1 /* XEmacs routines to deal with syntax tables; also word and list parsing.
2 Copyright (C) 1985-1994 Free Software Foundation, Inc.
3 Copyright (C) 1995 Sun Microsystems, Inc.
5 This file is part of XEmacs.
7 XEmacs is free software; you can redistribute it and/or modify it
8 under the terms of the GNU General Public License as published by the
9 Free Software Foundation; either version 2, or (at your option) any
12 XEmacs is distributed in the hope that it will be useful, but WITHOUT
13 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
14 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
17 You should have received a copy of the GNU General Public License
18 along with XEmacs; see the file COPYING. If not, write to
19 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
20 Boston, MA 02111-1307, USA. */
22 /* Synched up with: FSF 19.28. */
24 /* This file has been Mule-ized. */
33 /* Here is a comment from Ken'ichi HANDA <handa@etl.go.jp>
34 explaining the purpose of the Sextword syntax category:
36 Japanese words are not separated by spaces, which makes finding word
37 boundaries very difficult. Theoretically it's impossible without
38 using natural language processing techniques. But, by defining
39 pseudo-words as below (much simplified for letting you understand it
40 easily) for Japanese, we can have a convenient forward-word function
43 A Japanese word is a sequence of characters that consists of
44 zero or more Kanji characters followed by zero or more
47 Then, the problem is that now we can't say that a sequence of
48 word-constituents makes up a WORD. For instance, both Hiragana "A"
49 and Kanji "KAN" are word-constituents but the sequence of these two
50 letters can't be a single word.
52 So, we introduced Sextword for Japanese letters. A character of
53 Sextword is a word-constituent but a word boundary may exist between
54 two such characters. */
56 /* Mule 2.4 doesn't seem to have Sextword - I'm removing it -- mrb */
57 /* Recovered by tomo */
59 #define ST_COMMENT_STYLE 0x101
60 #define ST_STRING_STYLE 0x102
62 Lisp_Object Qsyntax_table;
63 int lookup_syntax_properties;
65 Lisp_Object Qsyntax_table_p;
67 int words_include_escapes;
69 int parse_sexp_ignore_comments;
71 /* The following two variables are provided to tell additional information
72 to the regex routines. We do it this way rather than change the
73 arguments to re_search_2() in an attempt to maintain some call
74 compatibility with other versions of the regex code. */
76 /* Tell the regex routines not to QUIT. Normally there is a QUIT
77 each iteration in re_search_2(). */
78 int no_quit_in_re_search;
80 /* Tell the regex routines which buffer to access for SYNTAX() lookups
82 struct buffer *regex_emacs_buffer;
84 /* In Emacs, this is the string or buffer in which we
85 are matching. It is used for looking up syntax properties. */
86 Lisp_Object regex_match_object;
88 Lisp_Object Vstandard_syntax_table;
90 Lisp_Object Vsyntax_designator_chars_string;
92 /* This is the internal form of the parse state used in parse-partial-sexp. */
94 struct lisp_parse_state
96 int depth; /* Depth at end of parsing */
97 Emchar instring; /* -1 if not within string, else desired terminator */
98 int incomment; /* Nonzero if within a comment at end of parsing */
99 int comstyle; /* comment style a=0, or b=1, or ST_COMMENT_STYLE */
100 int quoted; /* Nonzero if just after an escape char at end of
102 Bufpos thislevelstart;/* Char number of most recent start-of-expression
104 Bufpos prevlevelstart;/* Char number of start of containing expression */
105 Bufpos location; /* Char number at which parsing stopped */
106 int mindepth; /* Minimum depth seen while scanning */
107 Bufpos comstr_start; /* Position just after last comment/string starter */
108 Lisp_Object levelstarts; /* Char numbers of starts-of-expression
109 of levels (starting from outermost). */
112 /* These variables are a cache for finding the start of a defun.
113 find_start_pos is the place for which the defun start was found.
114 find_start_value is the defun start position found for it.
115 find_start_buffer is the buffer it was found in.
116 find_start_begv is the BEGV value when it was found.
117 find_start_modiff is the value of MODIFF when it was found. */
119 static Bufpos find_start_pos;
120 static Bufpos find_start_value;
121 static struct buffer *find_start_buffer;
122 static Bufpos find_start_begv;
123 static int find_start_modiff;
125 /* Find a defun-start that is the last one before POS (or nearly the last).
126 We record what we find, so that another call in the same area
127 can return the same value right away. */
130 find_defun_start (struct buffer *buf, Bufpos pos)
134 /* Use previous finding, if it's valid and applies to this inquiry. */
135 if (buf == find_start_buffer
136 /* Reuse the defun-start even if POS is a little farther on.
137 POS might be in the next defun, but that's ok.
138 Our value may not be the best possible, but will still be usable. */
139 && pos <= find_start_pos + 1000
140 && pos >= find_start_value
141 && BUF_BEGV (buf) == find_start_begv
142 && BUF_MODIFF (buf) == find_start_modiff)
143 return find_start_value;
145 /* Back up to start of line. */
146 tem = find_next_newline (buf, pos, -1);
148 SETUP_SYNTAX_CACHE (tem, 1);
149 while (tem > BUF_BEGV (buf))
151 UPDATE_SYNTAX_CACHE_BACKWARD(tem);
153 /* Open-paren at start of line means we found our defun-start. */
154 if (SYNTAX_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, tem)) == Sopen)
156 /* Move to beg of previous line. */
157 tem = find_next_newline (buf, tem, -2);
160 /* Record what we found, for the next try. */
161 find_start_value = tem;
162 find_start_buffer = buf;
163 find_start_modiff = BUF_MODIFF (buf);
164 find_start_begv = BUF_BEGV (buf);
165 find_start_pos = pos;
167 return find_start_value;
170 DEFUN ("syntax-table-p", Fsyntax_table_p, 1, 1, 0, /*
171 Return t if OBJECT is a syntax table.
172 Any vector of 256 elements will do.
176 return (CHAR_TABLEP (object)
177 && XCHAR_TABLE_TYPE (object) == CHAR_TABLE_TYPE_SYNTAX)
182 check_syntax_table (Lisp_Object obj, Lisp_Object default_)
186 while (NILP (Fsyntax_table_p (obj)))
187 obj = wrong_type_argument (Qsyntax_table_p, obj);
191 DEFUN ("syntax-table", Fsyntax_table, 0, 1, 0, /*
192 Return the current syntax table.
193 This is the one specified by the current buffer, or by BUFFER if it
198 return decode_buffer (buffer, 0)->syntax_table;
201 DEFUN ("standard-syntax-table", Fstandard_syntax_table, 0, 0, 0, /*
202 Return the standard syntax table.
203 This is the one used for new buffers.
207 return Vstandard_syntax_table;
210 DEFUN ("copy-syntax-table", Fcopy_syntax_table, 0, 1, 0, /*
211 Return a new syntax table which is a copy of SYNTAX-TABLE.
212 SYNTAX-TABLE defaults to the standard syntax table.
216 if (NILP (Vstandard_syntax_table))
217 return Fmake_char_table (Qsyntax);
219 syntax_table = check_syntax_table (syntax_table, Vstandard_syntax_table);
220 return Fcopy_char_table (syntax_table);
223 DEFUN ("set-syntax-table", Fset_syntax_table, 1, 2, 0, /*
224 Select SYNTAX-TABLE as the new syntax table for BUFFER.
225 BUFFER defaults to the current buffer if omitted.
227 (syntax_table, buffer))
229 struct buffer *buf = decode_buffer (buffer, 0);
230 syntax_table = check_syntax_table (syntax_table, Qnil);
231 buf->syntax_table = syntax_table;
232 buf->mirror_syntax_table = XCHAR_TABLE (syntax_table)->mirror_table;
233 /* Indicate that this buffer now has a specified syntax table. */
234 buf->local_var_flags |= XINT (buffer_local_flags.syntax_table);
238 /* The current syntax state */
239 struct syntax_cache syntax_cache;
243 Update syntax_cache to an appropriate setting for position POS
245 The sign of COUNT gives the relative position of POS wrt the
246 previously valid interval. (not currently used)
248 `syntax_cache.*_change' are the next and previous positions at
249 which syntax_code and c_s_t will need to be recalculated.
