Added command remapping.
[bpt/emacs.git] / src / syntax.c
CommitLineData
8489eb67 1/* GNU Emacs routines to deal with syntax tables; also word and list parsing.
b8855607 2 Copyright (C) 1985, 87, 93, 94, 95, 97, 1998, 1999 Free Software Foundation, Inc.
8489eb67
RS
3
4This file is part of GNU Emacs.
5
6GNU Emacs is free software; you can redistribute it and/or modify
7it under the terms of the GNU General Public License as published by
a4275ad1 8the Free Software Foundation; either version 2, or (at your option)
8489eb67
RS
9any later version.
10
11GNU Emacs is distributed in the hope that it will be useful,
12but WITHOUT ANY WARRANTY; without even the implied warranty of
13MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14GNU General Public License for more details.
15
16You should have received a copy of the GNU General Public License
17along with GNU Emacs; see the file COPYING. If not, write to
3b7ad313
EN
18the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19Boston, MA 02111-1307, USA. */
8489eb67
RS
20
21
18160b98 22#include <config.h>
8489eb67
RS
23#include <ctype.h>
24#include "lisp.h"
25#include "commands.h"
26#include "buffer.h"
93da5fff 27#include "charset.h"
e35f6ff7 28#include "keymap.h"
195d1361
RS
29
30/* Make syntax table lookup grant data in gl_state. */
31#define SYNTAX_ENTRY_VIA_PROPERTY
32
8489eb67 33#include "syntax.h"
195d1361
RS
34#include "intervals.h"
35
36/* We use these constants in place for comment-style and
37 string-ender-char to distinguish comments/strings started by
38 comment_fence and string_fence codes. */
39
40#define ST_COMMENT_STYLE (256 + 1)
41#define ST_STRING_STYLE (256 + 2)
93da5fff 42#include "category.h"
8489eb67 43
7bf5e9e4 44Lisp_Object Qsyntax_table_p, Qsyntax_table, Qscan_error;
8489eb67
RS
45
46int words_include_escapes;
195d1361 47int parse_sexp_lookup_properties;
8489eb67 48
bd25db08
KH
49/* Nonzero means `scan-sexps' treat all multibyte characters as symbol. */
50int multibyte_syntax_as_symbol;
51
8ea151b2
RS
52/* Used as a temporary in SYNTAX_ENTRY and other macros in syntax.h,
53 if not compiled with GCC. No need to mark it, since it is used
54 only very temporarily. */
55Lisp_Object syntax_temp;
56
f4ed767f
GM
57/* Non-zero means an open parenthesis in column 0 is always considered
58 to be the start of a defun. Zero means an open parenthesis in
59 column 0 has no special meaning. */
60
61int open_paren_in_column_0_is_defun_start;
62
e5d4f4dc
RS
63/* This is the internal form of the parse state used in parse-partial-sexp. */
64
65struct lisp_parse_state
66 {
195d1361
RS
67 int depth; /* Depth at end of parsing. */
68 int instring; /* -1 if not within string, else desired terminator. */
95ff8dfc 69 int incomment; /* -1 if in unnestable comment else comment nesting */
195d1361 70 int comstyle; /* comment style a=0, or b=1, or ST_COMMENT_STYLE. */
e5d4f4dc
RS
71 int quoted; /* Nonzero if just after an escape char at end of parsing */
72 int thislevelstart; /* Char number of most recent start-of-expression at current level */
73 int prevlevelstart; /* Char number of start of containing expression */
195d1361 74 int location; /* Char number at which parsing stopped. */
e5d4f4dc 75 int mindepth; /* Minimum depth seen while scanning. */
195d1361 76 int comstr_start; /* Position just after last comment/string starter. */
1a102a58
RS
77 Lisp_Object levelstarts; /* Char numbers of starts-of-expression
78 of levels (starting from outermost). */
e5d4f4dc
RS
79 };
80\f
37bef230
RS
81/* These variables are a cache for finding the start of a defun.
82 find_start_pos is the place for which the defun start was found.
83 find_start_value is the defun start position found for it.
6a140a74 84 find_start_value_byte is the corresponding byte position.
37bef230
RS
85 find_start_buffer is the buffer it was found in.
86 find_start_begv is the BEGV value when it was found.
87 find_start_modiff is the value of MODIFF when it was found. */
88
89static int find_start_pos;
90static int find_start_value;
6a140a74 91static int find_start_value_byte;
37bef230
RS
92static struct buffer *find_start_buffer;
93static int find_start_begv;
94static int find_start_modiff;
6a140a74
RS
95
96
97static int find_defun_start P_ ((int, int));
95ff8dfc 98static int back_comment P_ ((int, int, int, int, int, int *, int *));
6a140a74
RS
99static int char_quoted P_ ((int, int));
100static Lisp_Object skip_chars P_ ((int, int, Lisp_Object, Lisp_Object));
101static Lisp_Object scan_lists P_ ((int, int, int, int));
102static void scan_sexps_forward P_ ((struct lisp_parse_state *,
103 int, int, int, int,
104 int, Lisp_Object, int));
195d1361
RS
105\f
106
107struct gl_state_s gl_state; /* Global state of syntax parser. */
108
109INTERVAL interval_of ();
110#define INTERVALS_AT_ONCE 10 /* 1 + max-number of intervals
111 to scan to property-change. */
112
6a140a74
RS
113/* Update gl_state to an appropriate interval which contains CHARPOS. The
114 sign of COUNT give the relative position of CHARPOS wrt the previously
195d1361 115 valid interval. If INIT, only [be]_property fields of gl_state are
6a140a74 116 valid at start, the rest is filled basing on OBJECT.
195d1361 117
6a140a74 118 `gl_state.*_i' are the intervals, and CHARPOS is further in the search
195d1361
RS
119 direction than the intervals - or in an interval. We update the
120 current syntax-table basing on the property of this interval, and
6a140a74 121 update the interval to start further than CHARPOS - or be
195d1361 122 NULL_INTERVAL. We also update lim_property to be the next value of
6a140a74 123 charpos to call this subroutine again - or be before/after the
195d1361
RS
124 start/end of OBJECT. */
125
126void
6a140a74
RS
127update_syntax_table (charpos, count, init, object)
128 int charpos, count, init;
195d1361
RS
129 Lisp_Object object;
130{
131 Lisp_Object tmp_table;
4ffe723b 132 int cnt = 0, invalidate = 1;
195d1361
RS
133 INTERVAL i, oldi;
134
135 if (init)
136 {
bb0de084 137 gl_state.old_prop = Qnil;
195d1361
RS
138 gl_state.start = gl_state.b_property;
139 gl_state.stop = gl_state.e_property;
bb0de084
SM
140 i = interval_of (charpos, object);
141 gl_state.backward_i = gl_state.forward_i = i;
195d1361
RS
142 invalidate = 0;
143 if (NULL_INTERVAL_P (i))
144 return;
f902a008 145 /* interval_of updates only ->position of the return value, so
d80f4cc7 146 update the parents manually to speed up update_interval. */
bb0de084 147 while (!NULL_PARENT (i))
d80f4cc7
RS
148 {
149 if (AM_RIGHT_CHILD (i))
439d5cb4 150 INTERVAL_PARENT (i)->position = i->position
d80f4cc7 151 - LEFT_TOTAL_LENGTH (i) + TOTAL_LENGTH (i) /* right end */
439d5cb4
KR
152 - TOTAL_LENGTH (INTERVAL_PARENT (i))
153 + LEFT_TOTAL_LENGTH (INTERVAL_PARENT (i));
d80f4cc7 154 else
439d5cb4 155 INTERVAL_PARENT (i)->position = i->position - LEFT_TOTAL_LENGTH (i)
d80f4cc7 156 + TOTAL_LENGTH (i);
439d5cb4 157 i = INTERVAL_PARENT (i);
d80f4cc7
RS
158 }
159 i = gl_state.forward_i;
bb0de084 160 gl_state.b_property = i->position - gl_state.offset;
ee0cdb48 161 gl_state.e_property = INTERVAL_LAST_POS (i) - gl_state.offset;
195d1361
RS
162 goto update;
163 }
164 oldi = i = count > 0 ? gl_state.forward_i : gl_state.backward_i;
165
423e705d 166 /* We are guaranteed to be called with CHARPOS either in i,
6a140a74 167 or further off. */
195d1361
RS
168 if (NULL_INTERVAL_P (i))
169 error ("Error in syntax_table logic for to-the-end intervals");
6a140a74 170 else if (charpos < i->position) /* Move left. */
195d1361
RS
171 {
172 if (count > 0)
f902a008 173 error ("Error in syntax_table logic for intervals <-");
195d1361 174 /* Update the interval. */
6a140a74 175 i = update_interval (i, charpos);
bb0de084 176 if (INTERVAL_LAST_POS (i) != gl_state.b_property)
195d1361
RS
177 {
178 invalidate = 0;
195d1361 179 gl_state.forward_i = i;
ee0cdb48 180 gl_state.e_property = INTERVAL_LAST_POS (i) - gl_state.offset;
195d1361 181 }
423e705d 182 }
6a140a74 183 else if (charpos >= INTERVAL_LAST_POS (i)) /* Move right. */
195d1361
RS
184 {
185 if (count < 0)
f902a008 186 error ("Error in syntax_table logic for intervals ->");
195d1361 187 /* Update the interval. */
6a140a74 188 i = update_interval (i, charpos);
bb0de084 189 if (i->position != gl_state.e_property)
195d1361
RS
190 {
191 invalidate = 0;
195d1361 192 gl_state.backward_i = i;
bb0de084 193 gl_state.b_property = i->position - gl_state.offset;
195d1361
RS
194 }
195 }
195d1361
RS
196
197 update:
198 tmp_table = textget (i->plist, Qsyntax_table);
199
200 if (invalidate)
201 invalidate = !EQ (tmp_table, gl_state.old_prop); /* Need to invalidate? */
202
423e705d
SM
203 if (invalidate) /* Did not get to adjacent interval. */
204 { /* with the same table => */
205 /* invalidate the old range. */
195d1361
RS
206 if (count > 0)
207 {
208 gl_state.backward_i = i;
bb0de084
SM
209 gl_state.b_property = i->position - gl_state.offset;
210 }
211 else
195d1361 212 {
bb0de084 213 gl_state.forward_i = i;
ee0cdb48 214 gl_state.e_property = INTERVAL_LAST_POS (i) - gl_state.offset;
195d1361
RS
215 }
216 }
37bef230 217
bb0de084 218 if (!EQ (tmp_table, gl_state.old_prop))
195d1361 219 {
bb0de084
SM
220 gl_state.current_syntax_table = tmp_table;
221 gl_state.old_prop = tmp_table;
222 if (EQ (Fsyntax_table_p (tmp_table), Qt))
223 {
224 gl_state.use_global = 0;
225 }
226 else if (CONSP (tmp_table))
227 {
228 gl_state.use_global = 1;
229 gl_state.global_code = tmp_table;
230 }
231 else
232 {
233 gl_state.use_global = 0;
234 gl_state.current_syntax_table = current_buffer->syntax_table;
235 }
195d1361
RS
236 }
237
238 while (!NULL_INTERVAL_P (i))
239 {
240 if (cnt && !EQ (tmp_table, textget (i->plist, Qsyntax_table)))
241 {
242 if (count > 0)
bb0de084
SM
243 {
244 gl_state.e_property = i->position - gl_state.offset;
245 gl_state.forward_i = i;
246 }
247 else
248 {
249 gl_state.b_property = i->position + LENGTH (i) - gl_state.offset;
250 gl_state.backward_i = i;
251 }
252 return;
195d1361
RS
253 }
254 else if (cnt == INTERVALS_AT_ONCE)
255 {
256 if (count > 0)
bb0de084
SM
257 {
258 gl_state.e_property = i->position + LENGTH (i) - gl_state.offset;
259 gl_state.forward_i = i;
260 }
261 else
262 {
263 gl_state.b_property = i->position - gl_state.offset;
264 gl_state.backward_i = i;
265 }
266 return;
195d1361
RS
267 }
268 cnt++;
269 i = count > 0 ? next_interval (i) : previous_interval (i);
270 }
bb0de084
SM
271 eassert (NULL_INTERVAL_P (i)); /* This property goes to the end. */
272 if (count > 0)
273 gl_state.e_property = gl_state.stop;
274 else
275 gl_state.b_property = gl_state.start;
195d1361
RS
276}
277\f
6a140a74
RS
278/* Returns TRUE if char at CHARPOS is quoted.
279 Global syntax-table data should be set up already to be good at CHARPOS
280 or after. On return global syntax data is good for lookup at CHARPOS. */
195d1361
RS
281
282static int
6a140a74
RS
283char_quoted (charpos, bytepos)
284 register int charpos, bytepos;
195d1361
RS
285{
286 register enum syntaxcode code;
287 register int beg = BEGV;
288 register int quoted = 0;
6a140a74
RS
289 int orig = charpos;
290
291 DEC_BOTH (charpos, bytepos);
195d1361 292
6a140a74 293 while (bytepos >= beg)
195d1361 294 {
6a140a74
RS
295 UPDATE_SYNTAX_TABLE_BACKWARD (charpos);
296 code = SYNTAX (FETCH_CHAR (bytepos));
297 if (! (code == Scharquote || code == Sescape))
298 break;
299
300 DEC_BOTH (charpos, bytepos);
301 quoted = !quoted;
195d1361 302 }
6a140a74
RS
303
304 UPDATE_SYNTAX_TABLE (orig);
195d1361
RS
305 return quoted;
306}
6a140a74
RS
307
308/* Return the bytepos one character after BYTEPOS.
309 We assume that BYTEPOS is not at the end of the buffer. */
310
311INLINE int
312inc_bytepos (bytepos)
313 int bytepos;
314{
ef316cf0
RS
315 if (NILP (current_buffer->enable_multibyte_characters))
316 return bytepos + 1;
317
6a140a74
RS
318 INC_POS (bytepos);
319 return bytepos;
320}
321
322/* Return the bytepos one character before BYTEPOS.
323 We assume that BYTEPOS is not at the start of the buffer. */
324
325INLINE int
326dec_bytepos (bytepos)
327 int bytepos;
328{
ef316cf0
RS
329 if (NILP (current_buffer->enable_multibyte_characters))
330 return bytepos - 1;
331
6a140a74
RS
332 DEC_POS (bytepos);
333 return bytepos;
334}
195d1361 335\f
7cc80f0a
RS
336/* Return a defun-start position before before POS and not too far before.
337 It should be the last one before POS, or nearly the last.
338
339 When open_paren_in_column_0_is_defun_start is nonzero,
340 the beginning of every line is treated as a defun-start.
341
342 We record the information about where the scan started
343 and what its result was, so that another call in the same area
344 can return the same value very quickly.
195d1361
RS
345
346 There is no promise at which position the global syntax data is
347 valid on return from the subroutine, so the caller should explicitly
348 update the global data. */
37bef230
RS
349
350static int
6a140a74
RS
351find_defun_start (pos, pos_byte)
352 int pos, pos_byte;
37bef230 353{
6a140a74 354 int opoint = PT, opoint_byte = PT_BYTE;
37bef230
RS
355
356 /* Use previous finding, if it's valid and applies to this inquiry. */
357 if (current_buffer == find_start_buffer
358 /* Reuse the defun-start even if POS is a little farther on.
