1 /* GNU Emacs routines to deal with syntax tables; also word and list parsing.
2 Copyright (C) 1985, 1987, 1993 Free Software Foundation, Inc.
4 This file is part of GNU Emacs.
6 GNU Emacs is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
11 GNU Emacs is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with GNU Emacs; see the file COPYING. If not, write to
18 the Free Software Foundation, 675 Mass Ave, Cambridge, MA 02139, USA. */
28 Lisp_Object Qsyntax_table_p
;
30 static void scan_sexps_forward ();
31 static int char_quoted ();
33 int words_include_escapes
;
35 /* This is the internal form of the parse state used in parse-partial-sexp. */
37 struct lisp_parse_state
39 int depth
; /* Depth at end of parsing */
40 int instring
; /* -1 if not within string, else desired terminator. */
41 int incomment
; /* Nonzero if within a comment at end of parsing */
42 int comstyle
; /* comment style a=0, or b=1 */
43 int quoted
; /* Nonzero if just after an escape char at end of parsing */
44 int thislevelstart
; /* Char number of most recent start-of-expression at current level */
45 int prevlevelstart
; /* Char number of start of containing expression */
46 int location
; /* Char number at which parsing stopped. */
47 int mindepth
; /* Minimum depth seen while scanning. */
48 int comstart
; /* Position just after last comment starter. */
51 /* These variables are a cache for finding the start of a defun.
52 find_start_pos is the place for which the defun start was found.
53 find_start_value is the defun start position found for it.
54 find_start_buffer is the buffer it was found in.
55 find_start_begv is the BEGV value when it was found.
56 find_start_modiff is the value of MODIFF when it was found. */
58 static int find_start_pos
;
59 static int find_start_value
;
60 static struct buffer
*find_start_buffer
;
61 static int find_start_begv
;
62 static int find_start_modiff
;
64 /* Find a defun-start that is the last one before POS (or nearly the last).
65 We record what we find, so that another call in the same area
66 can return the same value right away. */
69 find_defun_start (pos
)
75 /* Use previous finding, if it's valid and applies to this inquiry. */
76 if (current_buffer
== find_start_buffer
77 /* Reuse the defun-start even if POS is a little farther on.
78 POS might be in the next defun, but that's ok.
79 Our value may not be the best possible, but will still be usable. */
80 && pos
<= find_start_pos
+ 1000
81 && pos
>= find_start_value
82 && BEGV
== find_start_begv
83 && MODIFF
== find_start_modiff
)
84 return find_start_value
;
86 /* Back up to start of line. */
87 tem
= scan_buffer ('\n', pos
, -1, &shortage
, 1);
91 /* Open-paren at start of line means we found our defun-start. */
92 if (SYNTAX (FETCH_CHAR (tem
)) == Sopen
)
94 /* Move to beg of previous line. */
95 tem
= scan_buffer ('\n', tem
, -2, &shortage
, 1);
98 /* Record what we found, for the next try. */
99 find_start_value
= tem
;
100 find_start_buffer
= current_buffer
;
101 find_start_modiff
= MODIFF
;
102 find_start_begv
= BEGV
;
103 find_start_pos
= pos
;
105 return find_start_value
;
108 DEFUN ("syntax-table-p", Fsyntax_table_p
, Ssyntax_table_p
, 1, 1, 0,
109 "Return t if ARG is a syntax table.\n\
110 Any vector of 256 elements will do.")
114 if (XTYPE (obj
) == Lisp_Vector
&& XVECTOR (obj
)->size
== 0400)
120 check_syntax_table (obj
)
123 register Lisp_Object tem
;
124 while (tem
= Fsyntax_table_p (obj
),
126 obj
= wrong_type_argument (Qsyntax_table_p
, obj
);
131 DEFUN ("syntax-table", Fsyntax_table
, Ssyntax_table
, 0, 0, 0,
132 "Return the current syntax table.\n\
133 This is the one specified by the current buffer.")
136 return current_buffer
->syntax_table
;
139 DEFUN ("standard-syntax-table", Fstandard_syntax_table
,
140 Sstandard_syntax_table
, 0, 0, 0,
141 "Return the standard syntax table.\n\
142 This is the one used for new buffers.")
145 return Vstandard_syntax_table
;
148 DEFUN ("copy-syntax-table", Fcopy_syntax_table
, Scopy_syntax_table
, 0, 1, 0,
149 "Construct a new syntax table and return it.\n\
150 It is a copy of the TABLE, which defaults to the standard syntax table.")
154 Lisp_Object size
, val
;
155 XFASTINT (size
) = 0400;
157 val
= Fmake_vector (size
, val
);
159 table
= check_syntax_table (table
);
160 else if (NILP (Vstandard_syntax_table
))
161 /* Can only be null during initialization */
163 else table
= Vstandard_syntax_table
;
165 bcopy (XVECTOR (table
)->contents
,
166 XVECTOR (val
)->contents
, 0400 * sizeof (Lisp_Object
));
170 DEFUN ("set-syntax-table", Fset_syntax_table
, Sset_syntax_table
, 1, 1, 0,
171 "Select a new syntax table for the current buffer.\n\
172 One argument, a syntax table.")
176 table
= check_syntax_table (table
);
177 current_buffer
->syntax_table
= table
;
178 /* Indicate that this buffer now has a specified syntax table. */
179 current_buffer
->local_var_flags
180 |= XFASTINT (buffer_local_flags
.syntax_table
);
184 /* Convert a letter which signifies a syntax code
185 into the code it signifies.