251 #### Currently this code uses 'get-char-property', which will
252 return the "last smallest" extent at a given position. In cases
253 where overlapping extents are defined, this code will simply use
254 whatever is returned by get-char-property.
256 It might be worth it at some point to merge provided syntax tables
257 outward to the current buffer. */
260 update_syntax_cache (int pos, int count, int init)
262 Lisp_Object tmp_table;
266 syntax_cache.prev_change = -1;
267 syntax_cache.next_change = -1;
270 if (pos > syntax_cache.prev_change &&
271 pos < syntax_cache.next_change)
277 if (NILP (syntax_cache.object) || EQ (syntax_cache.object, Qt))
279 int get_change_before = pos + 1;
281 tmp_table = Fget_char_property (make_int(pos), Qsyntax_table,
282 make_buffer (syntax_cache.buffer), Qnil);
283 syntax_cache.next_change =
284 XINT (Fnext_extent_change (make_int (pos > 0 ? pos : 1),
285 make_buffer (syntax_cache.buffer)));
287 if (get_change_before < 1)
288 get_change_before = 1;
289 else if (get_change_before > BUF_ZV (syntax_cache.buffer))
290 get_change_before = BUF_ZV (syntax_cache.buffer);
292 syntax_cache.prev_change =
293 XINT (Fprevious_extent_change (make_int (get_change_before),
294 make_buffer (syntax_cache.buffer)));
298 int get_change_before = pos + 1;
300 tmp_table = Fget_char_property (make_int(pos), Qsyntax_table,
301 syntax_cache.object, Qnil);
302 syntax_cache.next_change =
303 XINT (Fnext_extent_change (make_int (pos >= 0 ? pos : 0),
304 syntax_cache.object));
306 if (get_change_before < 0)
307 get_change_before = 0;
308 else if (get_change_before > XSTRING_LENGTH(syntax_cache.object))
309 get_change_before = XSTRING_LENGTH(syntax_cache.object);
311 syntax_cache.prev_change =
312 XINT (Fprevious_extent_change (make_int (pos >= 0 ? pos : 0),
313 syntax_cache.object));
316 if (EQ (Fsyntax_table_p (tmp_table), Qt))
318 syntax_cache.use_code = 0;
319 syntax_cache.current_syntax_table =
320 XCHAR_TABLE (tmp_table)->mirror_table;
322 else if (CONSP (tmp_table) && INTP (XCAR (tmp_table)))
324 syntax_cache.use_code = 1;
325 syntax_cache.syntax_code = XINT (XCAR(tmp_table));
329 syntax_cache.use_code = 0;
330 syntax_cache.current_syntax_table =
331 syntax_cache.buffer->mirror_syntax_table;
336 /* Convert a letter which signifies a syntax code
337 into the code it signifies.
338 This is used by modify-syntax-entry, and other things. */
340 const unsigned char syntax_spec_code[0400] =
341 { 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
342 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
343 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
344 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
345 (char) Swhitespace, 0377, (char) Sstring, 0377,
346 (char) Smath, 0377, 0377, (char) Squote,
347 (char) Sopen, (char) Sclose, 0377, 0377,
348 0377, (char) Swhitespace, (char) Spunct, (char) Scharquote,
349 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
350 0377, 0377, 0377, 0377,
351 (char) Scomment, 0377, (char) Sendcomment, 0377,
352 (char) Sinherit, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */
353 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
354 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
355 0377, 0377, 0377, 0377, (char) Sescape, 0377, 0377, (char) Ssymbol,
356 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* `, a, ... */
357 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
358 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
359 0377, 0377, 0377, 0377, (char) Sstring_fence, 0377, 0377, 0377
362 const unsigned char syntax_code_spec[] = " .w_()'\"$\\/<>@!|";
364 DEFUN ("syntax-designator-chars", Fsyntax_designator_chars, 0, 0, 0, /*
365 Return a string of the recognized syntax designator chars.
366 The chars are ordered by their internal syntax codes, which are
367 numbered starting at 0.
371 return Vsyntax_designator_chars_string;
374 DEFUN ("char-syntax", Fchar_syntax, 1, 2, 0, /*
375 Return the syntax code of CHARACTER, described by a character.
376 For example, if CHARACTER is a word constituent,
377 the character `?w' is returned.
378 The characters that correspond to various syntax codes
379 are listed in the documentation of `modify-syntax-entry'.
380 Optional second argument SYNTAX-TABLE defaults to the current buffer's
383 (character, syntax_table))
385 Lisp_Char_Table *mirrortab;
387 if (NILP (character))
389 character = make_char ('\000');
391 CHECK_CHAR_COERCE_INT (character);
392 syntax_table = check_syntax_table (syntax_table, current_buffer->syntax_table);
393 mirrortab = XCHAR_TABLE (XCHAR_TABLE (syntax_table)->mirror_table);
394 return make_char (syntax_code_spec[(int) SYNTAX (mirrortab, XCHAR (character))]);
400 charset_syntax (struct buffer *buf, Lisp_Object charset, int *multi_p_out)
403 /* #### get this right */
410 syntax_match (Lisp_Object syntax_table, Emchar ch)
412 Lisp_Object code = XCHAR_TABLE_VALUE_UNSAFE (syntax_table, ch);
413 Lisp_Object code2 = code;
417 if (SYNTAX_FROM_CODE (XINT (code2)) == Sinherit)
418 code = XCHAR_TABLE_VALUE_UNSAFE (Vstandard_syntax_table, ch);
420 return CONSP (code) ? XCDR (code) : Qnil;
423 DEFUN ("matching-paren", Fmatching_paren, 1, 2, 0, /*
424 Return the matching parenthesis of CHARACTER, or nil if none.
425 Optional second argument SYNTAX-TABLE defaults to the current buffer's
428 (character, syntax_table))
430 Lisp_Char_Table *mirrortab;
433 CHECK_CHAR_COERCE_INT (character);
434 syntax_table = check_syntax_table (syntax_table, current_buffer->syntax_table);
435 mirrortab = XCHAR_TABLE (XCHAR_TABLE (syntax_table)->mirror_table);
436 code = SYNTAX (mirrortab, XCHAR (character));
437 if (code == Sopen || code == Sclose || code == Sstring)
438 return syntax_match (syntax_table, XCHAR (character));
445 /* Return 1 if there is a word boundary between two word-constituent
446 characters C1 and C2 if they appear in this order, else return 0.