359 POS might be in the next defun, but that's ok.
360 Our value may not be the best possible, but will still be usable. */
361 && pos <= find_start_pos + 1000
362 && pos >= find_start_value
363 && BEGV == find_start_begv
364 && MODIFF == find_start_modiff)
365 return find_start_value;
366
367 /* Back up to start of line. */
6a140a74 368 scan_newline (pos, pos_byte, BEGV, BEGV_BYTE, -1, 1);
37bef230 369
195d1361
RS
370 /* We optimize syntax-table lookup for rare updates. Thus we accept
371 only those `^\s(' which are good in global _and_ text-property
372 syntax-tables. */
373 gl_state.current_syntax_table = current_buffer->syntax_table;
374 gl_state.use_global = 0;
4467eb94 375 if (open_paren_in_column_0_is_defun_start)
37bef230 376 {
4467eb94 377 while (PT > BEGV)
195d1361 378 {
4467eb94
GM
379 /* Open-paren at start of line means we may have found our
380 defun-start. */
381 if (SYNTAX (FETCH_CHAR (PT_BYTE)) == Sopen)
382 {
383 SETUP_SYNTAX_TABLE (PT + 1, -1); /* Try again... */
384 if (SYNTAX (FETCH_CHAR (PT_BYTE)) == Sopen)
385 break;
386 /* Now fallback to the default value. */
387 gl_state.current_syntax_table = current_buffer->syntax_table;
388 gl_state.use_global = 0;
389 }
390 /* Move to beg of previous line. */
391 scan_newline (PT, PT_BYTE, BEGV, BEGV_BYTE, -2, 1);
195d1361 392 }
37bef230
RS
393 }
394
395 /* Record what we found, for the next try. */
6a140a74
RS
396 find_start_value = PT;
397 find_start_value_byte = PT_BYTE;
37bef230
RS
398 find_start_buffer = current_buffer;
399 find_start_modiff = MODIFF;
400 find_start_begv = BEGV;
401 find_start_pos = pos;
402
6a140a74
RS
403 TEMP_SET_PT_BOTH (opoint, opoint_byte);
404
37bef230
RS
405 return find_start_value;
406}
407\f
f902a008
RS
408/* Return the SYNTAX_COMEND_FIRST of the character before POS, POS_BYTE. */
409
410static int
411prev_char_comend_first (pos, pos_byte)
412 int pos, pos_byte;
413{
414 int c, val;
415
416 DEC_BOTH (pos, pos_byte);
417 UPDATE_SYNTAX_TABLE_BACKWARD (pos);
418 c = FETCH_CHAR (pos_byte);
419 val = SYNTAX_COMEND_FIRST (c);
420 UPDATE_SYNTAX_TABLE_FORWARD (pos + 1);
421 return val;
422}
423
424/* Return the SYNTAX_COMSTART_FIRST of the character before POS, POS_BYTE. */
425
3f679f55
SM
426/* static int
427 * prev_char_comstart_first (pos, pos_byte)
428 * int pos, pos_byte;
429 * {
430 * int c, val;
431 *
432 * DEC_BOTH (pos, pos_byte);
433 * UPDATE_SYNTAX_TABLE_BACKWARD (pos);
434 * c = FETCH_CHAR (pos_byte);
435 * val = SYNTAX_COMSTART_FIRST (c);
436 * UPDATE_SYNTAX_TABLE_FORWARD (pos + 1);
437 * return val;
438 * } */
f902a008 439
6a140a74
RS
440/* Checks whether charpos FROM is at the end of a comment.
441 FROM_BYTE is the bytepos corresponding to FROM.
442 Do not move back before STOP.
443
444 Return a positive value if we find a comment ending at FROM/FROM_BYTE;
445 return -1 otherwise.
446
447 If successful, store the charpos of the comment's beginning
448 into *CHARPOS_PTR, and the bytepos into *BYTEPOS_PTR.
527a32d9
KH
449
450 Global syntax data remains valid for backward search starting at
451 the returned value (or at FROM, if the search was not successful). */
195d1361
RS
452
453static int
95ff8dfc 454back_comment (from, from_byte, stop, comnested, comstyle, charpos_ptr, bytepos_ptr)
6a140a74 455 int from, from_byte, stop;
95ff8dfc 456 int comnested, comstyle;
6a140a74 457 int *charpos_ptr, *bytepos_ptr;
195d1361
RS
458{
459 /* Look back, counting the parity of string-quotes,
460 and recording the comment-starters seen.
461 When we reach a safe place, assume that's not in a string;
462 then step the main scan to the earliest comment-starter seen
463 an even number of string quotes away from the safe place.
464
465 OFROM[I] is position of the earliest comment-starter seen
466 which is I+2X quotes from the comment-end.
467 PARITY is current parity of quotes from the comment end. */
3ee5041c 468 int string_style = -1; /* Presumed outside of any string. */
195d1361 469 int string_lossage = 0;
f7c436c1
SM
470 /* Not a real lossage: indicates that we have passed a matching comment
471 starter plus an non-matching comment-ender, meaning that any matching
472 comment-starter we might see later could be a false positive (hidden
473 inside another comment).
474 Test case: { a (* b } c (* d *) */
475 int comment_lossage = 0;
195d1361 476 int comment_end = from;
6a140a74 477 int comment_end_byte = from_byte;
195d1361 478 int comstart_pos = 0;
6a140a74 479 int comstart_byte;
eb35b628
RS
480 /* Place where the containing defun starts,
481 or 0 if we didn't come across it yet. */
482 int defun_start = 0;
483 int defun_start_byte = 0;
195d1361 484 register enum syntaxcode code;
95ff8dfc 485 int nesting = 1; /* current comment nesting */
ea315ed6 486 int c;
3f679f55
SM
487 int syntax = 0;
488
489 /* FIXME: A }} comment-ender style leads to incorrect behavior
490 in the case of {{ c }}} because we ignore the last two chars which are
491 assumed to be comment-enders although they aren't. */
195d1361
RS
492
493 /* At beginning of range to scan, we're outside of strings;
494 that determines quote parity to the comment-end. */
495 while (from != stop)
496 {
3f679f55
SM
497 int temp_byte, prev_syntax;
498 int com2start, com2end;
6a140a74 499
195d1361 500 /* Move back and examine a character. */
6a140a74 501 DEC_BOTH (from, from_byte);
195d1361
RS
502 UPDATE_SYNTAX_TABLE_BACKWARD (from);
503
3f679f55 504 prev_syntax = syntax;
6a140a74 505 c = FETCH_CHAR (from_byte);
3f679f55 506 syntax = SYNTAX_WITH_FLAGS (c);
195d1361
RS
507 code = SYNTAX (c);
508
3f679f55
SM
509 /* Check for 2-char comment markers. */
510 com2start = (SYNTAX_FLAGS_COMSTART_FIRST (syntax)
511 && SYNTAX_FLAGS_COMSTART_SECOND (prev_syntax)
512 && comstyle == SYNTAX_FLAGS_COMMENT_STYLE (prev_syntax)
513 && (SYNTAX_FLAGS_COMMENT_NESTED (prev_syntax)
514 || SYNTAX_FLAGS_COMMENT_NESTED (syntax)) == comnested);
515 com2end = (SYNTAX_FLAGS_COMEND_FIRST (syntax)
516 && SYNTAX_FLAGS_COMEND_SECOND (prev_syntax));
517
518 /* Nasty cases with overlapping 2-char comment markers:
519 - snmp-mode: -- c -- foo -- c --
520 --- c --
521 ------ c --
522 - c-mode: *||*
523 |* *|* *|
524 |*| |* |*|
525 /// */
526
527 /* If a 2-char comment sequence partly overlaps with another,
528 we don't try to be clever. */
529 if (from > stop && (com2end || com2start))
195d1361 530 {
3f679f55
SM
531 int next = from, next_byte = from_byte, next_c, next_syntax;
532 DEC_BOTH (next, next_byte);
533 UPDATE_SYNTAX_TABLE_BACKWARD (next);
534 next_c = FETCH_CHAR (next_byte);
535 next_syntax = SYNTAX_WITH_FLAGS (next_c);
536 if (((com2start || comnested)
537 && SYNTAX_FLAGS_COMEND_SECOND (syntax)
538 && SYNTAX_FLAGS_COMEND_FIRST (next_syntax))
539 || ((com2end || comnested)
540 && SYNTAX_FLAGS_COMSTART_SECOND (syntax)
541 && comstyle == SYNTAX_FLAGS_COMMENT_STYLE (syntax)
542 && SYNTAX_FLAGS_COMSTART_FIRST (next_syntax)))
543 goto lossage;
544 /* UPDATE_SYNTAX_TABLE_FORWARD (next + 1); */
f902a008 545 }
3f679f55
SM
546
547 if (com2start && comstart_pos == 0)
548 /* We're looking at a comment starter. But it might be a comment
549 ender as well (see snmp-mode). The first time we see one, we
550 need to consider it as a comment starter,
551 and the subsequent times as a comment ender. */
552 com2end = 0;
553
554 /* Turn a 2-char comment sequences into the appropriate syntax. */
555 if (com2end)
556 code = Sendcomment;
557 else if (com2start)
558 code = Scomment;
559 /* Ignore comment starters of a different style. */
560 else if (code == Scomment
561 && (comstyle != SYNTAX_FLAGS_COMMENT_STYLE (syntax)
562 || SYNTAX_FLAGS_COMMENT_NESTED (syntax) != comnested))
1aa963c8 563 continue;
195d1361 564
9828a477
KH
565 /* Ignore escaped characters, except comment-enders. */
566 if (code != Sendcomment && char_quoted (from, from_byte))
195d1361
RS
567 continue;
568
3ee5041c 569 switch (code)
195d1361 570 {
3ee5041c
SM
571 case Sstring_fence:
572 case Scomment_fence:
573 c = (code == Sstring_fence ? ST_STRING_STYLE : ST_COMMENT_STYLE);
574 case Sstring:
575 /* Track parity of quotes. */
576 if (string_style == -1)
577 /* Entering a string. */
578 string_style = c;
579 else if (string_style == c)
580 /* Leaving the string. */
581 string_style = -1;
582 else
583 /* If we have two kinds of string delimiters.
584 There's no way to grok this scanning backwards. */
195d1361 585 string_lossage = 1;
3ee5041c
SM
586 break;
587
588 case Scomment:
589 /* We've already checked that it is the relevant comstyle. */
f7c436c1 590 if (string_style != -1 || comment_lossage || string_lossage)
3ee5041c
SM
591 /* There are odd string quotes involved, so let's be careful.
592 Test case in Pascal: " { " a { " } */
593 goto lossage;
594
f7c436c1
SM
595 if (!comnested)
596 {
597 /* Record best comment-starter so far. */
598 comstart_pos = from;
599 comstart_byte = from_byte;
600 }
601 else if (--nesting <= 0)
95ff8dfc
RS
602 /* nested comments have to be balanced, so we don't need to
603 keep looking for earlier ones. We use here the same (slightly
604 incorrect) reasoning as below: since it is followed by uniform
605 paired string quotes, this comment-start has to be outside of
606 strings, else the comment-end itself would be inside a string. */
607 goto done;
3ee5041c
SM
608 break;
609
02010917 610 case Sendcomment:
3f679f55
SM
611 if (SYNTAX_FLAGS_COMMENT_STYLE (syntax) == comstyle
612 && (SYNTAX_FLAGS_COMMENT_NESTED (prev_syntax)
613 || SYNTAX_FLAGS_COMMENT_NESTED (syntax)) == comnested)
02010917
SM
614 /* This is the same style of comment ender as ours. */
615 {
616 if (comnested)
617 nesting++;
618 else
619 /* Anything before that can't count because it would match
620 this comment-ender rather than ours. */
621 from = stop; /* Break out of the loop. */
622 }
f7c436c1
SM
623 else if (comstart_pos != 0 || c != '\n')
624 /* We're mixing comment styles here, so we'd better be careful.
625 The (comstart_pos != 0 || c != '\n') check is not quite correct
626 (we should just always set comment_lossage), but removing it
627 would imply that any multiline comment in C would go through
628 lossage, which seems overkill.
629 The failure should only happen in the rare cases such as
630 { (* } *) */
631 comment_lossage = 1;
02010917 632 break;
195d1361 633
02010917
SM
634 case Sopen:
635 /* Assume a defun-start point is outside of strings. */
636 if (open_paren_in_column_0_is_defun_start
637 && (from == stop
638 || (temp_byte = dec_bytepos (from_byte),
639 FETCH_CHAR (temp_byte) == '\n')))
640 {
641 defun_start = from;
642 defun_start_byte = from_byte;
643 from = stop; /* Break out of the loop. */
644 }
eb35b628 645 break;
02010917
SM
646
647 default:
f7c436c1 648 break;
eb35b628 649 }
195d1361
RS
650 }
651
652 if (comstart_pos == 0)
653 {
654 from = comment_end;
6a140a74 655 from_byte = comment_end_byte;
195d1361
RS
656 UPDATE_SYNTAX_TABLE_FORWARD (comment_end - 1);
657 }
f7c436c1
SM
658 /* If comstart_pos is set and we get here (ie. didn't jump to `lossage'
659 or `done'), then we've found the beginning of the non-nested comment. */
660 else if (1) /* !comnested */
195d1361
RS
661 {
662 from = comstart_pos;
6a140a74 663 from_byte = comstart_byte;
195d1361
RS
664 /* Globals are correct now. */
665 }
666 else
667 {
3ee5041c
SM
668 struct lisp_parse_state state;
669 lossage:
195d1361
RS
670 /* We had two kinds of string delimiters mixed up
671 together. Decode this going forwards.
02010917 672 Scan fwd from a known safe place (beginning-of-defun)
195d1361
RS
673 to the one in question; this records where we
674 last passed a comment starter. */
eb35b628
RS
675 /* If we did not already find the defun start, find it now. */
676 if (defun_start == 0)
677 {
678 defun_start = find_defun_start (comment_end, comment_end_byte);
679 defun_start_byte = find_start_value_byte;
680 }
02010917 681 do
195d1361 682 {
02010917
SM
683 scan_sexps_forward (&state,
684 defun_start, defun_start_byte,
685 comment_end, -10000, 0, Qnil, 0);
686 defun_start = comment_end;
687 if (state.incomment == (comnested ? 1 : -1)
688 && state.comstyle == comstyle)
689 from = state.comstr_start;
690 else
691 {
692 from = comment_end;
693 if (state.incomment)
694 /* If comment_end is inside some other comment, maybe ours
695 is nested, so we need to try again from within the
696 surrounding comment. Example: { a (* " *) */
697 {
698 /* FIXME: We should advance by one or two chars. */
699 defun_start = state.comstr_start + 2;
700 defun_start_byte = CHAR_TO_BYTE (defun_start);
701 }
702 }
703 } while (defun_start < comment_end);
704
6a140a74 705 from_byte = CHAR_TO_BYTE (from);
195d1361
RS
706 UPDATE_SYNTAX_TABLE_FORWARD (from - 1);
707 }
708
95ff8dfc 709 done:
6a140a74
RS
710 *charpos_ptr = from;
711 *bytepos_ptr = from_byte;
712
1aa963c8 713 return (from == comment_end) ? -1 : from;
195d1361
RS
714}
715\f
8489eb67 716DEFUN ("syntax-table-p", Fsyntax_table_p, Ssyntax_table_p, 1, 1, 0,
fdb82f93
PJ
717 doc: /* Return t if OBJECT is a syntax table.
718Currently, any char-table counts as a syntax table. */)
719 (object)
2203e1e8 720 Lisp_Object object;
8489eb67 721{
2203e1e8 722 if (CHAR_TABLE_P (object)
e704cb4b 723 && EQ (XCHAR_TABLE (object)->purpose, Qsyntax_table))
8489eb67
RS
724 return Qt;
725 return Qnil;
726}
727
8ea151b2 728static void
8489eb67
RS
729check_syntax_table (obj)
730 Lisp_Object obj;
731{
d1be9f0f 732 if (!(CHAR_TABLE_P (obj)
e704cb4b 733 && EQ (XCHAR_TABLE (obj)->purpose, Qsyntax_table)))
d1be9f0f 734 wrong_type_argument (Qsyntax_table_p, obj);
8489eb67
RS
735}
736
8489eb67 737DEFUN ("syntax-table", Fsyntax_table, Ssyntax_table, 0, 0, 0,
fdb82f93
PJ
738 doc: /* Return the current syntax table.
739This is the one specified by the current buffer. */)
740 ()
8489eb67
RS
741{
742 return current_buffer->syntax_table;
743}
744
745DEFUN ("standard-syntax-table", Fstandard_syntax_table,
746 Sstandard_syntax_table, 0, 0, 0,
fdb82f93
PJ
747 doc: /* Return the standard syntax table.
748This is the one used for new buffers. */)
749 ()
8489eb67
RS
750{
751 return Vstandard_syntax_table;
752}
753
754DEFUN ("copy-syntax-table", Fcopy_syntax_table, Scopy_syntax_table, 0, 1, 0,
fdb82f93
PJ
755 doc: /* Construct a new syntax table and return it.
756It is a copy of the TABLE, which defaults to the standard syntax table. */)
757 (table)
8489eb67
RS
758 Lisp_Object table;
759{
8ea151b2
RS
760 Lisp_Object copy;
761
265a9e55 762 if (!NILP (table))
8ea151b2
RS
763 check_syntax_table (table);
764 else
765 table = Vstandard_syntax_table;
766
767 copy = Fcopy_sequence (table);
0f867324
RS
768
769 /* Only the standard syntax table should have a default element.
770 Other syntax tables should inherit from parents instead. */
771 XCHAR_TABLE (copy)->defalt = Qnil;
772
773 /* Copied syntax tables should all have parents.
