1 /****************************************************************************
2 * Copyright 2020 Thomas E. Dickey *
3 * Copyright 1998-2016,2017 Free Software Foundation, Inc. *
4 * *
5 * Permission is hereby granted, free of charge, to any person obtaining a *
6 * copy of this software and associated documentation files (the *
7 * "Software"), to deal in the Software without restriction, including *
8 * without limitation the rights to use, copy, modify, merge, publish, *
9 * distribute, distribute with modifications, sublicense, and/or sell *
10 * copies of the Software, and to permit persons to whom the Software is *
11 * furnished to do so, subject to the following conditions: *
12 * *
13 * The above copyright notice and this permission notice shall be included *
14 * in all copies or substantial portions of the Software. *
15 * *
16 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS *
17 * OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF *
18 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. *
19 * IN NO EVENT SHALL THE ABOVE COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, *
20 * DAMAGES OR OTHER LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR *
21 * OTHERWISE, ARISING FROM, OUT OF OR IN CONNECTION WITH THE SOFTWARE OR *
22 * THE USE OR OTHER DEALINGS IN THE SOFTWARE. *
23 * *
24 * Except as contained in this notice, the name(s) of the above copyright *
25 * holders shall not be used in advertising or otherwise to promote the *
26 * sale, use or other dealings in this Software without prior written *
27 * authorization. *
28 ****************************************************************************/
29
30 /****************************************************************************
31 * Author: Zeyd M. Ben-Halim <[email protected]> 1992,1995 *
32 * and: Eric S. Raymond <[email protected]> *
33 * and: Thomas E. Dickey 1996 on *
34 ****************************************************************************/
35
36 /* $FreeBSD$ */
37
38 /*
39 * comp_scan.c --- Lexical scanner for terminfo compiler.
40 *
41 * _nc_reset_input()
42 * _nc_get_token()
43 * _nc_panic_mode()
44 * int _nc_syntax;
45 * int _nc_curr_line;
46 * long _nc_curr_file_pos;
47 * long _nc_comment_start;
48 * long _nc_comment_end;
49 */
50
51 #include <curses.priv.h>
52
53 #include <ctype.h>
54 #include <tic.h>
55
56 MODULE_ID("$Id: comp_scan.c,v 1.109 2020/02/02 23:34:34 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 NCURSES_EXPORT_VAR (int) _nc_syntax = 0; /* termcap or terminfo? */
67 NCURSES_EXPORT_VAR (int) _nc_strict_bsd = 1; /* ncurses extended termcap? */
68 NCURSES_EXPORT_VAR (long) _nc_curr_file_pos = 0; /* file offset of current line */
69 NCURSES_EXPORT_VAR (long) _nc_comment_start = 0; /* start of comment range before name */
70 NCURSES_EXPORT_VAR (long) _nc_comment_end = 0; /* end of comment range before name */
71 NCURSES_EXPORT_VAR (long) _nc_start_line = 0; /* start line of current entry */
72
73 NCURSES_EXPORT_VAR (struct token) _nc_curr_token =
74 {
75 0, 0, 0
76 };
77
78 /*****************************************************************************
79 *
80 * Token-grabbing machinery
81 *
82 *****************************************************************************/
83
84 static bool first_column; /* See 'next_char()' below */
85 static bool had_newline;
86 static char separator; /* capability separator */
87 static int pushtype; /* type of pushback token */
88 static char *pushname;
89
90 #if NCURSES_EXT_FUNCS
91 NCURSES_EXPORT_VAR (bool) _nc_disable_period = FALSE; /* used by tic -a option */
92 #endif
93
94 /*****************************************************************************
95 *
96 * Character-stream handling
97 *
98 *****************************************************************************/
99
100 #define LEXBUFSIZ 1024
101
102 static char *bufptr; /* otherwise, the input buffer pointer */
103 static char *bufstart; /* start of buffer so we can compute offsets */
104 static FILE *yyin; /* scanner's input file descriptor */
105
106 /*
107 * _nc_reset_input()
108 *
109 * Resets the input-reading routines. Used on initialization,
110 * or after a seek has been done. Exactly one argument must be
111 * non-null.
