1 /* GNU Emacs routines to deal with syntax tables; also word and list parsing.
2 Copyright (C) 1985, 87, 93, 94, 95, 97, 1998, 1999 Free Software Foundation, Inc.
4 This file is part of GNU Emacs.
6 GNU Emacs is free software; you can redistribute it and/or modify
7 it under the terms of the GNU General Public License as published by
8 the Free Software Foundation; either version 2, or (at your option)
11 GNU Emacs is distributed in the hope that it will be useful,
12 but WITHOUT ANY WARRANTY; without even the implied warranty of
13 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
14 GNU General Public License for more details.
16 You should have received a copy of the GNU General Public License
17 along with GNU Emacs; see the file COPYING. If not, write to
18 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 Boston, MA 02111-1307, USA. */
27 #include "character.h"
30 /* Make syntax table lookup grant data in gl_state. */
31 #define SYNTAX_ENTRY_VIA_PROPERTY
34 #include "intervals.h"
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. */
40 #define ST_COMMENT_STYLE (256 + 1)
41 #define ST_STRING_STYLE (256 + 2)
44 Lisp_Object Qsyntax_table_p
, Qsyntax_table
, Qscan_error
;
46 int words_include_escapes
;
47 int parse_sexp_lookup_properties
;
49 /* Nonzero means `scan-sexps' treat all multibyte characters as symbol. */
50 int multibyte_syntax_as_symbol
;
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. */
55 Lisp_Object syntax_temp
;
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. */
61 int open_paren_in_column_0_is_defun_start
;
63 /* This is the internal form of the parse state used in parse-partial-sexp. */
65 struct lisp_parse_state
67 int depth
; /* Depth at end of parsing. */
68 int instring
; /* -1 if not within string, else desired terminator. */
69 int incomment
; /* -1 if in unnestable comment else comment nesting */
70 int comstyle
; /* comment style a=0, or b=1, or ST_COMMENT_STYLE. */
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 */
74 int location
; /* Char number at which parsing stopped. */
75 int mindepth
; /* Minimum depth seen while scanning. */
76 int comstr_start
; /* Position just after last comment/string starter. */
77 Lisp_Object levelstarts
; /* Char numbers of starts-of-expression
78 of levels (starting from outermost). */
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.
84 find_start_value_byte is the corresponding byte position.
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. */
89 static int find_start_pos
;
90 static int find_start_value
;
91 static int find_start_value_byte
;
92 static struct buffer
*find_start_buffer
;
93 static int find_start_begv
;
94 static int find_start_modiff
;
97 static int find_defun_start
P_ ((int, int));
98 static int back_comment
P_ ((int, int, int, int, int, int *, int *));
99 static int char_quoted
P_ ((int, int));
100 static Lisp_Object skip_chars
P_ ((int, int, Lisp_Object
, Lisp_Object
));
101 static Lisp_Object scan_lists
P_ ((int, int, int, int));
102 static void scan_sexps_forward
P_ ((struct lisp_parse_state
*,
104 int, Lisp_Object
, int));
107 struct gl_state_s gl_state
; /* Global state of syntax parser. */
109 INTERVAL
interval_of ();
110 #define INTERVALS_AT_ONCE 10 /* 1 + max-number of intervals
111 to scan to property-change. */
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
115 valid interval. If INIT, only [be]_property fields of gl_state are
116 valid at start, the rest is filled basing on OBJECT.
118 `gl_state.*_i' are the intervals, and CHARPOS is further in the search
119 direction than the intervals - or in an interval. We update the
120 current syntax-table basing on the property of this interval, and
121 update the interval to start further than CHARPOS - or be
122 NULL_INTERVAL. We also update lim_property to be the next value of
123 charpos to call this subroutine again - or be before/after the
124 start/end of OBJECT. */
127 update_syntax_table (charpos
, count
, init
, object
)
128 int charpos
, count
, init
;
131 Lisp_Object tmp_table
;
132 int cnt
= 0, invalidate
= 1;
137 gl_state
.old_prop
= Qnil
;
138 gl_state
.start
= gl_state
.b_property
;
139 gl_state
.stop
= gl_state
.e_property
;
140 i
= interval_of (charpos
, object
);
141 gl_state
.backward_i
= gl_state
.forward_i
= i
;
143 if (NULL_INTERVAL_P (i
))
145 /* interval_of updates only ->position of the return value, so
146 update the parents manually to speed up update_interval. */
147 while (!NULL_PARENT (i
))
149 if (AM_RIGHT_CHILD (i
))
150 INTERVAL_PARENT (i
)->position
= i
->position
151 - LEFT_TOTAL_LENGTH (i
) + TOTAL_LENGTH (i
) /* right end */
152 - TOTAL_LENGTH (INTERVAL_PARENT (i
))
153 + LEFT_TOTAL_LENGTH (INTERVAL_PARENT (i
));
155 INTERVAL_PARENT (i
)->position
= i
->position
- LEFT_TOTAL_LENGTH (i
)
157 i
= INTERVAL_PARENT (i
);
159 i
= gl_state
.forward_i
;
160 gl_state
.b_property
= i
->position
- gl_state
.offset
;
161 gl_state
.e_property
= INTERVAL_LAST_POS (i
) - gl_state
.offset
;
164 oldi
= i
= count
> 0 ? gl_state
.forward_i
: gl_state
.backward_i
;
166 /* We are guaranteed to be called with CHARPOS either in i,
168 if (NULL_INTERVAL_P (i
))
169 error ("Error in syntax_table logic for to-the-end intervals");
170 else if (charpos
< i
->position
) /* Move left. */
173 error ("Error in syntax_table logic for intervals <-");
174 /* Update the interval. */
175 i
= update_interval (i
, charpos
);
176 if (INTERVAL_LAST_POS (i
) != gl_state
.b_property
)
179 gl_state
.forward_i
= i
;
180 gl_state
.e_property
= INTERVAL_LAST_POS (i
) - gl_state
.offset
;
183 else if (charpos
>= INTERVAL_LAST_POS (i
)) /* Move right. */
186 error ("Error in syntax_table logic for intervals ->");
187 /* Update the interval. */
188 i
= update_interval (i
, charpos
);
189 if (i
->position
!= gl_state
.e_property
)
192 gl_state
.backward_i
= i
;
193 gl_state
.b_property
= i
->position
- gl_state
.offset
;
198 tmp_table
= textget (i
->plist
, Qsyntax_table
);
201 invalidate
= !EQ (tmp_table
, gl_state
.old_prop
); /* Need to invalidate? */
203 if (invalidate
) /* Did not get to adjacent interval. */
204 { /* with the same table => */
205 /* invalidate the old range. */
208 gl_state
.backward_i
= i
;
209 gl_state
.b_property
= i
->position
- gl_state
.offset
;
213 gl_state
.forward_i
= i
;
214 gl_state
.e_property
= INTERVAL_LAST_POS (i
) - gl_state
.offset
;
218 if (!EQ (tmp_table
, gl_state
.old_prop
))
220 gl_state
.current_syntax_table
= tmp_table
;
221 gl_state
.old_prop
= tmp_table
;
222 if (EQ (Fsyntax_table_p (tmp_table
), Qt
))
224 gl_state
.use_global
= 0;
226 else if (CONSP (tmp_table
))
228 gl_state
.use_global
= 1;
229 gl_state
.global_code
= tmp_table
;
233 gl_state
.use_global
= 0;
234 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
238 while (!NULL_INTERVAL_P (i
))
240 if (cnt
&& !EQ (tmp_table
, textget (i
->plist
, Qsyntax_table
)))
244 gl_state
.e_property
= i
->position
- gl_state
.offset
;
245 gl_state
.forward_i
= i
;
249 gl_state
.b_property
= i
->position
+ LENGTH (i
) - gl_state
.offset
;
250 gl_state
.backward_i
= i
;
254 else if (cnt
== INTERVALS_AT_ONCE
)
258 gl_state
.e_property
= i
->position
+ LENGTH (i
) - gl_state
.offset
;
259 gl_state
.forward_i
= i
;
263 gl_state
.b_property
= i
->position
- gl_state
.offset
;
264 gl_state
.backward_i
= i
;
269 i
= count
> 0 ? next_interval (i
) : previous_interval (i
);
271 eassert (NULL_INTERVAL_P (i
)); /* This property goes to the end. */
273 gl_state
.e_property
= gl_state
.stop
;
275 gl_state
.b_property
= gl_state
.start
;
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. */
283 char_quoted (charpos
, bytepos
)
284 register int charpos
, bytepos
;
286 register enum syntaxcode code
;
287 register int beg
= BEGV
;
288 register int quoted
= 0;
291 DEC_BOTH (charpos
, bytepos
);
293 while (bytepos
>= beg
)
295 UPDATE_SYNTAX_TABLE_BACKWARD (charpos
);
296 code
= SYNTAX (FETCH_CHAR (bytepos
));
297 if (! (code
== Scharquote
|| code
== Sescape
))
300 DEC_BOTH (charpos
, bytepos
);
304 UPDATE_SYNTAX_TABLE (orig
);
308 /* Return the bytepos one character after BYTEPOS.