186 This is used by modify-syntax-entry, and other things. */
188 unsigned char syntax_spec_code
[0400] =
189 { 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
190 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
191 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
192 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
193 (char) Swhitespace
, 0377, (char) Sstring
, 0377,
194 (char) Smath
, 0377, 0377, (char) Squote
,
195 (char) Sopen
, (char) Sclose
, 0377, 0377,
196 0377, (char) Swhitespace
, (char) Spunct
, (char) Scharquote
,
197 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
198 0377, 0377, 0377, 0377,
199 (char) Scomment
, 0377, (char) Sendcomment
, 0377,
200 (char) Sinherit
, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */
201 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
202 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword
,
203 0377, 0377, 0377, 0377, (char) Sescape
, 0377, 0377, (char) Ssymbol
,
204 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* `, a, ... */
205 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
206 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword
,
207 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377
210 /* Indexed by syntax code, give the letter that describes it. */
212 char syntax_code_spec
[14] =
214 ' ', '.', 'w', '_', '(', ')', '\'', '\"', '$', '\\', '/', '<', '>', '@'
217 DEFUN ("char-syntax", Fchar_syntax
, Schar_syntax
, 1, 1, 0,
218 "Return the syntax code of CHAR, described by a character.\n\
219 For example, if CHAR is a word constituent, the character `?w' is returned.\n\
220 The characters that correspond to various syntax codes\n\
221 are listed in the documentation of `modify-syntax-entry'.")
225 CHECK_NUMBER (ch
, 0);
226 return make_number (syntax_code_spec
[(int) SYNTAX (0xFF & XINT (ch
))]);
229 /* This comment supplies the doc string for modify-syntax-entry,
230 for make-docfile to see. We cannot put this in the real DEFUN
231 due to limits in the Unix cpp.
233 DEFUN ("modify-syntax-entry", foo, bar, 2, 3, 0,
234 "Set syntax for character CHAR according to string S.\n\
235 The syntax is changed only for table TABLE, which defaults to\n\
236 the current buffer's syntax table.\n\
237 The first character of S should be one of the following:\n\
238 Space or - whitespace syntax. w word constituent.\n\
239 _ symbol constituent. . punctuation.\n\
240 ( open-parenthesis. ) close-parenthesis.\n\
241 \" string quote. \\ escape.\n\
242 $ paired delimiter. ' expression quote or prefix operator.\n\
243 < comment starter. > comment ender.\n\
244 / character-quote. @ inherit from `standard-syntax-table'.\n\
246 Only single-character comment start and end sequences are represented thus.\n\
247 Two-character sequences are represented as described below.\n\
248 The second character of S is the matching parenthesis,\n\
249 used only if the first character is `(' or `)'.\n\
250 Any additional characters are flags.\n\
251 Defined flags are the characters 1, 2, 3, 4, b, and p.\n\
252 1 means C is the start of a two-char comment start sequence.\n\
253 2 means C is the second character of such a sequence.\n\
254 3 means C is the start of a two-char comment end sequence.\n\
255 4 means C is the second character of such a sequence.\n\
257 There can be up to two orthogonal comment sequences. This is to support\n\
258 language modes such as C++. By default, all comment sequences are of style\n\
259 a, but you can set the comment sequence style to b (on the second character\n\
260 of a comment-start, or the first character of a comment-end sequence) using\n\
262 b means C is part of comment sequence b.\n\
264 p means C is a prefix character for `backward-prefix-chars';\n\
265 such characters are treated as whitespace when they occur\n\
266 between expressions.")
270 DEFUN ("modify-syntax-entry", Fmodify_syntax_entry
, Smodify_syntax_entry
, 2, 3,
271 /* I really don't know why this is interactive
272 help-form should at least be made useful whilst reading the second arg
274 "cSet syntax for character: \nsSet syntax for %s to: ",
275 0 /* See immediately above */)
276 (c
, newentry
, syntax_table
)
277 Lisp_Object c
, newentry
, syntax_table
;
279 register unsigned char *p
, match
;
280 register enum syntaxcode code
;
284 CHECK_STRING (newentry
, 1);
285 if (NILP (syntax_table
))
286 syntax_table
= current_buffer
->syntax_table
;
288 syntax_table
= check_syntax_table (syntax_table
);
290 p
= XSTRING (newentry
)->data
;
291 code
= (enum syntaxcode
) syntax_spec_code
[*p
++];
292 if (((int) code
& 0377) == 0377)
293 error ("invalid syntax description letter: %c", c
);
297 if (match
== ' ') match
= 0;
299 XFASTINT (val
) = (match
<< 8) + (int) code
;
304 XFASTINT (val
) |= 1 << 16;
308 XFASTINT (val
) |= 1 << 17;
312 XFASTINT (val
) |= 1 << 18;
316 XFASTINT (val
) |= 1 << 19;
320 XFASTINT (val
) |= 1 << 20;
324 XFASTINT (val
) |= 1 << 21;
328 XVECTOR (syntax_table
)->contents
[0xFF & XINT (c
)] = val
;
333 /* Dump syntax table to buffer in human-readable format */
336 describe_syntax (value
)
339 register enum syntaxcode code
;
340 char desc
, match
, start1
, start2
, end1
, end2
, prefix
, comstyle
;
343 Findent_to (make_number (16), make_number (1));
345 if (XTYPE (value
) != Lisp_Int
)
347 insert_string ("invalid");
351 code
= (enum syntaxcode
) (XINT (value
) & 0377);
352 match
= (XINT (value
) >> 8) & 0377;
353 start1
= (XINT (value
) >> 16) & 1;
354 start2
= (XINT (value
) >> 17) & 1;
355 end1
= (XINT (value
) >> 18) & 1;
356 end2
= (XINT (value
) >> 19) & 1;
357 prefix
= (XINT (value
) >> 20) & 1;
358 comstyle
= (XINT (value
) >> 21) & 1;
360 if ((int) code
< 0 || (int) code
>= (int) Smax
)
362 insert_string ("invalid");
365 desc
= syntax_code_spec
[(int) code
