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 <[email protected]> 1992,1995 *
31 * and: Eric S. Raymond <[email protected]> *
32 * and: Thomas E. Dickey 1996 on *
33 ****************************************************************************/
34
35 /* $FreeBSD$ */
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 <tic.h>
54
55 MODULE_ID("$Id: comp_scan.c,v 1.102 2013/11/16 19:57:50 tom Exp $")
56
57 /*
58 * Maximum length of string capability we'll accept before raising an error.
59 * Yes, there is a real capability in /etc/termcap this long, an "is".
60 */
61 #define MAXCAPLEN 600
62
63 #define iswhite(ch) (ch == ' ' || ch == '\t')
64
65 NCURSES_EXPORT_VAR (int) _nc_syntax = 0; /* termcap or terminfo? */
66 NCURSES_EXPORT_VAR (int) _nc_strict_bsd = 1; /* ncurses extended termcap? */
67 NCURSES_EXPORT_VAR (long) _nc_curr_file_pos = 0; /* file offset of current line */
68 NCURSES_EXPORT_VAR (long) _nc_comment_start = 0; /* start of comment range before name */
69 NCURSES_EXPORT_VAR (long) _nc_comment_end = 0; /* end of comment range before name */
70 NCURSES_EXPORT_VAR (long) _nc_start_line = 0; /* start line of current entry */
71
72 NCURSES_EXPORT_VAR (struct token) _nc_curr_token =
73 {
74 0, 0, 0
75 };
76
77 /*****************************************************************************
78 *
79 * Token-grabbing machinery
80 *
81 *****************************************************************************/
82
83 static bool first_column; /* See 'next_char()' below */
84 static bool had_newline;
85 static char separator; /* capability separator */
86 static int pushtype; /* type of pushback token */
87 static char *pushname;
88
89 #if NCURSES_EXT_FUNCS
90 NCURSES_EXPORT_VAR (bool) _nc_disable_period = FALSE; /* used by tic -a option */
91 #endif
92
93 /*****************************************************************************
94 *
95 * Character-stream handling
96 *
97 *****************************************************************************/
98
99 #define LEXBUFSIZ 1024
100
101 static char *bufptr; /* otherwise, the input buffer pointer */
102 static char *bufstart; /* start of buffer so we can compute offsets */
103 static FILE *yyin; /* scanner's input file descriptor */
104
105 /*
106 * _nc_reset_input()
107 *
108 * Resets the input-reading routines. Used on initialization,
109 * or after a seek has been done. Exactly one argument must be
110 * non-null.
111 */
112
113 NCURSES_EXPORT(void)
_nc_reset_input(FILE * fp,char * buf)114 _nc_reset_input(FILE *fp, char *buf)
115 {
116 pushtype = NO_PUSHBACK;
117 if (pushname != 0)
118 pushname[0] = '\0';
119 yyin = fp;
120 bufstart = bufptr = buf;
121 _nc_curr_file_pos = 0L;
122 if (fp != 0)
123 _nc_curr_line = 0;
124 _nc_curr_col = 0;
125 }
126
127 /*
128 * int last_char()
129 *
130 * Returns the final nonblank character on the current input buffer
131 */
132 static int
last_char(int from_end)133 last_char(int from_end)
134 {
135 size_t len = strlen(bufptr);
136 int result = 0;
137
138 while (len--) {
139 if (!isspace(UChar(bufptr[len]))) {
140 if (from_end < (int) len)
141 result = bufptr[(int) len - from_end];
142 break;
143 }
144 }
145 return result;
146 }
147
148 /*
149 * int next_char()
150 *
151 * Returns the next character in the input stream. Comments and leading
152 * white space are stripped.
153 *
154 * The global state variable 'firstcolumn' is set TRUE if the character
155 * returned is from the first column of the input line.
156 *
157 * The global variable _nc_curr_line is incremented for each new line.
158 * The global variable _nc_curr_file_pos is set to the file offset of the
159 * beginning of each line.
160 */
161
162 static int
next_char(void)163 next_char(void)
164 {
165 static char *result;
166 static size_t allocated;
167 int the_char;
168
169 if (!yyin) {
170 if (result != 0) {
171 FreeAndNull(result);
172 FreeAndNull(pushname);
173 allocated = 0;
174 }
175 /*
176 * An string with an embedded null will truncate the input. This is
177 * intentional (we don't read binary files here).