309 We assume that BYTEPOS is not at the end of the buffer. */
312 inc_bytepos (bytepos
)
315 if (NILP (current_buffer
->enable_multibyte_characters
))
322 /* Return the bytepos one character before BYTEPOS.
323 We assume that BYTEPOS is not at the start of the buffer. */
326 dec_bytepos (bytepos
)
329 if (NILP (current_buffer
->enable_multibyte_characters
))
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.
339 When open_paren_in_column_0_is_defun_start is nonzero,
340 the beginning of every line is treated as a defun-start.
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.
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. */
351 find_defun_start (pos
, pos_byte
)
354 int opoint
= PT
, opoint_byte
= PT_BYTE
;
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
;
367 /* Back up to start of line. */
368 scan_newline (pos
, pos_byte
, BEGV
, BEGV_BYTE
, -1, 1);
370 /* We optimize syntax-table lookup for rare updates. Thus we accept
371 only those `^\s(' which are good in global _and_ text-property
373 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
374 gl_state
.use_global
= 0;
375 if (open_paren_in_column_0_is_defun_start
)
379 /* Open-paren at start of line means we may have found our
381 if (SYNTAX (FETCH_CHAR (PT_BYTE
)) == Sopen
)
383 SETUP_SYNTAX_TABLE (PT
+ 1, -1); /* Try again... */
384 if (SYNTAX (FETCH_CHAR (PT_BYTE
)) == Sopen
)
386 /* Now fallback to the default value. */
387 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
388 gl_state
.use_global
= 0;
390 /* Move to beg of previous line. */
391 scan_newline (PT
, PT_BYTE
, BEGV
, BEGV_BYTE
, -2, 1);
395 /* Record what we found, for the next try. */
396 find_start_value
= PT
;
397 find_start_value_byte
= PT_BYTE
;
398 find_start_buffer
= current_buffer
;
399 find_start_modiff
= MODIFF
;
400 find_start_begv
= BEGV
;
401 find_start_pos
= pos
;
403 TEMP_SET_PT_BOTH (opoint
, opoint_byte
);
405 return find_start_value
;
408 /* Return the SYNTAX_COMEND_FIRST of the character before POS, POS_BYTE. */
411 prev_char_comend_first (pos
, pos_byte
)
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);
424 /* Return the SYNTAX_COMSTART_FIRST of the character before POS, POS_BYTE. */
427 * prev_char_comstart_first (pos, pos_byte)
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);
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.
444 Return a positive value if we find a comment ending at FROM/FROM_BYTE;
447 If successful, store the charpos of the comment's beginning
448 into *CHARPOS_PTR, and the bytepos into *BYTEPOS_PTR.
450 Global syntax data remains valid for backward search starting at
451 the returned value (or at FROM, if the search was not successful). */
454 back_comment (from
, from_byte
, stop
, comnested
, comstyle
, charpos_ptr
, bytepos_ptr
)
455 int from
, from_byte
, stop
;
456 int comnested
, comstyle
;
457 int *charpos_ptr
, *bytepos_ptr
;
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.
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. */
468 int string_style
= -1; /* Presumed outside of any string. */
469 int string_lossage
= 0;
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;
476 int comment_end
= from
;
477 int comment_end_byte
= from_byte
;
478 int comstart_pos
= 0;
480 /* Place where the containing defun starts,
481 or 0 if we didn't come across it yet. */
483 int defun_start_byte
= 0;
484 register enum syntaxcode code
;
485 int nesting
= 1; /* current comment nesting */
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. */
493 /* At beginning of range to scan, we're outside of strings;
494 that determines quote parity to the comment-end. */
497 int temp_byte
, prev_syntax
;
498 int com2start
, com2end
;
500 /* Move back and examine a character. */
501 DEC_BOTH (from
, from_byte
);
502 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
504 prev_syntax
= syntax
;
505 c
= FETCH_CHAR (from_byte
);
506 syntax
= SYNTAX_WITH_FLAGS (c
);
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
));
518 /* Nasty cases with overlapping 2-char comment markers:
519 - snmp-mode: -- c -- foo -- c --
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
))
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
)))
544 /* UPDATE_SYNTAX_TABLE_FORWARD (next + 1); */
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. */
554 /* Turn a 2-char comment sequences into the appropriate syntax. */
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
))
565 /* Ignore escaped characters, except comment-enders. */
566 if (code
!= Sendcomment
&& char_quoted (from
, from_byte
))
573 c
= (code
== Sstring_fence
? ST_STRING_STYLE
: ST_COMMENT_STYLE
);
575 /* Track parity of quotes. */
576 if (string_style
== -1)
577 /* Entering a string. */
579 else if (string_style
== c
)
580 /* Leaving the string. */
583 /* If we have two kinds of string delimiters.
584 There's no way to grok this scanning backwards. */
589 /* We've already checked that it is the relevant comstyle. */
590 if (string_style
!= -1 || comment_lossage
|| string_lossage
)
591 /* There are odd string quotes involved, so let's be careful.
592 Test case in Pascal: " { " a { " } */
597 /* Record best comment-starter so far. */
599 comstart_byte
= from_byte
;
601 else if (--nesting
<= 0)
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. */
611 if (SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == comstyle
612 && (SYNTAX_FLAGS_COMMENT_NESTED (prev_syntax
)
613 || SYNTAX_FLAGS_COMMENT_NESTED (syntax
)) == comnested
)
614 /* This is the same style of comment ender as ours. */
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. */
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
635 /* Assume a defun-start point is outside of strings. */
636 if (open_paren_in_column_0_is_defun_start
638 || (temp_byte
= dec_bytepos (from_byte
),
639 FETCH_CHAR (temp_byte
) == '\n')))
642 defun_start_byte
= from_byte
;
643 from
= stop
; /* Break out of the loop. */
652 if (comstart_pos
== 0)
655 from_byte
= comment_end_byte
;
656 UPDATE_SYNTAX_TABLE_FORWARD (comment_end
- 1);
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 */
663 from_byte
= comstart_byte
;
664 /* Globals are correct now. */
668 struct lisp_parse_state state
;
670 /* We had two kinds of string delimiters mixed up
671 together. Decode this going forwards.
672 Scan fwd from a known safe place (beginning-of-defun)
673 to the one in question; this records where we
674 last passed a comment starter. */
675 /* If we did not already find the defun start, find it now. */
676 if (defun_start
== 0)
678 defun_start
= find_defun_start (comment_end
, comment_end_byte
);
679 defun_start_byte
= find_start_value_byte
;
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
;
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 (* " *) */
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
);
703 } while (defun_start
< comment_end
);
705 from_byte
= CHAR_TO_BYTE (from
);
706 UPDATE_SYNTAX_TABLE_FORWARD (from
- 1);
711 *bytepos_ptr
= from_byte
;
713 return (from
== comment_end
) ? -1 : from
;
716 DEFUN ("syntax-table-p", Fsyntax_table_p
, Ssyntax_table_p
, 1, 1, 0,
717 doc
: /* Return t if OBJECT is a syntax table.
718 Currently, any char-table counts as a syntax table. */)
722 if (CHAR_TABLE_P (object
)
723 && EQ (XCHAR_TABLE (object
)->purpose
, Qsyntax_table
))
729 check_syntax_table (obj
)
732 if (!(CHAR_TABLE_P (obj
)
733 && EQ (XCHAR_TABLE (obj
)->purpose
, Qsyntax_table
)))
734 wrong_type_argument (Qsyntax_table_p
, obj
);
737 DEFUN ("syntax-table", Fsyntax_table
, Ssyntax_table
, 0, 0, 0,
738 doc
: /* Return the current syntax table.