];
367 str
[0] = desc
, str
[1] = 0;
370 str
[0] = match
? match
: ' ';
389 insert_string ("\twhich means: ");
391 #ifdef SWITCH_ENUM_BUG
398 insert_string ("whitespace"); break;
400 insert_string ("punctuation"); break;
402 insert_string ("word"); break;
404 insert_string ("symbol"); break;
406 insert_string ("open"); break;
408 insert_string ("close"); break;
410 insert_string ("quote"); break;
412 insert_string ("string"); break;
414 insert_string ("math"); break;
416 insert_string ("escape"); break;
418 insert_string ("charquote"); break;
420 insert_string ("comment"); break;
422 insert_string ("endcomment"); break;
424 insert_string ("inherit"); break;
426 insert_string ("invalid");
432 insert_string (", matches ");
437 insert_string (",\n\t is the first character of a comment-start sequence");
439 insert_string (",\n\t is the second character of a comment-start sequence");
442 insert_string (",\n\t is the first character of a comment-end sequence");
444 insert_string (",\n\t is the second character of a comment-end sequence");
446 insert_string (" (comment style b)");
449 insert_string (",\n\t is a prefix character for `backward-prefix-chars'");
451 insert_string ("\n");
455 describe_syntax_1 (vector
)
458 struct buffer
*old
= current_buffer
;
459 set_buffer_internal (XBUFFER (Vstandard_output
));
460 describe_vector (vector
, Qnil
, describe_syntax
, 0, Qnil
);
461 set_buffer_internal (old
);
465 DEFUN ("describe-syntax", Fdescribe_syntax
, Sdescribe_syntax
, 0, 0, "",
466 "Describe the syntax specifications in the syntax table.\n\
467 The descriptions are inserted in a buffer, which is then displayed.")
470 internal_with_output_to_temp_buffer
471 ("*Help*", describe_syntax_1
, current_buffer
->syntax_table
);
476 /* Return the position across COUNT words from FROM.
477 If that many words cannot be found before the end of the buffer, return 0.
478 COUNT negative means scan backward and stop at word beginning. */
480 scan_words (from
, count
)
481 register int from
, count
;
483 register int beg
= BEGV
;
484 register int end
= ZV
;
499 code
= SYNTAX (FETCH_CHAR (from
));
500 if (words_include_escapes
501 && (code
== Sescape
|| code
== Scharquote
))
509 if (from
== end
) break;
510 code
= SYNTAX (FETCH_CHAR (from
));
511 if (!(words_include_escapes
512 && (code
== Sescape
|| code
== Scharquote
)))
528 code
= SYNTAX (FETCH_CHAR (from
- 1));
529 if (words_include_escapes
530 && (code
== Sescape
|| code
== Scharquote
))
538 if (from
== beg
) break;
539 code
= SYNTAX (FETCH_CHAR (from
- 1));
540 if (!(words_include_escapes
541 && (code
== Sescape
|| code
== Scharquote
)))
554 DEFUN ("forward-word", Fforward_word
, Sforward_word
, 1, 1, "p",
555 "Move point forward ARG words (backward if ARG is negative).\n\
556 Normally returns t.\n\
557 If an edge of the buffer is reached, point is left there\n\
558 and nil is returned.")
563 CHECK_NUMBER (count
, 0);
565 if (!(val
= scan_words (point
, XINT (count
))))
567 SET_PT (XINT (count
) > 0 ? ZV
: BEGV
);
574 DEFUN ("forward-comment", Fforward_comment
, Sforward_comment
, 1, 1, 0,
575 "Move forward across up to N comments. If N is negative, move backward.\n\
576 Stop scanning if we find something other than a comment or whitespace.\n\
577 Set point to where scanning stops.\n\
578 If N comments are found as expected, with nothing except whitespace\n\
579 between them, return t; otherwise return nil.")
586 register enum syntaxcode code
;
587 int comstyle
= 0; /* style of comment encountered */
591 CHECK_NUMBER (count
, 0);
592 count1
= XINT (count
);
604 c
= FETCH_CHAR (from
);
608 if (from
< stop
&& SYNTAX_COMSTART_FIRST (c
)
609 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from
)))
611 /* We have encountered a comment start sequence and we
612 are ignoring all text inside comments. We must record
613 the comment style this sequence begins so that later,
614 only a comment end of the same style actually ends
615 the comment section. */
617 comstyle
= SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
));
621 if (code
== Scomment
)
631 c
= FETCH_CHAR (from
);
632 if (SYNTAX (c
) == Sendcomment
633 && SYNTAX_COMMENT_STYLE (c
) == comstyle
)
634 /* we have encountered a comment end of the same style
635 as the comment sequence which began this comment
639 if (from
< stop
&& SYNTAX_COMEND_FIRST (c
)
640 && SYNTAX_COMEND_SECOND (FETCH_CHAR (from
))
641 && SYNTAX_COMMENT_STYLE (c
) == comstyle
)
642 /* we have encountered a comment end of the same style
643 as the comment sequence which began this comment
647 /* We have skipped one comment. */
650 else if (code
!= Swhitespace
&& code
!= Sendcomment
)
658 /* End of comment reached */
670 quoted
= char_quoted (from
);
673 c
= FETCH_CHAR (from
);
676 if (code
== Sendcomment
)
677 comstyle
= SYNTAX_COMMENT_STYLE (c
);
678 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
679 && SYNTAX_COMEND_FIRST (FETCH_CHAR (from
- 1))
680 && !char_quoted (from
- 1))
682 /* We must record the comment style encountered so that
683 later, we can match only the proper comment begin
684 sequence of the same style. */
686 comstyle
= SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
- 1));
690 if (code
== Sendcomment
&& !quoted
)
693 if (code
!= SYNTAX (c
))
694 /* For a two-char comment ender, we can assume
695 it does end a comment. So scan back in a simple way. */
697 if (from
!= stop
) from
--;
700 if (SYNTAX (c
= FETCH_CHAR (from
)) == Scomment
701 && SYNTAX_COMMENT_STYLE (c
) == comstyle
)
710 if (SYNTAX_COMSTART_SECOND (c
)
711 && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from
))
712 && SYNTAX_COMMENT_STYLE (c
) == comstyle
713 && !char_quoted (from
))
720 /* Look back, counting the parity of string-quotes,
721 and recording the comment-starters seen.