774 If we copied one with no parent, such as the standard syntax table,
775 use the standard syntax table as the copy's parent. */
776 if (NILP (XCHAR_TABLE (copy)->parent))
777 Fset_char_table_parent (copy, Vstandard_syntax_table);
8ea151b2 778 return copy;
8489eb67
RS
779}
780
781DEFUN ("set-syntax-table", Fset_syntax_table, Sset_syntax_table, 1, 1, 0,
fdb82f93
PJ
782 doc: /* Select a new syntax table for the current buffer.
783One argument, a syntax table. */)
784 (table)
8489eb67
RS
785 Lisp_Object table;
786{
4e4dfa78 787 int idx;
8ea151b2 788 check_syntax_table (table);
8489eb67
RS
789 current_buffer->syntax_table = table;
790 /* Indicate that this buffer now has a specified syntax table. */
f6cd0527
GM
791 idx = PER_BUFFER_VAR_IDX (syntax_table);
792 SET_PER_BUFFER_VALUE_P (current_buffer, idx, 1);
8489eb67
RS
793 return table;
794}
795\f
796/* Convert a letter which signifies a syntax code
797 into the code it signifies.
195d1361 798 This is used by modify-syntax-entry, and other things. */
8489eb67
RS
799
800unsigned char syntax_spec_code[0400] =
801 { 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
802 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
803 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
804 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
195d1361 805 (char) Swhitespace, (char) Scomment_fence, (char) Sstring, 0377,
8489eb67
RS
806 (char) Smath, 0377, 0377, (char) Squote,
807 (char) Sopen, (char) Sclose, 0377, 0377,
808 0377, (char) Swhitespace, (char) Spunct, (char) Scharquote,
809 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
810 0377, 0377, 0377, 0377,
811 (char) Scomment, 0377, (char) Sendcomment, 0377,
6cb71bf6 812 (char) Sinherit, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */
8489eb67
RS
813 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
814 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
815 0377, 0377, 0377, 0377, (char) Sescape, 0377, 0377, (char) Ssymbol,
816 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* `, a, ... */
817 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
818 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
195d1361 819 0377, 0377, 0377, 0377, (char) Sstring_fence, 0377, 0377, 0377
8489eb67
RS
820 };
821
195d1361 822/* Indexed by syntax code, give the letter that describes it. */
8489eb67 823
195d1361 824char syntax_code_spec[16] =
8489eb67 825 {
195d1361
RS
826 ' ', '.', 'w', '_', '(', ')', '\'', '\"', '$', '\\', '/', '<', '>', '@',
827 '!', '|'
8489eb67 828 };
93da5fff
KH
829
830/* Indexed by syntax code, give the object (cons of syntax code and
831 nil) to be stored in syntax table. Since these objects can be
832 shared among syntax tables, we generate them in advance. By
833 sharing objects, the function `describe-syntax' can give a more
834 compact listing. */
835static Lisp_Object Vsyntax_code_object;
836
8489eb67 837\f
8ea151b2
RS
838/* Look up the value for CHARACTER in syntax table TABLE's parent
839 and its parents. SYNTAX_ENTRY calls this, when TABLE itself has nil
840 for CHARACTER. It's actually used only when not compiled with GCC. */
841
842Lisp_Object
843syntax_parent_lookup (table, character)
844 Lisp_Object table;
845 int character;
846{
847 Lisp_Object value;
848
849 while (1)
850 {
851 table = XCHAR_TABLE (table)->parent;
852 if (NILP (table))
853 return Qnil;
854
855 value = XCHAR_TABLE (table)->contents[character];
856 if (!NILP (value))
857 return value;
858 }
859}
860
8489eb67 861DEFUN ("char-syntax", Fchar_syntax, Schar_syntax, 1, 1, 0,
fdb82f93
PJ
862 doc: /* Return the syntax code of CHARACTER, described by a character.
863For example, if CHARACTER is a word constituent,
864the character `w' is returned.
865The characters that correspond to various syntax codes
866are listed in the documentation of `modify-syntax-entry'. */)
867 (character)
2203e1e8 868 Lisp_Object character;
8489eb67 869{
8ea151b2 870 int char_int;
195d1361
RS
871 gl_state.current_syntax_table = current_buffer->syntax_table;
872
873 gl_state.use_global = 0;
b7826503 874 CHECK_NUMBER (character);
2203e1e8 875 char_int = XINT (character);
8ea151b2 876 return make_number (syntax_code_spec[(int) SYNTAX (char_int)]);
beefa22e
RS
877}
878
879DEFUN ("matching-paren", Fmatching_paren, Smatching_paren, 1, 1, 0,
fdb82f93
PJ
880 doc: /* Return the matching parenthesis of CHARACTER, or nil if none. */)
881 (character)
2203e1e8 882 Lisp_Object character;
beefa22e 883{
8ea151b2 884 int char_int, code;
195d1361
RS
885 gl_state.current_syntax_table = current_buffer->syntax_table;
886 gl_state.use_global = 0;
b7826503 887 CHECK_NUMBER (character);
2203e1e8 888 char_int = XINT (character);
8ea151b2 889 code = SYNTAX (char_int);
a8bd7cd8 890 if (code == Sopen || code == Sclose)
2e34157c 891 return SYNTAX_MATCH (char_int);
beefa22e 892 return Qnil;
8489eb67
RS
893}
894
c65adb44 895DEFUN ("string-to-syntax", Fstring_to_syntax, Sstring_to_syntax, 1, 1, 0,
fdb82f93
PJ
896 doc: /* Convert a syntax specification STRING into syntax cell form.
897STRING should be a string as it is allowed as argument of
898`modify-syntax-entry'. Value is the equivalent cons cell
899(CODE . MATCHING-CHAR) that can be used as value of a `syntax-table'
900text property. */)
901 (string)
c65adb44 902 Lisp_Object string;
8489eb67 903{
8ea151b2 904 register unsigned char *p;
8489eb67 905 register enum syntaxcode code;
247e20a8 906 int val;
8ea151b2 907 Lisp_Object match;
8489eb67 908
b7826503 909 CHECK_STRING (string);
8489eb67 910
c65adb44 911 p = XSTRING (string)->data;
8489eb67
RS
912 code = (enum syntaxcode) syntax_spec_code[*p++];
913 if (((int) code & 0377) == 0377)
2e34157c 914 error ("invalid syntax description letter: %c", p[-1]);
8489eb67 915
8ea151b2 916 if (code == Sinherit)
c65adb44 917 return Qnil;
8ea151b2
RS
918
919 if (*p)
d1be9f0f 920 {
93da5fff 921 int len;
7c19e8e1 922 int character = (STRING_CHAR_AND_LENGTH
c65adb44 923 (p, STRING_BYTES (XSTRING (string)) - 1, len));
93da5fff 924 XSETINT (match, character);
d1be9f0f
RS
925 if (XFASTINT (match) == ' ')
926 match = Qnil;
93da5fff 927 p += len;
d1be9f0f
RS
928 }
929 else
8ea151b2 930 match = Qnil;
8489eb67 931
8ea151b2 932 val = (int) code;
8489eb67
RS
933 while (*p)
934 switch (*p++)
935 {
936 case '1':
247e20a8 937 val |= 1 << 16;
8489eb67
RS
938 break;
939
940 case '2':
247e20a8 941 val |= 1 << 17;
8489eb67
RS
942 break;
943
944 case '3':
247e20a8 945 val |= 1 << 18;
8489eb67
RS
946 break;
947
948 case '4':
247e20a8 949 val |= 1 << 19;
8489eb67
RS
950 break;
951
952 case 'p':
247e20a8 953 val |= 1 << 20;
8489eb67 954 break;
e5d4f4dc
RS
955
956 case 'b':
247e20a8 957 val |= 1 << 21;
e5d4f4dc 958 break;
95ff8dfc
RS
959
960 case 'n':
961 val |= 1 << 22;
962 break;
8489eb67
RS
963 }
964
93da5fff 965 if (val < XVECTOR (Vsyntax_code_object)->size && NILP (match))
c65adb44 966 return XVECTOR (Vsyntax_code_object)->contents[val];
93da5fff
KH
967 else
968 /* Since we can't use a shared object, let's make a new one. */
c65adb44
SM
969 return Fcons (make_number (val), match);
970}
971
fdb82f93 972/* I really don't know why this is interactive
2c9e1900 973 help-form should at least be made useful whilst reading the second arg. */
c65adb44 974DEFUN ("modify-syntax-entry", Fmodify_syntax_entry, Smodify_syntax_entry, 2, 3,
c65adb44 975 "cSet syntax for character: \nsSet syntax for %s to: ",
a50a10a0 976 doc: /* Set syntax for character CHAR according to string NEWENTRY.
fdb82f93
PJ
977The syntax is changed only for table SYNTAX_TABLE, which defaults to
978 the current buffer's syntax table.
979The first character of NEWENTRY should be one of the following:
980 Space or - whitespace syntax. w word constituent.
981 _ symbol constituent. . punctuation.
982 ( open-parenthesis. ) close-parenthesis.
983 " string quote. \\ escape.
984 $ paired delimiter. ' expression quote or prefix operator.
985 < comment starter. > comment ender.
986 / character-quote. @ inherit from `standard-syntax-table'.
987 | generic string fence. ! generic comment fence.
988
989Only single-character comment start and end sequences are represented thus.
990Two-character sequences are represented as described below.
991The second character of NEWENTRY is the matching parenthesis,
992 used only if the first character is `(' or `)'.
993Any additional characters are flags.
994Defined flags are the characters 1, 2, 3, 4, b, p, and n.
a50a10a0
PJ
995 1 means CHAR is the start of a two-char comment start sequence.
996 2 means CHAR is the second character of such a sequence.
997 3 means CHAR is the start of a two-char comment end sequence.
998 4 means CHAR is the second character of such a sequence.
fdb82f93
PJ
999
1000There can be up to two orthogonal comment sequences. This is to support
1001language modes such as C++. By default, all comment sequences are of style
1002a, but you can set the comment sequence style to b (on the second character
1003of a comment-start, or the first character of a comment-end sequence) using
1004this flag:
a50a10a0
PJ
1005 b means CHAR is part of comment sequence b.
1006 n means CHAR is part of a nestable comment sequence.
fdb82f93 1007
a50a10a0 1008 p means CHAR is a prefix character for `backward-prefix-chars';
fdb82f93 1009 such characters are treated as whitespace when they occur
a50a10a0
PJ
1010 between expressions.
1011usage: (modify-syntax-entry CHAR NEWENTRY &optional SYNTAX-TABLE) */)
fdb82f93 1012 (c, newentry, syntax_table)
c65adb44
SM
1013 Lisp_Object c, newentry, syntax_table;
1014{
b7826503 1015 CHECK_NUMBER (c);
c65adb44
SM
1016
1017 if (NILP (syntax_table))
1018 syntax_table = current_buffer->syntax_table;
1019 else
1020 check_syntax_table (syntax_table);
8489eb67 1021
c65adb44 1022 SET_RAW_SYNTAX_ENTRY (syntax_table, XINT (c), Fstring_to_syntax (newentry));
8489eb67
RS
1023 return Qnil;
1024}
1025\f
1026/* Dump syntax table to buffer in human-readable format */
1027
62abe9cb
SM
1028DEFUN ("internal-describe-syntax-value", Finternal_describe_syntax_value,
1029 Sinternal_describe_syntax_value, 1, 1, 0,
1030 doc: /* Insert a description of the internal syntax description SYNTAX at point. */)
1031 (syntax)
1032 Lisp_Object syntax;
8489eb67
RS
1033{
1034 register enum syntaxcode code;
e6365855 1035 char desc, start1, start2, end1, end2, prefix, comstyle, comnested;
8489eb67 1036 char str[2];
62abe9cb 1037 Lisp_Object first, match_lisp, value = syntax;
8489eb67 1038
8ea151b2
RS
1039 if (NILP (value))
1040 {
62abe9cb
SM
1041 insert_string ("default");
1042 return syntax;
8ea151b2
RS
1043 }
1044
908b7fea
KH
1045 if (CHAR_TABLE_P (value))
1046 {
62abe9cb
SM
1047 insert_string ("deeper char-table ...");
1048 return syntax;
908b7fea
KH
1049 }
1050
8ea151b2
RS
1051 if (!CONSP (value))
1052 {
62abe9cb
SM
1053 insert_string ("invalid");
1054 return syntax;
8ea151b2
RS
1055 }
1056
c1d497be
KR
1057 first = XCAR (value);
1058 match_lisp = XCDR (value);
8ea151b2
RS
1059
1060 if (!INTEGERP (first) || !(NILP (match_lisp) || INTEGERP (match_lisp)))
8489eb67 1061 {
62abe9cb
SM
1062 insert_string ("invalid");
1063 return syntax;
8489eb67
RS
1064 }
1065
e704cb4b 1066 code = (enum syntaxcode) (XINT (first) & 0377);
8ea151b2
RS
1067 start1 = (XINT (first) >> 16) & 1;
1068 start2 = (XINT (first) >> 17) & 1;
1069 end1 = (XINT (first) >> 18) & 1;
1070 end2 = (XINT (first) >> 19) & 1;
1071 prefix = (XINT (first) >> 20) & 1;
1072 comstyle = (XINT (first) >> 21) & 1;
e6365855 1073 comnested = (XINT (first) >> 22) & 1;
8489eb67
RS
1074
1075 if ((int) code < 0 || (int) code >= (int) Smax)
1076 {
1077 insert_string ("invalid");
62abe9cb 1078 return syntax;
8489eb67
RS
1079 }
1080 desc = syntax_code_spec[(int) code];
1081
1082 str[0] = desc, str[1] = 0;
1083 insert (str, 1);
1084
93da5fff
KH
1085 if (NILP (match_lisp))
1086 insert (" ", 1);
1087 else
1088 insert_char (XINT (match_lisp));
8489eb67 1089
8489eb67
RS
1090 if (start1)
1091 insert ("1", 1);
1092 if (start2)
1093 insert ("2", 1);
1094
1095 if (end1)
1096 insert ("3", 1);
1097 if (end2)
1098 insert ("4", 1);
1099
1100 if (prefix)
1101 insert ("p", 1);
e5d4f4dc
RS
1102 if (comstyle)
1103 insert ("b", 1);
e6365855
SM
1104 if (comnested)
1105 insert ("n", 1);
8489eb67
RS
1106
1107 insert_string ("\twhich means: ");
1108
0220c518 1109 switch (SWITCH_ENUM_CAST (code))
8489eb67
RS
1110 {
1111 case Swhitespace:
1112 insert_string ("whitespace"); break;
1113 case Spunct:
1114 insert_string ("punctuation"); break;
1115 case Sword:
1116 insert_string ("word"); break;
1117 case Ssymbol:
1118 insert_string ("symbol"); break;
1119 case Sopen:
1120 insert_string ("open"); break;
1121 case Sclose:
1122 insert_string ("close"); break;
1123 case Squote:
d671382e 1124 insert_string ("prefix"); break;
8489eb67
RS
1125 case Sstring:
1126 insert_string ("string"); break;
1127 case Smath:
1128 insert_string ("math"); break;
1129 case Sescape:
1130 insert_string ("escape"); break;
1131 case Scharquote:
1132 insert_string ("charquote"); break;
1133 case Scomment:
1134 insert_string ("comment"); break;
1135 case Sendcomment:
1136 insert_string ("endcomment"); break;
d671382e
SM
1137 case Sinherit:
1138 insert_string ("inherit"); break;
1139 case Scomment_fence:
1140 insert_string ("comment fence"); break;
1141 case Sstring_fence:
1142 insert_string ("string fence"); break;
8489eb67
RS
1143 default:
1144 insert_string ("invalid");
62abe9cb 1145 return syntax;
8489eb67
RS
1146 }
1147
8ea151b2 1148 if (!NILP (match_lisp))
8489eb67
RS
1149 {
1150 insert_string (", matches ");
8ea151b2 1151 insert_char (XINT (match_lisp));
8489eb67
RS
1152 }
1153
1154 if (start1)
1155 insert_string (",\n\t is the first character of a comment-start sequence");
1156 if (start2)
1157 insert_string (",\n\t is the second character of a comment-start sequence");
1158
1159 if (end1)
1160 insert_string (",\n\t is the first character of a comment-end sequence");
1161 if (end2)
1162 insert_string (",\n\t is the second character of a comment-end sequence");
e5d4f4dc
RS
1163 if (comstyle)
1164 insert_string (" (comment style b)");
e6365855
SM
1165 if (comnested)
1166 insert_string (" (nestable)");
e5d4f4dc 1167
8489eb67
RS
1168 if (prefix)
1169 insert_string (",\n\t is a prefix character for `backward-prefix-chars'");
1170
62abe9cb
SM
1171 return syntax;
1172}
8489eb67 1173\f
195d1361
RS
1174int parse_sexp_ignore_comments;
1175
8489eb67
RS
1176/* Return the position across COUNT words from FROM.