739 This is the one specified by the current buffer. */)
742 return current_buffer
->syntax_table
;
745 DEFUN ("standard-syntax-table", Fstandard_syntax_table
,
746 Sstandard_syntax_table
, 0, 0, 0,
747 doc
: /* Return the standard syntax table.
748 This is the one used for new buffers. */)
751 return Vstandard_syntax_table
;
754 DEFUN ("copy-syntax-table", Fcopy_syntax_table
, Scopy_syntax_table
, 0, 1, 0,
755 doc
: /* Construct a new syntax table and return it.
756 It is a copy of the TABLE, which defaults to the standard syntax table. */)
763 check_syntax_table (table
);
765 table
= Vstandard_syntax_table
;
767 copy
= Fcopy_sequence (table
);
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
;
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
);
781 DEFUN ("set-syntax-table", Fset_syntax_table
, Sset_syntax_table
, 1, 1, 0,
782 doc
: /* Select a new syntax table for the current buffer.
783 One argument, a syntax table. */)
788 check_syntax_table (table
);
789 current_buffer
->syntax_table
= table
;
790 /* Indicate that this buffer now has a specified syntax table. */
791 idx
= PER_BUFFER_VAR_IDX (syntax_table
);
792 SET_PER_BUFFER_VALUE_P (current_buffer
, idx
, 1);
796 /* Convert a letter which signifies a syntax code
797 into the code it signifies.
798 This is used by modify-syntax-entry, and other things. */
800 unsigned 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,
805 (char) Swhitespace
, (char) Scomment_fence
, (char) Sstring
, 0377,
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,
812 (char) Sinherit
, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */
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
,
819 0377, 0377, 0377, 0377, (char) Sstring_fence
, 0377, 0377, 0377
822 /* Indexed by syntax code, give the letter that describes it. */
824 char syntax_code_spec
[16] =
826 ' ', '.', 'w', '_', '(', ')', '\'', '\"', '$', '\\', '/', '<', '>', '@',
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
835 static Lisp_Object Vsyntax_code_object
;
838 DEFUN ("char-syntax", Fchar_syntax
, Schar_syntax
, 1, 1, 0,
839 doc
: /* Return the syntax code of CHARACTER, described by a character.
840 For example, if CHARACTER is a word constituent,
841 the character `w' is returned.
842 The characters that correspond to various syntax codes
843 are listed in the documentation of `modify-syntax-entry'. */)
845 Lisp_Object character
;
848 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
850 gl_state
.use_global
= 0;
851 CHECK_NUMBER (character
);
852 char_int
= XINT (character
);
853 return make_number (syntax_code_spec
[(int) SYNTAX (char_int
)]);
856 DEFUN ("matching-paren", Fmatching_paren
, Smatching_paren
, 1, 1, 0,
857 doc
: /* Return the matching parenthesis of CHARACTER, or nil if none. */)
859 Lisp_Object character
;
862 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
863 gl_state
.use_global
= 0;
864 CHECK_NUMBER (character
);
865 char_int
= XINT (character
);
866 code
= SYNTAX (char_int
);
867 if (code
== Sopen
|| code
== Sclose
)
868 return SYNTAX_MATCH (char_int
);
872 DEFUN ("string-to-syntax", Fstring_to_syntax
, Sstring_to_syntax
, 1, 1, 0,
873 doc
: /* Convert a syntax specification STRING into syntax cell form.
874 STRING should be a string as it is allowed as argument of
875 `modify-syntax-entry'. Value is the equivalent cons cell
876 (CODE . MATCHING-CHAR) that can be used as value of a `syntax-table'
881 register unsigned char *p
;
882 register enum syntaxcode code
;
886 CHECK_STRING (string
);
888 p
= XSTRING (string
)->data
;
889 code
= (enum syntaxcode
) syntax_spec_code
[*p
++];
890 if (((int) code
& 0377) == 0377)
891 error ("invalid syntax description letter: %c", p
[-1]);
893 if (code
== Sinherit
)
899 int character
= (STRING_CHAR_AND_LENGTH
900 (p
, STRING_BYTES (XSTRING (string
)) - 1, len
));
901 XSETINT (match
, character
);
902 if (XFASTINT (match
) == ' ')
942 if (val
< XVECTOR (Vsyntax_code_object
)->size
&& NILP (match
))
943 return XVECTOR (Vsyntax_code_object
)->contents
[val
];
945 /* Since we can't use a shared object, let's make a new one. */
946 return Fcons (make_number (val
), match
);
949 /* I really don't know why this is interactive
950 help-form should at least be made useful whilst reading the second arg. */
951 DEFUN ("modify-syntax-entry", Fmodify_syntax_entry
, Smodify_syntax_entry
, 2, 3,
952 "cSet syntax for character: \nsSet syntax for %s to: ",
953 doc
: /* Set syntax for character CHAR according to string NEWENTRY.
954 The syntax is changed only for table SYNTAX_TABLE, which defaults to
955 the current buffer's syntax table.
956 CHAR may be a cons (MIN . MAX), in which case, syntaxes of all characters
957 in the range MIN and MAX are changed.
958 The first character of NEWENTRY should be one of the following:
959 Space or - whitespace syntax. w word constituent.
960 _ symbol constituent. . punctuation.
961 ( open-parenthesis. ) close-parenthesis.
962 " string quote. \\ escape.
963 $ paired delimiter. ' expression quote or prefix operator.
964 < comment starter. > comment ender.
965 / character-quote. @ inherit from `standard-syntax-table'.
966 | generic string fence. ! generic comment fence.
968 Only single-character comment start and end sequences are represented thus.
969 Two-character sequences are represented as described below.
970 The second character of NEWENTRY is the matching parenthesis,
971 used only if the first character is `(' or `)'.
972 Any additional characters are flags.
973 Defined flags are the characters 1, 2, 3, 4, b, p, and n.
974 1 means CHAR is the start of a two-char comment start sequence.
975 2 means CHAR is the second character of such a sequence.
976 3 means CHAR is the start of a two-char comment end sequence.
977 4 means CHAR is the second character of such a sequence.
979 There can be up to two orthogonal comment sequences. This is to support
980 language modes such as C++. By default, all comment sequences are of style
981 a, but you can set the comment sequence style to b (on the second character
982 of a comment-start, or the first character of a comment-end sequence) using
984 b means CHAR is part of comment sequence b.
985 n means CHAR is part of a nestable comment sequence.