447 There is no word boundary between two word-constituent ASCII
449 #define WORD_BOUNDARY_P(c1, c2) \
450 (!(CHAR_ASCII_P (c1) && CHAR_ASCII_P (c2)) \
451 && word_boundary_p (c1, c2))
453 extern int word_boundary_p (Emchar c1, Emchar c2);
456 /* Return the position across COUNT words from FROM.
457 If that many words cannot be found before the end of the buffer, return 0.
458 COUNT negative means scan backward and stop at word beginning. */
461 scan_words (struct buffer *buf, Bufpos from, int count)
463 Bufpos limit = count > 0 ? BUF_ZV (buf) : BUF_BEGV (buf);
465 enum syntaxcode code;
467 SETUP_SYNTAX_CACHE_FOR_BUFFER (buf, from, count);
469 /* #### is it really worth it to hand expand both cases? JV */
479 UPDATE_SYNTAX_CACHE_FORWARD (from);
480 ch0 = BUF_FETCH_CHAR (buf, from);
481 code = SYNTAX_FROM_CACHE (mirrortab, ch0);
484 if (words_include_escapes
485 && (code == Sescape || code == Scharquote))
493 while (from != limit)
495 UPDATE_SYNTAX_CACHE_FORWARD (from);
496 ch1 = BUF_FETCH_CHAR (buf, from);
497 code = SYNTAX_FROM_CACHE (mirrortab, ch1);
498 if (!(words_include_escapes
499 && (code == Sescape || code == Scharquote)))
502 || WORD_BOUNDARY_P (ch0, ch1)
523 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
524 ch1 = BUF_FETCH_CHAR (buf, from - 1);
525 code = SYNTAX_FROM_CACHE (mirrortab, ch1);
528 if (words_include_escapes
529 && (code == Sescape || code == Scharquote))
537 while (from != limit)
539 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
540 ch0 = BUF_FETCH_CHAR (buf, from - 1);
541 code = SYNTAX_FROM_CACHE (mirrortab, ch0);
543 if (!(words_include_escapes
544 && (code == Sescape || code == Scharquote)))
547 || WORD_BOUNDARY_P (ch0, ch1)
562 DEFUN ("forward-word", Fforward_word, 0, 2, "_p", /*
563 Move point forward COUNT words (backward if COUNT is negative).
564 Normally t is returned, but if an edge of the buffer is reached,
565 point is left there and nil is returned.
567 COUNT defaults to 1, and BUFFER defaults to the current buffer.
572 struct buffer *buf = decode_buffer (buffer, 0);
583 val = scan_words (buf, BUF_PT (buf), n);
586 BUF_SET_PT (buf, val);
591 BUF_SET_PT (buf, n > 0 ? BUF_ZV (buf) : BUF_BEGV (buf));
596 static void scan_sexps_forward (struct buffer *buf,
597 struct lisp_parse_state *,
598 Bufpos from, Bufpos end,
599 int targetdepth, int stopbefore,
600 Lisp_Object oldstate,
604 find_start_of_comment (struct buffer *buf, Bufpos from, Bufpos stop,
608 enum syntaxcode code;
610 /* Look back, counting the parity of string-quotes,
611 and recording the comment-starters seen.
612 When we reach a safe place, assume that's not in a string;
613 then step the main scan to the earliest comment-starter seen
614 an even number of string quotes away from the safe place.
616 OFROM[I] is position of the earliest comment-starter seen
617 which is I+2X quotes from the comment-end.
618 PARITY is current parity of quotes from the comment end. */
620 Emchar my_stringend = 0;
621 int string_lossage = 0;
622 Bufpos comment_end = from;
623 Bufpos comstart_pos = 0;
624 int comstart_parity = 0;
625 int styles_match_p = 0;
626 /* mask to match comment styles against; for ST_COMMENT_STYLE, this
627 will get set to SYNTAX_COMMENT_STYLE_B, but never get checked */
628 int mask = comstyle ? SYNTAX_COMMENT_STYLE_B : SYNTAX_COMMENT_STYLE_A;
630 /* At beginning of range to scan, we're outside of strings;
631 that determines quote parity to the comment-end. */
636 /* Move back and examine a character. */
638 UPDATE_SYNTAX_CACHE_BACKWARD (from);
640 c = BUF_FETCH_CHAR (buf, from);
641 code = SYNTAX_FROM_CACHE (mirrortab, c);
642 syncode = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
644 /* is this a 1-char comment end sequence? if so, try
645 to see if style matches previously extracted mask */
646 if (code == Sendcomment)
649 SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode) & mask;
652 /* or are we looking at a 1-char comment start sequence
653 of the style matching mask? */
654 else if (code == Scomment)
657 SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode) & mask;
660 /* otherwise, is this a 2-char comment end or start sequence? */
661 else if (from > stop)
664 /* 2-char comment end sequence? */
665 if (SYNTAX_CODE_END_SECOND_P (syncode))
668 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
670 SYNTAX_CODE_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, from - 1));
672 if (SYNTAX_CODES_END_P (prev_syncode, syncode))
676 SYNTAX_CODES_COMMENT_MASK_END (prev_syncode, syncode);
678 UPDATE_SYNTAX_CACHE_BACKWARD (from);
679 c = BUF_FETCH_CHAR (buf, from);
681 /* Found a comment-end sequence, so skip past the
682 check for a comment-start */
687 /* 2-char comment start sequence? */
688 if (SYNTAX_CODE_START_SECOND_P (syncode))
691 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
693 SYNTAX_CODE_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, from - 1));
695 if (SYNTAX_CODES_START_P (prev_syncode, syncode))
699 SYNTAX_CODES_COMMENT_MASK_START (prev_syncode, syncode);
701 UPDATE_SYNTAX_CACHE_BACKWARD (from);
702 c = BUF_FETCH_CHAR (buf, from);
707 /* Ignore escaped characters. */
708 if (char_quoted (buf, from))
711 /* Track parity of quotes. */
715 if (my_stringend == 0)
717 /* If we have two kinds of string delimiters.
718 There's no way to grok this scanning backwards. */
719 else if (my_stringend != c)
723 if (code == Sstring_fence || code == Scomment_fence)
726 if (my_stringend == 0)
728 code == Sstring_fence ? ST_STRING_STYLE : ST_COMMENT_STYLE;
729 /* If we have two kinds of string delimiters.
730 There's no way to grok this scanning backwards. */
731 else if (my_stringend != (code == Sstring_fence
732 ? ST_STRING_STYLE : ST_COMMENT_STYLE))
736 /* Record comment-starters according to that
737 quote-parity to the comment-end. */
738 if (code == Scomment && styles_match_p)
740 comstart_parity = parity;
744 /* If we find another earlier comment-ender,
745 any comment-starts earlier than that don't count
746 (because they go with the earlier comment-ender). */
747 if (code == Sendcomment && styles_match_p)
750 /* Assume a defun-start point is outside of strings. */
752 && (from == stop || BUF_FETCH_CHAR (buf, from - 1) == '\n'))
756 if (comstart_pos == 0)
758 /* If the earliest comment starter
759 is followed by uniform paired string quotes or none,
760 we know it can't be inside a string
761 since if it were then the comment ender would be inside one.