112 */
113
114 NCURSES_EXPORT(void)
_nc_reset_input(FILE * fp,char * buf)115 _nc_reset_input(FILE *fp, char *buf)
116 {
117 pushtype = NO_PUSHBACK;
118 if (pushname != 0)
119 pushname[0] = '\0';
120 yyin = fp;
121 bufstart = bufptr = buf;
122 _nc_curr_file_pos = 0L;
123 if (fp != 0)
124 _nc_curr_line = 0;
125 _nc_curr_col = 0;
126 }
127
128 /*
129 * int last_char()
130 *
131 * Returns the final nonblank character on the current input buffer
132 */
133 static int
last_char(int from_end)134 last_char(int from_end)
135 {
136 size_t len = strlen(bufptr);
137 int result = 0;
138
139 while (len--) {
140 if (!isspace(UChar(bufptr[len]))) {
141 if (from_end < (int) len)
142 result = bufptr[(int) len - from_end];
143 break;
144 }
145 }
146 return result;
147 }
148
149 /*
150 * int next_char()
151 *
152 * Returns the next character in the input stream. Comments and leading
153 * white space are stripped.
154 *
155 * The global state variable 'firstcolumn' is set TRUE if the character
156 * returned is from the first column of the input line.
157 *
158 * The global variable _nc_curr_line is incremented for each new line.
159 * The global variable _nc_curr_file_pos is set to the file offset of the
160 * beginning of each line.
161 */
162
163 static int
next_char(void)164 next_char(void)
165 {
166 static char *result;
167 static size_t allocated;
168 int the_char;
169
170 if (!yyin) {
171 if (result != 0) {
172 FreeAndNull(result);
173 FreeAndNull(pushname);
174 bufptr = 0;
175 bufstart = 0;
176 allocated = 0;
177 }
178 /*
179 * An string with an embedded null will truncate the input. This is
180 * intentional (we don't read binary files here).
181 */
182 if (bufptr == 0 || *bufptr == '\0')
183 return (EOF);
184 if (*bufptr == '\n') {
185 _nc_curr_line++;
186 _nc_curr_col = 0;
187 } else if (*bufptr == '\t') {
188 _nc_curr_col = (_nc_curr_col | 7);
189 }
190 } else if (!bufptr || !*bufptr) {
191 /*
192 * In theory this could be recoded to do its I/O one character at a
193 * time, saving the buffer space. In practice, this turns out to be
194 * quite hard to get completely right. Try it and see. If you
195 * succeed, don't forget to hack push_back() correspondingly.
196 */
197 size_t len;
198
199 do {
200 size_t used = 0;
201 bufstart = 0;
202 do {
203 if (used + (LEXBUFSIZ / 4) >= allocated) {
204 allocated += (allocated + LEXBUFSIZ);
205 result = typeRealloc(char, allocated, result);
206 if (result == 0)
207 return (EOF);
208 if (bufstart)
209 bufstart = result;
210 }
211 if (used == 0)
212 _nc_curr_file_pos = ftell(yyin);
213
214 if (fgets(result + used, (int) (allocated - used), yyin) != 0) {
215 bufstart = result;
216 if (used == 0) {
217 if (_nc_curr_line == 0
218 && IS_TIC_MAGIC(result)) {
219 _nc_err_abort("This is a compiled terminal description, not a source");
220 }
221 _nc_curr_line++;
222 _nc_curr_col = 0;
223 }
224 } else {
225 if (used != 0)
226 _nc_STRCAT(result, "\n", allocated);
227 }
228 if ((bufptr = bufstart) != 0) {
229 used = strlen(bufptr);
230 if (used == 0)
231 return (EOF);
232 while (iswhite(*bufptr)) {
233 if (*bufptr == '\t') {
234 _nc_curr_col = (_nc_curr_col | 7) + 1;
235 } else {
236 _nc_curr_col++;
237 }
238 bufptr++;
239 }
240
241 /*
242 * Treat a trailing <cr><lf> the same as a <newline> so we
243 * can read files on OS/2, etc.