987 p means CHAR is a prefix character for `backward-prefix-chars';
988 such characters are treated as whitespace when they occur
990 usage: (modify-syntax-entry CHAR NEWENTRY &optional SYNTAX-TABLE) */)
991 (c
, newentry
, syntax_table
)
992 Lisp_Object c
, newentry
, syntax_table
;
996 CHECK_CHARACTER (XCAR (c
));
997 CHECK_CHARACTER (XCDR (c
));
1000 CHECK_CHARACTER (c
);
1002 if (NILP (syntax_table
))
1003 syntax_table
= current_buffer
->syntax_table
;
1005 check_syntax_table (syntax_table
);
1007 newentry
= Fstring_to_syntax (newentry
);
1009 SET_RAW_SYNTAX_ENTRY_RANGE (syntax_table
, c
, newentry
);
1011 SET_RAW_SYNTAX_ENTRY (syntax_table
, XINT (c
), newentry
);
1015 /* Dump syntax table to buffer in human-readable format */
1017 DEFUN ("internal-describe-syntax-value", Finternal_describe_syntax_value
,
1018 Sinternal_describe_syntax_value
, 1, 1, 0,
1019 doc
: /* Insert a description of the internal syntax description SYNTAX at point. */)
1023 register enum syntaxcode code
;
1024 char desc
, start1
, start2
, end1
, end2
, prefix
, comstyle
, comnested
;
1026 Lisp_Object first
, match_lisp
, value
= syntax
;
1030 insert_string ("default");
1034 if (CHAR_TABLE_P (value
))
1036 insert_string ("deeper char-table ...");
1042 insert_string ("invalid");
1046 first
= XCAR (value
);
1047 match_lisp
= XCDR (value
);
1049 if (!INTEGERP (first
) || !(NILP (match_lisp
) || INTEGERP (match_lisp
)))
1051 insert_string ("invalid");
1055 code
= (enum syntaxcode
) (XINT (first
) & 0377);
1056 start1
= (XINT (first
) >> 16) & 1;
1057 start2
= (XINT (first
) >> 17) & 1;
1058 end1
= (XINT (first
) >> 18) & 1;
1059 end2
= (XINT (first
) >> 19) & 1;
1060 prefix
= (XINT (first
) >> 20) & 1;
1061 comstyle
= (XINT (first
) >> 21) & 1;
1062 comnested
= (XINT (first
) >> 22) & 1;
1064 if ((int) code
< 0 || (int) code
>= (int) Smax
)
1066 insert_string ("invalid");
1069 desc
= syntax_code_spec
[(int) code
];
1071 str
[0] = desc
, str
[1] = 0;
1074 if (NILP (match_lisp
))
1077 insert_char (XINT (match_lisp
));
1096 insert_string ("\twhich means: ");
1098 switch (SWITCH_ENUM_CAST (code
))
1101 insert_string ("whitespace"); break;
1103 insert_string ("punctuation"); break;
1105 insert_string ("word"); break;
1107 insert_string ("symbol"); break;
1109 insert_string ("open"); break;
1111 insert_string ("close"); break;
1113 insert_string ("prefix"); break;
1115 insert_string ("string"); break;
1117 insert_string ("math"); break;
1119 insert_string ("escape"); break;
1121 insert_string ("charquote"); break;
1123 insert_string ("comment"); break;
1125 insert_string ("endcomment"); break;
1127 insert_string ("inherit"); break;
1128 case Scomment_fence
:
1129 insert_string ("comment fence"); break;
1131 insert_string ("string fence"); break;
1133 insert_string ("invalid");
1137 if (!NILP (match_lisp
))
1139 insert_string (", matches ");
1140 insert_char (XINT (match_lisp
));
1144 insert_string (",\n\t is the first character of a comment-start sequence");
1146 insert_string (",\n\t is the second character of a comment-start sequence");
1149 insert_string (",\n\t is the first character of a comment-end sequence");
1151 insert_string (",\n\t is the second character of a comment-end sequence");
1153 insert_string (" (comment style b)");
1155 insert_string (" (nestable)");
1158 insert_string (",\n\t is a prefix character for `backward-prefix-chars'");
1163 int parse_sexp_ignore_comments
;
1165 /* Return the position across COUNT words from FROM.
1166 If that many words cannot be found before the end of the buffer, return 0.
1167 COUNT negative means scan backward and stop at word beginning. */
1170 scan_words (from
, count
)
1171 register int from
, count
;
1173 register int beg
= BEGV
;
1174 register int end
= ZV
;
1175 register int from_byte
= CHAR_TO_BYTE (from
);
1176 register enum syntaxcode code
;
1182 SETUP_SYNTAX_TABLE (from
, count
);
1193 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1194 ch0
= FETCH_CHAR (from_byte
);
1195 code
= SYNTAX (ch0
);
1196 INC_BOTH (from
, from_byte
);
1197 if (words_include_escapes
1198 && (code
== Sescape
|| code
== Scharquote
))
1203 /* Now CH0 is a character which begins a word and FROM is the
1204 position of the next character. */
1207 if (from
== end
) break;
1208 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1209 ch1
= FETCH_CHAR (from_byte
);
1210 code
= SYNTAX (ch1
);
1211 if (!(words_include_escapes
1212 && (code
== Sescape
|| code
== Scharquote
)))
1213 if (code
!= Sword
|| WORD_BOUNDARY_P (ch0
, ch1
))
1215 INC_BOTH (from
, from_byte
);
1229 DEC_BOTH (from
, from_byte
);
1230 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1231 ch1
= FETCH_CHAR (from_byte
);
1232 code
= SYNTAX (ch1
);
1233 if (words_include_escapes
1234 && (code
== Sescape
|| code
== Scharquote
))
1239 /* Now CH1 is a character which ends a word and FROM is the
1247 temp_byte
= dec_bytepos (from_byte
);
1248 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1249 ch0
= FETCH_CHAR (temp_byte
);
1250 code
= SYNTAX (ch0
);
1251 if (!(words_include_escapes
1252 && (code
== Sescape
|| code
== Scharquote
)))
1253 if (code
!= Sword
|| WORD_BOUNDARY_P (ch0
, ch1
))
1255 DEC_BOTH (from
, from_byte
);
1266 DEFUN ("forward-word", Fforward_word
, Sforward_word
, 1, 1, "p",
1267 doc
: /* Move point forward ARG words (backward if ARG is negative).
1269 If an edge of the buffer or a field boundary is reached, point is left there
1270 and the function returns nil. Field boundaries are not noticed if
1271 `inhibit-field-text-motion' is non-nil. */)
1276 CHECK_NUMBER (count
);
1278 val
= orig_val
= scan_words (PT
, XINT (count
));
1280 val
= XINT (count
) > 0 ? ZV
: BEGV
;
1282 /* Avoid jumping out of an input field. */
1283 val
= XFASTINT (Fconstrain_to_field (make_number (val
), make_number (PT
),
1287 return val
== orig_val
? Qt
: Qnil
;
1290 Lisp_Object
skip_chars ();
1292 DEFUN ("skip-chars-forward", Fskip_chars_forward
, Sskip_chars_forward
, 1, 2, 0,
1293 doc
: /* Move point forward, stopping before a char not in STRING, or at pos LIM.
1294 STRING is like the inside of a `[...]' in a regular expression
1295 except that `]' is never special and `\\' quotes `^', `-' or `\\'
1296 (but not as the end of a range; quoting is never needed there).
1297 Thus, with arg "a-zA-Z", this skips letters stopping before first nonletter.
1298 With arg "^a-zA-Z", skips nonletters stopping before first letter.
1299 Returns the distance traveled, either zero or positive. */)
1301 Lisp_Object string
, lim
;
1303 return skip_chars (1, 0, string
, lim
);
1306 DEFUN ("skip-chars-backward", Fskip_chars_backward
, Sskip_chars_backward
, 1, 2, 0,
1307 doc
: /* Move point backward, stopping after a char not in STRING, or at pos LIM.
1308 See `skip-chars-forward' for details.
1309 Returns the distance traveled, either zero or negative. */)
1311 Lisp_Object string
, lim
;
1313 return skip_chars (0, 0, string
, lim
);
1316 DEFUN ("skip-syntax-forward", Fskip_syntax_forward
, Sskip_syntax_forward
, 1, 2, 0,
1317 doc
: /* Move point forward across chars in specified syntax classes.
1318 SYNTAX is a string of syntax code characters.
1319 Stop before a char whose syntax is not in SYNTAX, or at position LIM.
1320 If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1321 This function returns the distance traveled, either zero or positive. */)
1323 Lisp_Object syntax
, lim
;
1325 return skip_chars (1, 1, syntax
, lim
);
1328 DEFUN ("skip-syntax-backward", Fskip_syntax_backward
, Sskip_syntax_backward
, 1, 2, 0,
1329 doc
: /* Move point backward across chars in specified syntax classes.
1330 SYNTAX is a string of syntax code characters.
1331 Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.
1332 If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1333 This function returns the distance traveled, either zero or negative. */)
1335 Lisp_Object syntax
, lim
;
1337 return skip_chars (0, 1, syntax
, lim
);
1341 skip_chars (forwardp
, syntaxp
, string
, lim
)
1342 int forwardp
, syntaxp
;
1343 Lisp_Object string
, lim
;
1345 register unsigned int c
;
1346 unsigned char fastmap
[0400];
1347 /* If SYNTAXP is 0, STRING may contain multi-byte form of characters
1348 of which codes don't fit in FASTMAP. In that case, set the
1349 ranges of characters in CHAR_RANGES. */
1351 int n_char_ranges
= 0;
1353 register int i
, i_byte
;
1354 int multibyte
= !NILP (current_buffer
->enable_multibyte_characters
);
1355 int string_multibyte
;
1360 CHECK_STRING (string
);
1361 char_ranges
= (int *) alloca (XSTRING (string
)->size
* (sizeof (int)) * 2);
1362 string_multibyte
= STRING_MULTIBYTE (string
);
1363 str
= XSTRING (string
)->data
;
1364 size_byte
= STRING_BYTES (XSTRING (string
));
1366 /* Adjust the multibyteness of the string to that of the buffer. */
1367 if (multibyte
!= string_multibyte
)
1372 nbytes
= count_size_as_multibyte (XSTRING (string
)->data
,
1373 XSTRING (string
)->size
);
1375 nbytes
= XSTRING (string
)->size
;
1376 if (nbytes
!= size_byte
)
1378 str
= (unsigned char *) alloca (nbytes
);
1379 copy_text (XSTRING (string
)->data
, str
, size_byte
,
1380 string_multibyte
, multibyte
);
1386 XSETINT (lim
, forwardp
? ZV
: BEGV
);
1388 CHECK_NUMBER_COERCE_MARKER (lim
);
1390 /* In any case, don't allow scan outside bounds of buffer. */
1391 if (XINT (lim
) > ZV
)
1392 XSETFASTINT (lim
, ZV
);
1393 if (XINT (lim
) < BEGV
)
1394 XSETFASTINT (lim
, BEGV
);
1396 bzero (fastmap
, sizeof fastmap
);
1400 if (i_byte
< size_byte
1401 && XSTRING (string
)->data
[0] == '^')
1403 negate
= 1; i_byte
++;
1406 /* Find the characters specified and set their elements of fastmap.
