1 /****************************************************************************
2 * Copyright (c) 1998-2002,2003 Free Software Foundation, Inc. *
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: *
12 * The above copyright notice and this permission notice shall be included *
13 * in all copies or substantial portions of the Software. *
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. *
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 *
27 ****************************************************************************/
29 /****************************************************************************
30 * Author: Zeyd M. Ben-Halim <zmbenhal@netcom.com> 1992,1995 *
31 * and: Eric S. Raymond <esr@snark.thyrsus.com> *
32 ****************************************************************************/
35 * comp_scan.c --- Lexical scanner for terminfo compiler.
42 * long _nc_curr_file_pos;
43 * long _nc_comment_start;
44 * long _nc_comment_end;
47 #include <curses.priv.h>
50 #include <term_entry.h>
53 MODULE_ID("$Id: comp_scan.c,v 1.66 2003/12/14 00:32:40 tom Exp $")
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".
61 #define iswhite(ch) (ch == ' ' || ch == '\t')
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 */
74 NCURSES_EXPORT_VAR(struct token)
80 /*****************************************************************************
82 * Token-grabbing machinery
84 *****************************************************************************/
86 static bool first_column; /* See 'next_char()' below */
87 static bool had_newline;
88 static char separator; /* capability separator */
89 static int pushtype; /* type of pushback token */
90 static char *pushname;
93 NCURSES_EXPORT_VAR(bool)
94 _nc_disable_period = FALSE; /* used by tic -a option */
97 static bool end_of_stream(void);
98 static int last_char(void);
99 static int next_char(void);
100 static long stream_pos(void);
101 static void push_back(char c);
103 /* Assume we may be looking at a termcap-style continuation */
105 eat_escaped_newline(int ch)
108 while ((ch = next_char()) == '\n' || iswhite(ch))
117 * Scans the input for the next token, storing the specifics in the
118 * global structure 'curr_token' and returning one of the following:
120 * NAMES A line beginning in column 1. 'name'
121 * will be set to point to everything up to but
122 * not including the first separator on the line.
123 * BOOLEAN An entry consisting of a name followed by
124 * a separator. 'name' will be set to point to
125 * the name of the capability.
126 * NUMBER An entry of the form
128 * 'name' will be set to point to the capability
129 * name and 'valnumber' to the number given.
130 * STRING An entry of the form
132 * 'name' is set to the capability name and
133 * 'valstring' to the string of characters, with
134 * input translations done.
135 * CANCEL An entry of the form
137 * 'name' is set to the capability name and
139 * EOF The end of the file has been reached.
141 * A `separator' is either a comma or a semicolon, depending on whether
142 * we are in termcap or terminfo mode.
147 _nc_get_token(bool silent)
149 static const char terminfo_punct[] = "@%&*!#";
159 int dot_flag = FALSE;
165 if (pushtype != NO_PUSHBACK) {
166 int retval = pushtype;
168 _nc_set_type(pushname != 0 ? pushname : "");
169 DEBUG(3, ("pushed-back token: `%s', class %d",
170 _nc_curr_token.tk_name, pushtype));
172 pushtype = NO_PUSHBACK;
176 /* currtok wasn't altered by _nc_push_token() */
180 if (end_of_stream()) {
188 token_start = stream_pos();
189 while ((ch = next_char()) == '\n' || iswhite(ch)) {
195 ch = eat_escaped_newline(ch);
200 /* if this is a termcap entry, skip a leading separator */
201 if (separator == ':' && ch == ':')
205 #if NCURSES_EXT_FUNCS
206 && !_nc_disable_period
210 DEBUG(8, ("dot-flag set"));
212 while ((ch = next_char()) == '.' || iswhite(ch))
221 /* have to make some punctuation chars legal for terminfo */
223 #if NCURSES_EXT_FUNCS
224 && !(ch == '.' && _nc_disable_period)
226 && !strchr(terminfo_punct, (char) ch)) {
228 _nc_warning("Illegal character (expected alphanumeric or %s) - %s",
229 terminfo_punct, unctrl((chtype) ch));
230 _nc_panic_mode(separator);
235 buffer = typeMalloc(char, MAX_ENTRY_SIZE);
241 _nc_comment_start = token_start;
242 _nc_comment_end = _nc_curr_file_pos;
243 _nc_start_line = _nc_curr_line;
248 while ((ch = next_char()) != '\n') {
250 _nc_err_abort(MSG_NO_MEMORY);
251 } else if (ch == '|') {
255 } else if (ch == ':' && last_char() != ',') {
256 _nc_syntax = SYN_TERMCAP;
259 } else if (ch == ',') {
260 _nc_syntax = SYN_TERMINFO;
263 * If we did not see a '|', then we found a name with no
264 * aliases or description.
