8, ("dot-flag set")); 420 421 while ((ch = next_char()) == '.' || iswhite(ch)) 422 continue; 423 } 424 425 if (ch == EOF) { 426 type = EOF; 427 goto end_of_token; 428 } 429 430 /* have to make some punctuation chars legal for terminfo */ 431 if (!isalnum(UChar(ch)) 432 #if NCURSES_EXT_FUNCS 433 && !(ch == '.' && _nc_disable_period) 434 #endif 435 && ((strchr) (terminfo_punct, (char) ch) == 0)) { 436 if (!silent) 437 _nc_warning("Illegal character (expected alphanumeric or %s) - '%s'", 438 terminfo_punct, unctrl(UChar(ch))); 439 _nc_panic_mode(separator); 440 goto start_token; 441 } 442 443 if (tok_buf == 0) 444 tok_buf = typeMalloc(char, TOK_BUF_SIZE); 445 446 #ifdef TRACE 447 old_line = _nc_curr_line; 448 old_col = _nc_curr_col; 449 #endif 450 tok_ptr = tok_buf; 451 AddCh(ch); 452 453 if (first_column) { 454 _nc_comment_start = token_start; 455 _nc_comment_end = _nc_curr_file_pos; 456 _nc_start_line = _nc_curr_line; 457 458 _nc_syntax = ERR; 459 after_name = 0; 460 after_list = 0; 461 while ((ch = next_char()) != '\n') { 462 if (ch == EOF) { 463 _nc_err_abort(MSG_NO_INPUTS); 464 } else if (ch == '|') { 465 after_list = tok_ptr; 466 if (after_name == 0) 467 after_name = tok_ptr; 468 } else if (ch == ':' && last_char(0) != ',') { 469 _nc_syntax = SYN_TERMCAP; 470 separator = ':'; 471 break; 472 } else if (ch == ',') { 473 _nc_syntax = SYN_TERMINFO; 474 separator = ','; 475 /* 476 * If we did not see a '|', then we found a name with no 477 * aliases or description. 478 */ 479 if (after_name == 0) 480 break; 481 /* 482 * We saw a comma, but are not entirely sure this is 483 * terminfo format, since we can still be parsing the 484 * description field (for either syntax). 485 * 486 * A properly formatted termcap line ends with either a 487 * colon, or a backslash after a colon. It is possible 488 * to have a backslash in the middle of a capability, but 489 * then there would be no leading whitespace on the next 490 * line - something we want to discourage. 491 */ 492 c0 = last_char(0); 493 c1 = last_char(1); 494 if (c1 != ':' && c0 != '\\' && c0 != ':') { 495 bool capability = FALSE; 496 497 /* 498 * Since it is not termcap, assume the line is terminfo 499 * format. However, the comma can be embedded in a 500 * description field. It also can be a separator 501 * between a description field and a capability. 502 * 503 * Improve the guess by checking if the next word after 504 * the comma does not look like a capability. In that 505 * case, extend the description past the comma. 506 */ 507 for (s = bufptr; isspace(UChar(*s)); ++s) { 508 ; 509 } 510 if (islower(UChar(*s))) { 511 char *name = s; 512 while (isalnum(UChar(*s))) { 513 ++s; 514 } 515 if (*s == '#' || *s == '=' || *s == '@') { 516 /* 517 * Checking solely with syntax allows us to 518 * support extended capabilities with string 519 * values. 520 */ 521 capability = TRUE; 522 } else if (*s == ',') { 523 c0 = *s; 524 *s = '\0'; 525 /* 526 * Otherwise, we can handle predefined boolean 527 * capabilities, still aided by syntax. 528 */ 529 if (_nc_find_entry(name, 530 _nc_get_hash_table(FALSE))) { 531 capability = TRUE; 532 } 533 *s = (char) c0; 534 } 535 } 536 if (capability) { 537 break; 538 } 539 } 540 } else 541 ch = eat_escaped_newline(ch); 542 543 if (OkToAdd()) { 544 AddCh(ch); 545 } else { 546 break; 547 } 548 } 549 *tok_ptr = '\0'; 550 if (_nc_syntax == ERR) { 551 /* 552 * Grrr...what we ought to do here is barf, complaining that 553 * the entry is malformed. But because a couple of name fields 554 * in the 8.2 termcap file end with |\, we just have to assume 555 * it's termcap syntax. 