1407 If syntaxp, each character counts as itself.
1408 Otherwise, handle backslashes and ranges specially. */
1410 if (size_byte
== XSTRING (string
)->size
)
1411 while (i_byte
< size_byte
)
1416 fastmap
[syntax_spec_code
[c
]] = 1;
1421 if (i_byte
== size_byte
)
1426 if (i_byte
< size_byte
1427 && str
[i_byte
] == '-')
1431 /* Skip over the dash. */
1434 if (i_byte
== size_byte
)
1437 /* Get the end of the range. */
1447 while (i_byte
< size_byte
)
1449 c
= STRING_CHAR_AND_LENGTH (str
+ i_byte
, size_byte
-i_byte
, len
);
1453 fastmap
[syntax_spec_code
[c
& 0377]] = 1;
1458 if (i_byte
== size_byte
)
1461 c
= STRING_CHAR_AND_LENGTH (str
+i_byte
, size_byte
-i_byte
, len
);
1464 if (i_byte
< size_byte
1465 && str
[i_byte
] == '-')
1469 /* Skip over the dash. */
1472 if (i_byte
== size_byte
)
1475 /* Get the end of the range. */
1476 c2
=STRING_CHAR_AND_LENGTH (str
+ i_byte
, size_byte
-i_byte
, len
);
1479 if (ASCII_CHAR_P (c
))
1480 while (c
<= c2
&& c
< 0x80)
1484 char_ranges
[n_char_ranges
++] = c
;
1485 char_ranges
[n_char_ranges
++] = c2
;
1490 if (ASCII_CHAR_P (c
))
1494 char_ranges
[n_char_ranges
++] = c
;
1495 char_ranges
[n_char_ranges
++] = c
;
1501 /* If ^ was the first character, complement the fastmap. */
1503 for (i
= 0; i
< sizeof fastmap
; i
++)
1507 int start_point
= PT
;
1509 int pos_byte
= PT_BYTE
;
1514 SETUP_SYNTAX_TABLE (pos
, forwardp
? 1 : -1);
1519 if (pos
< XINT (lim
))
1520 while (fastmap
[(int) SYNTAX (FETCH_CHAR (pos_byte
))])
1522 /* Since we already checked for multibyteness,
1523 avoid using INC_BOTH which checks again. */
1526 if (pos
>= XINT (lim
))
1528 UPDATE_SYNTAX_TABLE_FORWARD (pos
);
1533 while (pos
< XINT (lim
)
1534 && fastmap
[(int) SYNTAX (FETCH_BYTE (pos
))])
1537 UPDATE_SYNTAX_TABLE_FORWARD (pos
);
1545 while (pos
> XINT (lim
))
1547 int savepos
= pos_byte
;
1548 /* Since we already checked for multibyteness,
1549 avoid using DEC_BOTH which checks again. */
1552 UPDATE_SYNTAX_TABLE_BACKWARD (pos
);
1553 if (!fastmap
[(int) SYNTAX (FETCH_CHAR (pos_byte
))])
1563 if (pos
> XINT (lim
))
1564 while (fastmap
[(int) SYNTAX (FETCH_BYTE (pos
- 1))])
1567 if (pos
<= XINT (lim
))
1569 UPDATE_SYNTAX_TABLE_BACKWARD (pos
- 1);
1579 while (pos
< XINT (lim
))
1581 c
= FETCH_MULTIBYTE_CHAR (pos_byte
);
1582 if (ASCII_CHAR_P (c
))
1589 /* If we are looking at a multibyte character,
1590 we must look up the character in the table
1591 CHAR_RANGES. If there's no data in the
1592 table, that character is not what we want to
1595 /* The following code do the right thing even if
1596 n_char_ranges is zero (i.e. no data in
1598 for (i
= 0; i
< n_char_ranges
; i
+= 2)
1599 if (c
>= char_ranges
[i
] && c
<= char_ranges
[i
+ 1])
1601 if (!(negate
^ (i
< n_char_ranges
)))
1604 INC_BOTH (pos
, pos_byte
);
1607 while (pos
< XINT (lim
) && fastmap
[FETCH_BYTE (pos
)])
1613 while (pos
> XINT (lim
))
1615 int prev_pos_byte
= pos_byte
;
1617 DEC_POS (prev_pos_byte
);
1618 c
= FETCH_MULTIBYTE_CHAR (prev_pos_byte
);
1619 if (SINGLE_BYTE_CHAR_P (c
))
1626 /* See the comment in the previous similar code. */
1627 for (i
= 0; i
< n_char_ranges
; i
+= 2)
1628 if (c
>= char_ranges
[i
] && c
<= char_ranges
[i
+ 1])
1630 if (!(negate
^ (i
< n_char_ranges
)))
1634 pos_byte
= prev_pos_byte
;
1637 while (pos
> XINT (lim
) && fastmap
[FETCH_BYTE (pos
- 1)])
1645 SET_PT_BOTH (pos
, pos_byte
);
1648 return make_number (PT
- start_point
);
1652 /* Jump over a comment, assuming we are at the beginning of one.
1653 FROM is the current position.
1654 FROM_BYTE is the bytepos corresponding to FROM.
1655 Do not move past STOP (a charpos).
1656 The comment over which we have to jump is of style STYLE
1657 (either SYNTAX_COMMENT_STYLE(foo) or ST_COMMENT_STYLE).
1658 NESTING should be positive to indicate the nesting at the beginning
1659 for nested comments and should be zero or negative else.
1660 ST_COMMENT_STYLE cannot be nested.
1661 PREV_SYNTAX is the SYNTAX_WITH_FLAGS of the previous character
1662 (or 0 If the search cannot start in the middle of a two-character).
1664 If successful, return 1 and store the charpos of the comment's end
1665 into *CHARPOS_PTR and the corresponding bytepos into *BYTEPOS_PTR.
1666 Else, return 0 and store the charpos STOP into *CHARPOS_PTR, the
1667 corresponding bytepos into *BYTEPOS_PTR and the current nesting
1668 (as defined for state.incomment) in *INCOMMENT_PTR.
1670 The comment end is the last character of the comment rather than the
1671 character just after the comment.
1673 Global syntax data is assumed to initially be valid for FROM and
1674 remains valid for forward search starting at the returned position. */
1677 forw_comment (from
, from_byte
, stop
, nesting
, style
, prev_syntax
,
1678 charpos_ptr
, bytepos_ptr
, incomment_ptr
)
1679 int from
, from_byte
, stop
;
1680 int nesting
, style
, prev_syntax
;
1681 int *charpos_ptr
, *bytepos_ptr
, *incomment_ptr
;
1684 register enum syntaxcode code
;
1685 register int syntax
;
1687 if (nesting
<= 0) nesting
= -1;
1689 /* Enter the loop in the middle so that we find
1690 a 2-char comment ender if we start in the middle of it. */
1691 syntax
= prev_syntax
;
1692 if (syntax
!= 0) goto forw_incomment
;
1698 *incomment_ptr
= nesting
;
1699 *charpos_ptr
= from
;
1700 *bytepos_ptr
= from_byte
;
1703 c
= FETCH_CHAR (from_byte
);
1704 syntax
= SYNTAX_WITH_FLAGS (c
);
1705 code
= syntax
& 0xff;
1706 if (code
== Sendcomment
1707 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
1708 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ?