269 * If we see a comma, we assume this is terminfo unless we
270 * subsequently run into a colon. But we don't stop
271 * looking for a colon until hitting a newline. This
272 * allows commas to be embedded in description fields of
276 ch = eat_escaped_newline(ch);
281 if (_nc_syntax == ERR) {
283 * Grrr...what we ought to do here is barf, complaining that
284 * the entry is malformed. But because a couple of name fields
285 * in the 8.2 termcap file end with |\, we just have to assume
286 * it's termcap syntax.
288 _nc_syntax = SYN_TERMCAP;
290 } else if (_nc_syntax == SYN_TERMINFO) {
291 /* throw away trailing /, *$/ */
292 for (--ptr; iswhite(*ptr) || *ptr == ','; ptr--)
298 * This is the soonest we have the terminal name fetched. Set up
299 * for following warning messages. If there's no '|', then there
302 if (after_name != 0) {
305 _nc_set_type(buffer);
310 * Compute the boundary between the aliases and the description
311 * field for syntax-checking purposes.
313 if (after_list != 0) {
315 if (*after_list == '\0')
316 _nc_warning("empty longname field");
317 else if (strchr(after_list, ' ') == 0)
318 _nc_warning("older tic versions may treat the description field as an alias");
321 after_list = buffer + strlen(buffer);
322 DEBUG(1, ("missing description"));
326 * Whitespace in a name field other than the long name can confuse
327 * rdist and some termcap tools. Slashes are a no-no. Other
328 * special characters can be dangerous due to shell expansion.
330 for (s = buffer; s < after_list; ++s) {
331 if (isspace(UChar(*s))) {
333 _nc_warning("whitespace in name or alias field");
335 } else if (*s == '/') {
337 _nc_warning("slashes aren't allowed in names or aliases");
339 } else if (strchr("$[]!*?", *s)) {
341 _nc_warning("dubious character `%c' in name or alias field", *s);
346 _nc_curr_token.tk_name = buffer;
349 if (had_newline && _nc_syntax == SYN_TERMCAP) {
350 _nc_warning("Missing backslash before newline");
353 while ((ch = next_char()) != EOF) {
355 if (_nc_syntax == SYN_TERMINFO) {
358 } else { /* allow ';' for "k;" */
371 _nc_err_abort("Separator inconsistent with syntax");
372 _nc_curr_token.tk_name = buffer;
376 if ((ch = next_char()) != separator && !silent)
377 _nc_warning("Missing separator after `%s', have %s",
378 buffer, unctrl((chtype) ch));
379 _nc_curr_token.tk_name = buffer;
385 while (isalnum(ch = next_char())) {
386 numbuf[found++] = ch;
387 if (found >= sizeof(numbuf) - 1)
390 numbuf[found] = '\0';
391 number = strtol(numbuf, &numchk, 0);
393 if (numchk == numbuf)
394 _nc_warning("no value given for `%s'", buffer);
395 if ((*numchk != '\0') || (ch != separator))
396 _nc_warning("Missing separator");
398 _nc_curr_token.tk_name = buffer;
399 _nc_curr_token.tk_valnumber = number;
404 ch = _nc_trans_string(ptr, buffer + MAX_ENTRY_SIZE);
405 if (!silent && ch != separator)
406 _nc_warning("Missing separator");
407 _nc_curr_token.tk_name = buffer;
408 _nc_curr_token.tk_valstring = ptr;
416 /* just to get rid of the compiler warning */
419 _nc_warning("Illegal character - %s", unctrl((chtype) ch));
421 } /* end else (first_column == FALSE) */