178 */
179 if (bufptr == 0 || *bufptr == '\0')
180 return (EOF);
181 if (*bufptr == '\n') {
182 _nc_curr_line++;
183 _nc_curr_col = 0;
184 } else if (*bufptr == '\t') {
185 _nc_curr_col = (_nc_curr_col | 7);
186 }
187 } else if (!bufptr || !*bufptr) {
188 /*
189 * In theory this could be recoded to do its I/O one character at a
190 * time, saving the buffer space. In practice, this turns out to be
191 * quite hard to get completely right. Try it and see. If you
192 * succeed, don't forget to hack push_back() correspondingly.
193 */
194 size_t used;
195 size_t len;
196
197 do {
198 bufstart = 0;
199 used = 0;
200 do {
201 if (used + (LEXBUFSIZ / 4) >= allocated) {
202 allocated += (allocated + LEXBUFSIZ);
203 result = typeRealloc(char, allocated, result);
204 if (result == 0)
205 return (EOF);
206 if (bufstart)
207 bufstart = result;
208 }
209 if (used == 0)
210 _nc_curr_file_pos = ftell(yyin);
211
212 if (fgets(result + used, (int) (allocated - used), yyin) != 0) {
213 bufstart = result;
214 if (used == 0) {
215 if (_nc_curr_line == 0
216 && IS_TIC_MAGIC(result)) {
217 _nc_err_abort("This is a compiled terminal description, not a source");
218 }
219 _nc_curr_line++;
220 _nc_curr_col = 0;
221 }
222 } else {
223 if (used != 0)
224 _nc_STRCAT(result, "\n", allocated);
225 }
226 if ((bufptr = bufstart) != 0) {
227 used = strlen(bufptr);
228 while (iswhite(*bufptr)) {
229 if (*bufptr == '\t') {
230 _nc_curr_col = (_nc_curr_col | 7) + 1;
231 } else {
232 _nc_curr_col++;
233 }
234 bufptr++;
235 }
236
237 /*
238 * Treat a trailing <cr><lf> the same as a <newline> so we
239 * can read files on OS/2, etc.
240 */
241 if ((len = strlen(bufptr)) > 1) {
242 if (bufptr[len - 1] == '\n'
243 && bufptr[len - 2] == '\r') {
244 len--;
245 bufptr[len - 1] = '\n';
246 bufptr[len] = '\0';
247 }
248 }
249 } else {
250 return (EOF);
251 }
252 } while (bufptr[len - 1] != '\n'); /* complete a line */
253 } while (result[0] == '#'); /* ignore comments */
254 } else if (*bufptr == '\t') {
255 _nc_curr_col = (_nc_curr_col | 7);
256 }
257
258 first_column = (bufptr == bufstart);
259 if (first_column)
260 had_newline = FALSE;
261
262 _nc_curr_col++;
263 the_char = *bufptr++;
264 return UChar(the_char);
265 }
266
267 static void
push_back(int c)268 push_back(int c)
269 /* push a character back onto the input stream */
270 {
271 if (bufptr == bufstart)
272 _nc_syserr_abort("Can't backspace off beginning of line");
273 *--bufptr = (char) c;
274 _nc_curr_col--;
275 }
276
277 static long
stream_pos(void)278 stream_pos(void)
279 /* return our current character position in the input stream */
280 {
281 return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0));
282 }
283
284 static bool
end_of_stream(void)285 end_of_stream(void)
286 /* are we at end of input? */
287 {
288 return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0'))
289 ? TRUE : FALSE);
290 }
291
292 /* Assume we may be looking at a termcap-style continuation */
293 static NCURSES_INLINE int
eat_escaped_newline(int ch)294 eat_escaped_newline(int ch)
295 {
296 if (ch == '\\')
297 while ((ch = next_char()) == '\n' || iswhite(ch))
298 continue;
299 return ch;
300 }
301
302 #define TOK_BUF_SIZE MAX_ENTRY_SIZE
303
304 #define OkToAdd() \
305 ((tok_ptr - tok_buf) < (TOK_BUF_SIZE - 2))
306
307 #define AddCh(ch) \
308 *tok_ptr++ = (char) ch; \
309 *tok_ptr = '\0'
310
311 static char *tok_buf;
312
313 /*
314 * int
315 * get_token()
316 *
317 * Scans the input for the next token, storing the specifics in the
318 * global structure 'curr_token' and returning one of the following:
319 *
320 * NAMES A line beginning in column 1. 'name'
321 * will be set to point to everything up to but
322 * not including the first separator on the line.