1709 (nesting
> 0 && --nesting
== 0) : nesting
< 0))
1710 /* we have encountered a comment end of the same style
1711 as the comment sequence which began this comment
1714 if (code
== Scomment_fence
1715 && style
== ST_COMMENT_STYLE
)
1716 /* we have encountered a comment end of the same style
1717 as the comment sequence which began this comment
1722 && SYNTAX_FLAGS_COMMENT_NESTED (syntax
)
1723 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
)
1724 /* we have encountered a nested comment of the same style
1725 as the comment sequence which began this comment section */
1727 INC_BOTH (from
, from_byte
);
1728 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1731 if (from
< stop
&& SYNTAX_FLAGS_COMEND_FIRST (syntax
)
1732 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
1733 && (c1
= FETCH_CHAR (from_byte
),
1734 SYNTAX_COMEND_SECOND (c1
))
1735 && ((SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ||
1736 SYNTAX_COMMENT_NESTED (c1
)) ? nesting
> 0 : nesting
< 0))
1739 /* we have encountered a comment end of the same style
1740 as the comment sequence which began this comment
1745 INC_BOTH (from
, from_byte
);
1746 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1751 && SYNTAX_FLAGS_COMSTART_FIRST (syntax
)
1752 && (c1
= FETCH_CHAR (from_byte
),
1753 SYNTAX_COMMENT_STYLE (c1
) == style
1754 && SYNTAX_COMSTART_SECOND (c1
))
1755 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ||
1756 SYNTAX_COMMENT_NESTED (c1
)))
1757 /* we have encountered a nested comment of the same style
1758 as the comment sequence which began this comment
1761 INC_BOTH (from
, from_byte
);
1762 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1766 *charpos_ptr
= from
;
1767 *bytepos_ptr
= from_byte
;
1771 DEFUN ("forward-comment", Fforward_comment
, Sforward_comment
, 1, 1, 0,
1772 doc
: /* Move forward across up to N comments. If N is negative, move backward.
1773 Stop scanning if we find something other than a comment or whitespace.
1774 Set point to where scanning stops.
1775 If N comments are found as expected, with nothing except whitespace
1776 between them, return t; otherwise return nil. */)
1784 register enum syntaxcode code
;
1785 int comstyle
= 0; /* style of comment encountered */
1786 int comnested
= 0; /* whether the comment is nestable or not */
1789 int out_charpos
, out_bytepos
;
1792 CHECK_NUMBER (count
);
1793 count1
= XINT (count
);
1794 stop
= count1
> 0 ? ZV
: BEGV
;
1800 from_byte
= PT_BYTE
;
1802 SETUP_SYNTAX_TABLE (from
, count1
);
1811 SET_PT_BOTH (from
, from_byte
);
1815 c
= FETCH_CHAR (from_byte
);
1817 comstart_first
= SYNTAX_COMSTART_FIRST (c
);
1818 comnested
= SYNTAX_COMMENT_NESTED (c
);
1819 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1820 INC_BOTH (from
, from_byte
);
1821 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1822 if (from
< stop
&& comstart_first
1823 && (c1
= FETCH_CHAR (from_byte
),
1824 SYNTAX_COMSTART_SECOND (c1
)))
1826 /* We have encountered a comment start sequence and we
1827 are ignoring all text inside comments. We must record
1828 the comment style this sequence begins so that later,
1829 only a comment end of the same style actually ends
1830 the comment section. */
1832 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
1833 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
1834 INC_BOTH (from
, from_byte
);
1835 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1838 while (code
== Swhitespace
|| (code
== Sendcomment
&& c
== '\n'));
1840 if (code
== Scomment_fence
)
1841 comstyle
= ST_COMMENT_STYLE
;
1842 else if (code
!= Scomment
)
1845 DEC_BOTH (from
, from_byte
);
1846 SET_PT_BOTH (from
, from_byte
);
1849 /* We're at the start of a comment. */
1850 found
= forw_comment (from
, from_byte
, stop
, comnested
, comstyle
, 0,
1851 &out_charpos
, &out_bytepos
, &dummy
);
1852 from
= out_charpos
; from_byte
= out_bytepos
;
1856 SET_PT_BOTH (from
, from_byte
);
1859 INC_BOTH (from
, from_byte
);
1860 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1861 /* We have skipped one comment. */
1873 SET_PT_BOTH (BEGV
, BEGV_BYTE
);
1878 DEC_BOTH (from
, from_byte
);
1879 /* char_quoted does UPDATE_SYNTAX_TABLE_BACKWARD (from). */
1880 quoted
= char_quoted (from
, from_byte
);
1881 c
= FETCH_CHAR (from_byte
);
1884 comnested
= SYNTAX_COMMENT_NESTED (c
);
1885 if (code
== Sendcomment
)
1886 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1887 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
1888 && prev_char_comend_first (from
, from_byte
)
1889 && !char_quoted (from
- 1, dec_bytepos (from_byte
)))
1891 /* We must record the comment style encountered so that
1892 later, we can match only the proper comment begin
1893 sequence of the same style. */
1894 DEC_BOTH (from
, from_byte
);
1896 /* Calling char_quoted, above, set up global syntax position
1897 at the new value of FROM. */
1898 c1
= FETCH_CHAR (from_byte
);
1899 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
1900 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
1903 if (code
== Scomment_fence
)
1905 /* Skip until first preceding unquoted comment_fence. */
1906 int found
= 0, ini
= from
, ini_byte
= from_byte
;
1910 DEC_BOTH (from
, from_byte
);
1913 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1914 c
= FETCH_CHAR (from_byte
);
1915 if (SYNTAX (c
) == Scomment_fence
1916 && !char_quoted (from
, from_byte
))
1924 from
= ini
; /* Set point to ini + 1. */
1925 from_byte
= ini_byte
;
1929 else if (code
== Sendcomment
)
1931 found
= back_comment (from
, from_byte
, stop
, comnested
, comstyle
,
1932 &out_charpos
, &out_bytepos
);
1936 /* This end-of-line is not an end-of-comment.