422 } /* end else (ch != EOF) */
427 if (dot_flag == TRUE)
428 DEBUG(8, ("Commented out "));
430 if (_nc_tracing >= DEBUG_LEVEL(7)) {
433 _tracef("Token: Boolean; name='%s'",
434 _nc_curr_token.tk_name);
438 _tracef("Token: Number; name='%s', value=%d",
439 _nc_curr_token.tk_name,
440 _nc_curr_token.tk_valnumber);
444 _tracef("Token: String; name='%s', value=%s",
445 _nc_curr_token.tk_name,
446 _nc_visbuf(_nc_curr_token.tk_valstring));
450 _tracef("Token: Cancel; name='%s'",
451 _nc_curr_token.tk_name);
456 _tracef("Token: Names; value='%s'",
457 _nc_curr_token.tk_name);
461 _tracef("Token: End of file");
465 _nc_warning("Bad token type");
470 if (dot_flag == TRUE) /* if commented out, use the next one */
471 type = _nc_get_token(silent);
473 DEBUG(3, ("token: `%s', class %d",
474 _nc_curr_token.tk_name != 0 ? _nc_curr_token.tk_name :
485 * Reads characters using next_char() until encountering a separator, nl,
486 * or end-of-file. The returned value is the character which caused
487 * reading to stop. The following translations are done on the input:
489 * ^X goes to ctrl-X (i.e. X & 037)
490 * {\E,\n,\r,\b,\t,\f} go to
491 * {ESCAPE,newline,carriage-return,backspace,tab,formfeed}
492 * {\^,\\} go to {carat,backslash}
493 * \ddd (for ddd = up to three octal digits) goes to the character ddd
501 _nc_trans_string(char *ptr, char *last)
506 chtype ch, last_ch = '\0';
507 bool ignored = FALSE;
508 bool long_warning = FALSE;
510 while ((ch = c = next_char()) != (chtype) separator && c != EOF) {
511 if (ptr == (last - 1))
513 if ((_nc_syntax == SYN_TERMCAP) && c == '\n')
515 if (ch == '^' && last_ch != '%') {
516 ch = c = next_char();
518 _nc_err_abort(MSG_NO_INPUTS);
520 if (!(is7bits(ch) && isprint(ch))) {
521 _nc_warning("Illegal ^ character - %s", unctrl(ch));
526 _nc_warning("Allow ^? as synonym for \\177");
528 if ((ch &= 037) == 0)
530 *(ptr++) = (char) (ch);
532 } else if (ch == '\\') {
533 ch = c = next_char();
535 _nc_err_abort(MSG_NO_INPUTS);
537 if (ch >= '0' && ch <= '7') {
539 for (i = 0; i < 2; i++) {
540 ch = c = next_char();
542 _nc_err_abort(MSG_NO_INPUTS);
544 if (c < '0' || c > '7') {
546 _nc_warning("Non-octal digit `%c' in \\ sequence", c);
547 /* allow the digit; it'll do less harm */
554 number = number * 8 + c - '0';
559 *(ptr++) = (char) number;
616 _nc_warning("Illegal character %s in \\ sequence",
620 *(ptr++) = (char) ch;
621 } /* endswitch (ch) */
622 } /* endelse (ch < '0' || ch > '7') */
624 /* end else if (ch == '\\') */
625 else if (ch == '\n' && (_nc_syntax == SYN_TERMINFO)) {
626 /* newlines embedded in a terminfo string are ignored */
629 *(ptr++) = (char) ch;
638 if (count > MAXCAPLEN && !long_warning) {
639 _nc_warning("Very long string found. Missing separator?");
652 * Push a token of given type so that it will be reread by the next
657 _nc_push_token(int tokclass)
660 * This implementation is kind of bogus, it will fail if we ever do more
661 * than one pushback at a time between get_token() calls. It relies on the
662 * fact that _nc_curr_token is static storage that nothing but
663 * _nc_get_token() touches.