1177 If that many words cannot be found before the end of the buffer, return 0.
1178 COUNT negative means scan backward and stop at word beginning. */
1179
dfcf069d 1180int
8489eb67
RS
1181scan_words (from, count)
1182 register int from, count;
1183{
1184 register int beg = BEGV;
1185 register int end = ZV;
6a140a74 1186 register int from_byte = CHAR_TO_BYTE (from);
93da5fff
KH
1187 register enum syntaxcode code;
1188 int ch0, ch1;
8489eb67
RS
1189
1190 immediate_quit = 1;
1191 QUIT;
1192
195d1361
RS
1193 SETUP_SYNTAX_TABLE (from, count);
1194
8489eb67
RS
1195 while (count > 0)
1196 {
1197 while (1)
1198 {
1199 if (from == end)
1200 {
1201 immediate_quit = 0;
1202 return 0;
1203 }
195d1361 1204 UPDATE_SYNTAX_TABLE_FORWARD (from);
6a140a74 1205 ch0 = FETCH_CHAR (from_byte);
93da5fff 1206 code = SYNTAX (ch0);
6a140a74 1207 INC_BOTH (from, from_byte);
8489eb67
RS
1208 if (words_include_escapes
1209 && (code == Sescape || code == Scharquote))
1210 break;
1211 if (code == Sword)
1212 break;
8489eb67 1213 }
93da5fff
KH
1214 /* Now CH0 is a character which begins a word and FROM is the
1215 position of the next character. */
8489eb67
RS
1216 while (1)
1217 {
1218 if (from == end) break;
195d1361 1219 UPDATE_SYNTAX_TABLE_FORWARD (from);
6a140a74 1220 ch1 = FETCH_CHAR (from_byte);
93da5fff 1221 code = SYNTAX (ch1);
8489eb67
RS
1222 if (!(words_include_escapes
1223 && (code == Sescape || code == Scharquote)))
93da5fff 1224 if (code != Sword || WORD_BOUNDARY_P (ch0, ch1))
8489eb67 1225 break;
6a140a74 1226 INC_BOTH (from, from_byte);
93da5fff 1227 ch0 = ch1;
8489eb67
RS
1228 }
1229 count--;
1230 }
1231 while (count < 0)
1232 {
1233 while (1)
1234 {
1235 if (from == beg)
1236 {
1237 immediate_quit = 0;
1238 return 0;
1239 }
6a140a74 1240 DEC_BOTH (from, from_byte);
195d1361 1241 UPDATE_SYNTAX_TABLE_BACKWARD (from);
6a140a74 1242 ch1 = FETCH_CHAR (from_byte);
93da5fff 1243 code = SYNTAX (ch1);
8489eb67
RS
1244 if (words_include_escapes
1245 && (code == Sescape || code == Scharquote))
1246 break;
1247 if (code == Sword)
1248 break;
8489eb67 1249 }
93da5fff
KH
1250 /* Now CH1 is a character which ends a word and FROM is the
1251 position of it. */
8489eb67
RS
1252 while (1)
1253 {
6a140a74
RS
1254 int temp_byte;
1255
1256 if (from == beg)
1257 break;
1258 temp_byte = dec_bytepos (from_byte);
195d1361 1259 UPDATE_SYNTAX_TABLE_BACKWARD (from);
6a140a74 1260 ch0 = FETCH_CHAR (temp_byte);
93da5fff 1261 code = SYNTAX (ch0);
8489eb67
RS
1262 if (!(words_include_escapes
1263 && (code == Sescape || code == Scharquote)))
93da5fff 1264 if (code != Sword || WORD_BOUNDARY_P (ch0, ch1))
8489eb67 1265 break;
6a140a74 1266 DEC_BOTH (from, from_byte);
93da5fff 1267 ch1 = ch0;
8489eb67
RS
1268 }
1269 count++;
1270 }
1271
1272 immediate_quit = 0;
1273
1274 return from;
1275}
1276
1277DEFUN ("forward-word", Fforward_word, Sforward_word, 1, 1, "p",
fdb82f93
PJ
1278 doc: /* Move point forward ARG words (backward if ARG is negative).
1279Normally returns t.
1280If an edge of the buffer or a field boundary is reached, point is left there
1281and the function returns nil. Field boundaries are not noticed if
1282`inhibit-field-text-motion' is non-nil. */)
1283 (count)
8489eb67
RS
1284 Lisp_Object count;
1285{
2c6ea900 1286 int orig_val, val;
b7826503 1287 CHECK_NUMBER (count);
8489eb67 1288
2c6ea900
RS
1289 val = orig_val = scan_words (PT, XINT (count));
1290 if (! orig_val)
1291 val = XINT (count) > 0 ? ZV : BEGV;
5878ee6f 1292
b8855607 1293 /* Avoid jumping out of an input field. */
64aa4eb1 1294 val = XFASTINT (Fconstrain_to_field (make_number (val), make_number (PT),
070fe99c 1295 Qt, Qnil, Qnil));
64aa4eb1 1296
8489eb67 1297 SET_PT (val);
e6d8341f 1298 return val == orig_val ? Qt : Qnil;
8489eb67
RS
1299}
1300\f
195d1361
RS
1301Lisp_Object skip_chars ();
1302
1303DEFUN ("skip-chars-forward", Fskip_chars_forward, Sskip_chars_forward, 1, 2, 0,
fdb82f93
PJ
1304 doc: /* Move point forward, stopping before a char not in STRING, or at pos LIM.
1305STRING is like the inside of a `[...]' in a regular expression
1306except that `]' is never special and `\\' quotes `^', `-' or `\\'
1307 (but not as the end of a range; quoting is never needed there).
1308Thus, with arg "a-zA-Z", this skips letters stopping before first nonletter.
1309With arg "^a-zA-Z", skips nonletters stopping before first letter.
1310Returns the distance traveled, either zero or positive. */)
1311 (string, lim)
195d1361
RS
1312 Lisp_Object string, lim;
1313{
1314 return skip_chars (1, 0, string, lim);
1315}
1316
1317DEFUN ("skip-chars-backward", Fskip_chars_backward, Sskip_chars_backward, 1, 2, 0,
fdb82f93
PJ
1318 doc: /* Move point backward, stopping after a char not in STRING, or at pos LIM.
1319See `skip-chars-forward' for details.
1320Returns the distance traveled, either zero or negative. */)
1321 (string, lim)
195d1361
RS
1322 Lisp_Object string, lim;
1323{
1324 return skip_chars (0, 0, string, lim);
1325}
1326
1327DEFUN ("skip-syntax-forward", Fskip_syntax_forward, Sskip_syntax_forward, 1, 2, 0,
fdb82f93
PJ
1328 doc: /* Move point forward across chars in specified syntax classes.
1329SYNTAX is a string of syntax code characters.
1330Stop before a char whose syntax is not in SYNTAX, or at position LIM.
1331If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1332This function returns the distance traveled, either zero or positive. */)
1333 (syntax, lim)
195d1361
RS
1334 Lisp_Object syntax, lim;
1335{
1336 return skip_chars (1, 1, syntax, lim);
1337}
1338
1339DEFUN ("skip-syntax-backward", Fskip_syntax_backward, Sskip_syntax_backward, 1, 2, 0,
fdb82f93
PJ
1340 doc: /* Move point backward across chars in specified syntax classes.
1341SYNTAX is a string of syntax code characters.
1342Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.
1343If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1344This function returns the distance traveled, either zero or negative. */)
1345 (syntax, lim)
195d1361
RS
1346 Lisp_Object syntax, lim;
1347{
1348 return skip_chars (0, 1, syntax, lim);
1349}
1350
6a140a74 1351static Lisp_Object
195d1361
RS
1352skip_chars (forwardp, syntaxp, string, lim)
1353 int forwardp, syntaxp;
1354 Lisp_Object string, lim;
1355{
195d1361
RS
1356 register unsigned int c;
1357 unsigned char fastmap[0400];
7e68b0ea 1358 /* If SYNTAXP is 0, STRING may contain multi-byte form of characters
9690d026
KH
1359 of which codes don't fit in FASTMAP. In that case, set the
1360 ranges of characters in CHAR_RANGES. */
5df4982e 1361 int *char_ranges;
7e68b0ea 1362 int n_char_ranges = 0;
195d1361 1363 int negate = 0;
4101e6fe 1364 register int i, i_byte;
7e68b0ea 1365 int multibyte = !NILP (current_buffer->enable_multibyte_characters);
1674d9a2
RS
1366 int string_multibyte;
1367 int size_byte;
82d497fc
KH
1368 unsigned char *str;
1369 int len;
195d1361 1370
b7826503 1371 CHECK_STRING (string);
5df4982e 1372 char_ranges = (int *) alloca (XSTRING (string)->size * (sizeof (int)) * 2);
1674d9a2 1373 string_multibyte = STRING_MULTIBYTE (string);
82d497fc 1374 str = XSTRING (string)->data;
1674d9a2 1375 size_byte = STRING_BYTES (XSTRING (string));
195d1361 1376
82d497fc
KH
1377 /* Adjust the multibyteness of the string to that of the buffer. */
1378 if (multibyte != string_multibyte)
1379 {
1380 int nbytes;
1381
1382 if (multibyte)
1383 nbytes = count_size_as_multibyte (XSTRING (string)->data,
1384 XSTRING (string)->size);
1385 else
1386 nbytes = XSTRING (string)->size;
1387 if (nbytes != size_byte)
1388 {
1389 str = (unsigned char *) alloca (nbytes);
e39091b6 1390 copy_text (XSTRING (string)->data, str, size_byte,
82d497fc
KH
1391 string_multibyte, multibyte);
1392 size_byte = nbytes;
1393 }
1394 }
1395
195d1361
RS
1396 if (NILP (lim))
1397 XSETINT (lim, forwardp ? ZV : BEGV);
1398 else
b7826503 1399 CHECK_NUMBER_COERCE_MARKER (lim);
195d1361
RS
1400
1401 /* In any case, don't allow scan outside bounds of buffer. */
195d1361
RS
1402 if (XINT (lim) > ZV)
1403 XSETFASTINT (lim, ZV);
1404 if (XINT (lim) < BEGV)
1405 XSETFASTINT (lim, BEGV);
1406
195d1361
RS
1407 bzero (fastmap, sizeof fastmap);
1408
82d497fc 1409 i_byte = 0;
4101e6fe 1410
13090112 1411 if (i_byte < size_byte
4101e6fe 1412 && XSTRING (string)->data[0] == '^')
195d1361 1413 {
82d497fc 1414 negate = 1; i_byte++;
195d1361
RS
1415 }
1416
1417 /* Find the characters specified and set their elements of fastmap.
1418 If syntaxp, each character counts as itself.
1419 Otherwise, handle backslashes and ranges specially. */
1420
13090112 1421 while (i_byte < size_byte)
195d1361 1422 {
82d497fc
KH
1423 c = STRING_CHAR_AND_LENGTH (str + i_byte, size_byte - i_byte, len);
1424 i_byte += len;
4101e6fe 1425
195d1361 1426 if (syntaxp)
4101e6fe 1427 fastmap[syntax_spec_code[c & 0377]] = 1;
195d1361
RS
1428 else
1429 {
1430 if (c == '\\')
1431 {
13090112 1432 if (i_byte == size_byte)
4101e6fe
RS
1433 break;
1434
82d497fc
KH
1435 c = STRING_CHAR_AND_LENGTH (str+i_byte, size_byte-i_byte, len);
1436 i_byte += len;
195d1361 1437 }
13090112 1438 if (i_byte < size_byte
82d497fc 1439 && str[i_byte] == '-')
195d1361 1440 {
9690d026 1441 unsigned int c2;
4101e6fe
RS
1442
1443 /* Skip over the dash. */
82d497fc 1444 i_byte++;
4101e6fe 1445
13090112 1446 if (i_byte == size_byte)
4101e6fe
RS
1447 break;
1448
1449 /* Get the end of the range. */
82d497fc
KH
1450 c2 =STRING_CHAR_AND_LENGTH (str+i_byte, size_byte-i_byte, len);
1451 i_byte += len;
7e68b0ea 1452
4101e6fe 1453 if (SINGLE_BYTE_CHAR_P (c))
06274af5
KH
1454 {
1455 if (! SINGLE_BYTE_CHAR_P (c2))
e39091b6 1456 {
7ae339a7
KH
1457 /* Handle a range starting with a character of
1458 less than 256, and ending with a character of
1459 not less than 256. Split that into two
1460 ranges, the low one ending at 0377, and the
1461 high one starting at the smallest character
1462 in the charset of C2 and ending at C2. */
e39091b6
KH
1463 int charset = CHAR_CHARSET (c2);
1464 int c1 = MAKE_CHAR (charset, 0, 0);
1465
e39091b6
KH
1466 char_ranges[n_char_ranges++] = c1;
1467 char_ranges[n_char_ranges++] = c2;
465cf84c 1468 c2 = 0377;
e39091b6 1469 }
06274af5
KH
1470 while (c <= c2)
1471 {
1472 fastmap[c] = 1;
1473 c++;
1474 }
1475 }
9690d026 1476 else if (c <= c2) /* Both C and C2 are multibyte char. */
195d1361 1477 {
9690d026
KH
1478 char_ranges[n_char_ranges++] = c;
1479 char_ranges[n_char_ranges++] = c2;
195d1361 1480 }
195d1361
RS
1481 }
1482 else
7e68b0ea 1483 {
e39091b6
KH
1484 if (SINGLE_BYTE_CHAR_P (c))
1485 fastmap[c] = 1;
1486 else
7e68b0ea 1487 {
4101e6fe
RS
1488 char_ranges[n_char_ranges++] = c;
1489 char_ranges[n_char_ranges++] = c;
7e68b0ea
RS
1490 }
1491 }
195d1361
RS
1492 }
1493 }
1494
9690d026 1495 /* If ^ was the first character, complement the fastmap. */
195d1361
RS
1496 if (negate)
1497 for (i = 0; i < sizeof fastmap; i++)
9690d026 1498 fastmap[i] ^= 1;
195d1361
RS
1499
1500 {
1501 int start_point = PT;
1502 int pos = PT;
6a140a74 1503 int pos_byte = PT_BYTE;
195d1361
RS
1504
1505 immediate_quit = 1;
1506 if (syntaxp)
1507 {
1508 SETUP_SYNTAX_TABLE (pos, forwardp ? 1 : -1);
1509 if (forwardp)
1510 {
7e68b0ea 1511 if (multibyte)
195d1361 1512 {
ee0cdb48
KH
1513 if (pos < XINT (lim))
1514 while (fastmap[(int) SYNTAX (FETCH_CHAR (pos_byte))])
1515 {
1516 /* Since we already checked for multibyteness,
1517 avoid using INC_BOTH which checks again. */
1518 INC_POS (pos_byte);
1519 pos++;
1520 if (pos >= XINT (lim))
1521 break;
1522 UPDATE_SYNTAX_TABLE_FORWARD (pos);
1523 }
7e68b0ea
RS
1524 }
1525 else
1526 {
1527 while (pos < XINT (lim)
1528 && fastmap[(int) SYNTAX (FETCH_BYTE (pos))])
1529 {
1530 pos++;
1531 UPDATE_SYNTAX_TABLE_FORWARD (pos);
1532 }
195d1361
RS
1533 }
1534 }
1535 else
1536 {
7e68b0ea 1537 if (multibyte)
195d1361 1538 {
7e68b0ea
RS
1539 while (pos > XINT (lim))
1540 {
6a140a74 1541 int savepos = pos_byte;
ef316cf0
RS
1542 /* Since we already checked for multibyteness,
1543 avoid using DEC_BOTH which checks again. */
1544 pos--;
1545 DEC_POS (pos_byte);
4c8535eb 1546 UPDATE_SYNTAX_TABLE_BACKWARD (pos);
6a140a74 1547 if (!fastmap[(int) SYNTAX (FETCH_CHAR (pos_byte))])
7e68b0ea 1548 {
6a140a74
RS
1549 pos++;
1550 pos_byte = savepos;
7e68b0ea
RS
1551 break;
1552 }
7e68b0ea
RS
1553 }
1554 }
1555 else
1556 {
ee0cdb48
KH
1557 if (pos > XINT (lim))
1558 while (fastmap[(int) SYNTAX (FETCH_BYTE (pos - 1))])
1559 {
1560 pos--;
1561 if (pos <= XINT (lim))
4c8535eb 1562 break;
ee0cdb48
KH
1563 UPDATE_SYNTAX_TABLE_BACKWARD (pos - 1);
1564 }
195d1361
RS
1565 }
1566 }
1567 }
1568 else
1569 {
1570 if (forwardp)
1571 {
7e68b0ea 1572 if (multibyte)
9690d026 1573 while (pos < XINT (lim))
7e68b0ea 1574 {
9690d026
KH
1575 c = FETCH_MULTIBYTE_CHAR (pos_byte);
1576 if (SINGLE_BYTE_CHAR_P (c))
7e68b0ea 1577 {
9690d026
KH
1578 if (!fastmap[c])
1579 break;
1580 }
1581 else
1582 {
1583 /* If we are looking at a multibyte character,
1584 we must look up the character in the table
1585 CHAR_RANGES. If there's no data in the
1586 table, that character is not what we want to
1587 skip. */
1588
e39091b6
KH
1589 /* The following code do the right thing even if
1590 n_char_ranges is zero (i.e. no data in
1591 CHAR_RANGES). */
7e68b0ea 1592 for (i = 0; i < n_char_ranges; i += 2)
e39091b6 1593 if (c >= char_ranges[i] && c <= char_ranges[i + 1])
7e68b0ea
RS
1594 break;
1595 if (!(negate ^ (i < n_char_ranges)))
1596 break;
7e68b0ea 1597 }
e39091b6 1598 INC_BOTH (pos, pos_byte);
7e68b0ea
RS
1599 }
1600 else
1601 while (pos < XINT (lim) && fastmap[FETCH_BYTE (pos)])
1602 pos++;
195d1361
RS
1603 }
1604 else
1605 {
7e68b0ea
RS
1606 if (multibyte)
1607 while (pos > XINT (lim))
1608 {
e39091b6
KH
1609 int prev_pos_byte = pos_byte;
1610
1611 DEC_POS (prev_pos_byte);
9690d026
KH
1612 c = FETCH_MULTIBYTE_CHAR (prev_pos_byte);
1613 if (SINGLE_BYTE_CHAR_P (c))
1614 {
1615 if (!fastmap[c])
1616 break;
1617 }
1618 else
7e68b0ea 1619 {
9690d026 1620 /* See the comment in the previous similar code. */
e39091b6
KH
1621 for (i = 0; i < n_char_ranges; i += 2)
1622 if (c >= char_ranges[i] && c <= char_ranges[i + 1])
1623 break;
1624 if (!(negate ^ (i < n_char_ranges)))
1625 break;
7e68b0ea 1626 }
e39091b6
KH
1627 pos--;
1628 pos_byte = prev_pos_byte;
7e68b0ea
RS
1629 }
1630 else
1631 while (pos > XINT (lim) && fastmap[FETCH_BYTE (pos - 1)])
1632 pos--;
195d1361
RS
1633 }
1634 }
7e68b0ea 1635
6a140a74
RS
1636#if 0 /* Not needed now that a position in mid-character
1637 cannot be specified in Lisp. */
7e68b0ea
RS
1638 if (multibyte
1639 /* INC_POS or DEC_POS might have moved POS over LIM. */
1640 && (forwardp ? (pos > XINT (lim)) : (pos < XINT (lim))))
1641 pos = XINT (lim);
6a140a74 1642#endif
7e68b0ea 1643
6a140a74
RS
1644 if (! multibyte)
1645 pos_byte = pos;
1646
1647 SET_PT_BOTH (pos, pos_byte);
195d1361
RS
1648 immediate_quit = 0;
1649
1650 return make_number (PT - start_point);
1651 }
1652}
1653\f
95ff8dfc
RS
1654/* Jump over a comment, assuming we are at the beginning of one.