244 */
245 if ((len = strlen(bufptr)) > 1) {
246 if (bufptr[len - 1] == '\n'
247 && bufptr[len - 2] == '\r') {
248 len--;
249 bufptr[len - 1] = '\n';
250 bufptr[len] = '\0';
251 }
252 }
253 } else {
254 return (EOF);
255 }
256 } while (bufptr[len - 1] != '\n'); /* complete a line */
257 } while (result[0] == '#'); /* ignore comments */
258 } else if (*bufptr == '\t') {
259 _nc_curr_col = (_nc_curr_col | 7);
260 }
261
262 first_column = (bufptr == bufstart);
263 if (first_column)
264 had_newline = FALSE;
265
266 _nc_curr_col++;
267 the_char = *bufptr++;
268 return UChar(the_char);
269 }
270
271 static void
push_back(int c)272 push_back(int c)
273 /* push a character back onto the input stream */
274 {
275 if (bufptr == bufstart)
276 _nc_syserr_abort("Can't backspace off beginning of line");
277 *--bufptr = (char) c;
278 _nc_curr_col--;
279 }
280
281 static long
stream_pos(void)282 stream_pos(void)
283 /* return our current character position in the input stream */
284 {
285 return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0));
286 }
287
288 static bool
end_of_stream(void)289 end_of_stream(void)
290 /* are we at end of input? */
291 {
292 return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0'))
293 ? TRUE : FALSE);
294 }
295
296 /* Assume we may be looking at a termcap-style continuation */
297 static NCURSES_INLINE int
eat_escaped_newline(int ch)298 eat_escaped_newline(int ch)
299 {
300 if (ch == '\\')
301 while ((ch = next_char()) == '\n' || iswhite(ch))
302 continue;
303 return ch;
304 }
305
306 #define TOK_BUF_SIZE MAX_ENTRY_SIZE
307
308 #define OkToAdd() \
309 ((tok_ptr - tok_buf) < (TOK_BUF_SIZE - 2))
310
311 #define AddCh(ch) \
312 *tok_ptr++ = (char) ch; \
313 *tok_ptr = '\0'
314
315 static char *tok_buf;
316
317 /*
318 * int
319 * get_token()
320 *
321 * Scans the input for the next token, storing the specifics in the
322 * global structure 'curr_token' and returning one of the following:
323 *
324 * NAMES A line beginning in column 1. 'name'
325 * will be set to point to everything up to but
326 * not including the first separator on the line.
327 * BOOLEAN An entry consisting of a name followed by
328 * a separator. 'name' will be set to point to
329 * the name of the capability.
330 * NUMBER An entry of the form
331 * name#digits,
332 * 'name' will be set to point to the capability
333 * name and 'valnumber' to the number given.
334 * STRING An entry of the form
335 * name=characters,
336 * 'name' is set to the capability name and
337 * 'valstring' to the string of characters, with
338 * input translations done.
339 * CANCEL An entry of the form
340 * name@,
341 * 'name' is set to the capability name and
342 * 'valnumber' to -1.
343 * EOF The end of the file has been reached.
344 *
345 * A `separator' is either a comma or a semicolon, depending on whether
346 * we are in termcap or terminfo mode.