762 So it does start a comment. Skip back to it. */
763 else if (comstart_parity == 0 && !string_lossage)
767 /* We had two kinds of string delimiters mixed up
768 together. Decode this going forwards.
769 Scan fwd from the previous comment ender
770 to the one in question; this records where we
771 last passed a comment starter. */
773 struct lisp_parse_state state;
774 scan_sexps_forward (buf, &state, find_defun_start (buf, comment_end),
775 comment_end - 1, -10000, 0, Qnil, 0);
777 from = state.comstr_start;
779 /* We can't grok this as a comment; scan it normally. */
781 UPDATE_SYNTAX_CACHE_FORWARD (from - 1);
787 find_end_of_comment (struct buffer *buf, Bufpos from, Bufpos stop, int comstyle)
791 /* mask to match comment styles against; for ST_COMMENT_STYLE, this
792 will get set to SYNTAX_COMMENT_STYLE_B, but never get checked */
793 int mask = comstyle ? SYNTAX_COMMENT_STYLE_B : SYNTAX_COMMENT_STYLE_A;
795 /* This is only called by functions which have already set up the
796 syntax_cache and are keeping it up-to-date */
804 UPDATE_SYNTAX_CACHE_FORWARD (from);
805 c = BUF_FETCH_CHAR (buf, from);
807 /* Test for generic comments */
808 if (comstyle == ST_COMMENT_STYLE)
810 if (SYNTAX_FROM_CACHE (mirrortab, c) == Scomment_fence)
813 UPDATE_SYNTAX_CACHE_FORWARD (from);
817 continue; /* No need to test other comment styles in a
822 if (SYNTAX_FROM_CACHE (mirrortab, c) == Sendcomment
823 && SYNTAX_CODE_MATCHES_1CHAR_P
824 (SYNTAX_CODE_FROM_CACHE (mirrortab, c), mask))
825 /* we have encountered a comment end of the same style
826 as the comment sequence which began this comment
830 UPDATE_SYNTAX_CACHE_FORWARD (from);
834 prev_code = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
836 UPDATE_SYNTAX_CACHE_FORWARD (from);
838 && SYNTAX_CODES_MATCH_END_P
840 SYNTAX_CODE_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, from)),
844 /* we have encountered a comment end of the same style
845 as the comment sequence which began this comment
849 UPDATE_SYNTAX_CACHE_FORWARD (from);
857 /* #### between FSF 19.23 and 19.28 there are some changes to the logic
858 in this function (and minor changes to find_start_of_comment(),
859 above, which is part of Fforward_comment() in FSF). Attempts to port
860 that logic made this function break, so I'm leaving it out. If anyone
861 ever complains about this function not working properly, take a look
862 at those changes. --ben */
864 DEFUN ("forward-comment", Fforward_comment, 0, 2, 0, /*
865 Move forward across up to COUNT comments, or backwards if COUNT is negative.
866 Stop scanning if we find something other than a comment or whitespace.
867 Set point to where scanning stops.
868 If COUNT comments are found as expected, with nothing except whitespace
869 between them, return t; otherwise return nil.
870 Point is set in either case.
871 COUNT defaults to 1, and BUFFER defaults to the current buffer.
878 enum syntaxcode code;
881 struct buffer *buf = decode_buffer (buffer, 0);
893 SETUP_SYNTAX_CACHE (from, n);
901 int comstyle = 0; /* mask for finding matching comment style */
903 if (char_quoted (buf, from))
909 UPDATE_SYNTAX_CACHE_FORWARD (from);
910 c = BUF_FETCH_CHAR (buf, from);
911 code = SYNTAX_FROM_CACHE (mirrortab, c);
912 syncode = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
914 if (code == Scomment)
916 /* we have encountered a single character comment start
917 sequence, and we are ignoring all text inside comments.
918 we must record the comment style this character begins
919 so that later, only a comment end of the same style actually
920 ends the comment section */
921 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
922 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
925 else if (code == Scomment_fence)
929 comstyle = ST_COMMENT_STYLE;
933 && SYNTAX_CODE_START_FIRST_P (syncode))
936 UPDATE_SYNTAX_CACHE_FORWARD (from + 1);
938 SYNTAX_CODE_FROM_CACHE (mirrortab,
939 BUF_FETCH_CHAR (buf, from + 1));
941 if (SYNTAX_CODES_START_P (syncode, next_syncode))
943 /* we have encountered a 2char comment start sequence and we
944 are ignoring all text inside comments. we must record
945 the comment style this sequence begins so that later,
946 only a comment end of the same style actually ends
947 the comment section */
950 SYNTAX_CODES_COMMENT_MASK_START (syncode, next_syncode)
951 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
956 if (code == Scomment)
958 Bufpos newfrom = find_end_of_comment (buf, from, stop, comstyle);
961 /* we stopped because from==stop */
962 BUF_SET_PT (buf, stop);
967 /* We have skipped one comment. */
970 else if (code != Swhitespace
971 && code != Sendcomment
972 && code != Scomment )
974 BUF_SET_PT (buf, from);
980 /* End of comment reached */
988 stop = BUF_BEGV (buf);
991 int comstyle = 0; /* mask for finding matching comment style */
994 if (char_quoted (buf, from))
1000 c = BUF_FETCH_CHAR (buf, from);
1001 code = SYNTAX_FROM_CACHE (mirrortab, c);
1002 syncode = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
1004 if (code == Sendcomment)
1006 /* we have found a single char end comment. we must record
1007 the comment style encountered so that later, we can match
1008 only the proper comment begin sequence of the same style */
1009 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
1010 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1013 else if (code == Scomment_fence)
1016 comstyle = ST_COMMENT_STYLE;
1019 else if (from > stop
1020 && SYNTAX_CODE_END_SECOND_P (syncode))
1023 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
1025 SYNTAX_CODE_FROM_CACHE (mirrortab,
1026 BUF_FETCH_CHAR (buf, from - 1));
1027 if (SYNTAX_CODES_END_P (prev_syncode, syncode))
1029 /* We must record the comment style encountered so that
1030 later, we can match only the proper comment begin
1031 sequence of the same style. */
1033 comstyle = SYNTAX_CODES_COMMENT_MASK_END
1034 (prev_syncode, syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1039 if (code == Sendcomment)
1041 from = find_start_of_comment (buf, from, stop, comstyle);
1045 else if (code != Swhitespace
1047 && code != Sendcomment)
1049 BUF_SET_PT (buf, from + 1);
1057 BUF_SET_PT (buf, from);
1063 scan_lists (struct buffer *buf, Bufpos from, int count, int depth,
1064 int sexpflag, int noerror)
1070 enum syntaxcode code;
1072 int min_depth = depth; /* Err out if depth gets less than this. */
1074 if (depth > 0) min_depth = 0;
1076 SETUP_SYNTAX_CACHE_FOR_BUFFER (buf, from, count);