323 * BOOLEAN An entry consisting of a name followed by
324 * a separator. 'name' will be set to point to
325 * the name of the capability.
326 * NUMBER An entry of the form
327 * name#digits,
328 * 'name' will be set to point to the capability
329 * name and 'valnumber' to the number given.
330 * STRING An entry of the form
331 * name=characters,
332 * 'name' is set to the capability name and
333 * 'valstring' to the string of characters, with
334 * input translations done.
335 * CANCEL An entry of the form
336 * name@,
337 * 'name' is set to the capability name and
338 * 'valnumber' to -1.
339 * EOF The end of the file has been reached.
340 *
341 * A `separator' is either a comma or a semicolon, depending on whether
342 * we are in termcap or terminfo mode.
343 *
344 */
345
346 NCURSES_EXPORT(int)
_nc_get_token(bool silent)347 _nc_get_token(bool silent)
348 {
349 static const char terminfo_punct[] = "@%&*!#";
350
351 char *after_name; /* after primary name */
352 char *after_list; /* after primary and alias list */
353 char *numchk;
354 char *tok_ptr;
355 char *s;
356 char numbuf[80];
357 int ch, c0, c1;
358 int dot_flag = FALSE;
359 int type;
360 long number;
361 long token_start;
362 unsigned found;
363 #ifdef TRACE
364 int old_line;
365 int old_col;
366 #endif
367
368 if (pushtype != NO_PUSHBACK) {
369 int retval = pushtype;
370
371 _nc_set_type(pushname != 0 ? pushname : "");
372 DEBUG(3, ("pushed-back token: `%s', class %d",
373 _nc_curr_token.tk_name, pushtype));
374
375 pushtype = NO_PUSHBACK;
376 if (pushname != 0)
377 pushname[0] = '\0';
378
379 /* currtok wasn't altered by _nc_push_token() */
380 return (retval);
381 }
382
383 if (end_of_stream()) {
384 yyin = 0;
385 (void) next_char(); /* frees its allocated memory */
386 if (tok_buf != 0) {
387 if (_nc_curr_token.tk_name == tok_buf)
388 _nc_curr_token.tk_name = 0;
389 }
390 return (EOF);
391 }
392
393 start_token:
394 token_start = stream_pos();
395 while ((ch = next_char()) == '\n' || iswhite(ch)) {
396 if (ch == '\n')
397 had_newline = TRUE;
398 continue;
399 }
400
401 ch = eat_escaped_newline(ch);
402 _nc_curr_token.tk_valstring = 0;
403
404 #ifdef TRACE
405 old_line = _nc_curr_line;
406 old_col = _nc_curr_col;
407 #endif
408 if (ch == EOF)
409 type = EOF;
410 else {
411 /* if this is a termcap entry, skip a leading separator */
412 if (separator == ':' && ch == ':')
413 ch = next_char();
414
415 if (ch == '.'
416 #if NCURSES_EXT_FUNCS
417 && !_nc_disable_period
418 #endif
419 ) {
420 dot_flag = TRUE;
421 DEBUG(8, ("dot-flag set"));
422
423 while ((ch = next_char()) == '.' || iswhite(ch))
424 continue;
425 }
426
427 if (ch == EOF) {
428 type = EOF;
429 goto end_of_token;
430 }
431
432 /* have to make some punctuation chars legal for terminfo */
433 if (!isalnum(UChar(ch))
434 #if NCURSES_EXT_FUNCS
435 && !(ch == '.' && _nc_disable_period)
436 #endif
437 && ((strchr) (terminfo_punct, (char) ch) == 0)) {
438 if (!silent)
439 _nc_warning("Illegal character (expected alphanumeric or %s) - '%s'",
440 terminfo_punct, unctrl(UChar(ch)));
441 _nc_panic_mode(separator);
442 goto start_token;
443 }
444
445 if (tok_buf == 0)
446 tok_buf = typeMalloc(char, TOK_BUF_SIZE);
447
448 #ifdef TRACE
449 old_line = _nc_curr_line;
450 old_col = _nc_curr_col;
451 #endif
452 tok_ptr = tok_buf;
453 AddCh(ch);
454
455 if (first_column) {
456 _nc_comment_start = token_start;
457 _nc_comment_end = _nc_curr_file_pos;
458 _nc_start_line = _nc_curr_line;
459
460 _nc_syntax = ERR;
461 after_name = 0;
462 after_list = 0;
463 while ((ch = next_char()) != '\n') {
464 if (ch == EOF) {
465 _nc_err_abort(MSG_NO_INPUTS);
466 } else if (ch == '|') {
467 after_list = tok_ptr;
468 if (after_name == 0)
469 after_name = tok_ptr;
470 } else if (ch == ':' && last_char(0) != ',') {
471 _nc_syntax = SYN_TERMCAP;
472 separator = ':';
473 break;
474 } else if (ch == ',') {
475 _nc_syntax = SYN_TERMINFO;
476 separator = ',';
477 /*
478 * If we did not see a '|', then we found a name with no
479 * aliases or description.