347 *
348 */
349
350 NCURSES_EXPORT(int)
_nc_get_token(bool silent)351 _nc_get_token(bool silent)
352 {
353 static const char terminfo_punct[] = "@%&*!#";
354
355 char *after_name; /* after primary name */
356 char *after_list; /* after primary and alias list */
357 char *numchk;
358 char *tok_ptr;
359 char *s;
360 char numbuf[80];
361 int ch, c0, c1;
362 int dot_flag = FALSE;
363 int type;
364 long number;
365 long token_start;
366 unsigned found;
367 #ifdef TRACE
368 int old_line;
369 int old_col;
370 #endif
371
372 if (pushtype != NO_PUSHBACK) {
373 int retval = pushtype;
374
375 _nc_set_type(pushname != 0 ? pushname : "");
376 DEBUG(3, ("pushed-back token: `%s', class %d",
377 _nc_curr_token.tk_name, pushtype));
378
379 pushtype = NO_PUSHBACK;
380 if (pushname != 0)
381 pushname[0] = '\0';
382
383 /* currtok wasn't altered by _nc_push_token() */
384 return (retval);
385 }
386
387 if (end_of_stream()) {
388 yyin = 0;
389 (void) next_char(); /* frees its allocated memory */
390 if (tok_buf != 0) {
391 if (_nc_curr_token.tk_name == tok_buf)
392 _nc_curr_token.tk_name = 0;
393 }
394 return (EOF);
395 }
396
397 start_token:
398 token_start = stream_pos();
399 while ((ch = next_char()) == '\n' || iswhite(ch)) {
400 if (ch == '\n')
401 had_newline = TRUE;
402 continue;
403 }
404
405 ch = eat_escaped_newline(ch);
406 _nc_curr_token.tk_valstring = 0;
407
408 #ifdef TRACE
409 old_line = _nc_curr_line;
410 old_col = _nc_curr_col;
411 #endif
412 if (ch == EOF)
413 type = EOF;
414 else {
415 /* if this is a termcap entry, skip a leading separator */
416 if (separator == ':' && ch == ':')
417 ch = next_char();
418
419 if (ch == '.'
420 #if NCURSES_EXT_FUNCS
421 && !_nc_disable_period
422 #endif
423 ) {
424 dot_flag = TRUE;
425 DEBUG(8, ("dot-flag set"));
426
427 while ((ch = next_char()) == '.' || iswhite(ch))
428 continue;
429 }
430
431 if (ch == EOF) {
432 type = EOF;
433 goto end_of_token;
434 }
435
436 /* have to make some punctuation chars legal for terminfo */
437 if (!isalnum(UChar(ch))
438 #if NCURSES_EXT_FUNCS
439 && !(ch == '.' && _nc_disable_period)
440 #endif
441 && ((strchr) (terminfo_punct, (char) ch) == 0)) {
442 if (!silent)
443 _nc_warning("Illegal character (expected alphanumeric or %s) - '%s'",
444 terminfo_punct, unctrl(UChar(ch)));
445 _nc_panic_mode(separator);
446 goto start_token;
447 }
448
449 if (tok_buf == 0)
450 tok_buf = typeMalloc(char, TOK_BUF_SIZE);
451
452 #ifdef TRACE
453 old_line = _nc_curr_line;
454 old_col = _nc_curr_col;
455 #endif
456 tok_ptr = tok_buf;
457 AddCh(ch);
458
459 if (first_column) {
460 _nc_comment_start = token_start;
461 _nc_comment_end = _nc_curr_file_pos;
462 _nc_start_line = _nc_curr_line;
463
464 _nc_syntax = ERR;
465 after_name = 0;
466 after_list = 0;
467 while ((ch = next_char()) != '\n') {
468 if (ch == EOF) {
469 _nc_err_abort(MSG_NO_INPUTS);
470 } else if (ch == '|') {
471 after_list = tok_ptr;
472 if (after_name == 0)
473 after_name = tok_ptr;
474 } else if (ch == ':' && last_char(0) != ',') {
475 _nc_syntax = SYN_TERMCAP;
476 separator = ':';
477 break;
478 } else if (ch == ',') {
479 _nc_syntax = SYN_TERMINFO;
480 separator = ',';
481 /*
482 * If we did not see a '|', then we found a name with no
483 * aliases or description.
484 */
485 if (after_name == 0)
486 break;
487 /*
488 * We saw a comma, but are not entirely sure this is
489 * terminfo format, since we can still be parsing the
490 * description field (for either syntax).
491 *
492 * A properly formatted termcap line ends with either a
493 * colon, or a backslash after a colon. It is possible
494 * to have a backslash in the middle of a capability, but
495 * then there would be no leading whitespace on the next
496 * line - something we want to discourage.