722 When we reach a safe place, assume that's not in a string;
723 then step the main scan to the earliest comment-starter seen
724 an even number of string quotes away from the safe place.
726 OFROM[I] is position of the earliest comment-starter seen
727 which is I+2X quotes from the comment-end.
728 PARITY is current parity of quotes from the comment end. */
731 char my_stringend
= 0;
732 int string_lossage
= 0;
733 int comment_end
= from
;
734 int comstart_pos
= 0;
735 int comstart_parity
= 0;
737 /* At beginning of range to scan, we're outside of strings;
738 that determines quote parity to the comment-end. */
741 /* Move back and examine a character. */
744 c
= FETCH_CHAR (from
);
747 /* If this char is the second of a 2-char comment sequence,
748 back up and give the pair the appropriate syntax. */
749 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
750 && SYNTAX_COMEND_FIRST (FETCH_CHAR (from
- 1)))
756 else if (from
> stop
&& SYNTAX_COMSTART_SECOND (c
)
757 && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from
- 1))
758 && comstyle
== SYNTAX_COMMENT_STYLE (c
))
764 /* Ignore escaped characters. */
765 if (char_quoted (from
))
768 /* Track parity of quotes. */
772 if (my_stringend
== 0)
774 /* If we have two kinds of string delimiters.
775 There's no way to grok this scanning backwards. */
776 else if (my_stringend
!= c
)
780 /* Record comment-starters according to that
781 quote-parity to the comment-end. */
782 if (code
== Scomment
)
784 comstart_parity
= parity
;
788 /* If we find another earlier comment-ender,
789 any comment-starts earlier than that don't count
790 (because they go with the earlier comment-ender). */
791 if (code
== Sendcomment
792 && SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
)) == comstyle
)
795 /* Assume a defun-start point is outside of strings. */
797 && (from
== stop
|| FETCH_CHAR (from
- 1) == '\n'))
801 if (comstart_pos
== 0)
803 /* If the earliest comment starter
804 is followed by uniform paired string quotes or none,
805 we know it can't be inside a string
806 since if it were then the comment ender would be inside one.
807 So it does start a comment. Skip back to it. */
808 else if (comstart_parity
== 0 && !string_lossage
)
812 /* We had two kinds of string delimiters mixed up
813 together. Decode this going forwards.
814 Scan fwd from the previous comment ender
815 to the one in question; this records where we
816 last passed a comment starter. */
817 struct lisp_parse_state state
;
818 scan_sexps_forward (&state
, find_defun_start (comment_end
),
819 comment_end
- 1, -10000, 0, Qnil
, 0);
821 from
= state
.comstart
;
823 /* We can't grok this as a comment; scan it normally. */
827 /* We have skipped one comment. */
830 else if ((code
!= Swhitespace
&& code
!= Scomment
) || quoted
)
846 int parse_sexp_ignore_comments
;
849 scan_lists (from
, count
, depth
, sexpflag
)
851 int count
, depth
, sexpflag
;
859 register enum syntaxcode code
;
860 int min_depth
= depth
; /* Err out if depth gets less than this. */
861 int comstyle
= 0; /* style of comment encountered */
863 if (depth
> 0) min_depth
= 0;
873 c
= FETCH_CHAR (from
);
876 if (from
< stop
&& SYNTAX_COMSTART_FIRST (c
)
877 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from
))
878 && parse_sexp_ignore_comments
)
880 /* we have encountered a comment start sequence and we
881 are ignoring all text inside comments. we must record
882 the comment style this sequence begins so that later,
883 only a comment end of the same style actually ends
884 the comment section */
886 comstyle
= SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
));
890 if (SYNTAX_PREFIX (c
))
893 #ifdef SWITCH_ENUM_BUG
901 if (from
== stop
) goto lose
;
903 /* treat following character as a word constituent */
906 if (depth
|| !sexpflag
) break;
907 /* This word counts as a sexp; return at end of it. */
910 #ifdef SWITCH_ENUM_BUG
911 switch ((int) SYNTAX (FETCH_CHAR (from
)))
913 switch (SYNTAX (FETCH_CHAR (from
)))
919 if (from
== stop
) goto lose
;
933 if (!parse_sexp_ignore_comments
) break;
936 if (from
== stop
) goto done
;
937 c
= FETCH_CHAR (from
);
938 if (SYNTAX (c
) == Sendcomment
939 && SYNTAX_COMMENT_STYLE (c
) == comstyle
)
940 /* we have encountered a comment end of the same style
941 as the comment sequence which began this comment
945 if (from
< stop
&& SYNTAX_COMEND_FIRST (c
)
946 && SYNTAX_COMEND_SECOND (FETCH_CHAR (from
))
947 && SYNTAX_COMMENT_STYLE (c
) == comstyle
)
948 /* we have encountered a comment end of the same style
949 as the comment sequence which began this comment
958 if (from
!= stop
&& c
== FETCH_CHAR (from
))
968 if (!++depth