556 */ 557 _nc_syntax = SYN_TERMCAP; 558 separator = ':'; 559 } else if (_nc_syntax == SYN_TERMINFO) { 560 /* throw away trailing /, *$/ */ 561 for (--tok_ptr; 562 iswhite(*tok_ptr) || *tok_ptr == ','; 563 tok_ptr--) 564 continue; 565 tok_ptr[1] = '\0'; 566 } 567 568 /* 569 * This is the soonest we have the terminal name fetched. Set up 570 * for following warning messages. If there's no '|', then there 571 * is no description. 572 */ 573 if (after_name != 0) { 574 ch = *after_name; 575 *after_name = '\0'; 576 _nc_set_type(tok_buf); 577 *after_name = (char) ch; 578 } 579 580 /* 581 * Compute the boundary between the aliases and the description 582 * field for syntax-checking purposes. 583 */ 584 if (after_list != 0) { 585 if (!silent) { 586 if (*after_list == '\0') 587 _nc_warning("empty longname field"); 588 #ifndef DRAGONFLY_NATIVE 589 else if (strchr(after_list, ' ') == 0) 590 _nc_warning("older tic versions may treat the description field as an alias"); 591 #endif 592 } 593 } else { 594 after_list = tok_buf + strlen(tok_buf); 595 DEBUG(1, ("missing description")); 596 } 597 598 /* 599 * Whitespace in a name field other than the long name can confuse 600 * rdist and some termcap tools. Slashes are a no-no. Other 601 * special characters can be dangerous due to shell expansion. 602 */ 603 for (s = tok_buf; s < after_list; ++s) { 604 if (isspace(UChar(*s))) { 605 if (!silent) 606 _nc_warning("whitespace in name or alias field"); 607 break; 608 } else if (*s == '/') { 609 if (!silent) 610 _nc_warning("slashes aren't allowed in names or aliases"); 611 break; 612 } else if (strchr("$[]!*?", *s)) { 613 if (!silent) 614 _nc_warning("dubious character `%c' in name or alias field", *s); 615 break; 616 } 617 } 618 619 _nc_curr_token.tk_name = tok_buf; 620 type = NAMES; 621 } else { 622 if (had_newline && _nc_syntax == SYN_TERMCAP) { 623 _nc_warning("Missing backslash before newline"); 624 had_newline = FALSE; 625 } 626 while ((ch = next_char()) != EOF) { 627 if (!isalnum(UChar(ch))) { 628 if (_nc_syntax == SYN_TERMINFO) { 629 if (ch != '_') 630 break; 631 } else { /* allow ';' for "k;" */ 632 if (ch != ';') 633 break; 634 } 635 } 636 if (OkToAdd()) { 637 AddCh(ch); 638 } else { 639 ch = EOF; 640 break; 641 } 642 } 643 644 *tok_ptr++ = '\0'; /* separate name/value in buffer */ 645 switch (ch) { 646 case ',': 647 case ':': 648 if (ch != separator) 649 _nc_err_abort("Separator inconsistent with syntax"); 650 _nc_curr_token.tk_name = tok_buf; 651 type = BOOLEAN; 652 break; 653 case '@': 654 if ((ch = next_char()) != separator && !silent) 655 _nc_warning("Missing separator after `%s', have %s", 656 tok_buf, unctrl(UChar(ch))); 657 _nc_curr_token.tk_name = tok_buf; 658 type = CANCEL; 659 break; 660 661 case '#': 662 found = 0; 663 while (isalnum(ch = next_char())) { 664 numbuf[found++] = (char) ch; 665 if (found >= sizeof(numbuf) - 1) 666 break; 667 } 668 numbuf[found] = '\0'; 669 number = strtol(numbuf, &numchk, 0); 670 if (!silent) { 671 if (numchk == numbuf) 672 _nc_warning("no value given for `%s'", tok_buf); 673 if ((*numchk != '\0') || (ch != separator)) 674 _nc_warning("Missing separator"); 675 } 676 _nc_curr_token.tk_name = tok_buf; 677 _nc_curr_token.tk_valnumber = (int) number; 678 type = NUMBER; 679 break; 680 681 case '=': 682 ch = _nc_trans_string(tok_ptr, tok_buf + TOK_BUF_SIZE); 683 if (!silent && ch != separator) 684 _nc_warning("Missing