497 */
498 c0 = last_char(0);
499 c1 = last_char(1);
500 if (c1 != ':' && c0 != '\\' && c0 != ':') {
501 bool capability = FALSE;
502
503 /*
504 * Since it is not termcap, assume the line is terminfo
505 * format. However, the comma can be embedded in a
506 * description field. It also can be a separator
507 * between a description field and a capability.
508 *
509 * Improve the guess by checking if the next word after
510 * the comma does not look like a capability. In that
511 * case, extend the description past the comma.
512 */
513 for (s = bufptr; isspace(UChar(*s)); ++s) {
514 ;
515 }
516 if (islower(UChar(*s))) {
517 char *name = s;
518 while (isalnum(UChar(*s))) {
519 ++s;
520 }
521 if (*s == '#' || *s == '=' || *s == '@') {
522 /*
523 * Checking solely with syntax allows us to
524 * support extended capabilities with string
525 * values.
526 */
527 capability = TRUE;
528 } else if (*s == ',') {
529 c0 = *s;
530 *s = '\0';
531 /*
532 * Otherwise, we can handle predefined boolean
533 * capabilities, still aided by syntax.
534 */
535 if (_nc_find_entry(name,
536 _nc_get_hash_table(FALSE))) {
537 capability = TRUE;
538 }
539 *s = (char) c0;
540 }
541 }
542 if (capability) {
543 break;
544 }
545 }
546 } else
547 ch = eat_escaped_newline(ch);
548
549 if (OkToAdd()) {
550 AddCh(ch);
551 } else {
552 break;
553 }
554 }
555 *tok_ptr = '\0';
556 if (_nc_syntax == ERR) {
557 /*
558 * Grrr...what we ought to do here is barf, complaining that
559 * the entry is malformed. But because a couple of name fields
560 * in the 8.2 termcap file end with |\, we just have to assume
561 * it's termcap syntax.
562 */
563 _nc_syntax = SYN_TERMCAP;
564 separator = ':';
565 } else if (_nc_syntax == SYN_TERMINFO) {
566 /* throw away trailing /, *$/ */
567 for (--tok_ptr;
568 iswhite(*tok_ptr) || *tok_ptr == ',';
569 tok_ptr--)
570 continue;
571 tok_ptr[1] = '\0';
572 }
573
574 /*
575 * This is the soonest we have the terminal name fetched. Set up
576 * for following warning messages. If there's no '|', then there
577 * is no description.
578 */
579 if (after_name != 0) {
580 ch = *after_name;
581 *after_name = '\0';
582 _nc_set_type(tok_buf);
583 *after_name = (char) ch;
584 }
585
586 /*
587 * Compute the boundary between the aliases and the description
588 * field for syntax-checking purposes.
589 */
590 if (after_list != 0) {
591 if (!silent) {
592 if (*after_list == '\0')
593 _nc_warning("empty longname field");
594 #ifndef FREEBSD_NATIVE
595 else if (strchr(after_list, ' ') == 0)
596 _nc_warning("older tic versions may treat the description field as an alias");
597 #endif
598 }
599 } else {
600 after_list = tok_buf + strlen(tok_buf);
601 DEBUG(1, ("missing description"));
602 }
603
604 /*
605 * Whitespace in a name field other than the long name can confuse
606 * rdist and some termcap tools. Slashes are a no-no. Other
607 * special characters can be dangerous due to shell expansion.