1655 FROM is the current position.
1656 FROM_BYTE is the bytepos corresponding to FROM.
1657 Do not move past STOP (a charpos).
1658 The comment over which we have to jump is of style STYLE
1659 (either SYNTAX_COMMENT_STYLE(foo) or ST_COMMENT_STYLE).
1660 NESTING should be positive to indicate the nesting at the beginning
1661 for nested comments and should be zero or negative else.
1662 ST_COMMENT_STYLE cannot be nested.
1663 PREV_SYNTAX is the SYNTAX_WITH_FLAGS of the previous character
1664 (or 0 If the search cannot start in the middle of a two-character).
1665
1666 If successful, return 1 and store the charpos of the comment's end
1667 into *CHARPOS_PTR and the corresponding bytepos into *BYTEPOS_PTR.
1668 Else, return 0 and store the charpos STOP into *CHARPOS_PTR, the
1669 corresponding bytepos into *BYTEPOS_PTR and the current nesting
1670 (as defined for state.incomment) in *INCOMMENT_PTR.
1671
1672 The comment end is the last character of the comment rather than the
1673 character just after the comment.
1674
1675 Global syntax data is assumed to initially be valid for FROM and
1676 remains valid for forward search starting at the returned position. */
1677
1678static int
1679forw_comment (from, from_byte, stop, nesting, style, prev_syntax,
1680 charpos_ptr, bytepos_ptr, incomment_ptr)
1681 int from, from_byte, stop;
1682 int nesting, style, prev_syntax;
1683 int *charpos_ptr, *bytepos_ptr, *incomment_ptr;
1684{
1685 register int c, c1;
1686 register enum syntaxcode code;
1687 register int syntax;
1688
1689 if (nesting <= 0) nesting = -1;
1690
1691 /* Enter the loop in the middle so that we find
1692 a 2-char comment ender if we start in the middle of it. */
1693 syntax = prev_syntax;
1694 if (syntax != 0) goto forw_incomment;
1695
1696 while (1)
1697 {
1698 if (from == stop)
1699 {
1700 *incomment_ptr = nesting;
1701 *charpos_ptr = from;
1702 *bytepos_ptr = from_byte;
1703 return 0;
1704 }
1705 c = FETCH_CHAR (from_byte);
1706 syntax = SYNTAX_WITH_FLAGS (c);
1707 code = syntax & 0xff;
1708 if (code == Sendcomment
1709 && SYNTAX_FLAGS_COMMENT_STYLE (syntax) == style
abf8a9ff
SM
1710 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax) ?
1711 (nesting > 0 && --nesting == 0) : nesting < 0))
95ff8dfc
RS
1712 /* we have encountered a comment end of the same style
1713 as the comment sequence which began this comment
1714 section */
1715 break;
1716 if (code == Scomment_fence
1717 && style == ST_COMMENT_STYLE)
1718 /* we have encountered a comment end of the same style
1719 as the comment sequence which began this comment
1720 section. */
1721 break;
1722 if (nesting > 0
1723 && code == Scomment
abf8a9ff 1724 && SYNTAX_FLAGS_COMMENT_NESTED (syntax)
95ff8dfc
RS
1725 && SYNTAX_FLAGS_COMMENT_STYLE (syntax) == style)
1726 /* we have encountered a nested comment of the same style
1727 as the comment sequence which began this comment section */
1728 nesting++;
1729 INC_BOTH (from, from_byte);
1730 UPDATE_SYNTAX_TABLE_FORWARD (from);
1731
1732 forw_incomment:
1733 if (from < stop && SYNTAX_FLAGS_COMEND_FIRST (syntax)
1734 && SYNTAX_FLAGS_COMMENT_STYLE (syntax) == style
1735 && (c1 = FETCH_CHAR (from_byte),
abf8a9ff
SM
1736 SYNTAX_COMEND_SECOND (c1))
1737 && ((SYNTAX_FLAGS_COMMENT_NESTED (syntax) ||
1738 SYNTAX_COMMENT_NESTED (c1)) ? nesting > 0 : nesting < 0))
4ffe723b
GM
1739 {
1740 if (--nesting <= 0)
1741 /* we have encountered a comment end of the same style
1742 as the comment sequence which began this comment
1743 section */
1744 break;
1745 else
1746 {
1747 INC_BOTH (from, from_byte);
1748 UPDATE_SYNTAX_TABLE_FORWARD (from);
1749 }
1750 }
95ff8dfc
RS
1751 if (nesting > 0
1752 && from < stop
1753 && SYNTAX_FLAGS_COMSTART_FIRST (syntax)
1754 && (c1 = FETCH_CHAR (from_byte),
1755 SYNTAX_COMMENT_STYLE (c1) == style
abf8a9ff
SM
1756 && SYNTAX_COMSTART_SECOND (c1))
1757 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax) ||
1758 SYNTAX_COMMENT_NESTED (c1)))
95ff8dfc
RS
1759 /* we have encountered a nested comment of the same style
1760 as the comment sequence which began this comment
1761 section */
1762 {
1763 INC_BOTH (from, from_byte);
1764 UPDATE_SYNTAX_TABLE_FORWARD (from);
1765 nesting++;
1766 }
1767 }
1768 *charpos_ptr = from;
1769 *bytepos_ptr = from_byte;
1770 return 1;
1771}
1772
b3cfe0c8 1773DEFUN ("forward-comment", Fforward_comment, Sforward_comment, 1, 1, 0,
fdb82f93
PJ
1774 doc: /* Move forward across up to N comments. If N is negative, move backward.
1775Stop scanning if we find something other than a comment or whitespace.
1776Set point to where scanning stops.
1777If N comments are found as expected, with nothing except whitespace
1778between them, return t; otherwise return nil. */)
1779 (count)
840f481c 1780 Lisp_Object count;
b3cfe0c8
RS
1781{
1782 register int from;
6a140a74 1783 int from_byte;
b3cfe0c8 1784 register int stop;
8ea151b2 1785 register int c, c1;
b3cfe0c8
RS
1786 register enum syntaxcode code;
1787 int comstyle = 0; /* style of comment encountered */
95ff8dfc 1788 int comnested = 0; /* whether the comment is nestable or not */
be720845 1789 int found;
840f481c 1790 int count1;
6a140a74 1791 int out_charpos, out_bytepos;
95ff8dfc 1792 int dummy;
840f481c 1793
b7826503 1794 CHECK_NUMBER (count);
840f481c 1795 count1 = XINT (count);
195d1361 1796 stop = count1 > 0 ? ZV : BEGV;
b3cfe0c8
RS
1797
1798 immediate_quit = 1;
1799 QUIT;
1800
1801 from = PT;
6a140a74 1802 from_byte = PT_BYTE;
b3cfe0c8 1803
195d1361 1804 SETUP_SYNTAX_TABLE (from, count1);
840f481c 1805 while (count1 > 0)
b3cfe0c8 1806 {
04882296 1807 do
b3cfe0c8 1808 {
f902a008
RS
1809 int comstart_first;
1810
04882296
KH
1811 if (from == stop)
1812 {
ef316cf0 1813 SET_PT_BOTH (from, from_byte);
b7e6e612 1814 immediate_quit = 0;
04882296
KH
1815 return Qnil;
1816 }
6a140a74 1817 c = FETCH_CHAR (from_byte);
b3cfe0c8 1818 code = SYNTAX (c);
f902a008 1819 comstart_first = SYNTAX_COMSTART_FIRST (c);
95ff8dfc 1820 comnested = SYNTAX_COMMENT_NESTED (c);
e6365855 1821 comstyle = SYNTAX_COMMENT_STYLE (c);
6a140a74 1822 INC_BOTH (from, from_byte);
f902a008 1823 UPDATE_SYNTAX_TABLE_FORWARD (from);
f902a008 1824 if (from < stop && comstart_first
6a140a74 1825 && (c1 = FETCH_CHAR (from_byte),
8ea151b2 1826 SYNTAX_COMSTART_SECOND (c1)))
b3cfe0c8 1827 {
7fc8191e
RS
1828 /* We have encountered a comment start sequence and we
1829 are ignoring all text inside comments. We must record
b3cfe0c8
RS
1830 the comment style this sequence begins so that later,
1831 only a comment end of the same style actually ends
7fc8191e 1832 the comment section. */
b3cfe0c8 1833 code = Scomment;
8ea151b2 1834 comstyle = SYNTAX_COMMENT_STYLE (c1);
95ff8dfc 1835 comnested = comnested || SYNTAX_COMMENT_NESTED (c1);
6a140a74 1836 INC_BOTH (from, from_byte);
95ff8dfc 1837 UPDATE_SYNTAX_TABLE_FORWARD (from);
b3cfe0c8 1838 }
04882296 1839 }
abf8a9ff 1840 while (code == Swhitespace || (code == Sendcomment && c == '\n'));
6a140a74 1841
da71ab91
KH
1842 if (code == Scomment_fence)
1843 comstyle = ST_COMMENT_STYLE;
1844 else if (code != Scomment)
04882296
KH
1845 {
1846 immediate_quit = 0;
6a140a74 1847 DEC_BOTH (from, from_byte);
ef316cf0 1848 SET_PT_BOTH (from, from_byte);
04882296
KH
1849 return Qnil;
1850 }
1851 /* We're at the start of a comment. */
95ff8dfc
RS
1852 found = forw_comment (from, from_byte, stop, comnested, comstyle, 0,
1853 &out_charpos, &out_bytepos, &dummy);
1854 from = out_charpos; from_byte = out_bytepos;
1855 if (!found)
04882296 1856 {
95ff8dfc
RS
1857 immediate_quit = 0;
1858 SET_PT_BOTH (from, from_byte);
1859 return Qnil;
b3cfe0c8 1860 }
95ff8dfc
RS
1861 INC_BOTH (from, from_byte);
1862 UPDATE_SYNTAX_TABLE_FORWARD (from);
04882296 1863 /* We have skipped one comment. */
840f481c 1864 count1--;
b3cfe0c8
RS
1865 }
1866
840f481c 1867 while (count1 < 0)
b3cfe0c8 1868 {
b9145dbb 1869 while (1)
b3cfe0c8 1870 {
c65adb44 1871 int quoted;
f902a008 1872
b9145dbb
RS
1873 if (from <= stop)
1874 {
6a140a74 1875 SET_PT_BOTH (BEGV, BEGV_BYTE);
b9145dbb
RS
1876 immediate_quit = 0;
1877 return Qnil;
1878 }
b3cfe0c8 1879
6a140a74 1880 DEC_BOTH (from, from_byte);
f902a008 1881 /* char_quoted does UPDATE_SYNTAX_TABLE_BACKWARD (from). */
6a140a74 1882 quoted = char_quoted (from, from_byte);
6a140a74 1883 c = FETCH_CHAR (from_byte);
b3cfe0c8
RS
1884 code = SYNTAX (c);
1885 comstyle = 0;
95ff8dfc 1886 comnested = SYNTAX_COMMENT_NESTED (c);
7fc8191e
RS
1887 if (code == Sendcomment)
1888 comstyle = SYNTAX_COMMENT_STYLE (c);
b3cfe0c8 1889 if (from > stop && SYNTAX_COMEND_SECOND (c)
f902a008 1890 && prev_char_comend_first (from, from_byte)
89c6809a 1891 && !char_quoted (from - 1, dec_bytepos (from_byte)))
b3cfe0c8 1892 {
7fc8191e 1893 /* We must record the comment style encountered so that
b3cfe0c8 1894 later, we can match only the proper comment begin
7fc8191e 1895 sequence of the same style. */
6a140a74 1896 DEC_BOTH (from, from_byte);
f902a008
RS
1897 code = Sendcomment;
1898 /* Calling char_quoted, above, set up global syntax position
1899 at the new value of FROM. */
95ff8dfc
RS
1900 c1 = FETCH_CHAR (from_byte);
1901 comstyle = SYNTAX_COMMENT_STYLE (c1);
1902 comnested = comnested || SYNTAX_COMMENT_NESTED (c1);
b3cfe0c8
RS
1903 }
1904
195d1361
RS
1905 if (code == Scomment_fence)
1906 {
1907 /* Skip until first preceding unquoted comment_fence. */
6a140a74 1908 int found = 0, ini = from, ini_byte = from_byte;
195d1361 1909
6a140a74 1910 while (1)
195d1361 1911 {
6a140a74
RS
1912 DEC_BOTH (from, from_byte);
1913 if (from == stop)
1914 break;
195d1361 1915 UPDATE_SYNTAX_TABLE_BACKWARD (from);
6a140a74
RS
1916 c = FETCH_CHAR (from_byte);
1917 if (SYNTAX (c) == Scomment_fence
1918 && !char_quoted (from, from_byte))
195d1361
RS
1919 {
1920 found = 1;
1921 break;
1922 }
1923 }
1924 if (found == 0)
1925 {
1926 from = ini; /* Set point to ini + 1. */
6a140a74 1927 from_byte = ini_byte;
195d1361
RS
1928 goto leave;
1929 }
1930 }
1931 else if (code == Sendcomment)
b3cfe0c8 1932 {
95ff8dfc 1933 found = back_comment (from, from_byte, stop, comnested, comstyle,
6a140a74 1934 &out_charpos, &out_bytepos);
1aa963c8
SM
1935 if (found == -1)
1936 {
abf8a9ff
SM
1937 if (c == '\n')
1938 /* This end-of-line is not an end-of-comment.