1081 stop = BUF_ZV (buf);
1084 int comstyle = 0; /* mask for finding matching comment style */
1086 UPDATE_SYNTAX_CACHE_FORWARD (from);
1087 c = BUF_FETCH_CHAR (buf, from);
1088 code = SYNTAX_FROM_CACHE (mirrortab, c);
1089 syncode = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
1092 /* a 1-char comment start sequence */
1093 if (code == Scomment && parse_sexp_ignore_comments)
1095 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode) ==
1096 SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1099 /* else, a 2-char comment start sequence? */
1100 else if (from < stop
1101 && SYNTAX_CODE_START_FIRST_P (syncode)
1102 && parse_sexp_ignore_comments)
1105 UPDATE_SYNTAX_CACHE_FORWARD (from);
1107 SYNTAX_CODE_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, from));
1109 if (SYNTAX_CODES_START_P (syncode, next_syncode))
1111 /* we have encountered a comment start sequence and we
1112 are ignoring all text inside comments. we must record
1113 the comment style this sequence begins so that later,
1114 only a comment end of the same style actually ends
1115 the comment section */
1117 comstyle = SYNTAX_CODES_COMMENT_MASK_START
1118 (syncode, next_syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1122 UPDATE_SYNTAX_CACHE_FORWARD (from);
1124 if (SYNTAX_CODE_PREFIX (syncode))
1131 if (from == stop) goto lose;
1133 /* treat following character as a word constituent */
1136 if (depth || !sexpflag) break;
1137 /* This word counts as a sexp; return at end of it. */
1140 UPDATE_SYNTAX_CACHE_FORWARD (from);
1141 switch (SYNTAX_FROM_CACHE (mirrortab,
1142 BUF_FETCH_CHAR (buf, from)))
1147 if (from == stop) goto lose;
1160 case Scomment_fence:
1161 comstyle = ST_COMMENT_STYLE;
1163 if (!parse_sexp_ignore_comments)
1165 UPDATE_SYNTAX_CACHE_FORWARD (from);
1168 find_end_of_comment (buf, from, stop, comstyle);
1171 /* we stopped because from == stop in search forward */
1184 if (from != stop && c == BUF_FETCH_CHAR (buf, from))
1194 if (!++depth) goto done;
1199 if (!--depth) goto done;
1200 if (depth < min_depth)
1204 error ("Containing expression ends prematurely");
1213 if (code != Sstring_fence)
1215 /* XEmacs change: call syntax_match on character */
1216 Emchar ch = BUF_FETCH_CHAR (buf, from - 1);
1217 Lisp_Object stermobj =
1218 syntax_match (syntax_cache.current_syntax_table, ch);
1220 if (CHARP (stermobj))
1221 stringterm = XCHAR (stermobj);
1226 stringterm = '\0'; /* avoid compiler warnings */
1232 UPDATE_SYNTAX_CACHE_FORWARD (from);
1233 c = BUF_FETCH_CHAR (buf, from);
1236 : SYNTAX_FROM_CACHE (mirrortab, c) == Sstring_fence)
1239 switch (SYNTAX_FROM_CACHE (mirrortab, c))
1251 if (!depth && sexpflag) goto done;
1260 /* Reached end of buffer. Error if within object,
1261 return nil if between */
1262 if (depth) goto lose;
1266 /* End of object reached */
1275 stop = BUF_BEGV (buf);
1278 int comstyle = 0; /* mask for finding matching comment style */
1281 UPDATE_SYNTAX_CACHE_BACKWARD (from);
1282 quoted = char_quoted (buf, from);
1286 UPDATE_SYNTAX_CACHE_BACKWARD (from);
1289 c = BUF_FETCH_CHAR (buf, from);
1290 code = SYNTAX_FROM_CACHE (mirrortab, c);
1291 syncode = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
1293 if (code == Sendcomment && parse_sexp_ignore_comments)
1295 /* we have found a single char end comment. we must record
1296 the comment style encountered so that later, we can match
1297 only the proper comment begin sequence of the same style */
1298 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
1299 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1302 else if (from > stop
1303 && SYNTAX_CODE_END_SECOND_P (syncode)
1304 && !char_quoted (buf, from - 1)
1305 && parse_sexp_ignore_comments)
1308 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
1309 prev_syncode = SYNTAX_CODE_FROM_CACHE
1310 (mirrortab, BUF_FETCH_CHAR (buf, from - 1));
1312 if (SYNTAX_CODES_END_P (prev_syncode, syncode))
1314 /* we must record the comment style encountered so that
1315 later, we can match only the proper comment begin
1316 sequence of the same style */
1318 comstyle = SYNTAX_CODES_COMMENT_MASK_END
1319 (prev_syncode, syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1324 if (SYNTAX_CODE_PREFIX (syncode))
1327 switch (quoted ? Sword : code)
1331 if (depth || !sexpflag) break;
1332 /* This word counts as a sexp; count object finished after
1336 UPDATE_SYNTAX_CACHE_BACKWARD (from);
1337 quoted = char_quoted (buf, from - 1);
1343 SYNTAX_FROM_CACHE (mirrortab,
1344 BUF_FETCH_CHAR (buf, from - 1)))
1346 || syncode == Ssymbol
1347 || syncode == Squote))
1356 if (from != stop && c == BUF_FETCH_CHAR (buf, from - 1))
1366 if (!++depth) goto done2;
1371 if (!--depth) goto done2;
1372 if (depth < min_depth)
1376 error ("Containing expression ends prematurely");
1380 case Scomment_fence:
1381 comstyle = ST_COMMENT_STYLE;
1383 if (parse_sexp_ignore_comments)
1384 from = find_start_of_comment (buf, from, stop, comstyle);
1392 if (code != Sstring_fence)
1394 /* XEmacs change: call syntax_match() on character */
1395 Emchar ch = BUF_FETCH_CHAR (buf, from);
1396 Lisp_Object stermobj =
1397 syntax_match (syntax_cache.current_syntax_table, ch);
1399 if (CHARP (stermobj))
1400 stringterm = XCHAR (stermobj);
1405 stringterm = '\0'; /* avoid compiler warnings */
1409 if (from == stop) goto lose;
1411 UPDATE_SYNTAX_CACHE_BACKWARD (from - 1);
1412 c = BUF_FETCH_CHAR (buf, from - 1);
1414 if ((code == Sstring
1416 : SYNTAX_FROM_CACHE (mirrortab, c) == Sstring_fence)
1417 && !char_quoted (buf, from - 1))
1425 if (!depth && sexpflag) goto done2;
1431 /* Reached start of buffer. Error if within object,
1432 return nil if between */
1433 if (depth) goto lose;
1442 return (make_int (from));
1446 error ("Unbalanced parentheses");
1451 char_quoted (struct buffer *buf, Bufpos pos)
1453 enum syntaxcode code;
1454 Bufpos beg = BUF_BEGV (buf);
1456 Bufpos startpos = pos;
1460 UPDATE_SYNTAX_CACHE_BACKWARD (pos - 1);
1461 code = SYNTAX_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, pos - 1));
1463 if (code != Scharquote && code != Sescape)
1469 UPDATE_SYNTAX_CACHE (startpos);
1473 DEFUN ("scan-lists", Fscan_lists, 3, 5, 0, /*
1474 Scan from character number FROM by COUNT lists.
1475 Returns the character number of the position thus found.
1477 If DEPTH is nonzero, paren depth begins counting from that value,
1478 only places where the depth in parentheses becomes zero
1479 are candidates for stopping; COUNT such places are counted.