480 */
481 if (after_name == 0)
482 break;
483 /*
484 * We saw a comma, but are not entirely sure this is
485 * terminfo format, since we can still be parsing the
486 * description field (for either syntax).
487 *
488 * A properly formatted termcap line ends with either a
489 * colon, or a backslash after a colon. It is possible
490 * to have a backslash in the middle of a capability, but
491 * then there would be no leading whitespace on the next
492 * line - something we want to discourage.
493 */
494 c0 = last_char(0);
495 c1 = last_char(1);
496 if (c1 != ':' && c0 != '\\' && c0 != ':') {
497 bool capability = FALSE;
498
499 /*
500 * Since it is not termcap, assume the line is terminfo
501 * format. However, the comma can be embedded in a
502 * description field. It also can be a separator
503 * between a description field and a capability.
504 *
505 * Improve the guess by checking if the next word after
506 * the comma does not look like a capability. In that
507 * case, extend the description past the comma.
508 */
509 for (s = bufptr; isspace(UChar(*s)); ++s) {
510 ;
511 }
512 if (islower(UChar(*s))) {
513 char *name = s;
514 while (isalnum(UChar(*s))) {
515 ++s;
516 }
517 if (*s == '#' || *s == '=' || *s == '@') {
518 /*
519 * Checking solely with syntax allows us to
520 * support extended capabilities with string
521 * values.
522 */
523 capability = TRUE;
524 } else if (*s == ',') {
525 c0 = *s;
526 *s = '\0';
527 /*
528 * Otherwise, we can handle predefined boolean
529 * capabilities, still aided by syntax.
530 */
531 if (_nc_find_entry(name,
532 _nc_get_hash_table(FALSE))) {
533 capability = TRUE;
534 }
535 *s = (char) c0;
536 }
537 }
538 if (capability) {
539 break;
540 }
541 }
542 } else
543 ch = eat_escaped_newline(ch);
544
545 if (OkToAdd()) {
546 AddCh(ch);
547 } else {
548 break;
549 }
550 }
551 *tok_ptr = '\0';
552 if (_nc_syntax == ERR) {
553 /*
554 * Grrr...what we ought to do here is barf, complaining that
555 * the entry is malformed. But because a couple of name fields
556 * in the 8.2 termcap file end with |\, we just have to assume
557 * it's termcap syntax.
558 */
559 _nc_syntax = SYN_TERMCAP;
560 separator = ':';
561 } else if (_nc_syntax == SYN_TERMINFO) {
562 /* throw away trailing /, *$/ */
563 for (--tok_ptr;
564 iswhite(*tok_ptr) || *tok_ptr == ',';
565 tok_ptr--)
566 continue;
567 tok_ptr[1] = '\0';
568 }
569
570 /*
571 * This is the soonest we have the terminal name fetched. Set up
572 * for following warning messages. If there's no '|', then there
573 * is no description.
574 */
575 if (after_name != 0) {
576 ch = *after_name;
577 *after_name = '\0';
578 _nc_set_type(tok_buf);
579 *after_name = (char) ch;
580 }
581
582 /*
583 * Compute the boundary between the aliases and the description
584 * field for syntax-checking purposes.