) goto done
;
973 if (!--depth
) goto done
;
974 if (depth
< min_depth
)
975 error ("Containing expression ends prematurely");
979 stringterm
= FETCH_CHAR (from
- 1);
982 if (from
>= stop
) goto lose
;
983 if (FETCH_CHAR (from
) == stringterm
) break;
984 #ifdef SWITCH_ENUM_BUG
985 switch ((int) SYNTAX (FETCH_CHAR (from
)))
987 switch (SYNTAX (FETCH_CHAR (from
)))
997 if (!depth
&& sexpflag
) goto done
;
1002 /* Reached end of buffer. Error if within object, return nil if between */
1003 if (depth
) goto lose
;
1008 /* End of object reached */
1019 if (quoted
= char_quoted (from
))
1021 c
= FETCH_CHAR (from
);
1024 if (code
== Sendcomment
)
1025 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1026 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
1027 && SYNTAX_COMEND_FIRST (FETCH_CHAR (from
- 1))
1028 && !char_quoted (from
- 1)
1029 && parse_sexp_ignore_comments
)
1031 /* we must record the comment style encountered so that
1032 later, we can match only the proper comment begin
1033 sequence of the same style */
1035 comstyle
= SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
- 1));
1039 if (SYNTAX_PREFIX (c
))
1042 #ifdef SWITCH_ENUM_BUG
1043 switch ((int) (quoted
? Sword
: code
))
1045 switch (quoted
? Sword
: code
)
1050 if (depth
|| !sexpflag
) break;
1051 /* This word counts as a sexp; count object finished after passing it. */
1054 quoted
= char_quoted (from
- 1);
1057 if (! (quoted
|| SYNTAX (FETCH_CHAR (from
- 1)) == Sword
1058 || SYNTAX (FETCH_CHAR (from
- 1)) == Ssymbol
1059 || SYNTAX (FETCH_CHAR (from
- 1)) == Squote
))
1068 if (from
!= stop
&& c
== FETCH_CHAR (from
- 1))
1078 if (!++depth
) goto done2
;
1083 if (!--depth
) goto done2
;
1084 if (depth
< min_depth
)
1085 error ("Containing expression ends prematurely");
1089 if (!parse_sexp_ignore_comments
)
1092 if (code
!= SYNTAX (c
))
1093 /* For a two-char comment ender, we can assume
1094 it does end a comment. So scan back in a simple way. */
1096 if (from
!= stop
) from
--;
1099 if (SYNTAX (c
= FETCH_CHAR (from
)) == Scomment
1100 && SYNTAX_COMMENT_STYLE (c
) == comstyle
)
1102 if (from
== stop
) goto done
;
1104 if (SYNTAX_COMSTART_SECOND (c
)
1105 && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from
))
1106 && SYNTAX_COMMENT_STYLE (c
) == comstyle
1107 && !char_quoted (from
))
1114 /* Look back, counting the parity of string-quotes,
1115 and recording the comment-starters seen.
1116 When we reach a safe place, assume that's not in a string;
1117 then step the main scan to the earliest comment-starter seen
1118 an even number of string quotes away from the safe place.
1120 OFROM[I] is position of the earliest comment-starter seen
1121 which is I+2X quotes from the comment-end.
1122 PARITY is current parity of quotes from the comment end. */
1125 char my_stringend
= 0;
1126 int string_lossage
= 0;
1127 int comment_end
= from
;
1128 int comstart_pos
= 0;
1129 int comstart_parity
= 0;
1131 /* At beginning of range to scan, we're outside of strings;
1132 that determines quote parity to the comment-end. */
1133 while (from
!= stop
)
1135 /* Move back and examine a character. */
1138 c
= FETCH_CHAR (from
);
1141 /* If this char is the second of a 2-char comment sequence,
1142 back up and give the pair the appropriate syntax. */
1143 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
1144 && SYNTAX_COMEND_FIRST (FETCH_CHAR (from
- 1)))
1150 else if (from
> stop
&& SYNTAX_COMSTART_SECOND (c
)
1151 && SYNTAX_COMSTART_FIRST (FETCH_CHAR (from
- 1))
1152 && comstyle
== SYNTAX_COMMENT_STYLE (c
))
1158 /* Ignore escaped characters. */
1159 if (char_quoted (from
))
1162 /* Track parity of quotes. */
1163 if (code
== Sstring
)
1166 if (my_stringend
== 0)
1168 /* If we have two kinds of string delimiters.
1169 There's no way to grok this scanning backwards. */
1170 else if (my_stringend
!= c
)
1174 /* Record comment-starters according to that
1175 quote-parity to the comment-end. */
1176 if (code
== Scomment
)
1178 comstart_parity
= parity
;
1179 comstart_pos
= from
;
1182 /* If we find another earlier comment-ender,
1183 any comment-starts earlier than that don't count
1184 (because they go with the earlier comment-ender). */
1185 if (code
== Sendcomment
1186 && SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
)) == comstyle
)
1189 /* Assume a defun-start point is outside of strings. */
1191 && (from
== stop
|| FETCH_CHAR (from
- 1) == '\n'))
1195 if (comstart_pos
== 0)
1197 /* If the earliest comment starter
1198 is followed by uniform paired string quotes or none,
1199 we know it can't be inside a string
1200 since if it were then the comment ender would be inside one.