1939 Treat it like a whitespace.
1940 CC-mode (and maybe others) relies on this behavior. */
1941 ;
1942 else
1943 {
1944 /* Failure: we should go back to the end of this
1945 not-quite-endcomment. */
1946 if (SYNTAX(c) != code)
1947 /* It was a two-char Sendcomment. */
1948 INC_BOTH (from, from_byte);
1949 goto leave;
1950 }
1aa963c8 1951 }
2b927d02 1952 else
abf8a9ff
SM
1953 {
1954 /* We have skipped one comment. */
1955 from = out_charpos, from_byte = out_bytepos;
1956 break;
1957 }
b3cfe0c8 1958 }
bb0de084 1959 else if (code != Swhitespace || quoted)
b3cfe0c8 1960 {
195d1361 1961 leave:
b3cfe0c8 1962 immediate_quit = 0;
6a140a74 1963 INC_BOTH (from, from_byte);
ef316cf0 1964 SET_PT_BOTH (from, from_byte);
b3cfe0c8
RS
1965 return Qnil;
1966 }
1967 }
1968
840f481c 1969 count1++;
b3cfe0c8
RS
1970 }
1971
ef316cf0 1972 SET_PT_BOTH (from, from_byte);
b3cfe0c8
RS
1973 immediate_quit = 0;
1974 return Qt;
1975}
1976\f
bd25db08 1977/* Return syntax code of character C if C is a single byte character
db9f48c3 1978 or `multibyte_symbol_p' is zero. Otherwise, return Ssymbol. */
bd25db08
KH
1979
1980#define SYNTAX_WITH_MULTIBYTE_CHECK(c) \
1981 ((SINGLE_BYTE_CHAR_P (c) || !multibyte_symbol_p) \
1982 ? SYNTAX (c) : Ssymbol)
1983
6a140a74 1984static Lisp_Object
8489eb67
RS
1985scan_lists (from, count, depth, sexpflag)
1986 register int from;
1987 int count, depth, sexpflag;
1988{
1989 Lisp_Object val;
195d1361 1990 register int stop = count > 0 ? ZV : BEGV;
93da5fff
KH
1991 register int c, c1;
1992 int stringterm;
8489eb67
RS
1993 int quoted;
1994 int mathexit = 0;
93da5fff 1995 register enum syntaxcode code, temp_code;
195d1361 1996 int min_depth = depth; /* Err out if depth gets less than this. */
e5d4f4dc 1997 int comstyle = 0; /* style of comment encountered */
95ff8dfc 1998 int comnested = 0; /* whether the comment is nestable or not */
93da5fff 1999 int temp_pos;
7bf5e9e4 2000 int last_good = from;
195d1361 2001 int found;
c3907a7e 2002 int from_byte;
6a140a74 2003 int out_bytepos, out_charpos;
95ff8dfc 2004 int temp, dummy;
bd25db08 2005 int multibyte_symbol_p = sexpflag && multibyte_syntax_as_symbol;
8489eb67
RS
2006
2007 if (depth > 0) min_depth = 0;
2008
c3907a7e
RS
2009 if (from > ZV) from = ZV;
2010 if (from < BEGV) from = BEGV;
2011
2012 from_byte = CHAR_TO_BYTE (from);
2013
8489eb67
RS
2014 immediate_quit = 1;
2015 QUIT;
2016
195d1361 2017 SETUP_SYNTAX_TABLE (from, count);
8489eb67
RS
2018 while (count > 0)
2019 {
8489eb67
RS
2020 while (from < stop)
2021 {
f902a008 2022 int comstart_first, prefix;
195d1361 2023 UPDATE_SYNTAX_TABLE_FORWARD (from);
6a140a74 2024 c = FETCH_CHAR (from_byte);
bd25db08 2025 code = SYNTAX_WITH_MULTIBYTE_CHECK (c);
f902a008 2026 comstart_first = SYNTAX_COMSTART_FIRST (c);
95ff8dfc 2027 comnested = SYNTAX_COMMENT_NESTED (c);
e6365855 2028 comstyle = SYNTAX_COMMENT_STYLE (c);
f902a008 2029 prefix = SYNTAX_PREFIX (c);
7bf5e9e4
RS
2030 if (depth == min_depth)
2031 last_good = from;
6a140a74 2032 INC_BOTH (from, from_byte);
195d1361 2033 UPDATE_SYNTAX_TABLE_FORWARD (from);
f902a008 2034 if (from < stop && comstart_first
6a140a74 2035 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from_byte))
8489eb67 2036 && parse_sexp_ignore_comments)
e5d4f4dc
RS
2037 {
2038 /* we have encountered a comment start sequence and we
195d1361 2039 are ignoring all text inside comments. We must record
e5d4f4dc
RS
2040 the comment style this sequence begins so that later,
2041 only a comment end of the same style actually ends
2042 the comment section */
2043 code = Scomment;
95ff8dfc
RS
2044 c1 = FETCH_CHAR (from_byte);
2045 comstyle = SYNTAX_COMMENT_STYLE (c1);
2046 comnested = comnested || SYNTAX_COMMENT_NESTED (c1);
6a140a74 2047 INC_BOTH (from, from_byte);
f902a008 2048 UPDATE_SYNTAX_TABLE_FORWARD (from);
e5d4f4dc
RS
2049 }
2050
f902a008 2051 if (prefix)
8489eb67
RS
2052 continue;
2053
0220c518 2054 switch (SWITCH_ENUM_CAST (code))
8489eb67
RS
2055 {
2056 case Sescape:
2057 case Scharquote:
2058 if (from == stop) goto lose;
6a140a74 2059 INC_BOTH (from, from_byte);
8489eb67
RS
2060 /* treat following character as a word constituent */
2061 case Sword:
2062 case Ssymbol:
2063 if (depth || !sexpflag) break;
195d1361 2064 /* This word counts as a sexp; return at end of it. */
8489eb67
RS
2065 while (from < stop)
2066 {
195d1361 2067 UPDATE_SYNTAX_TABLE_FORWARD (from);
7c7ca3d2
RS
2068
2069 /* Some compilers can't handle this inside the switch. */
bd25db08
KH
2070 c = FETCH_CHAR (from_byte);
2071 temp = SYNTAX_WITH_MULTIBYTE_CHECK (c);
7c7ca3d2 2072 switch (temp)
8489eb67
RS
2073 {
2074 case Scharquote:
2075 case Sescape:
6a140a74 2076 INC_BOTH (from, from_byte);
8489eb67
RS
2077 if (from == stop) goto lose;
2078 break;
2079 case Sword:
2080 case Ssymbol:
2081 case Squote:
2082 break;
2083 default:
2084 goto done;
2085 }
6a140a74 2086 INC_BOTH (from, from_byte);
8489eb67
RS
2087 }
2088 goto done;
2089
195d1361 2090 case Scomment_fence:
95ff8dfc
RS
2091 comstyle = ST_COMMENT_STYLE;
2092 /* FALLTHROUGH */
2093 case Scomment:
8489eb67 2094 if (!parse_sexp_ignore_comments) break;
95ff8dfc
RS
2095 UPDATE_SYNTAX_TABLE_FORWARD (from);
2096 found = forw_comment (from, from_byte, stop,
2097 comnested, comstyle, 0,
2098 &out_charpos, &out_bytepos, &dummy);
2099 from = out_charpos, from_byte = out_bytepos;
2100 if (!found)
8489eb67 2101 {
95ff8dfc
RS
2102 if (depth == 0)
2103 goto done;
2104 goto lose;
8489eb67 2105 }
95ff8dfc
RS
2106 INC_BOTH (from, from_byte);
2107 UPDATE_SYNTAX_TABLE_FORWARD (from);
8489eb67
RS
2108 break;
2109
2110 case Smath:
2111 if (!sexpflag)
2112 break;
6a140a74
RS
2113 if (from != stop && c == FETCH_CHAR (from_byte))
2114 {
2115 INC_BOTH (from, from_byte);
2116 }
8489eb67
RS
2117 if (mathexit)
2118 {
2119 mathexit = 0;
2120 goto close1;
2121 }
2122 mathexit = 1;
2123
2124 case Sopen:
2125 if (!++depth) goto done;
2126 break;
2127
2128 case Sclose:
2129 close1:
2130 if (!--depth) goto done;
2131 if (depth < min_depth)
7bf5e9e4
RS
2132 Fsignal (Qscan_error,
2133 Fcons (build_string ("Containing expression ends prematurely"),
2134 Fcons (make_number (last_good),
2135 Fcons (make_number (from), Qnil))));
8489eb67
RS
2136 break;
2137
2138 case Sstring:
195d1361 2139 case Sstring_fence:
ef316cf0 2140 temp_pos = dec_bytepos (from_byte);
93da5fff 2141 stringterm = FETCH_CHAR (temp_pos);
8489eb67
RS
2142 while (1)
2143 {
2144 if (from >= stop) goto lose;
195d1361 2145 UPDATE_SYNTAX_TABLE_FORWARD (from);
bd25db08 2146 c = FETCH_CHAR (from_byte);
42ebfa31
SM
2147 if (code == Sstring
2148 ? (c == stringterm
2149 && SYNTAX_WITH_MULTIBYTE_CHECK (c) == Sstring)
bd25db08 2150 : SYNTAX_WITH_MULTIBYTE_CHECK (c) == Sstring_fence)
195d1361 2151 break;
7c7ca3d2
RS
2152
2153 /* Some compilers can't handle this inside the switch. */
bd25db08 2154 temp = SYNTAX_WITH_MULTIBYTE_CHECK (c);
7c7ca3d2 2155 switch (temp)
8489eb67
RS
2156 {
2157 case Scharquote:
2158 case Sescape:
6a140a74 2159 INC_BOTH (from, from_byte);
8489eb67 2160 }
6a140a74 2161 INC_BOTH (from, from_byte);
8489eb67 2162 }
6a140a74 2163 INC_BOTH (from, from_byte);
8489eb67
RS
2164 if (!depth && sexpflag) goto done;
2165 break;
2166 }
2167 }
2168
2169 /* Reached end of buffer. Error if within object, return nil if between */
2170 if (depth) goto lose;
2171
2172 immediate_quit = 0;
2173 return Qnil;
2174
2175 /* End of object reached */
2176 done:
2177 count--;
2178 }
2179
2180 while (count < 0)
2181 {
8489eb67
RS
2182 while (from > stop)
2183 {
6a140a74 2184 DEC_BOTH (from, from_byte);
195d1361 2185 UPDATE_SYNTAX_TABLE_BACKWARD (from);
6a140a74 2186 c = FETCH_CHAR (from_byte);
bd25db08 2187 code = SYNTAX_WITH_MULTIBYTE_CHECK (c);
7bf5e9e4
RS
2188 if (depth == min_depth)
2189 last_good = from;
7fc8191e 2190 comstyle = 0;
95ff8dfc 2191 comnested = SYNTAX_COMMENT_NESTED (c);
7fc8191e
RS
2192 if (code == Sendcomment)
2193 comstyle = SYNTAX_COMMENT_STYLE (c);
8489eb67 2194 if (from > stop && SYNTAX_COMEND_SECOND (c)
1674d9a2 2195 && prev_char_comend_first (from, from_byte)
8489eb67 2196 && parse_sexp_ignore_comments)
e5d4f4dc 2197 {
f902a008 2198 /* We must record the comment style encountered so that
e5d4f4dc 2199 later, we can match only the proper comment begin
f902a008 2200 sequence of the same style. */
6a140a74 2201 DEC_BOTH (from, from_byte);
f902a008
RS
2202 UPDATE_SYNTAX_TABLE_BACKWARD (from);
2203 code = Sendcomment;
95ff8dfc
RS
2204 c1 = FETCH_CHAR (from_byte);
2205 comstyle = SYNTAX_COMMENT_STYLE (c1);
2206 comnested = comnested || SYNTAX_COMMENT_NESTED (c1);
e5d4f4dc
RS
2207 }
2208
9828a477 2209 /* Quoting turns anything except a comment-ender
e6365855 2210 into a word character. Note that this cannot be true
f902a008 2211 if we decremented FROM in the if-statement above. */
9828a477
KH
2212 if (code != Sendcomment && char_quoted (from, from_byte))
2213 code = Sword;
2214 else if (SYNTAX_PREFIX (c))
8489eb67
RS
2215 continue;
2216
9828a477 2217 switch (SWITCH_ENUM_CAST (code))
8489eb67
RS
2218 {
2219 case Sword:
2220 case Ssymbol:
9828a477
KH
2221 case Sescape:
2222 case Scharquote:
8489eb67 2223 if (depth || !sexpflag) break;
195d1361
RS
2224 /* This word counts as a sexp; count object finished
2225 after passing it. */
8489eb67
RS
2226 while (from > stop)
2227 {
6a140a74 2228 temp_pos = from_byte;
ef316cf0
RS
2229 if (! NILP (current_buffer->enable_multibyte_characters))
2230 DEC_POS (temp_pos);
2231 else
2232 temp_pos--;
6a140a74 2233 UPDATE_SYNTAX_TABLE_BACKWARD (from - 1);
9828a477 2234 c1 = FETCH_CHAR (temp_pos);
bd25db08 2235 temp_code = SYNTAX_WITH_MULTIBYTE_CHECK (c1);
9828a477
KH
2236 /* Don't allow comment-end to be quoted. */
2237 if (temp_code == Sendcomment)
2238 goto done2;
6a140a74 2239 quoted = char_quoted (from - 1, temp_pos);
8489eb67 2240 if (quoted)
93da5fff 2241 {
6a140a74 2242 DEC_BOTH (from, from_byte);
ef316cf0 2243 temp_pos = dec_bytepos (temp_pos);
6a140a74 2244 UPDATE_SYNTAX_TABLE_BACKWARD (from - 1);
93da5fff
KH
2245 }
2246 c1 = FETCH_CHAR (temp_pos);
bd25db08 2247 temp_code = SYNTAX_WITH_MULTIBYTE_CHECK (c1);
93da5fff
KH
2248 if (! (quoted || temp_code == Sword
2249 || temp_code == Ssymbol
2250 || temp_code == Squote))
8489eb67 2251 goto done2;
6a140a74 2252 DEC_BOTH (from, from_byte);
8489eb67
RS
2253 }
2254 goto done2;
2255
2256 case Smath:
2257 if (!sexpflag)
2258 break;
ef316cf0 2259 temp_pos = dec_bytepos (from_byte);
6a140a74 2260 UPDATE_SYNTAX_TABLE_BACKWARD (from - 1);
93da5fff 2261 if (from != stop && c == FETCH_CHAR (temp_pos))
6a140a74 2262 DEC_BOTH (from, from_byte);
8489eb67
RS
2263 if (mathexit)
2264 {
2265 mathexit = 0;
2266 goto open2;
2267 }
2268 mathexit = 1;
2269
2270 case Sclose:
2271 if (!++depth) goto done2;
2272 break;
2273
2274 case Sopen:
2275 open2:
2276 if (!--depth) goto done2;
2277 if (depth < min_depth)
7bf5e9e4
RS
2278 Fsignal (Qscan_error,
2279 Fcons (build_string ("Containing expression ends prematurely"),
2280 Fcons (make_number (last_good),
2281 Fcons (make_number (from), Qnil))));
8489eb67
RS
2282 break;
2283
2284 case Sendcomment:
2285 if (!parse_sexp_ignore_comments)
2286 break;
95ff8dfc 2287 found = back_comment (from, from_byte, stop, comnested, comstyle,
6a140a74 2288 &out_charpos, &out_bytepos);
1aa963c8
SM
2289 /* FIXME: if found == -1, then it really wasn't a comment-end.
2290 For single-char Sendcomment, we can't do much about it apart
2291 from skipping the char.