1937 Treat it like a whitespace.
1938 CC-mode (and maybe others) relies on this behavior. */
1942 /* Failure: we should go back to the end of this
1943 not-quite-endcomment. */
1944 if (SYNTAX(c
) != code
)
1945 /* It was a two-char Sendcomment. */
1946 INC_BOTH (from
, from_byte
);
1952 /* We have skipped one comment. */
1953 from
= out_charpos
, from_byte
= out_bytepos
;
1957 else if (code
!= Swhitespace
|| quoted
)
1961 INC_BOTH (from
, from_byte
);
1962 SET_PT_BOTH (from
, from_byte
);
1970 SET_PT_BOTH (from
, from_byte
);
1975 /* Return syntax code of character C if C is a single byte character
1976 or `multibyte_symbol_p' is zero. Otherwise, return Ssymbol. */
1978 #define SYNTAX_WITH_MULTIBYTE_CHECK(c) \
1979 ((SINGLE_BYTE_CHAR_P (c) || !multibyte_symbol_p) \
1980 ? SYNTAX (c) : Ssymbol)
1983 scan_lists (from
, count
, depth
, sexpflag
)
1985 int count
, depth
, sexpflag
;
1988 register int stop
= count
> 0 ? ZV
: BEGV
;
1993 register enum syntaxcode code
, temp_code
;
1994 int min_depth
= depth
; /* Err out if depth gets less than this. */
1995 int comstyle
= 0; /* style of comment encountered */
1996 int comnested
= 0; /* whether the comment is nestable or not */
1998 int last_good
= from
;
2001 int out_bytepos
, out_charpos
;
2003 int multibyte_symbol_p
= sexpflag
&& multibyte_syntax_as_symbol
;
2005 if (depth
> 0) min_depth
= 0;
2007 if (from
> ZV
) from
= ZV
;
2008 if (from
< BEGV
) from
= BEGV
;
2010 from_byte
= CHAR_TO_BYTE (from
);
2015 SETUP_SYNTAX_TABLE (from
, count
);
2020 int comstart_first
, prefix
;
2021 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2022 c
= FETCH_CHAR (from_byte
);
2023 code
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2024 comstart_first
= SYNTAX_COMSTART_FIRST (c
);
2025 comnested
= SYNTAX_COMMENT_NESTED (c
);
2026 comstyle
= SYNTAX_COMMENT_STYLE (c
);
2027 prefix
= SYNTAX_PREFIX (c
);
2028 if (depth
== min_depth
)
2030 INC_BOTH (from
, from_byte
);
2031 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2032 if (from
< stop
&& comstart_first
2033 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from_byte
))
2034 && parse_sexp_ignore_comments
)
2036 /* we have encountered a comment start sequence and we
2037 are ignoring all text inside comments. We must record
2038 the comment style this sequence begins so that later,
2039 only a comment end of the same style actually ends
2040 the comment section */
2042 c1
= FETCH_CHAR (from_byte
);
2043 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2044 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2045 INC_BOTH (from
, from_byte
);
2046 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2052 switch (SWITCH_ENUM_CAST (code
))
2056 if (from
== stop
) goto lose
;
2057 INC_BOTH (from
, from_byte
);
2058 /* treat following character as a word constituent */
2061 if (depth
|| !sexpflag
) break;
2062 /* This word counts as a sexp; return at end of it. */
2065 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2067 /* Some compilers can't handle this inside the switch. */
2068 c
= FETCH_CHAR (from_byte
);
2069 temp
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2074 INC_BOTH (from
, from_byte
);
2075 if (from
== stop
) goto lose
;
2084 INC_BOTH (from
, from_byte
);
2088 case Scomment_fence
:
2089 comstyle
= ST_COMMENT_STYLE
;
2092 if (!parse_sexp_ignore_comments
) break;
2093 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2094 found
= forw_comment (from
, from_byte
, stop
,
2095 comnested
, comstyle
, 0,
2096 &out_charpos
, &out_bytepos
, &dummy
);
2097 from
= out_charpos
, from_byte
= out_bytepos
;
2104 INC_BOTH (from
, from_byte
);
2105 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2111 if (from
!= stop
&& c
== FETCH_CHAR (from_byte
))
2113 INC_BOTH (from
, from_byte
);
2123 if (!++depth
) goto done
;
2128 if (!--depth
) goto done
;
2129 if (depth
< min_depth
)
2130 Fsignal (Qscan_error
,
2131 Fcons (build_string ("Containing expression ends prematurely"),
2132 Fcons (make_number (last_good
),
2133 Fcons (make_number (from
), Qnil
))));
2138 temp_pos
= dec_bytepos (from_byte
);
2139 stringterm
= FETCH_CHAR (temp_pos
);
2142 if (from
>= stop
) goto lose
;
2143 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2144 c
= FETCH_CHAR (from_byte
);
2147 && SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring
)
2148 : SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring_fence
)
2151 /* Some compilers can't handle this inside the switch. */
2152 temp
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2157 INC_BOTH (from
, from_byte
);
2159 INC_BOTH (from
, from_byte
);
2161 INC_BOTH (from
, from_byte
);
2162 if (!depth
&& sexpflag
) goto done
;
2167 /* Reached end of buffer. Error if within object, return nil if between */
2168 if (depth
) goto lose
;
2173 /* End of object reached */
2182 DEC_BOTH (from
, from_byte
);
2183 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2184 c
= FETCH_CHAR (from_byte
);
2185 code
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2186 if (depth
== min_depth
)
2189 comnested
= SYNTAX_COMMENT_NESTED (c
);
2190 if (code
== Sendcomment
)
2191 comstyle
= SYNTAX_COMMENT_STYLE (c
);
2192 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
2193 && prev_char_comend_first (from
, from_byte
)
2194 && parse_sexp_ignore_comments
)
2196 /* We must record the comment style encountered so that
2197 later, we can match only the proper comment begin
2198 sequence of the same style. */
2199 DEC_BOTH (from
, from_byte
);
2200 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2202 c1
= FETCH_CHAR (from_byte
);
2203 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2204 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2207 /* Quoting turns anything except a comment-ender
2208 into a word character. Note that this cannot be true
2209 if we decremented FROM in the if-statement above. */
2210 if (code
!= Sendcomment
&& char_quoted (from
, from_byte
))
2212 else if (SYNTAX_PREFIX (c
))
2215 switch (SWITCH_ENUM_CAST (code
))
2221 if (depth
|| !sexpflag
) break;
2222 /* This word counts as a sexp; count object finished
2223 after passing it. */
2226 temp_pos
= from_byte
;
2227 if (! NILP (current_buffer
->enable_multibyte_characters
))
2231 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2232 c1
= FETCH_CHAR (temp_pos
);
2233 temp_code
= SYNTAX_WITH_MULTIBYTE_CHECK (c1
);
2234 /* Don't allow comment-end to be quoted. */
2235 if (temp_code
== Sendcomment
)
2237 quoted
= char_quoted (from
- 1, temp_pos
);
2240 DEC_BOTH (from
, from_byte
);
2241 temp_pos
= dec_bytepos (temp_pos
);
2242 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2244 c1
= FETCH_CHAR (temp_pos
);
2245 temp_code
= SYNTAX_WITH_MULTIBYTE_CHECK (c1
);
2246 if (! (quoted
|| temp_code
== Sword
2247 || temp_code
== Ssymbol
2248 || temp_code
== Squote
))
2250 DEC_BOTH (from
, from_byte
);
2257 temp_pos
= dec_bytepos (from_byte
);
2258 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2259 if (from
!= stop
&& c
== FETCH_CHAR (temp_pos
))
2260 DEC_BOTH (from
, from_byte
);
2269 if (!++depth
) goto done2
;
2274 if (!--depth
) goto done2
;
2275 if (depth
< min_depth
)
2276 Fsignal (Qscan_error
,
2277 Fcons (build_string ("Containing expression ends prematurely"),
2278 Fcons (make_number (last_good
),
2279 Fcons (make_number (from
), Qnil
))));
2283 if (!parse_sexp_ignore_comments
)
2285 found
= back_comment (from
, from_byte
, stop
, comnested
, comstyle
,
2286 &out_charpos
, &out_bytepos
);
2287 /* FIXME: if found == -1, then it really wasn't a comment-end.
2288 For single-char Sendcomment, we can't do much about it apart
2289 from skipping the char.
2290 For 2-char endcomments, we could try again, taking both
2291 chars as separate entities, but it's a lot of trouble
2292 for very little gain, so we don't bother either. -sm */
2294 from
= out_charpos
, from_byte
= out_bytepos
;
2297 case Scomment_fence
:
2301 DEC_BOTH (from
, from_byte
);
2302 if (from
== stop
) goto lose
;
2303 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2304 if (!char_quoted (from
, from_byte
)
2305 && (c
= FETCH_CHAR (from_byte
),
2306 SYNTAX_WITH_MULTIBYTE_CHECK (c
) == code
))
2309 if (code
== Sstring_fence
&& !depth
&& sexpflag
) goto done2
;
2313 stringterm
= FETCH_CHAR (from_byte
);
2316 if (from
== stop
) goto lose
;
2317 temp_pos
= from_byte
;
2318 if (! NILP (current_buffer
->enable_multibyte_characters
))
2322 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2323 if (!char_quoted (from
- 1, temp_pos
)
2324 && stringterm
== (c
= FETCH_CHAR (temp_pos
))
2325 && SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring
)
2327 DEC_BOTH (from
, from_byte
);
2329 DEC_BOTH (from
, from_byte
);
2330 if (!depth
&& sexpflag
) goto done2
;
2335 /* Reached start of buffer. Error if within object, return nil if between */
2336 if (depth
) goto lose
;
2347 XSETFASTINT (val
, from
);
2351 Fsignal (Qscan_error
,
2352 Fcons (build_string ("Unbalanced parentheses"),
2353 Fcons (make_number (last_good
),
2354 Fcons (make_number (from
), Qnil
))));
2359 DEFUN ("scan-lists", Fscan_lists
, Sscan_lists
, 3, 3, 0,
2360 doc
: /* Scan from character number FROM by COUNT lists.
2361 Returns the character number of the position thus found.
2363 If DEPTH is nonzero, paren depth begins counting from that value,
2364 only places where the depth in parentheses becomes zero
2365 are candidates for stopping; COUNT such places are counted.
2366 Thus, a positive value for DEPTH means go out levels.
2368 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2370 If the beginning or end of (the accessible part of) the buffer is reached
2371 and the depth is wrong, an error is signaled.