1201 So it does start a comment. Skip back to it. */
1202 else if (comstart_parity
== 0 && !string_lossage
)
1203 from
= comstart_pos
;
1206 /* We had two kinds of string delimiters mixed up
1207 together. Decode this going forwards.
1208 Scan fwd from the previous comment ender
1209 to the one in question; this records where we
1210 last passed a comment starter. */
1211 struct lisp_parse_state state
;
1212 scan_sexps_forward (&state
, find_defun_start (comment_end
),
1213 comment_end
- 1, -10000, 0, Qnil
, 0);
1214 if (state
.incomment
)
1215 from
= state
.comstart
;
1217 /* We can't grok this as a comment; scan it normally. */
1224 stringterm
= FETCH_CHAR (from
);
1227 if (from
== stop
) goto lose
;
1228 if (!char_quoted (from
- 1)
1229 && stringterm
== FETCH_CHAR (from
- 1))
1234 if (!depth
&& sexpflag
) goto done2
;
1239 /* Reached start of buffer. Error if within object, return nil if between */
1240 if (depth
) goto lose
;
1251 XFASTINT (val
) = from
;
1255 error ("Unbalanced parentheses");
1263 register enum syntaxcode code
;
1264 register int beg
= BEGV
;
1265 register int quoted
= 0;
1268 && ((code
= SYNTAX (FETCH_CHAR (pos
- 1))) == Scharquote
1269 || code
== Sescape
))
1270 pos
--, quoted
= !quoted
;
1274 DEFUN ("scan-lists", Fscan_lists
, Sscan_lists
, 3, 3, 0,
1275 "Scan from character number FROM by COUNT lists.\n\
1276 Returns the character number of the position thus found.\n\
1278 If DEPTH is nonzero, paren depth begins counting from that value,\n\
1279 only places where the depth in parentheses becomes zero\n\
1280 are candidates for stopping; COUNT such places are counted.\n\
1281 Thus, a positive value for DEPTH means go out levels.\n\
1283 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.\n\
1285 If the beginning or end of (the accessible part of) the buffer is reached\n\
1286 and the depth is wrong, an error is signaled.\n\
1287 If the depth is right but the count is not used up, nil is returned.")
1288 (from
, count
, depth
)
1289 Lisp_Object from
, count
, depth
;
1291 CHECK_NUMBER (from
, 0);
1292 CHECK_NUMBER (count
, 1);
1293 CHECK_NUMBER (depth
, 2);
1295 return scan_lists (XINT (from
), XINT (count
), XINT (depth
), 0);
1298 DEFUN ("scan-sexps", Fscan_sexps
, Sscan_sexps
, 2, 2, 0,
1299 "Scan from character number FROM by COUNT balanced expressions.\n\
1300 If COUNT is negative, scan backwards.\n\
1301 Returns the character number of the position thus found.\n\
1303 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.\n\
1305 If the beginning or end of (the accessible part of) the buffer is reached\n\
1306 in the middle of a parenthetical grouping, an error is signaled.\n\
1307 If the beginning or end is reached between groupings\n\
1308 but before count is used up, nil is returned.")
1310 Lisp_Object from
, count
;
1312 CHECK_NUMBER (from
, 0);
1313 CHECK_NUMBER (count
, 1);
1315 return scan_lists (XINT (from
), XINT (count
), 0, 1);
1318 DEFUN ("backward-prefix-chars", Fbackward_prefix_chars
, Sbackward_prefix_chars
,
1320 "Move point backward over any number of chars with prefix syntax.\n\
1321 This includes chars with \"quote\" or \"prefix\" syntax (' or p).")
1327 while (pos
> beg
&& !char_quoted (pos
- 1)
1328 && (SYNTAX (FETCH_CHAR (pos
- 1)) == Squote
1329 || SYNTAX_PREFIX (FETCH_CHAR (pos
- 1))))
1337 /* Parse forward from FROM to END,
1338 assuming that FROM has state OLDSTATE (nil means FROM is start of function),
1339 and return a description of the state of the parse at END.
1340 If STOPBEFORE is nonzero, stop at the start of an atom.
1341 If COMMENTSTOP is nonzero, stop at the start of a comment. */
1344 scan_sexps_forward (stateptr
, from
, end
, targetdepth
,
1345 stopbefore
, oldstate
, commentstop
)
1346 struct lisp_parse_state
*stateptr
;
1348 int end
, targetdepth
, stopbefore
;
1349 Lisp_Object oldstate
;
1352 struct lisp_parse_state state
;
1354 register enum syntaxcode code
;
1355 struct level
{ int last
, prev
; };
1356 struct level levelstart
[100];
1357 register struct level
*curlevel
= levelstart
;
1358 struct level
*endlevel
= levelstart
+ 100;
1360 register int depth
; /* Paren depth of current scanning location.