667 pushname = typeMalloc(char, MAX_NAME_SIZE + 1);
668 _nc_get_type(pushname);
670 DEBUG(3, ("pushing token: `%s', class %d",
671 _nc_curr_token.tk_name, pushtype));
675 * Panic mode error recovery - skip everything until a "ch" is found.
678 _nc_panic_mode(char ch)
691 /*****************************************************************************
693 * Character-stream handling
695 *****************************************************************************/
697 #define LEXBUFSIZ 1024
699 static char *bufptr; /* otherwise, the input buffer pointer */
700 static char *bufstart; /* start of buffer so we can compute offsets */
701 static FILE *yyin; /* scanner's input file descriptor */
706 * Resets the input-reading routines. Used on initialization,
707 * or after a seek has been done. Exactly one argument must be
712 _nc_reset_input(FILE *fp, char *buf)
714 pushtype = NO_PUSHBACK;
718 bufstart = bufptr = buf;
719 _nc_curr_file_pos = 0L;
728 * Returns the final nonblank character on the current input buffer
733 size_t len = strlen(bufptr);
735 if (!isspace(UChar(bufptr[len])))
744 * Returns the next character in the input stream. Comments and leading
745 * white space are stripped.
747 * The global state variable 'firstcolumn' is set TRUE if the character
748 * returned is from the first column of the input line.
750 * The global variable _nc_curr_line is incremented for each new line.
751 * The global variable _nc_curr_file_pos is set to the file offset of the
752 * beginning of each line.
760 * An string with an embedded null will truncate the input. This is
761 * intentional (we don't read binary files here).
765 if (*bufptr == '\n') {
769 } else if (!bufptr || !*bufptr) {
771 * In theory this could be recoded to do its I/O one character at a
772 * time, saving the buffer space. In practice, this turns out to be
773 * quite hard to get completely right. Try it and see. If you
774 * succeed, don't forget to hack push_back() correspondingly.
777 static size_t allocated;
785 if (used + (LEXBUFSIZ / 4) >= allocated) {
786 allocated += (allocated + LEXBUFSIZ);
787 result = typeRealloc(char, allocated, result);
792 _nc_curr_file_pos = ftell(yyin);
794 if (fgets(result + used, allocated - used, yyin) != 0) {
802 strcat(result, "\n");
804 if ((bufptr = bufstart) != 0) {
805 used = strlen(bufptr);
806 while (iswhite(*bufptr))
810 * Treat a trailing <cr><lf> the same as a <newline> so we
811 * can read files on OS/2, etc.
813 if ((len = strlen(bufptr)) > 1) {
814 if (bufptr[len - 1] == '\n'
815 && bufptr[len - 2] == '\r') {
817 bufptr[len - 1] = '\n';
824 } while (bufptr[len - 1] != '\n'); /* complete a line */
825 } while (result[0] == '#'); /* ignore comments */
828 first_column = (bufptr == bufstart);
838 /* push a character back onto the input stream */
840 if (bufptr == bufstart)
841 _nc_syserr_abort("Can't backspace off beginning of line");
847 /* return our current character position in the input stream */
849 return (yyin ? ftell(yyin) : (bufptr ? bufptr - bufstart : 0));
854 /* are we at end of input? */
856 return ((yyin ? feof(yyin) : (bufptr && *bufptr == '\0'))