separator"); 685 _nc_curr_token.tk_name = tok_buf; 686 _nc_curr_token.tk_valstring = tok_ptr; 687 type = STRING; 688 break; 689 690 case EOF: 691 type = EOF; 692 break; 693 default: 694 /* just to get rid of the compiler warning */ 695 type = UNDEF; 696 if (!silent) 697 _nc_warning("Illegal character - '%s'", unctrl(UChar(ch))); 698 } 699 } /* end else (first_column == FALSE) */ 700 } /* end else (ch != EOF) */ 701 702 end_of_token: 703 704 #ifdef TRACE 705 if (dot_flag == TRUE) 706 DEBUG(8, ("Commented out ")); 707 708 if (_nc_tracing >= DEBUG_LEVEL(8)) { 709 _tracef("parsed %d.%d to %d.%d", 710 old_line, old_col, 711 _nc_curr_line, _nc_curr_col); 712 } 713 if (_nc_tracing >= DEBUG_LEVEL(7)) { 714 switch (type) { 715 case BOOLEAN: 716 _tracef("Token: Boolean; name='%s'", 717 _nc_curr_token.tk_name); 718 break; 719 720 case NUMBER: 721 _tracef("Token: Number; name='%s', value=%d", 722 _nc_curr_token.tk_name, 723 _nc_curr_token.tk_valnumber); 724 break; 725 726 case STRING: 727 _tracef("Token: String; name='%s', value=%s", 728 _nc_curr_token.tk_name, 729 _nc_visbuf(_nc_curr_token.tk_valstring)); 730 break; 731 732 case CANCEL: 733 _tracef("Token: Cancel; name='%s'", 734 _nc_curr_token.tk_name); 735 break; 736 737 case NAMES: 738 739 _tracef("Token: Names; value='%s'", 740 _nc_curr_token.tk_name); 741 break; 742 743 case EOF: 744 _tracef("Token: End of file"); 745 break; 746 747 default: 748 _nc_warning("Bad token type"); 749 } 750 } 751 #endif 752 753 if (dot_flag == TRUE) /* if commented out, use the next one */ 754 type = _nc_get_token(silent); 755 756 DEBUG(3, ("token: `%s', class %d", 757 ((_nc_curr_token.tk_name != 0) 758 ? _nc_curr_token.tk_name 759 : "<null>"), 760 type)); 761 762 return (type); 763 } 764 765 /* 766 * char 767 * trans_string(ptr) 768 * 769 * Reads characters using next_char() until encountering a separator, nl, 770 * or end-of-file. The returned value is the character which caused 771 * reading to stop. The following translations are done on the input: 772 * 773 * ^X goes to ctrl-X (i.e. X & 037) 774 * {\E,\n,\r,\b,\t,\f} go to 775 * {ESCAPE,newline,carriage-return,backspace,tab,formfeed} 776 * {\^,\\} go to {carat,backslash} 777 * \ddd (for ddd = up to three octal digits) goes to the character ddd 778 * 779 * \e == \E 780 * \0 == \200 781 * 782 */ 783 784 NCURSES_EXPORT(int) 785 _nc_trans_string(char *ptr, char *last) 786 { 787 int count = 0; 788 int number = 0; 789 int i, c; 790 int last_ch = '\0'; 791 bool ignored = FALSE; 792 bool long_warning = FALSE; 793 794 while ((c = next_char()) != separator && c != EOF) { 795 if (ptr >= (last - 1)) { 796 if (c != EOF) { 797 while ((c = next_char()) != separator && c != EOF) { 798 ; 799 } 800 } 801 break; 802 } 803 if ((_nc_syntax == SYN_TERMCAP) && c == '\n') 804 break; 805 if (c == '^' && last_ch != '%') { 806 c = next_char(); 807 if (c == EOF) 808 _nc_err_abort(MSG_NO_INPUTS); 809 810 if (!(is7bits(c) && isprint(c))) { 811 _nc_warning("Illegal ^ character - '%s'", unctrl(UChar(c))); 812 } 813 if (c == '?' && (_nc_syntax != SYN_TERMCAP)) { 814 *(ptr++) = '\177'; 815 if (_nc_tracing) 816 _nc_warning("Allow ^? as synonym for \\177"); 817 } else { 818 if ((c &= 037) == 0) 819 c = 128; 820 *(ptr++) = (char) (c); 821 } 822 } else if (c == '\\') { 823 bool strict_bsd = ((_nc_syntax == SYN_TERMCAP) && _nc_strict_bsd); 824 825 c = next_char(); 826 if (c == EOF) 827 _nc_err_abort(MSG_NO_INPUTS); 828 829 #define isoctal(c) ((c) >= '0' && (c) <= '7') 830 831 if (isoctal(c) || (strict_bsd && isdigit(c))) { 832 number = c - '0'; 833 for (i = 0; i < 2; i++) { 834 c = next_char(); 835 if (c == EOF) 836 _nc_err_abort(MSG_NO_INPUTS); 837 838 if (!isoctal(c)) { 839 if (isdigit(c)) { 840 if (!strict_bsd) { 841 _nc_warning("Non-octal digit `%c' in \\ sequence", c); 842 /* allow the digit; it'll do less harm */ 843 } 844 } else { 845 push_back(c); 846 break; 847 } 848 } 849 850 number = number * 8 + c - '0'; 851 } 852 853 number = UChar(number); 854 if (number == 0 && !strict_bsd) 855 number = 0200; 856 *(ptr++) = (char) number; 857 } else { 858 switch (c) { 859 case 'E': 860 *(ptr++) = '\033'; 861 break; 862 863 case 'n': 864 *(ptr++) = '\n'; 865 break; 866 867 case 'r': 868 *(ptr++) = '\r'; 869 break; 870 871 case 'b': 872 *(ptr++) = '\010'; 873 break; 874 875 case 'f': 876 *(ptr++) = '\014'; 877 break; 878 879 case 't': 880 *(ptr++) = '\t'; 881 break; 882 883 case '\\': 884 *(ptr++) = '\\'; 885 break; 886 887 case '^': 888 *(ptr++) = '^'; 889 break; 890 891 case ',': 892 *(ptr++) = ','; 893 break; 894 895 case '\n': 896 continue; 897 898 default: 899 if ((_nc_syntax == SYN_TERMINFO) || !_nc_strict_bsd) { 900 switch (c) { 901 case 'a': 902 c = '\007'; 903 break; 904 case 'e': 905 c = '\033'; 906 break; 907 case 'l': 908 c = '\n'; 909 break; 910 case 's': 911 c = ' '; 912 break; 913 case ':': 914 c = ':'; 915 break; 916 default: 917 _nc_warning("Illegal character '%s' in \\ sequence", 918 unctrl(UChar(c))); 919 break; 920 } 921 } 922 /* FALLTHRU */ 923 case '|': 924 *(ptr++) = (char) c; 925 } /* endswitch (c) */ 926 } /* endelse (c < '0' || c > '7') */ 927 } 928 /* end else if (c == '\\') */ 929 else if (c == '\n' && (_nc_syntax == SYN_TERMINFO)) { 930 /* 931 * Newlines embedded in a terminfo string are ignored, provided 932 * that the next line begins with whitespace. 933 */ 934 ignored = TRUE; 935 } else { 936 *(ptr++) = (char) c; 937 } 938 939 if (!ignored) { 940 if (_nc_curr_col <= 1) { 941 push_back(c); 942 c = '\n'; 943 break; 944 } 945 last_ch = c; 946 count++; 947 } 948 ignored = FALSE; 949 950 if (count > MAXCAPLEN && !long_warning) { 951 _nc_warning("Very long string found. Missing separator?"); 952 long_warning = TRUE; 953 } 954 } /* end while */ 955 956 *ptr = '\0'; 957 958 return (c); 959 } 960 961 /* 962 * _nc_push_token() 963 * 964 * Push a token of given type so that it will be reread by the next 965 * get_token() call. 966 */ 967 968 NCURSES_EXPORT(void) 969 _nc_push_token(int tokclass) 970 { 971 /* 972 * This implementation is kind of bogus, it will fail if we ever do more 973 * than one pushback at a time between get_token() calls. It relies on the 974 * fact that _nc_curr_token is static storage that nothing but 975 * _nc_get_token() touches. 976 */ 977 pushtype = tokclass; 978 if (pushname == 0) 979 pushname = typeMalloc(char, MAX_NAME_SIZE + 1); 980 _nc_get_type(pushname); 981 982 DEBUG(3, ("pushing token: `%s', class %d", 983 ((_nc_curr_token.tk_name != 0) 984 ? _nc_curr_token.tk_name 985 : "<null>"), 986 pushtype)); 987 } 988 989 /* 990 * Panic mode error recovery - skip everything until a "ch" is found. 991 */ 992 NCURSES_EXPORT(void) 993 _nc_panic_mode(char ch) 994 { 995 int c; 996 997 for (;;) { 998 c = next_char(); 999 if (c == ch) 1000 return; 1001 if (c == EOF) 1002 return; 1003 } 1004 } 1005 1006 #if NO_LEAKS 1007 NCURSES_EXPORT(void) 1008 _nc_comp_scan_leaks(void) 1009 { 1010 if (pushname != 0) { 1011 FreeAndNull(pushname); 1012 } 1013 if (tok_buf != 0) { 1014 FreeAndNull(tok_buf); 1015 } 1016 } 1017 #endif 1018