1361 level - levelstart equals this except
1362 when the depth becomes negative. */
1363 int mindepth
; /* Lowest DEPTH value seen. */
1364 int start_quoted
= 0; /* Nonzero means starting after a char quote */
1370 if (NILP (oldstate
))
1373 state
.instring
= -1;
1374 state
.incomment
= 0;
1375 state
.comstyle
= 0; /* comment style a by default */
1379 tem
= Fcar (oldstate
);
1385 oldstate
= Fcdr (oldstate
);
1386 oldstate
= Fcdr (oldstate
);
1387 oldstate
= Fcdr (oldstate
);
1388 tem
= Fcar (oldstate
);
1389 state
.instring
= !NILP (tem
) ? XINT (tem
) : -1;
1391 oldstate
= Fcdr (oldstate
);
1392 tem
= Fcar (oldstate
);
1393 state
.incomment
= !NILP (tem
);
1395 oldstate
= Fcdr (oldstate
);
1396 tem
= Fcar (oldstate
);
1397 start_quoted
= !NILP (tem
);
1399 /* if the eight element of the list is nil, we are in comment
1400 style a. if it is non-nil, we are in comment style b */
1401 oldstate
= Fcdr (oldstate
);
1402 oldstate
= Fcdr (oldstate
);
1403 tem
= Fcar (oldstate
);
1404 state
.comstyle
= !NILP (tem
);
1409 curlevel
->prev
= -1;
1410 curlevel
->last
= -1;
1412 /* Enter the loop at a place appropriate for initial state. */
1414 if (state
.incomment
) goto startincomment
;
1415 if (state
.instring
>= 0)
1417 if (start_quoted
) goto startquotedinstring
;
1420 if (start_quoted
) goto startquoted
;
1424 code
= SYNTAX (FETCH_CHAR (from
));
1426 if (code
== Scomment
)
1427 state
.comstart
= from
-1;
1429 else if (from
< end
&& SYNTAX_COMSTART_FIRST (FETCH_CHAR (from
- 1))
1430 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from
)))
1432 /* Record the comment style we have entered so that only
1433 the comment-end sequence of the same style actually
1434 terminates the comment section. */
1436 state
.comstyle
= SYNTAX_COMMENT_STYLE (FETCH_CHAR (from
));
1437 state
.comstart
= from
-1;
1441 if (SYNTAX_PREFIX (FETCH_CHAR (from
- 1)))
1443 #ifdef SWITCH_ENUM_BUG
1451 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
1452 curlevel
->last
= from
- 1;
1454 if (from
== end
) goto endquoted
;
1457 /* treat following character as a word constituent */
1460 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
1461 curlevel
->last
= from
- 1;
1465 #ifdef SWITCH_ENUM_BUG
1466 switch ((int) SYNTAX (FETCH_CHAR (from
)))
1468 switch (SYNTAX (FETCH_CHAR (from
)))
1474 if (from
== end
) goto endquoted
;
1486 curlevel
->prev
= curlevel
->last
;
1490 state
.incomment
= 1;
1496 if (from
== end
) goto done
;
1497 prev
= FETCH_CHAR (from
);
1498 if (SYNTAX (prev
) == Sendcomment
1499 && SYNTAX_COMMENT_STYLE (prev
) == state
.comstyle
)
1500 /* Only terminate the comment section if the endcomment
1501 of the same style as the start sequence has been
1505 if (from
< end
&& SYNTAX_COMEND_FIRST (prev
)
1506 && SYNTAX_COMEND_SECOND (FETCH_CHAR (from
))
1507 && SYNTAX_COMMENT_STYLE (prev
) == state
.comstyle
)
1508 /* Only terminate the comment section if the end-comment
1509 sequence of the same style as the start sequence has
1510 been encountered. */
1513 state
.incomment
= 0;
1514 state
.comstyle
= 0; /* reset the comment style */
1518 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
1520 /* curlevel++->last ran into compiler bug on Apollo */
1521 curlevel
->last
= from
- 1;
1522 if (++curlevel
== endlevel
)
1523 error ("Nesting too deep for parser");
1524 curlevel
->prev
= -1;
1525 curlevel
->last
= -1;
1526 if (!--targetdepth
) goto done
;
1531 if (depth
< mindepth
)
1533 if (curlevel
!= levelstart
)
1535 curlevel
->prev
= curlevel
->last
;
1536 if (!++targetdepth
) goto done
;
1540 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
1541 curlevel
->last
= from
- 1;
1542 state
.instring
= FETCH_CHAR (from
- 1);
1546 if (from
>= end
) goto done
;
1547 if (FETCH_CHAR (from
) == state
.instring
) break;
1548 #ifdef SWITCH_ENUM_BUG
1549 switch ((int) SYNTAX (FETCH_CHAR (from
)))
1551 switch (SYNTAX (FETCH_CHAR (from
)))
1557 startquotedinstring
:
1558 if (from
>= end
) goto endquoted
;
1562 state
.instring
= -1;
1563 curlevel
->prev
= curlevel
->last
;
1573 stop
: /* Here if stopping before start of sexp. */
1574 from
--; /* We have just fetched the char that starts it; */
1575 goto done
; /* but return the position before it. */
1580 state
.depth
= depth
;
1581 state
.mindepth
= mindepth
;
1582 state
.thislevelstart
= curlevel
->prev
;
1583 state
.prevlevelstart
1584 = (curlevel
== levelstart
) ? -1 : (curlevel
- 1)->last
;
1585 state
.location
= from
;
1591 /* This comment supplies the doc string for parse-partial-sexp,
1592 for make-docfile to see. We cannot put this in the real DEFUN
1593 due to limits in the Unix cpp.