608 */
609 for (s = tok_buf; s < after_list; ++s) {
610 if (isspace(UChar(*s))) {
611 if (!silent)
612 _nc_warning("whitespace in name or alias field");
613 break;
614 } else if (*s == '/') {
615 if (!silent)
616 _nc_warning("slashes aren't allowed in names or aliases");
617 break;
618 } else if (strchr("$[]!*?", *s)) {
619 if (!silent)
620 _nc_warning("dubious character `%c' in name or alias field", *s);
621 break;
622 }
623 }
624
625 _nc_curr_token.tk_name = tok_buf;
626 type = NAMES;
627 } else {
628 if (had_newline && _nc_syntax == SYN_TERMCAP) {
629 _nc_warning("Missing backslash before newline");
630 had_newline = FALSE;
631 }
632 while ((ch = next_char()) != EOF) {
633 if (!isalnum(UChar(ch))) {
634 if (_nc_syntax == SYN_TERMINFO) {
635 if (ch != '_')
636 break;
637 } else { /* allow ';' for "k;" */
638 if (ch != ';')
639 break;
640 }
641 }
642 if (OkToAdd()) {
643 AddCh(ch);
644 } else {
645 ch = EOF;
646 break;
647 }
648 }
649
650 *tok_ptr++ = '\0'; /* separate name/value in buffer */
651 switch (ch) {
652 case ',':
653 case ':':
654 if (ch != separator)
655 _nc_err_abort("Separator inconsistent with syntax");
656 _nc_curr_token.tk_name = tok_buf;
657 type = BOOLEAN;
658 break;
659 case '@':
660 if ((ch = next_char()) != separator && !silent)
661 _nc_warning("Missing separator after `%s', have %s",
662 tok_buf, unctrl(UChar(ch)));
663 _nc_curr_token.tk_name = tok_buf;
664 type = CANCEL;
665 break;
666
667 case '#':
668 found = 0;
669 while (isalnum(ch = next_char())) {
670 numbuf[found++] = (char) ch;
671 if (found >= sizeof(numbuf) - 1)
672 break;
673 }
674 numbuf[found] = '\0';
675 number = strtol(numbuf, &numchk, 0);
676 if (!silent) {
677 if (numchk == numbuf)
678 _nc_warning("no value given for `%s'", tok_buf);
679 if ((*numchk != '\0') || (ch != separator))
680 _nc_warning("Missing separator for `%s'", tok_buf);
681 if (number < 0)
682 _nc_warning("value of `%s' cannot be negative", tok_buf);
683 if (number > MAX_OF_TYPE(NCURSES_INT2)) {
684 _nc_warning("limiting value of `%s' from %#lx to %#x",
685 tok_buf,
686 number, MAX_OF_TYPE(NCURSES_INT2));
687 number = MAX_OF_TYPE(NCURSES_INT2);
688 }
689 }
690 _nc_curr_token.tk_name = tok_buf;
691 _nc_curr_token.tk_valnumber = (int) number;
692 type = NUMBER;
693 break;
694
695 case '=':
696 ch = _nc_trans_string(tok_ptr, tok_buf + TOK_BUF_SIZE);
697 if (!silent && ch != separator)
698 _nc_warning("Missing separator");
699 _nc_curr_token.tk_name = tok_buf;
700 _nc_curr_token.tk_valstring = tok_ptr;
701 type = STRING;
702 break;
703
704 case EOF:
705 type = EOF;
706 break;
707 default:
708 /* just to get rid of the compiler warning */
709 type = UNDEF;
710 if (!silent)
711 _nc_warning("Illegal character - '%s'", unctrl(UChar(ch)));
712 }
713 } /* end else (first_column == FALSE) */
714 } /* end else (ch != EOF) */
715
716 end_of_token:
717
718 #ifdef TRACE
719 if (dot_flag == TRUE)
720 DEBUG(8, ("Commented out "));
721
722 if (_nc_tracing >= DEBUG_LEVEL(8)) {
723 _tracef("parsed %d.%d to %d.%d",
724 old_line, old_col,
725 _nc_curr_line, _nc_curr_col);
726 }
727 if (_nc_tracing >= DEBUG_LEVEL(7)) {
728 switch (type) {
729 case BOOLEAN:
730 _tracef("Token: Boolean; name='%s'",
731 _nc_curr_token.tk_name);
732 break;
733
734 case NUMBER:
735 _tracef("Token: Number; name='%s', value=%d",
736 _nc_curr_token.tk_name,
737 _nc_curr_token.tk_valnumber);
738 break;
739
740 case STRING:
741 _tracef("Token: String; name='%s', value=%s",
742 _nc_curr_token.tk_name,
743 _nc_visbuf(_nc_curr_token.tk_valstring));
744 break;
745
746 case CANCEL:
747 _tracef("Token: Cancel; name='%s'",
748 _nc_curr_token.tk_name);
749 break;
750
751 case NAMES:
752
753 _tracef("Token: Names; value='%s'",
754 _nc_curr_token.tk_name);
755 break;
756
757 case EOF:
758 _tracef("Token: End of file");
759 break;
760
761 default:
762 _nc_warning("Bad token type");
763 }
764 }
765 #endif
766
767 if (dot_flag == TRUE) /* if commented out, use the next one */
768 type = _nc_get_token(silent);
769
770 DEBUG(3, ("token: `%s', class %d",
771 ((_nc_curr_token.tk_name != 0)
772 ? _nc_curr_token.tk_name
773 : "<null>"),
774 type));
775
776 return (type);
777 }
778
779 /*
780 * char
781 * trans_string(ptr)
782 *
783 * Reads characters using next_char() until encountering a separator, nl,
784 * or end-of-file. The returned value is the character which caused
785 * reading to stop. The following translations are done on the input:
786 *
787 * ^X goes to ctrl-X (i.e. X & 037)
788 * {\E,\n,\r,\b,\t,\f} go to
789 * {ESCAPE,newline,carriage-return,backspace,tab,formfeed}
790 * {\^,\\} go to {carat,backslash}
791 * \ddd (for ddd = up to three octal digits) goes to the character ddd
792 *
793 * \e == \E
794 * \0 == \200
795 *
796 */
797
798 NCURSES_EXPORT(int)
_nc_trans_string(char * ptr,char * last)799 _nc_trans_string(char *ptr, char *last)
800 {
801 int count = 0;
802 int number = 0;
803 int i, c;
804 int last_ch = '\0';
805 bool ignored = FALSE;
806 bool long_warning = FALSE;
807
808 while ((c = next_char()) != separator && c != EOF) {
809 if (ptr >= (last - 1)) {
810 if (c != EOF) {
811 while ((c = next_char()) != separator && c != EOF) {
812 ;
813 }
814 }
815 break;
816 }
817 if ((_nc_syntax == SYN_TERMCAP) && c == '\n')
818 break;
819 if (c == '^' && last_ch != '%') {
820 c = next_char();
821 if (c == EOF)
822 _nc_err_abort(MSG_NO_INPUTS);
823
824 if (!(is7bits(c) && isprint(c))) {
825 _nc_warning("Illegal ^ character - '%s'", unctrl(UChar(c)));
826 }
827 if (c == '?' && (_nc_syntax != SYN_TERMCAP)) {
828 *(ptr++) = '\177';
829 } else {
830 if ((c &= 037) == 0)
831 c = 128;
832 *(ptr++) = (char) (c);
833 }
834 } else if (c == '\\') {
835 bool strict_bsd = ((_nc_syntax == SYN_TERMCAP) && _nc_strict_bsd);
836
837 c = next_char();
838 if (c == EOF)
839 _nc_err_abort(MSG_NO_INPUTS);
840
841 if (isoctal(c) || (strict_bsd && isdigit(c))) {
842 number = c - '0';
843 for (i = 0; i < 2; i++) {
844 c = next_char();
845 if (c == EOF)
846 _nc_err_abort(MSG_NO_INPUTS);
847
848 if (!isoctal(c)) {
849 if (isdigit(c)) {
850 if (!strict_bsd) {
851 _nc_warning("Non-octal digit `%c' in \\ sequence", c);
852 /* allow the digit; it'll do less harm */
853 }
854 } else {
855 push_back(c);
856 break;
857 }
858 }
859
860 number = number * 8 + c - '0';
861 }
862
863 number = UChar(number);