585 */
586 if (after_list != 0) {
587 if (!silent) {
588 if (*after_list == '\0')
589 _nc_warning("empty longname field");
590 #ifndef FREEBSD_NATIVE
591 else if (strchr(after_list, ' ') == 0)
592 _nc_warning("older tic versions may treat the description field as an alias");
593 #endif
594 }
595 } else {
596 after_list = tok_buf + strlen(tok_buf);
597 DEBUG(1, ("missing description"));
598 }
599
600 /*
601 * Whitespace in a name field other than the long name can confuse
602 * rdist and some termcap tools. Slashes are a no-no. Other
603 * special characters can be dangerous due to shell expansion.
604 */
605 for (s = tok_buf; s < after_list; ++s) {
606 if (isspace(UChar(*s))) {
607 if (!silent)
608 _nc_warning("whitespace in name or alias field");
609 break;
610 } else if (*s == '/') {
611 if (!silent)
612 _nc_warning("slashes aren't allowed in names or aliases");
613 break;
614 } else if (strchr("$[]!*?", *s)) {
615 if (!silent)
616 _nc_warning("dubious character `%c' in name or alias field", *s);
617 break;
618 }
619 }
620
621 _nc_curr_token.tk_name = tok_buf;
622 type = NAMES;
623 } else {
624 if (had_newline && _nc_syntax == SYN_TERMCAP) {
625 _nc_warning("Missing backslash before newline");
626 had_newline = FALSE;
627 }
628 while ((ch = next_char()) != EOF) {
629 if (!isalnum(UChar(ch))) {
630 if (_nc_syntax == SYN_TERMINFO) {
631 if (ch != '_')
632 break;
633 } else { /* allow ';' for "k;" */
634 if (ch != ';')
635 break;
636 }
637 }
638 if (OkToAdd()) {
639 AddCh(ch);
640 } else {
641 ch = EOF;
642 break;
643 }
644 }
645
646 *tok_ptr++ = '\0'; /* separate name/value in buffer */
647 switch (ch) {
648 case ',':
649 case ':':
650 if (ch != separator)
651 _nc_err_abort("Separator inconsistent with syntax");
652 _nc_curr_token.tk_name = tok_buf;
653 type = BOOLEAN;
654 break;
655 case '@':
656 if ((ch = next_char()) != separator && !silent)
657 _nc_warning("Missing separator after `%s', have %s",
658 tok_buf, unctrl(UChar(ch)));
659 _nc_curr_token.tk_name = tok_buf;
660 type = CANCEL;
661 break;
662
663 case '#':
664 found = 0;
665 while (isalnum(ch = next_char())) {
666 numbuf[found++] = (char) ch;
667 if (found >= sizeof(numbuf) - 1)
668 break;
669 }
670 numbuf[found] = '\0';
671 number = strtol(numbuf, &numchk, 0);
672 if (!silent) {
673 if (numchk == numbuf)
674 _nc_warning("no value given for `%s'", tok_buf);
675 if ((*numchk != '\0') || (ch != separator))
676 _nc_warning("Missing separator");
677 }
678 _nc_curr_token.tk_name = tok_buf;
679 _nc_curr_token.tk_valnumber = (int) number;
680 type = NUMBER;
681 break;
682
683 case '=':
684 ch = _nc_trans_string(tok_ptr, tok_buf + TOK_BUF_SIZE);
685 if (!silent && ch != separator)
686 _nc_warning("Missing separator");
687 _nc_curr_token.tk_name = tok_buf;
688 _nc_curr_token.tk_valstring = tok_ptr;
689 type = STRING;
690 break;
691
692 case EOF:
693 type = EOF;
694 break;
695 default:
696 /* just to get rid of the compiler warning */
697 type = UNDEF;
698 if (!silent)
699 _nc_warning("Illegal character - '%s'", unctrl(UChar(ch)));
700 }
701 } /* end else (first_column == FALSE) */
702 } /* end else (ch != EOF) */
703
704 end_of_token:
705
706 #ifdef TRACE
707 if (dot_flag == TRUE)
708 DEBUG(8, ("Commented out "));
709
710 if (_nc_tracing >= DEBUG_LEVEL(8)) {