1595 DEFUN ("parse-partial-sexp", Ffoo, Sfoo, 2, 6, 0,
1596 "Parse Lisp syntax starting at FROM until TO; return status of parse at TO.\n\
1597 Parsing stops at TO or when certain criteria are met;\n\
1598 point is set to where parsing stops.\n\
1599 If fifth arg STATE is omitted or nil,\n\
1600 parsing assumes that FROM is the beginning of a function.\n\
1601 Value is a list of eight elements describing final state of parsing:\n\
1602 0. depth in parens.\n\
1603 1. character address of start of innermost containing list; nil if none.\n\
1604 2. character address of start of last complete sexp terminated.\n\
1605 3. non-nil if inside a string.\n\
1606 (it is the character that will terminate the string.)\n\
1607 4. t if inside a comment.\n\
1608 5. t if following a quote character.\n\
1609 6. the minimum paren-depth encountered during this scan.\n\
1610 7. t if in a comment of style `b'.\n\
1611 If third arg TARGETDEPTH is non-nil, parsing stops if the depth\n\
1612 in parentheses becomes equal to TARGETDEPTH.\n\
1613 Fourth arg STOPBEFORE non-nil means stop when come to\n\
1614 any character that starts a sexp.\n\
1615 Fifth arg STATE is an eight-list like what this function returns.\n\
1616 It is used to initialize the state of the parse. Its second and third
1617 elements are ignored.
1618 Sixth args COMMENTSTOP non-nil means stop at the start of a comment.")
1619 (from, to, targetdepth, stopbefore, state, commentstop)
1622 DEFUN ("parse-partial-sexp", Fparse_partial_sexp
, Sparse_partial_sexp
, 2, 6, 0,
1623 0 /* See immediately above */)
1624 (from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
)
1625 Lisp_Object from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
;
1627 struct lisp_parse_state state
;
1630 if (!NILP (targetdepth
))
1632 CHECK_NUMBER (targetdepth
, 3);
1633 target
= XINT (targetdepth
);
1636 target
= -100000; /* We won't reach this depth */
1638 validate_region (&from
, &to
);
1639 scan_sexps_forward (&state
, XINT (from
), XINT (to
),
1640 target
, !NILP (stopbefore
), oldstate
,
1641 !NILP (commentstop
));
1643 SET_PT (state
.location
);
1645 return Fcons (make_number (state
.depth
),
1646 Fcons (state
.prevlevelstart
< 0 ? Qnil
: make_number (state
.prevlevelstart
),
1647 Fcons (state
.thislevelstart
< 0 ? Qnil
: make_number (state
.thislevelstart
),
1648 Fcons (state
.instring
>= 0 ? make_number (state
.instring
) : Qnil
,
1649 Fcons (state
.incomment
? Qt
: Qnil
,
1650 Fcons (state
.quoted
? Qt
: Qnil
,
1651 Fcons (make_number (state
.mindepth
),
1652 Fcons (state
.comstyle
? Qt
: Qnil
,
1659 register struct Lisp_Vector
*v
;
1661 /* Set this now, so first buffer creation can refer to it. */
1662 /* Make it nil before calling copy-syntax-table
1663 so that copy-syntax-table will know not to try to copy from garbage */
1664 Vstandard_syntax_table
= Qnil
;
1665 Vstandard_syntax_table
= Fcopy_syntax_table (Qnil
);
1667 v
= XVECTOR (Vstandard_syntax_table
);
1669 for (i
= 'a'; i
<= 'z'; i
++)
1670 XFASTINT (v
->contents
[i
]) = (int) Sword
;
1671 for (i
= 'A'; i
<= 'Z'; i
++)
1672 XFASTINT (v
->contents
[i
]) = (int) Sword
;
1673 for (i
= '0'; i
<= '9'; i
++)
1674 XFASTINT (v
->contents
[i
]) = (int) Sword
;
1675 XFASTINT (v
->contents
['$']) = (int) Sword
;
1676 XFASTINT (v
->contents
['%']) = (int) Sword
;
1678 XFASTINT (v
->contents
['(']) = (int) Sopen
+ (')' << 8);
1679 XFASTINT (v
->contents
[')']) = (int) Sclose
+ ('(' << 8);
1680 XFASTINT (v
->contents
['[']) = (int) Sopen
+ (']' << 8);
1681 XFASTINT (v
->contents
[']']) = (int) Sclose
+ ('[' << 8);
1682 XFASTINT (v
->contents
['{']) = (int) Sopen
+ ('}' << 8);
1683 XFASTINT (v
->contents
['}']) = (int) Sclose
+ ('{' << 8);
1684 XFASTINT (v
->contents
['"']) = (int) Sstring
;
1685 XFASTINT (v
->contents
['\\']) = (int) Sescape
;
1687 for (i
= 0; i
< 10; i
++)
1688 XFASTINT (v
->contents
["_-+*/&|<>="[i
]]) = (int) Ssymbol
;
1690 for (i
= 0; i
< 12; i
++)
1691 XFASTINT (v
->contents
[".,;:?!#@~^'`"[i
]]) = (int) Spunct
;
1696 Qsyntax_table_p
= intern ("syntax-table-p");
1697 staticpro (&Qsyntax_table_p
);
1699 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments
,
1700 "Non-nil means `forward-sexp', etc., should treat comments as whitespace.");
1702 words_include_escapes
= 0;
1703 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes
,
1704 "Non-nil means `forward-word', etc., should treat escape chars part of words.");
1706 defsubr (&Ssyntax_table_p
);
1707 defsubr (&Ssyntax_table
);
1708 defsubr (&Sstandard_syntax_table
);
1709 defsubr (&Scopy_syntax_table
);
1710 defsubr (&Sset_syntax_table
);
1711 defsubr (&Schar_syntax
);
1712 defsubr (&Smodify_syntax_entry
);
1713 defsubr (&Sdescribe_syntax
);
1715 defsubr (&Sforward_word
);
1717 defsubr (&Sforward_comment
);
1718 defsubr (&Sscan_lists
);
1719 defsubr (&Sscan_sexps
);
1720 defsubr (&Sbackward_prefix_chars
);
1721 defsubr (&Sparse_partial_sexp
);