864 if (number == 0 && !strict_bsd)
865 number = 0200;
866 *(ptr++) = (char) number;
867 } else {
868 switch (c) {
869 case 'E':
870 *(ptr++) = '\033';
871 break;
872
873 case 'n':
874 *(ptr++) = '\n';
875 break;
876
877 case 'r':
878 *(ptr++) = '\r';
879 break;
880
881 case 'b':
882 *(ptr++) = '\010';
883 break;
884
885 case 'f':
886 *(ptr++) = '\014';
887 break;
888
889 case 't':
890 *(ptr++) = '\t';
891 break;
892
893 case '\\':
894 *(ptr++) = '\\';
895 break;
896
897 case '^':
898 *(ptr++) = '^';
899 break;
900
901 case ',':
902 *(ptr++) = ',';
903 break;
904
905 case '\n':
906 continue;
907
908 default:
909 if ((_nc_syntax == SYN_TERMINFO) || !_nc_strict_bsd) {
910 switch (c) {
911 case 'a':
912 c = '\007';
913 break;
914 case 'e':
915 c = '\033';
916 break;
917 case 'l':
918 c = '\n';
919 break;
920 case 's':
921 c = ' ';
922 break;
923 case ':':
924 c = ':';
925 break;
926 default:
927 _nc_warning("Illegal character '%s' in \\ sequence",
928 unctrl(UChar(c)));
929 break;
930 }
931 }
932 /* FALLTHRU */
933 case '|':
934 *(ptr++) = (char) c;
935 } /* endswitch (c) */
936 } /* endelse (c < '0' || c > '7') */
937 }
938 /* end else if (c == '\\') */
939 else if (c == '\n' && (_nc_syntax == SYN_TERMINFO)) {
940 /*
941 * Newlines embedded in a terminfo string are ignored, provided
942 * that the next line begins with whitespace.
943 */
944 ignored = TRUE;
945 } else {
946 *(ptr++) = (char) c;
947 }
948
949 if (!ignored) {
950 if (_nc_curr_col <= 1) {
951 push_back(c);
952 c = '\n';
953 break;
954 }
955 last_ch = c;
956 count++;
957 }
958 ignored = FALSE;
959
960 if (count > MAXCAPLEN && !long_warning) {
961 _nc_warning("Very long string found. Missing separator?");
962 long_warning = TRUE;
963 }
964 } /* end while */
965
966 *ptr = '\0';
967
968 return (c);
969 }
970
971 /*
972 * _nc_push_token()
973 *
974 * Push a token of given type so that it will be reread by the next
975 * get_token() call.
976 */
977
978 NCURSES_EXPORT(void)
_nc_push_token(int tokclass)979 _nc_push_token(int tokclass)
980 {
981 /*
982 * This implementation is kind of bogus, it will fail if we ever do more
983 * than one pushback at a time between get_token() calls. It relies on the
984 * fact that _nc_curr_token is static storage that nothing but
985 * _nc_get_token() touches.
986 */
987 pushtype = tokclass;
988 if (pushname == 0)
989 pushname = typeMalloc(char, MAX_NAME_SIZE + 1);
990 _nc_get_type(pushname);
991
992 DEBUG(3, ("pushing token: `%s', class %d",
993 ((_nc_curr_token.tk_name != 0)
994 ? _nc_curr_token.tk_name
995 : "<null>"),
996 pushtype));
997 }
998
999 /*
1000 * Panic mode error recovery - skip everything until a "ch" is found.
1001 */
1002 NCURSES_EXPORT(void)
_nc_panic_mode(char ch)1003 _nc_panic_mode(char ch)
1004 {
1005 for (;;) {
1006 int c = next_char();
1007 if (c == ch)
1008 return;
1009 if (c == EOF)
1010 return;
1011 }
1012 }
1013
1014 #if NO_LEAKS
1015 NCURSES_EXPORT(void)
_nc_comp_scan_leaks(void)1016 _nc_comp_scan_leaks(void)
1017 {
1018 if (pushname != 0) {
1019 FreeAndNull(pushname);
1020 }
1021 if (tok_buf != 0) {
1022 FreeAndNull(tok_buf);
1023 }
1024 }
1025 #endif
1026