2292 For 2-char endcomments, we could try again, taking both
2293 chars as separate entities, but it's a lot of trouble
2294 for very little gain, so we don't bother either. -sm */
6a140a74
RS
2295 if (found != -1)
2296 from = out_charpos, from_byte = out_bytepos;
8489eb67
RS
2297 break;
2298
195d1361
RS
2299 case Scomment_fence:
2300 case Sstring_fence:
2301 while (1)
2302 {
6a140a74 2303 DEC_BOTH (from, from_byte);
195d1361
RS
2304 if (from == stop) goto lose;
2305 UPDATE_SYNTAX_TABLE_BACKWARD (from);
6a140a74 2306 if (!char_quoted (from, from_byte)
bd25db08
KH
2307 && (c = FETCH_CHAR (from_byte),
2308 SYNTAX_WITH_MULTIBYTE_CHECK (c) == code))
195d1361
RS
2309 break;
2310 }
2311 if (code == Sstring_fence && !depth && sexpflag) goto done2;
2312 break;
2313
8489eb67 2314 case Sstring:
6a140a74 2315 stringterm = FETCH_CHAR (from_byte);
8489eb67
RS
2316 while (1)
2317 {
2318 if (from == stop) goto lose;
6a140a74 2319 temp_pos = from_byte;
ef316cf0
RS
2320 if (! NILP (current_buffer->enable_multibyte_characters))
2321 DEC_POS (temp_pos);
2322 else
2323 temp_pos--;
6a140a74
RS
2324 UPDATE_SYNTAX_TABLE_BACKWARD (from - 1);
2325 if (!char_quoted (from - 1, temp_pos)
42ebfa31
SM
2326 && stringterm == (c = FETCH_CHAR (temp_pos))
2327 && SYNTAX_WITH_MULTIBYTE_CHECK (c) == Sstring)
8489eb67 2328 break;
6a140a74 2329 DEC_BOTH (from, from_byte);
8489eb67 2330 }
6a140a74 2331 DEC_BOTH (from, from_byte);
8489eb67
RS
2332 if (!depth && sexpflag) goto done2;
2333 break;
2334 }
2335 }
2336
2337 /* Reached start of buffer. Error if within object, return nil if between */
2338 if (depth) goto lose;
2339
2340 immediate_quit = 0;
2341 return Qnil;
2342
2343 done2:
2344 count++;
2345 }
2346
2347
2348 immediate_quit = 0;
1e142fb7 2349 XSETFASTINT (val, from);
8489eb67
RS
2350 return val;
2351
2352 lose:
7bf5e9e4
RS
2353 Fsignal (Qscan_error,
2354 Fcons (build_string ("Unbalanced parentheses"),
2355 Fcons (make_number (last_good),
2356 Fcons (make_number (from), Qnil))));
2357
8489eb67
RS
2358 /* NOTREACHED */
2359}
2360
8489eb67 2361DEFUN ("scan-lists", Fscan_lists, Sscan_lists, 3, 3, 0,
fdb82f93
PJ
2362 doc: /* Scan from character number FROM by COUNT lists.
2363Returns the character number of the position thus found.
2364
2365If DEPTH is nonzero, paren depth begins counting from that value,
2366only places where the depth in parentheses becomes zero
2367are candidates for stopping; COUNT such places are counted.
2368Thus, a positive value for DEPTH means go out levels.
2369
2370Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2371
2372If the beginning or end of (the accessible part of) the buffer is reached
2373and the depth is wrong, an error is signaled.
2374If the depth is right but the count is not used up, nil is returned. */)
2375 (from, count, depth)
8489eb67
RS
2376 Lisp_Object from, count, depth;
2377{
b7826503
PJ
2378 CHECK_NUMBER (from);
2379 CHECK_NUMBER (count);
2380 CHECK_NUMBER (depth);
8489eb67
RS
2381
2382 return scan_lists (XINT (from), XINT (count), XINT (depth), 0);
2383}
2384
2385DEFUN ("scan-sexps", Fscan_sexps, Sscan_sexps, 2, 2, 0,
fdb82f93
PJ
2386 doc: /* Scan from character number FROM by COUNT balanced expressions.
2387If COUNT is negative, scan backwards.
2388Returns the character number of the position thus found.
2389
2390Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2391
2392If the beginning or end of (the accessible part of) the buffer is reached
2393in the middle of a parenthetical grouping, an error is signaled.
2394If the beginning or end is reached between groupings
2395but before count is used up, nil is returned. */)
2396 (from, count)
8489eb67
RS
2397 Lisp_Object from, count;
2398{
b7826503
PJ
2399 CHECK_NUMBER (from);
2400 CHECK_NUMBER (count);
8489eb67
RS
2401
2402 return scan_lists (XINT (from), XINT (count), 0, 1);
2403}
2404
2405DEFUN ("backward-prefix-chars", Fbackward_prefix_chars, Sbackward_prefix_chars,
fdb82f93
PJ
2406 0, 0, 0,
2407 doc: /* Move point backward over any number of chars with prefix syntax.
2408This includes chars with "quote" or "prefix" syntax (' or p). */)
2409 ()
8489eb67
RS
2410{
2411 int beg = BEGV;
6a140a74
RS
2412 int opoint = PT;
2413 int opoint_byte = PT_BYTE;
6ec8bbd2 2414 int pos = PT;
6a140a74 2415 int pos_byte = PT_BYTE;
93da5fff 2416 int c;
93da5fff 2417
f902a008 2418 if (pos <= beg)
195d1361 2419 {
f902a008
RS
2420 SET_PT_BOTH (opoint, opoint_byte);
2421
2422 return Qnil;
195d1361 2423 }
8489eb67 2424
f902a008
RS
2425 SETUP_SYNTAX_TABLE (pos, -1);
2426
6a140a74
RS
2427 DEC_BOTH (pos, pos_byte);
2428
1fd3172d 2429 while (!char_quoted (pos, pos_byte)
195d1361 2430 /* Previous statement updates syntax table. */
6a140a74 2431 && ((c = FETCH_CHAR (pos_byte), SYNTAX (c) == Squote)
93da5fff
KH
2432 || SYNTAX_PREFIX (c)))
2433 {
1fd3172d
RS
2434 opoint = pos;
2435 opoint_byte = pos_byte;
2436
2437 if (pos + 1 > beg)
2438 DEC_BOTH (pos, pos_byte);
93da5fff 2439 }
8489eb67 2440
6a140a74 2441 SET_PT_BOTH (opoint, opoint_byte);
8489eb67
RS
2442
2443 return Qnil;
2444}
2445\f
6a140a74 2446/* Parse forward from FROM / FROM_BYTE to END,
e5d4f4dc
RS
2447 assuming that FROM has state OLDSTATE (nil means FROM is start of function),
2448 and return a description of the state of the parse at END.
c81a3712 2449 If STOPBEFORE is nonzero, stop at the start of an atom.
644ea4df
RS
2450 If COMMENTSTOP is 1, stop at the start of a comment.
2451 If COMMENTSTOP is -1, stop at the start or end of a comment,
2452 after the beginning of a string, or after the end of a string. */
8489eb67 2453
340f92b5 2454static void
6a140a74 2455scan_sexps_forward (stateptr, from, from_byte, end, targetdepth,
c81a3712 2456 stopbefore, oldstate, commentstop)
e5d4f4dc 2457 struct lisp_parse_state *stateptr;
8489eb67
RS
2458 register int from;
2459 int end, targetdepth, stopbefore;
2460 Lisp_Object oldstate;
c81a3712 2461 int commentstop;
8489eb67
RS
2462{
2463 struct lisp_parse_state state;
2464
2465 register enum syntaxcode code;
95ff8dfc
RS
2466 int c1;
2467 int comnested;
8489eb67
RS
2468 struct level { int last, prev; };
2469 struct level levelstart[100];
2470 register struct level *curlevel = levelstart;
2471 struct level *endlevel = levelstart + 100;
8489eb67
RS
2472 register int depth; /* Paren depth of current scanning location.
2473 level - levelstart equals this except
2474 when the depth becomes negative. */
2475 int mindepth; /* Lowest DEPTH value seen. */
2476 int start_quoted = 0; /* Nonzero means starting after a char quote */
2477 Lisp_Object tem;
93da5fff 2478 int prev_from; /* Keep one character before FROM. */
6a140a74 2479 int prev_from_byte;
1fd3172d 2480 int prev_from_syntax;
195d1361
RS
2481 int boundary_stop = commentstop == -1;
2482 int nofence;
95ff8dfc
RS
2483 int found;
2484 int out_bytepos, out_charpos;
7c7ca3d2 2485 int temp;
93da5fff
KH
2486
2487 prev_from = from;
6a140a74
RS
2488 prev_from_byte = from_byte;
2489 if (from != BEGV)
2490 DEC_BOTH (prev_from, prev_from_byte);
93da5fff
KH
2491
2492 /* Use this macro instead of `from++'. */
6a140a74
RS
2493#define INC_FROM \
2494do { prev_from = from; \
2495 prev_from_byte = from_byte; \
1fd3172d
RS
2496 prev_from_syntax \
2497 = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte)); \
ef316cf0 2498 INC_BOTH (from, from_byte); \
1fd3172d 2499 UPDATE_SYNTAX_TABLE_FORWARD (from); \
6a140a74 2500 } while (0)
8489eb67
RS
2501
2502 immediate_quit = 1;
2503 QUIT;
2504
265a9e55 2505 if (NILP (oldstate))
8489eb67
RS
2506 {
2507 depth = 0;
2508 state.instring = -1;
2509 state.incomment = 0;
195d1361
RS
2510 state.comstyle = 0; /* comment style a by default. */
2511 state.comstr_start = -1; /* no comment/string seen. */
8489eb67
RS
2512 }
2513 else
2514 {
2515 tem = Fcar (oldstate);
265a9e55 2516 if (!NILP (tem))
8489eb67
RS
2517 depth = XINT (tem);
2518 else
2519 depth = 0;
2520
2521 oldstate = Fcdr (oldstate);
2522 oldstate = Fcdr (oldstate);
2523 oldstate = Fcdr (oldstate);
2524 tem = Fcar (oldstate);
195d1361 2525 /* Check whether we are inside string_fence-style string: */
e76f1c44
GM
2526 state.instring = (!NILP (tem)
2527 ? (INTEGERP (tem) ? XINT (tem) : ST_STRING_STYLE)
2528 : -1);
8489eb67
RS
2529
2530 oldstate = Fcdr (oldstate);
2531 tem = Fcar (oldstate);
e76f1c44
GM
2532 state.incomment = (!NILP (tem)
2533 ? (INTEGERP (tem) ? XINT (tem) : -1)
95ff8dfc 2534 : 0);
8489eb67
RS
2535
2536 oldstate = Fcdr (oldstate);
2537 tem = Fcar (oldstate);
265a9e55 2538 start_quoted = !NILP (tem);
e5d4f4dc 2539
95ff8dfc 2540 /* if the eighth element of the list is nil, we are in comment
195d1361
RS
2541 style a. If it is non-nil, we are in comment style b */
2542 oldstate = Fcdr (oldstate);
e5d4f4dc 2543 oldstate = Fcdr (oldstate);
195d1361 2544 tem = Fcar (oldstate);
e76f1c44
GM
2545 state.comstyle = NILP (tem) ? 0 : (EQ (tem, Qsyntax_table)
2546 ? ST_COMMENT_STYLE : 1);
195d1361 2547
e5d4f4dc 2548 oldstate = Fcdr (oldstate);
e5d4f4dc 2549 tem = Fcar (oldstate);
195d1361 2550 state.comstr_start = NILP (tem) ? -1 : XINT (tem) ;
1a102a58
RS
2551 oldstate = Fcdr (oldstate);
2552 tem = Fcar (oldstate);
2553 while (!NILP (tem)) /* >= second enclosing sexps. */
2554 {
2555 /* curlevel++->last ran into compiler bug on Apollo */
2556 curlevel->last = XINT (Fcar (tem));
2557 if (++curlevel == endlevel)
02010917 2558 curlevel--; /* error ("Nesting too deep for parser"); */
1a102a58
RS
2559 curlevel->prev = -1;
2560 curlevel->last = -1;
2561 tem = Fcdr (tem);
2562 }
8489eb67
RS
2563 }
2564 state.quoted = 0;
2565 mindepth = depth;
2566
2567 curlevel->prev = -1;
2568 curlevel->last = -1;
2569
326b283b
RS
2570 SETUP_SYNTAX_TABLE (prev_from, 1);
2571 prev_from_syntax = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte));
2572 UPDATE_SYNTAX_TABLE_FORWARD (from);
2573
195d1361 2574 /* Enter the loop at a place appropriate for initial state. */
8489eb67 2575
326b283b
RS
2576 if (state.incomment)
2577 goto startincomment;
8489eb67
RS
2578 if (state.instring >= 0)
2579 {
195d1361 2580 nofence = state.instring != ST_STRING_STYLE;
326b283b
RS
2581 if (start_quoted)
2582 goto startquotedinstring;
8489eb67
RS
2583 goto startinstring;
2584 }
326b283b
RS
2585 else if (start_quoted)
2586 goto startquoted;
1fd3172d 2587
ccf89641 2588#if 0 /* This seems to be redundant with the identical code above. */
1fd3172d
RS
2589 SETUP_SYNTAX_TABLE (prev_from, 1);
2590 prev_from_syntax = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte));
2591 UPDATE_SYNTAX_TABLE_FORWARD (from);
ccf89641 2592#endif
1fd3172d 2593
8489eb67
RS
2594 while (from < end)
2595 {
93da5fff 2596 INC_FROM;
1fd3172d 2597 code = prev_from_syntax & 0xff;
4c920633 2598
8f9dc2ed 2599 if (code == Scomment)
95ff8dfc 2600 {
d355bd8a 2601 state.comstyle = SYNTAX_FLAGS_COMMENT_STYLE (prev_from_syntax);
95ff8dfc
RS
2602 state.incomment = (SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax) ?
2603 1 : -1);
2604 state.comstr_start = prev_from;
2605 }
4c920633 2606 else if (code == Scomment_fence)
e5d4f4dc
RS
2607 {
2608 /* Record the comment style we have entered so that only
2609 the comment-end sequence of the same style actually
2610 terminates the comment section. */
95ff8dfc
RS
2611 state.comstyle = ST_COMMENT_STYLE;
2612 state.incomment = -1;
195d1361 2613 state.comstr_start = prev_from;
e5d4f4dc 2614 code = Scomment;
e5d4f4dc 2615 }
4c920633 2616 else if (from < end)
1fd3172d 2617 if (SYNTAX_FLAGS_COMSTART_FIRST (prev_from_syntax))
95ff8dfc
RS
2618 if (c1 = FETCH_CHAR (from_byte),
2619 SYNTAX_COMSTART_SECOND (c1))
1fd3172d 2620 /* Duplicate code to avoid a complex if-expression
4c920633
RS
2621 which causes trouble for the SGI compiler. */
2622 {
2623 /* Record the comment style we have entered so that only
2624 the comment-end sequence of the same style actually
2625 terminates the comment section. */
95ff8dfc
RS
2626 state.comstyle = SYNTAX_COMMENT_STYLE (FETCH_CHAR (from_byte));
2627 comnested = SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax);
2628 comnested = comnested || SYNTAX_COMMENT_NESTED (c1);
2629 state.incomment = comnested ? 1 : -1;
4c920633 2630 state.comstr_start = prev_from;
95ff8dfc 2631 INC_FROM;
4c920633
RS
2632 code = Scomment;
2633 }
e5d4f4dc 2634
1fd3172d 2635 if (SYNTAX_FLAGS_PREFIX (prev_from_syntax))
8489eb67 2636 continue;
0220c518 2637 switch (SWITCH_ENUM_CAST (code))
8489eb67
RS
2638 {
2639 case Sescape:
2640 case Scharquote:
2641 if (stopbefore) goto stop; /* this arg means stop at sexp start */
93da5fff 2642 curlevel->last = prev_from;
8489eb67
RS
2643 startquoted:
2644 if (from == end) goto endquoted;
93da5fff 2645 INC_FROM;
8489eb67
RS
2646 goto symstarted;
2647 /* treat following character as a word constituent */
2648 case Sword:
2649 case Ssymbol:
2650 if (stopbefore) goto stop; /* this arg means stop at sexp start */
93da5fff 2651 curlevel->last = prev_from;
8489eb67
RS
2652 symstarted:
2653 while (from < end)
2654 {
7c7ca3d2
RS
2655 /* Some compilers can't handle this inside the switch. */
2656 temp = SYNTAX (FETCH_CHAR (from_byte));
2657 switch (temp)
8489eb67
RS
2658 {
2659 case Scharquote:
2660 case Sescape:
93da5fff 2661 INC_FROM;
8489eb67
RS
2662 if (from == end) goto endquoted;
2663 break;
2664 case Sword:
2665 case Ssymbol:
2666 case Squote:
2667 break;
2668 default:
2669 goto symdone;
2670 }
93da5fff 2671 INC_FROM;
8489eb67
RS
2672 }
2673 symdone:
2674 curlevel->prev = curlevel->last;
2675 break;
2676
2677 case Scomment:
195d1361 2678 if (commentstop || boundary_stop) goto done;
d355bd8a
SM
2679 startincomment:
2680 /* The (from == BEGV) test was to enter the loop in the middle so
95ff8dfc 2681 that we find a 2-char comment ender even if we start in the
e6365855 2682 middle of it. We don't want to do that if we're just at the
d355bd8a 2683 beginning of the comment (think of (*) ... (*)). */
95ff8dfc
RS
2684 found = forw_comment (from, from_byte, end,
2685 state.incomment, state.comstyle,
e6365855
SM
2686 (from == BEGV || from < state.comstr_start + 3)
2687 ? 0 : prev_from_syntax,
95ff8dfc
RS
2688 &out_charpos, &out_bytepos, &state.incomment);
2689 from = out_charpos; from_byte = out_bytepos;
2690 /* Beware! prev_from and friends are invalid now.