711 _tracef("parsed %d.%d to %d.%d",
712 old_line, old_col,
713 _nc_curr_line, _nc_curr_col);
714 }
715 if (_nc_tracing >= DEBUG_LEVEL(7)) {
716 switch (type) {
717 case BOOLEAN:
718 _tracef("Token: Boolean; name='%s'",
719 _nc_curr_token.tk_name);
720 break;
721
722 case NUMBER:
723 _tracef("Token: Number; name='%s', value=%d",
724 _nc_curr_token.tk_name,
725 _nc_curr_token.tk_valnumber);
726 break;
727
728 case STRING:
729 _tracef("Token: String; name='%s', value=%s",
730 _nc_curr_token.tk_name,
731 _nc_visbuf(_nc_curr_token.tk_valstring));
732 break;
733
734 case CANCEL:
735 _tracef("Token: Cancel; name='%s'",
736 _nc_curr_token.tk_name);
737 break;
738
739 case NAMES:
740
741 _tracef("Token: Names; value='%s'",
742 _nc_curr_token.tk_name);
743 break;
744
745 case EOF:
746 _tracef("Token: End of file");
747 break;
748
749 default:
750 _nc_warning("Bad token type");
751 }
752 }
753 #endif
754
755 if (dot_flag == TRUE) /* if commented out, use the next one */
756 type = _nc_get_token(silent);
757
758 DEBUG(3, ("token: `%s', class %d",
759 ((_nc_curr_token.tk_name != 0)
760 ? _nc_curr_token.tk_name
761 : "<null>"),
762 type));
763
764 return (type);
765 }
766
767 /*
768 * char
769 * trans_string(ptr)
770 *
771 * Reads characters using next_char() until encountering a separator, nl,
772 * or end-of-file. The returned value is the character which caused
773 * reading to stop. The following translations are done on the input:
774 *
775 * ^X goes to ctrl-X (i.e. X & 037)
776 * {\E,\n,\r,\b,\t,\f} go to
777 * {ESCAPE,newline,carriage-return,backspace,tab,formfeed}
778 * {\^,\\} go to {carat,backslash}
779 * \ddd (for ddd = up to three octal digits) goes to the character ddd
780 *
781 * \e == \E
782 * \0 == \200
783 *
784 */
785
786 NCURSES_EXPORT(int)
_nc_trans_string(char * ptr,char * last)787 _nc_trans_string(char *ptr, char *last)
788 {
789 int count = 0;
790 int number = 0;
791 int i, c;
792 int last_ch = '\0';
793 bool ignored = FALSE;
794 bool long_warning = FALSE;
795
796 while ((c = next_char()) != separator && c != EOF) {
797 if (ptr >= (last - 1)) {
798 if (c != EOF) {
799 while ((c = next_char()) != separator && c != EOF) {
800 ;
801 }
802 }
803 break;
804 }
805 if ((_nc_syntax == SYN_TERMCAP) && c == '\n')
806 break;
807 if (c == '^' && last_ch != '%') {
808 c = next_char();
809 if (c == EOF)
810 _nc_err_abort(MSG_NO_INPUTS);
811
812 if (!(is7bits(c) && isprint(c))) {
813 _nc_warning("Illegal ^ character - '%s'", unctrl(UChar(c)));
814 }
815 if (c == '?' && (_nc_syntax != SYN_TERMCAP)) {
816 *(ptr++) = '\177';
817 if (_nc_tracing)
818 _nc_warning("Allow ^? as synonym for \\177");
819 } else {
820 if ((c &= 037) == 0)
821 c = 128;
822 *(ptr++) = (char) (c);
823 }
824 } else if (c == '\\') {
825 bool strict_bsd = ((_nc_syntax == SYN_TERMCAP) && _nc_strict_bsd);
826
827 c = next_char();
828 if (c == EOF)
829 _nc_err_abort(MSG_NO_INPUTS);
830
831 #define isoctal(c) ((c) >= '0' && (c) <= '7')
832
833 if (isoctal(c) || (strict_bsd && isdigit(c))) {
834 number = c - '0';
835 for (i = 0; i < 2; i++) {
836 c = next_char();
837 if (c == EOF)
838 _nc_err_abort(MSG_NO_INPUTS);
839
840 if (!isoctal(c)) {
841 if (isdigit(c)) {
842 if (!strict_bsd) {
843 _nc_warning("Non-octal digit `%c' in \\ sequence", c);
844 /* allow the digit; it'll do less harm */