1480 Thus, a positive value for DEPTH means go out levels.
1482 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
1484 If the beginning or end of (the accessible part of) the buffer is reached
1485 and the depth is wrong, an error is signaled.
1486 If the depth is right but the count is not used up, nil is returned.
1488 If optional arg BUFFER is non-nil, scanning occurs in that buffer instead
1489 of in the current buffer.
1491 If optional arg NOERROR is non-nil, scan-lists will return nil instead of
1492 signalling an error.
1494 (from, count, depth, buffer, noerror))
1501 buf = decode_buffer (buffer, 0);
1503 return scan_lists (buf, XINT (from), XINT (count), XINT (depth), 0,
1507 DEFUN ("scan-sexps", Fscan_sexps, 2, 4, 0, /*
1508 Scan from character number FROM by COUNT balanced expressions.
1509 If COUNT is negative, scan backwards.
1510 Returns the character number of the position thus found.
1512 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
1514 If the beginning or end of (the accessible part of) the buffer is reached
1515 in the middle of a parenthetical grouping, an error is signaled.
1516 If the beginning or end is reached between groupings
1517 but before count is used up, nil is returned.
1519 If optional arg BUFFER is non-nil, scanning occurs in that buffer instead
1520 of in the current buffer.
1522 If optional arg NOERROR is non-nil, scan-sexps will return nil instead of
1523 signalling an error.
1525 (from, count, buffer, noerror))
1527 struct buffer *buf = decode_buffer (buffer, 0);
1531 return scan_lists (buf, XINT (from), XINT (count), 0, 1, !NILP (noerror));
1534 DEFUN ("backward-prefix-chars", Fbackward_prefix_chars, 0, 1, 0, /*
1535 Move point backward over any number of chars with prefix syntax.
1536 This includes chars with "quote" or "prefix" syntax (' or p).
1538 Optional arg BUFFER defaults to the current buffer.
1542 struct buffer *buf = decode_buffer (buffer, 0);
1543 Bufpos beg = BUF_BEGV (buf);
1544 Bufpos pos = BUF_PT (buf);
1546 Lisp_Char_Table *mirrortab = XCHAR_TABLE (buf->mirror_syntax_table);
1548 Emchar c = '\0'; /* initialize to avoid compiler warnings */
1551 SETUP_SYNTAX_CACHE_FOR_BUFFER (buf, pos, -1);
1553 while (pos > beg && !char_quoted (buf, pos - 1)
1554 /* Previous statement updates syntax table. */
1555 && (SYNTAX_FROM_CACHE (mirrortab, c = BUF_FETCH_CHAR (buf, pos - 1)) == Squote
1556 || SYNTAX_CODE_PREFIX (SYNTAX_CODE_FROM_CACHE (mirrortab, c))))
1559 BUF_SET_PT (buf, pos);
1564 /* Parse forward from FROM to END,
1565 assuming that FROM has state OLDSTATE (nil means FROM is start of function),
1566 and return a description of the state of the parse at END.
1567 If STOPBEFORE is nonzero, stop at the start of an atom.
1568 If COMMENTSTOP is nonzero, stop at the start of a comment. */
1571 scan_sexps_forward (struct buffer *buf, struct lisp_parse_state *stateptr,
1572 Bufpos from, Bufpos end,
1573 int targetdepth, int stopbefore,
1574 Lisp_Object oldstate,
1577 struct lisp_parse_state state;
1579 enum syntaxcode code;
1580 struct level { int last, prev; };
1581 struct level levelstart[100];
1582 struct level *curlevel = levelstart;
1583 struct level *endlevel = levelstart + 100;
1584 int depth; /* Paren depth of current scanning location.
1585 level - levelstart equals this except
1586 when the depth becomes negative. */
1587 int mindepth; /* Lowest DEPTH value seen. */
1588 int start_quoted = 0; /* Nonzero means starting after a char quote */
1589 int boundary_stop = commentstop == -1;
1592 SETUP_SYNTAX_CACHE (from, 1);
1593 if (NILP (oldstate))
1596 state.instring = -1;
1597 state.incomment = 0;
1598 state.comstyle = 0; /* comment style a by default */
1599 state.comstr_start = -1; /* no comment/string seen. */
1603 tem = Fcar (oldstate); /* elt 0, depth */
1609 oldstate = Fcdr (oldstate);
1610 oldstate = Fcdr (oldstate);
1611 oldstate = Fcdr (oldstate);
1612 tem = Fcar (oldstate); /* elt 3, instring */
1613 state.instring = ( !NILP (tem)
1614 ? ( INTP (tem) ? XINT (tem) : ST_STRING_STYLE)
1617 oldstate = Fcdr (oldstate);
1618 tem = Fcar (oldstate); /* elt 4, incomment */
1619 state.incomment = !NILP (tem);
1621 oldstate = Fcdr (oldstate);
1622 tem = Fcar (oldstate); /* elt 5, follows-quote */
1623 start_quoted = !NILP (tem);
1625 /* if the eighth element of the list is nil, we are in comment style
1626 a; if it is t, we are in comment style b; if it is 'syntax-table,
1627 we are in a generic comment */
1628 oldstate = Fcdr (oldstate);
1629 oldstate = Fcdr (oldstate);
1630 tem = Fcar (oldstate); /* elt 7, comment style a/b/fence */
1631 state.comstyle = NILP (tem) ? 0 : ( EQ (tem, Qsyntax_table)
1632 ? ST_COMMENT_STYLE : 1 );
1634 oldstate = Fcdr (oldstate); /* elt 8, start of last comment/string */
1635 tem = Fcar (oldstate);
1636 state.comstr_start = NILP (tem) ? -1 : XINT (tem);
1638 /* elt 9, char numbers of starts-of-expression of levels
1639 (starting from outermost). */
1640 oldstate = Fcdr (oldstate);
1641 tem = Fcar (oldstate); /* elt 9, intermediate data for
1642 continuation of parsing (subject
1644 while (!NILP (tem)) /* >= second enclosing sexps. */
1646 curlevel->last = XINT (Fcar (tem));
1647 if (++curlevel == endlevel)
1648 error ("Nesting too deep for parser");
1649 curlevel->prev = -1;
1650 curlevel->last = -1;
1657 curlevel->prev = -1;
1658 curlevel->last = -1;
1660 /* Enter the loop at a place appropriate for initial state. */
1662 if (state.incomment) goto startincomment;
1663 if (state.instring >= 0)
1665 if (start_quoted) goto startquotedinstring;
1668 if (start_quoted) goto startquoted;
1677 UPDATE_SYNTAX_CACHE_FORWARD (from);
1678 c = BUF_FETCH_CHAR (buf, from);
1679 code = SYNTAX_FROM_CACHE (mirrortab, c);
1680 syncode = SYNTAX_CODE_FROM_CACHE (mirrortab, c);
1683 /* record the comment style we have entered so that only the
1684 comment-ender sequence (or single char) of the same style
1685 actually terminates the comment section. */
1686 if (code == Scomment)
1689 SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
1690 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1691 state.comstr_start = from - 1;
1694 /* a generic comment delimiter? */
1695 else if (code == Scomment_fence)
1697 state.comstyle = ST_COMMENT_STYLE;
1698 state.comstr_start = from - 1;
1702 else if (from < end &&
1703 SYNTAX_CODE_START_FIRST_P (syncode))
1706 UPDATE_SYNTAX_CACHE_FORWARD (from);
1708 SYNTAX_CODE_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, from));