2691 Luckily, the `done' doesn't use them and the INC_FROM
2692 sets them to a sane value without looking at them. */
2693 if (!found) goto done;
d355bd8a 2694 INC_FROM;
8489eb67 2695 state.incomment = 0;
e5d4f4dc 2696 state.comstyle = 0; /* reset the comment style */
195d1361 2697 if (boundary_stop) goto done;
8489eb67
RS
2698 break;
2699
2700 case Sopen:
2701 if (stopbefore) goto stop; /* this arg means stop at sexp start */
2702 depth++;
2703 /* curlevel++->last ran into compiler bug on Apollo */
93da5fff 2704 curlevel->last = prev_from;
8489eb67 2705 if (++curlevel == endlevel)
02010917 2706 curlevel--; /* error ("Nesting too deep for parser"); */
8489eb67
RS
2707 curlevel->prev = -1;
2708 curlevel->last = -1;
30844415 2709 if (targetdepth == depth) goto done;
8489eb67
RS
2710 break;
2711
2712 case Sclose:
2713 depth--;
2714 if (depth < mindepth)
2715 mindepth = depth;
2716 if (curlevel != levelstart)
2717 curlevel--;
2718 curlevel->prev = curlevel->last;
30844415 2719 if (targetdepth == depth) goto done;
8489eb67
RS
2720 break;
2721
2722 case Sstring:
195d1361
RS
2723 case Sstring_fence:
2724 state.comstr_start = from - 1;
8489eb67 2725 if (stopbefore) goto stop; /* this arg means stop at sexp start */
93da5fff 2726 curlevel->last = prev_from;
195d1361 2727 state.instring = (code == Sstring
6a140a74 2728 ? (FETCH_CHAR (prev_from_byte))
195d1361
RS
2729 : ST_STRING_STYLE);
2730 if (boundary_stop) goto done;
8489eb67 2731 startinstring:
195d1361 2732 {
644ea4df 2733 nofence = state.instring != ST_STRING_STYLE;
195d1361 2734
644ea4df
RS
2735 while (1)
2736 {
2737 int c;
195d1361 2738
644ea4df 2739 if (from >= end) goto done;
6a140a74 2740 c = FETCH_CHAR (from_byte);
7c7ca3d2
RS
2741 /* Some compilers can't handle this inside the switch. */
2742 temp = SYNTAX (c);
ccf89641
KH
2743
2744 /* Check TEMP here so that if the char has
2745 a syntax-table property which says it is NOT
2746 a string character, it does not end the string. */
2747 if (nofence && c == state.instring && temp == Sstring)
2748 break;
2749
7c7ca3d2 2750 switch (temp)
644ea4df
RS
2751 {
2752 case Sstring_fence:
2753 if (!nofence) goto string_end;
2754 break;
2755 case Scharquote:
2756 case Sescape:
2757 INC_FROM;
2758 startquotedinstring:
2759 if (from >= end) goto endquoted;
195d1361 2760 }
644ea4df
RS
2761 INC_FROM;
2762 }
195d1361
RS
2763 }
2764 string_end:
8489eb67
RS
2765 state.instring = -1;
2766 curlevel->prev = curlevel->last;
93da5fff 2767 INC_FROM;
195d1361 2768 if (boundary_stop) goto done;
8489eb67
RS
2769 break;
2770
2771 case Smath:
2772 break;
2773 }
2774 }
2775 goto done;
2776
2777 stop: /* Here if stopping before start of sexp. */
93da5fff 2778 from = prev_from; /* We have just fetched the char that starts it; */
8489eb67
RS
2779 goto done; /* but return the position before it. */
2780
2781 endquoted:
2782 state.quoted = 1;
2783 done:
2784 state.depth = depth;
2785 state.mindepth = mindepth;
2786 state.thislevelstart = curlevel->prev;
2787 state.prevlevelstart
2788 = (curlevel == levelstart) ? -1 : (curlevel - 1)->last;
2789 state.location = from;
1a102a58
RS
2790 state.levelstarts = Qnil;
2791 while (--curlevel >= levelstart)
2792 state.levelstarts = Fcons (make_number (curlevel->last),
2793 state.levelstarts);
8489eb67
RS
2794 immediate_quit = 0;
2795
e5d4f4dc 2796 *stateptr = state;
8489eb67
RS
2797}
2798
c81a3712 2799DEFUN ("parse-partial-sexp", Fparse_partial_sexp, Sparse_partial_sexp, 2, 6, 0,
fdb82f93
PJ
2800 doc: /* Parse Lisp syntax starting at FROM until TO; return status of parse at TO.
2801Parsing stops at TO or when certain criteria are met;
2802 point is set to where parsing stops.
2803If fifth arg OLDSTATE is omitted or nil,
2804 parsing assumes that FROM is the beginning of a function.
2805Value is a list of ten elements describing final state of parsing:
2806 0. depth in parens.
2807 1. character address of start of innermost containing list; nil if none.
2808 2. character address of start of last complete sexp terminated.
2809 3. non-nil if inside a string.
2810 (it is the character that will terminate the string,
2811 or t if the string should be terminated by a generic string delimiter.)
2812 4. nil if outside a comment, t if inside a non-nestable comment,
2813 else an integer (the current comment nesting).
2814 5. t if following a quote character.
2815 6. the minimum paren-depth encountered during this scan.
2816 7. t if in a comment of style b; symbol `syntax-table' if the comment
2817 should be terminated by a generic comment delimiter.
2818 8. character address of start of comment or string; nil if not in one.
2819 9. Intermediate data for continuation of parsing (subject to change).
2820If third arg TARGETDEPTH is non-nil, parsing stops if the depth
2821in parentheses becomes equal to TARGETDEPTH.
2822Fourth arg STOPBEFORE non-nil means stop when come to
2823 any character that starts a sexp.
2824Fifth arg OLDSTATE is a nine-element list like what this function returns.
2825 It is used to initialize the state of the parse. Elements number 1, 2, 6
2826 and 8 are ignored; you can leave off element 8 (the last) entirely.
2827Sixth arg COMMENTSTOP non-nil means stop at the start of a comment.
2828 If it is symbol `syntax-table', stop after the start of a comment or a
2829 string, or after end of a comment or a string. */)
2830 (from, to, targetdepth, stopbefore, oldstate, commentstop)
c81a3712 2831 Lisp_Object from, to, targetdepth, stopbefore, oldstate, commentstop;
8489eb67
RS
2832{
2833 struct lisp_parse_state state;
2834 int target;
2835
265a9e55 2836 if (!NILP (targetdepth))
8489eb67 2837 {
b7826503 2838 CHECK_NUMBER (targetdepth);
8489eb67
RS
2839 target = XINT (targetdepth);
2840 }
2841 else
2842 target = -100000; /* We won't reach this depth */
2843
2844 validate_region (&from, &to);
6a140a74
RS
2845 scan_sexps_forward (&state, XINT (from), CHAR_TO_BYTE (XINT (from)),
2846 XINT (to),
c81a3712 2847 target, !NILP (stopbefore), oldstate,
195d1361
RS
2848 (NILP (commentstop)
2849 ? 0 : (EQ (commentstop, Qsyntax_table) ? -1 : 1)));
8489eb67
RS
2850
2851 SET_PT (state.location);
2852
2853 return Fcons (make_number (state.depth),
2854 Fcons (state.prevlevelstart < 0 ? Qnil : make_number (state.prevlevelstart),
2855 Fcons (state.thislevelstart < 0 ? Qnil : make_number (state.thislevelstart),
195d1361
RS
2856 Fcons (state.instring >= 0
2857 ? (state.instring == ST_STRING_STYLE
2858 ? Qt : make_number (state.instring)) : Qnil,
95ff8dfc
RS
2859 Fcons (state.incomment < 0 ? Qt :
2860 (state.incomment == 0 ? Qnil :
2861 make_number (state.incomment)),
8489eb67 2862 Fcons (state.quoted ? Qt : Qnil,
bff37d2a
RS
2863 Fcons (make_number (state.mindepth),
2864 Fcons ((state.comstyle
2865 ? (state.comstyle == ST_COMMENT_STYLE
2866 ? Qsyntax_table : Qt) :
2867 Qnil),
0beaf54f
DL
2868 Fcons (((state.incomment
2869 || (state.instring >= 0))
bff37d2a
RS
2870 ? make_number (state.comstr_start)
2871 : Qnil),
1a102a58 2872 Fcons (state.levelstarts, Qnil))))))))));
8489eb67
RS
2873}
2874\f
dfcf069d 2875void
8489eb67
RS
2876init_syntax_once ()
2877{
78f9a1f7 2878 register int i, c;
8ea151b2 2879 Lisp_Object temp;
8489eb67 2880
5ebaddf5
RS
2881 /* This has to be done here, before we call Fmake_char_table. */
2882 Qsyntax_table = intern ("syntax-table");
2883 staticpro (&Qsyntax_table);
2884
2885 /* Intern this now in case it isn't already done.
2886 Setting this variable twice is harmless.
2887 But don't staticpro it here--that is done in alloc.c. */
2888 Qchar_table_extra_slots = intern ("char-table-extra-slots");
2889
93da5fff 2890 /* Create objects which can be shared among syntax tables. */
42ebfa31 2891 Vsyntax_code_object = Fmake_vector (make_number (Smax), Qnil);
93da5fff
KH
2892 for (i = 0; i < XVECTOR (Vsyntax_code_object)->size; i++)
2893 XVECTOR (Vsyntax_code_object)->contents[i]
2894 = Fcons (make_number (i), Qnil);
2895
5ebaddf5
RS
2896 /* Now we are ready to set up this property, so we can
2897 create syntax tables. */
2898 Fput (Qsyntax_table, Qchar_table_extra_slots, make_number (0));
2899
93da5fff 2900 temp = XVECTOR (Vsyntax_code_object)->contents[(int) Swhitespace];
8489eb67 2901
5ebaddf5 2902 Vstandard_syntax_table = Fmake_char_table (Qsyntax_table, temp);
8489eb67 2903
93da5fff 2904 temp = XVECTOR (Vsyntax_code_object)->contents[(int) Sword];
8489eb67 2905 for (i = 'a'; i <= 'z'; i++)
8ea151b2 2906 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp);
8489eb67 2907 for (i = 'A'; i <= 'Z'; i++)
8ea151b2 2908 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp);
8489eb67 2909 for (i = '0'; i <= '9'; i++)
8ea151b2
RS
2910 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp);
2911
2912 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '$', temp);
2913 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '%', temp);
2914
2915 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '(',
2916 Fcons (make_number (Sopen), make_number (')')));
2917 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, ')',
2918 Fcons (make_number (Sclose), make_number ('(')));
2919 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '[',
2920 Fcons (make_number (Sopen), make_number (']')));
2921 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, ']',
2922 Fcons (make_number (Sclose), make_number ('[')));
2923 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '{',
2924 Fcons (make_number (Sopen), make_number ('}')));
2925 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '}',
2926 Fcons (make_number (Sclose), make_number ('{')));
2927 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '"',
2928 Fcons (make_number ((int) Sstring), Qnil));
2929 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '\\',
2930 Fcons (make_number ((int) Sescape), Qnil));
2931
93da5fff 2932 temp = XVECTOR (Vsyntax_code_object)->contents[(int) Ssymbol];
8489eb67 2933 for (i = 0; i < 10; i++)
78f9a1f7
KH
2934 {
2935 c = "_-+*/&|<>="[i];
2936 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, c, temp);
2937 }
8489eb67 2938
93da5fff 2939 temp = XVECTOR (Vsyntax_code_object)->contents[(int) Spunct];
8489eb67 2940 for (i = 0; i < 12; i++)
78f9a1f7
KH
2941 {
2942 c = ".,;:?!#@~^'`"[i];
2943 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, c, temp);
2944 }
bd25db08
KH
2945
2946 /* All multibyte characters have syntax `word' by default. */
2947 temp = XVECTOR (Vsyntax_code_object)->contents[(int) Sword];
2948 for (i = CHAR_TABLE_SINGLE_BYTE_SLOTS; i < CHAR_TABLE_ORDINARY_SLOTS; i++)
2949 XCHAR_TABLE (Vstandard_syntax_table)->contents[i] = temp;
8489eb67
RS
2950}
2951
dfcf069d 2952void
8489eb67
RS
2953syms_of_syntax ()
2954{
2955 Qsyntax_table_p = intern ("syntax-table-p");
2956 staticpro (&Qsyntax_table_p);
2957
93da5fff
KH
2958 staticpro (&Vsyntax_code_object);
2959
7bf5e9e4
RS
2960 Qscan_error = intern ("scan-error");
2961 staticpro (&Qscan_error);
2962 Fput (Qscan_error, Qerror_conditions,
f3be100f 2963 Fcons (Qscan_error, Fcons (Qerror, Qnil)));
7bf5e9e4
RS
2964 Fput (Qscan_error, Qerror_message,
2965 build_string ("Scan error"));
2966
8489eb67 2967 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments,
fdb82f93 2968 doc: /* Non-nil means `forward-sexp', etc., should treat comments as whitespace. */);
8489eb67 2969
195d1361 2970 DEFVAR_BOOL ("parse-sexp-lookup-properties", &parse_sexp_lookup_properties,
fdb82f93
PJ
2971 doc: /* Non-nil means `forward-sexp', etc., obey `syntax-table' property.
2972Otherwise, that text property is simply ignored.
2973See the info node `(elisp)Syntax Properties' for a description of the
2974`syntax-table' property. */);
195d1361 2975
8489eb67
RS
2976 words_include_escapes = 0;
2977 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes,
fdb82f93 2978 doc: /* Non-nil means `forward-word', etc., should treat escape chars part of words. */);
8489eb67 2979
bd25db08 2980 DEFVAR_BOOL ("multibyte-syntax-as-symbol", &multibyte_syntax_as_symbol,
fdb82f93 2981 doc: /* Non-nil means `scan-sexps' treats all multibyte characters as symbol. */);
bd25db08
KH
2982 multibyte_syntax_as_symbol = 0;
2983
f4ed767f
GM
2984 DEFVAR_BOOL ("open-paren-in-column-0-is-defun-start",
2985 &open_paren_in_column_0_is_defun_start,
fdb82f93 2986 doc: /* Non-nil means an open paren in column 0 denotes the start of a defun. */);
f4ed767f
GM
2987 open_paren_in_column_0_is_defun_start = 1;
2988
8489eb67
RS
2989 defsubr (&Ssyntax_table_p);
2990 defsubr (&Ssyntax_table);
2991 defsubr (&Sstandard_syntax_table);
2992 defsubr (&Scopy_syntax_table);
2993 defsubr (&Sset_syntax_table);
2994 defsubr (&Schar_syntax);
beefa22e 2995 defsubr (&Smatching_paren);
c65adb44 2996 defsubr (&Sstring_to_syntax);
8489eb67 2997 defsubr (&Smodify_syntax_entry);
62abe9cb 2998 defsubr (&Sinternal_describe_syntax_value);
8489eb67
RS
2999
3000 defsubr (&Sforward_word);
3001
195d1361
RS
3002 defsubr (&Sskip_chars_forward);
3003 defsubr (&Sskip_chars_backward);
3004 defsubr (&Sskip_syntax_forward);
3005 defsubr (&Sskip_syntax_backward);
3006
b3cfe0c8 3007 defsubr (&Sforward_comment);
8489eb67
RS
3008 defsubr (&Sscan_lists);
3009 defsubr (&Sscan_sexps);
3010 defsubr (&Sbackward_prefix_chars);
3011 defsubr (&Sparse_partial_sexp);
3012}