845 }
846 } else {
847 push_back(c);
848 break;
849 }
850 }
851
852 number = number * 8 + c - '0';
853 }
854
855 number = UChar(number);
856 if (number == 0 && !strict_bsd)
857 number = 0200;
858 *(ptr++) = (char) number;
859 } else {
860 switch (c) {
861 case 'E':
862 *(ptr++) = '\033';
863 break;
864
865 case 'n':
866 *(ptr++) = '\n';
867 break;
868
869 case 'r':
870 *(ptr++) = '\r';
871 break;
872
873 case 'b':
874 *(ptr++) = '\010';
875 break;
876
877 case 'f':
878 *(ptr++) = '\014';
879 break;
880
881 case 't':
882 *(ptr++) = '\t';
883 break;
884
885 case '\\':
886 *(ptr++) = '\\';
887 break;
888
889 case '^':
890 *(ptr++) = '^';
891 break;
892
893 case ',':
894 *(ptr++) = ',';
895 break;
896
897 case '\n':
898 continue;
899
900 default:
901 if ((_nc_syntax == SYN_TERMINFO) || !_nc_strict_bsd) {
902 switch (c) {
903 case 'a':
904 c = '\007';
905 break;
906 case 'e':
907 c = '\033';
908 break;
909 case 'l':
910 c = '\n';
911 break;
912 case 's':
913 c = ' ';
914 break;
915 case ':':
916 c = ':';
917 break;
918 default:
919 _nc_warning("Illegal character '%s' in \\ sequence",
920 unctrl(UChar(c)));
921 break;
922 }
923 }
924 /* FALLTHRU */
925 case '|':
926 *(ptr++) = (char) c;
927 } /* endswitch (c) */
928 } /* endelse (c < '0' || c > '7') */
929 }
930 /* end else if (c == '\\') */
931 else if (c == '\n' && (_nc_syntax == SYN_TERMINFO)) {
932 /*
933 * Newlines embedded in a terminfo string are ignored, provided
934 * that the next line begins with whitespace.
935 */
936 ignored = TRUE;
937 } else {
938 *(ptr++) = (char) c;
939 }
940
941 if (!ignored) {
942 if (_nc_curr_col <= 1) {
943 push_back(c);
944 c = '\n';
945 break;
946 }
947 last_ch = c;
948 count++;
949 }
950 ignored = FALSE;
951
952 if (count > MAXCAPLEN && !long_warning) {
953 _nc_warning("Very long string found. Missing separator?");
954 long_warning = TRUE;
955 }
956 } /* end while */
957
958 *ptr = '\0';
959
960 return (c);
961 }
962
963 /*
964 * _nc_push_token()
965 *
966 * Push a token of given type so that it will be reread by the next
967 * get_token() call.
968 */
969
970 NCURSES_EXPORT(void)
_nc_push_token(int tokclass)971 _nc_push_token(int tokclass)
972 {
973 /*
974 * This implementation is kind of bogus, it will fail if we ever do more
975 * than one pushback at a time between get_token() calls. It relies on the
976 * fact that _nc_curr_token is static storage that nothing but
977 * _nc_get_token() touches.
978 */
979 pushtype = tokclass;
980 if (pushname == 0)
981 pushname = typeMalloc(char, MAX_NAME_SIZE + 1);
982 _nc_get_type(pushname);
983
984 DEBUG(3, ("pushing token: `%s', class %d",
985 ((_nc_curr_token.tk_name != 0)
986 ? _nc_curr_token.tk_name
987 : "<null>"),
988 pushtype));
989 }
990
991 /*
992 * Panic mode error recovery - skip everything until a "ch" is found.
993 */
994 NCURSES_EXPORT(void)
_nc_panic_mode(char ch)995 _nc_panic_mode(char ch)
996 {
997 int c;
998
999 for (;;) {
1000 c = next_char();
1001 if (c == ch)
1002 return;
1003 if (c == EOF)
1004 return;
1005 }
1006 }
1007
1008 #if NO_LEAKS
1009 NCURSES_EXPORT(void)
_nc_comp_scan_leaks(void)1010 _nc_comp_scan_leaks(void)
1011 {
1012 if (pushname != 0) {
1013 FreeAndNull(pushname);
1014 }
1015 if (tok_buf != 0) {
1016 FreeAndNull(tok_buf);
1017 }
1018 }
1019 #endif
1020