1710 if (SYNTAX_CODES_START_P (syncode, next_syncode))
1713 state.comstyle = SYNTAX_CODES_COMMENT_MASK_START
1714 (syncode, next_syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
1715 state.comstr_start = from - 1;
1717 UPDATE_SYNTAX_CACHE_FORWARD (from);
1721 if (SYNTAX_CODE_PREFIX (syncode))
1727 if (stopbefore) goto stop; /* this arg means stop at sexp start */
1728 curlevel->last = from - 1;
1730 if (from == end) goto endquoted;
1733 /* treat following character as a word constituent */
1736 if (stopbefore) goto stop; /* this arg means stop at sexp start */
1737 curlevel->last = from - 1;
1741 UPDATE_SYNTAX_CACHE_FORWARD (from);
1742 switch (SYNTAX_FROM_CACHE (mirrortab, BUF_FETCH_CHAR (buf, from)))
1747 if (from == end) goto endquoted;
1759 curlevel->prev = curlevel->last;
1763 state.incomment = 1;
1764 if (commentstop || boundary_stop) goto done;
1766 if (commentstop == 1)
1768 UPDATE_SYNTAX_CACHE_FORWARD (from);
1770 Bufpos newfrom = find_end_of_comment (buf, from, end, state.comstyle);
1773 /* we terminated search because from == end */
1779 state.incomment = 0;
1780 state.comstyle = 0; /* reset the comment style */
1781 if (boundary_stop) goto done;
1785 if (stopbefore) goto stop; /* this arg means stop at sexp start */
1787 /* curlevel++->last ran into compiler bug on Apollo */
1788 curlevel->last = from - 1;
1789 if (++curlevel == endlevel)
1790 error ("Nesting too deep for parser");
1791 curlevel->prev = -1;
1792 curlevel->last = -1;
1793 if (targetdepth == depth) goto done;
1798 if (depth < mindepth)
1800 if (curlevel != levelstart)
1802 curlevel->prev = curlevel->last;
1803 if (targetdepth == depth) goto done;
1808 state.comstr_start = from - 1;
1809 if (stopbefore) goto stop; /* this arg means stop at sexp start */
1810 curlevel->last = from - 1;
1811 if (code == Sstring_fence)
1813 state.instring = ST_STRING_STYLE;
1817 /* XEmacs change: call syntax_match() on character */
1818 Emchar ch = BUF_FETCH_CHAR (buf, from - 1);
1819 Lisp_Object stermobj =
1820 syntax_match (syntax_cache.current_syntax_table, ch);
1822 if (CHARP (stermobj))
1823 state.instring = XCHAR (stermobj);
1825 state.instring = ch;
1827 if (boundary_stop) goto done;
1831 enum syntaxcode temp_code;
1833 if (from >= end) goto done;
1835 UPDATE_SYNTAX_CACHE_FORWARD (from);
1836 c = BUF_FETCH_CHAR (buf, from);
1837 temp_code = SYNTAX_FROM_CACHE (mirrortab, c);
1840 state.instring != ST_STRING_STYLE &&
1841 temp_code == Sstring &&
1842 c == state.instring) break;
1847 if (state.instring == ST_STRING_STYLE)
1854 startquotedinstring:
1855 if (from >= end) goto endquoted;
1864 state.instring = -1;
1865 curlevel->prev = curlevel->last;
1867 if (boundary_stop) goto done;
1877 case Scomment_fence:
1885 stop: /* Here if stopping before start of sexp. */
1886 from--; /* We have just fetched the char that starts it; */
1887 goto done; /* but return the position before it. */
1892 state.depth = depth;
1893 state.mindepth = mindepth;
1894 state.thislevelstart = curlevel->prev;
1895 state.prevlevelstart
1896 = (curlevel == levelstart) ? -1 : (curlevel - 1)->last;
1897 state.location = from;
1898 state.levelstarts = Qnil;
1899 while (--curlevel >= levelstart)
1900 state.levelstarts = Fcons (make_int (curlevel->last),
1906 DEFUN ("parse-partial-sexp", Fparse_partial_sexp, 2, 7, 0, /*
1907 Parse Lisp syntax starting at FROM until TO; return status of parse at TO.
1908 Parsing stops at TO or when certain criteria are met;
1909 point is set to where parsing stops.
1910 If fifth arg OLDSTATE is omitted or nil,
1911 parsing assumes that FROM is the beginning of a function.
1912 Value is a list of nine elements describing final state of parsing:
1914 1. character address of start of innermost containing list; nil if none.
1915 2. character address of start of last complete sexp terminated.
1916 3. non-nil if inside a string.
1917 (It is the character that will terminate the string,
1918 or t if the string should be terminated by an explicit
1919 `syntax-table' property.)
1920 4. t if inside a comment.
1921 5. t if following a quote character.
1922 6. the minimum paren-depth encountered during this scan.
1923 7. nil if in comment style a, or not in a comment; t if in comment style b;
1924 `syntax-table' if given by an explicit `syntax-table' property.
1925 8. character address of start of last comment or string; nil if none.
1926 9. Intermediate data for continuation of parsing (subject to change).
1927 If third arg TARGETDEPTH is non-nil, parsing stops if the depth
1928 in parentheses becomes equal to TARGETDEPTH.
1929 Fourth arg STOPBEFORE non-nil means stop when come to
1930 any character that starts a sexp.
1931 Fifth arg OLDSTATE is a nine-element list like what this function returns.
1932 It is used to initialize the state of the parse. Its second and third
1933 elements are ignored.
1934 Sixth arg COMMENTSTOP non-nil means stop at the start of a comment. If it
1935 is `syntax-table', stop after the start of a comment or a string, or after
1936 the end of a comment or string.
1938 (from, to, targetdepth, stopbefore, oldstate, commentstop, buffer))
1940 struct lisp_parse_state state;
1943 struct buffer *buf = decode_buffer (buffer, 0);
1946 if (!NILP (targetdepth))
1948 CHECK_INT (targetdepth);
1949 target = XINT (targetdepth);
1952 target = -100000; /* We won't reach this depth */
1954 get_buffer_range_char (buf, from, to, &start, &end, 0);
1955 scan_sexps_forward (buf, &state, start, end,
1956 target, !NILP (stopbefore), oldstate,
1958 ? 0 : (EQ (commentstop, Qsyntax_table) ? -1 : 1)));
1959 BUF_SET_PT (buf, state.location);
1963 val = Fcons (state.levelstarts, val);
1964 val = Fcons ((state.incomment || (state.instring >= 0))
1965 ? make_int (state.comstr_start) : Qnil, val);
1966 val = Fcons (state.comstyle ? (state.comstyle == ST_COMMENT_STYLE
1967 ? Qsyntax_table : Qt) : Qnil, val);
1968 val = Fcons (make_int (state.mindepth), val);
1969 val = Fcons (state.quoted ? Qt : Qnil, val);
1970 val = Fcons (state.incomment ? Qt : Qnil, val);
1971 val = Fcons (state.instring < 0
1973 : (state.instring == ST_STRING_STYLE
1974 ? Qt : make_int (state.instring)), val);
1975 val = Fcons (state.thislevelstart < 0 ? Qnil : make_int (state.thislevelstart), val);
1976 val = Fcons (state.prevlevelstart < 0 ? Qnil : make_int (state.prevlevelstart), val);
1977 val = Fcons (make_int (state.depth), val);
1983 /* Updating of the mirror syntax table.
1985 Each syntax table has a corresponding mirror table in it.
1986 Whenever we make a change to a syntax table, we call
1987 update_syntax_table() on it.
1989 #### We really only need to map over the changed range.
1991 If we change the standard syntax table, we need to map over
1992 all tables because any of them could be inheriting from the
1993 standard syntax table.
1995 When `set-syntax-table' is called, we set the buffer's mirror
1996 syntax table as well.
2001 Lisp_Object mirrortab;
2006 cmst_mapfun (struct chartab_range *range, Lisp_Object val, void *arg)
2008 struct cmst_arg *closure = (struct cmst_arg *) arg;
2012 if (SYNTAX_FROM_CODE (XINT (val)) == Sinherit
2013 && closure->check_inherit)
2015 struct cmst_arg recursive;
2017 recursive.mirrortab = closure->mirrortab;
2018 recursive.check_inherit = 0;
2019 map_char_table (XCHAR_TABLE (Vstandard_syntax_table), range,
2020 cmst_mapfun, &recursive);
2023 put_char_table (XCHAR_TABLE (closure->mirrortab), range, val);
2028 update_just_this_syntax_table (Lisp_Char_Table *ct)
2030 struct chartab_range range;
2031 struct cmst_arg arg;
2033 arg.mirrortab = ct->mirror_table;
2034 arg.check_inherit = (CHAR_TABLEP (Vstandard_syntax_table)
2035 && ct != XCHAR_TABLE (Vstandard_syntax_table));
2036 range.type = CHARTAB_RANGE_ALL;
2037 map_char_table (ct, &range, cmst_mapfun, &arg);
2040 /* Called from chartab.c when a change is made to a syntax table.
2041 If this is the standard syntax table, we need to recompute
2042 *all* syntax tables (yuck). Otherwise we just recompute this
2046 update_syntax_table (Lisp_Char_Table *ct)
2048 /* Don't be stymied at startup. */
2049 if (CHAR_TABLEP (Vstandard_syntax_table)
2050 && ct == XCHAR_TABLE (Vstandard_syntax_table))
2054 for (syntab = Vall_syntax_tables; !NILP (syntab);
2055 syntab = XCHAR_TABLE (syntab)->next_table)
2056 update_just_this_syntax_table (XCHAR_TABLE (syntab));
2059 update_just_this_syntax_table (ct);
2063 /************************************************************************/
2064 /* initialization */
2065 /************************************************************************/
2068 syms_of_syntax (void)
2070 defsymbol (&Qsyntax_table_p, "syntax-table-p");
2071 defsymbol (&Qsyntax_table, "syntax-table");
2073 DEFSUBR (Fsyntax_table_p);
2074 DEFSUBR (Fsyntax_table);
2075 DEFSUBR (Fstandard_syntax_table);
2076 DEFSUBR (Fcopy_syntax_table);
2077 DEFSUBR (Fset_syntax_table);
2078 DEFSUBR (Fsyntax_designator_chars);
2079 DEFSUBR (Fchar_syntax);
2080 DEFSUBR (Fmatching_paren);
2081 /* DEFSUBR (Fmodify_syntax_entry); now in Lisp. */
2082 /* DEFSUBR (Fdescribe_syntax); now in Lisp. */
2084 DEFSUBR (Fforward_word);
2086 DEFSUBR (Fforward_comment);
2087 DEFSUBR (Fscan_lists);
2088 DEFSUBR (Fscan_sexps);
2089 DEFSUBR (Fbackward_prefix_chars);
2090 DEFSUBR (Fparse_partial_sexp);
2094 vars_of_syntax (void)
2096 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments /*
2097 Non-nil means `forward-sexp', etc., should treat comments as whitespace.
2099 parse_sexp_ignore_comments = 0;
2101 DEFVAR_BOOL ("lookup-syntax-properties", &lookup_syntax_properties /*
2102 Non-nil means `forward-sexp', etc., grant `syntax-table' property.
2103 The value of this property should be either a syntax table, or a cons
2104 of the form (SYNTAXCODE . MATCHCHAR), SYNTAXCODE being the numeric
2105 syntax code, MATCHCHAR being nil or the character to match (which is
2106 relevant only for open/close type.
2108 lookup_syntax_properties = 1;
2110 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes /*
2111 Non-nil means `forward-word', etc., should treat escape chars part of words.
2113 words_include_escapes = 0;
2115 no_quit_in_re_search = 0;
2119 define_standard_syntax (const char *p, enum syntaxcode syn)
2122 Fput_char_table (make_char (*p), make_int (syn), Vstandard_syntax_table);
2126 complex_vars_of_syntax (void)
2130 /* Set this now, so first buffer creation can refer to it. */
2131 /* Make it nil before calling copy-syntax-table
2132 so that copy-syntax-table will know not to try to copy from garbage */
2133 Vstandard_syntax_table = Qnil;
2134 Vstandard_syntax_table = Fcopy_syntax_table (Qnil);
2135 staticpro (&Vstandard_syntax_table);
2137 Vsyntax_designator_chars_string = make_string_nocopy (syntax_code_spec,
2139 staticpro (&Vsyntax_designator_chars_string);
2141 fill_char_table (XCHAR_TABLE (Vstandard_syntax_table), make_int (Spunct));
2143 for (i = 0; i <= 32; i++) /* Control 0 plus SPACE */
2144 Fput_char_table (make_char (i), make_int (Swhitespace),
2145 Vstandard_syntax_table);
2146 for (i = 127; i <= 159; i++) /* DEL plus Control 1 */
2147 Fput_char_table (make_char (i), make_int (Swhitespace),
2148 Vstandard_syntax_table);
2150 define_standard_syntax ("abcdefghijklmnopqrstuvwxyz"
2151 "ABCDEFGHIJKLMNOPQRSTUVWXYZ"
2154 define_standard_syntax ("\"", Sstring);
2155 define_standard_syntax ("\\", Sescape);
2156 define_standard_syntax ("_-+*/&|<>=", Ssymbol);
2157 define_standard_syntax (".,;:?!#@~^'`", Spunct);
2159 for (p = "()[]{}"; *p; p+=2)
2161 Fput_char_table (make_char (p[0]),
2162 Fcons (make_int (Sopen), make_char (p[1])),
2163 Vstandard_syntax_table);
2164 Fput_char_table (make_char (p[1]),
2165 Fcons (make_int (Sclose), make_char (p[0])),
2166 Vstandard_syntax_table);