2372 If the depth is right but the count is not used up, nil is returned. */)
2373 (from
, count
, depth
)
2374 Lisp_Object from
, count
, depth
;
2376 CHECK_NUMBER (from
);
2377 CHECK_NUMBER (count
);
2378 CHECK_NUMBER (depth
);
2380 return scan_lists (XINT (from
), XINT (count
), XINT (depth
), 0);
2383 DEFUN ("scan-sexps", Fscan_sexps
, Sscan_sexps
, 2, 2, 0,
2384 doc
: /* Scan from character number FROM by COUNT balanced expressions.
2385 If COUNT is negative, scan backwards.
2386 Returns the character number of the position thus found.
2388 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2390 If the beginning or end of (the accessible part of) the buffer is reached
2391 in the middle of a parenthetical grouping, an error is signaled.
2392 If the beginning or end is reached between groupings
2393 but before count is used up, nil is returned. */)
2395 Lisp_Object from
, count
;
2397 CHECK_NUMBER (from
);
2398 CHECK_NUMBER (count
);
2400 return scan_lists (XINT (from
), XINT (count
), 0, 1);
2403 DEFUN ("backward-prefix-chars", Fbackward_prefix_chars
, Sbackward_prefix_chars
,
2405 doc
: /* Move point backward over any number of chars with prefix syntax.
2406 This includes chars with "quote" or "prefix" syntax (' or p). */)
2411 int opoint_byte
= PT_BYTE
;
2413 int pos_byte
= PT_BYTE
;
2418 SET_PT_BOTH (opoint
, opoint_byte
);
2423 SETUP_SYNTAX_TABLE (pos
, -1);
2425 DEC_BOTH (pos
, pos_byte
);
2427 while (!char_quoted (pos
, pos_byte
)
2428 /* Previous statement updates syntax table. */
2429 && ((c
= FETCH_CHAR (pos_byte
), SYNTAX (c
) == Squote
)
2430 || SYNTAX_PREFIX (c
)))
2433 opoint_byte
= pos_byte
;
2436 DEC_BOTH (pos
, pos_byte
);
2439 SET_PT_BOTH (opoint
, opoint_byte
);
2444 /* Parse forward from FROM / FROM_BYTE to END,
2445 assuming that FROM has state OLDSTATE (nil means FROM is start of function),
2446 and return a description of the state of the parse at END.
2447 If STOPBEFORE is nonzero, stop at the start of an atom.
2448 If COMMENTSTOP is 1, stop at the start of a comment.
2449 If COMMENTSTOP is -1, stop at the start or end of a comment,
2450 after the beginning of a string, or after the end of a string. */
2453 scan_sexps_forward (stateptr
, from
, from_byte
, end
, targetdepth
,
2454 stopbefore
, oldstate
, commentstop
)
2455 struct lisp_parse_state
*stateptr
;
2457 int end
, targetdepth
, stopbefore
;
2458 Lisp_Object oldstate
;
2461 struct lisp_parse_state state
;
2463 register enum syntaxcode code
;
2466 struct level
{ int last
, prev
; };
2467 struct level levelstart
[100];
2468 register struct level
*curlevel
= levelstart
;
2469 struct level
*endlevel
= levelstart
+ 100;
2470 register int depth
; /* Paren depth of current scanning location.
2471 level - levelstart equals this except
2472 when the depth becomes negative. */
2473 int mindepth
; /* Lowest DEPTH value seen. */
2474 int start_quoted
= 0; /* Nonzero means starting after a char quote */
2476 int prev_from
; /* Keep one character before FROM. */
2478 int prev_from_syntax
;
2479 int boundary_stop
= commentstop
== -1;
2482 int out_bytepos
, out_charpos
;
2486 prev_from_byte
= from_byte
;
2488 DEC_BOTH (prev_from
, prev_from_byte
);
2490 /* Use this macro instead of `from++'. */
2492 do { prev_from = from; \
2493 prev_from_byte = from_byte; \
2495 = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte)); \
2496 INC_BOTH (from, from_byte); \
2497 UPDATE_SYNTAX_TABLE_FORWARD (from); \
2503 if (NILP (oldstate
))
2506 state
.instring
= -1;
2507 state
.incomment
= 0;
2508 state
.comstyle
= 0; /* comment style a by default. */
2509 state
.comstr_start
= -1; /* no comment/string seen. */
2513 tem
= Fcar (oldstate
);
2519 oldstate
= Fcdr (oldstate
);
2520 oldstate
= Fcdr (oldstate
);
2521 oldstate
= Fcdr (oldstate
);
2522 tem
= Fcar (oldstate
);
2523 /* Check whether we are inside string_fence-style string: */
2524 state
.instring
= (!NILP (tem
)
2525 ? (INTEGERP (tem
) ? XINT (tem
) : ST_STRING_STYLE
)
2528 oldstate
= Fcdr (oldstate
);
2529 tem
= Fcar (oldstate
);
2530 state
.incomment
= (!NILP (tem
)
2531 ? (INTEGERP (tem
) ? XINT (tem
) : -1)
2534 oldstate
= Fcdr (oldstate
);
2535 tem
= Fcar (oldstate
);
2536 start_quoted
= !NILP (tem
);
2538 /* if the eighth element of the list is nil, we are in comment
2539 style a. If it is non-nil, we are in comment style b */
2540 oldstate
= Fcdr (oldstate
);
2541 oldstate
= Fcdr (oldstate
);
2542 tem
= Fcar (oldstate
);
2543 state
.comstyle
= NILP (tem
) ? 0 : (EQ (tem
, Qsyntax_table
)
2544 ? ST_COMMENT_STYLE
: 1);
2546 oldstate
= Fcdr (oldstate
);
2547 tem
= Fcar (oldstate
);
2548 state
.comstr_start
= NILP (tem
) ? -1 : XINT (tem
) ;
2549 oldstate
= Fcdr (oldstate
);
2550 tem
= Fcar (oldstate
);
2551 while (!NILP (tem
)) /* >= second enclosing sexps. */
2553 /* curlevel++->last ran into compiler bug on Apollo */
2554 curlevel
->last
= XINT (Fcar (tem
));
2555 if (++curlevel
== endlevel
)
2556 curlevel
--; /* error ("Nesting too deep for parser"); */
2557 curlevel
->prev
= -1;
2558 curlevel
->last
= -1;
2565 curlevel
->prev
= -1;
2566 curlevel
->last
= -1;
2568 SETUP_SYNTAX_TABLE (prev_from
, 1);
2569 prev_from_syntax
= SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte
));
2570 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2572 /* Enter the loop at a place appropriate for initial state. */
2574 if (state
.incomment
)
2575 goto startincomment
;
2576 if (state
.instring
>= 0)
2578 nofence
= state
.instring
!= ST_STRING_STYLE
;
2580 goto startquotedinstring
;
2583 else if (start_quoted
)
2586 #if 0 /* This seems to be redundant with the identical code above. */
2587 SETUP_SYNTAX_TABLE (prev_from
, 1);
2588 prev_from_syntax
= SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte
));
2589 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2595 code
= prev_from_syntax
& 0xff;
2597 if (code
== Scomment
)
2599 state
.comstyle
= SYNTAX_FLAGS_COMMENT_STYLE (prev_from_syntax
);
2600 state
.incomment
= (SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax
) ?
2602 state
.comstr_start
= prev_from
;
2604 else if (code
== Scomment_fence
)
2606 /* Record the comment style we have entered so that only
2607 the comment-end sequence of the same style actually
2608 terminates the comment section. */
2609 state
.comstyle
= ST_COMMENT_STYLE
;
2610 state
.incomment
= -1;
2611 state
.comstr_start
= prev_from
;
2614 else if (from
< end
)
2615 if (SYNTAX_FLAGS_COMSTART_FIRST (prev_from_syntax
))
2616 if (c1
= FETCH_CHAR (from_byte
),
2617 SYNTAX_COMSTART_SECOND (c1
))
2618 /* Duplicate code to avoid a complex if-expression
2619 which causes trouble for the SGI compiler. */
2621 /* Record the comment style we have entered so that only
2622 the comment-end sequence of the same style actually
2623 terminates the comment section. */
2624 state
.comstyle
= SYNTAX_COMMENT_STYLE (FETCH_CHAR (from_byte
));
2625 comnested
= SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax
);
2626 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2627 state
.incomment
= comnested
? 1 : -1;
2628 state
.comstr_start
= prev_from
;
2633 if (SYNTAX_FLAGS_PREFIX (prev_from_syntax
))
2635 switch (SWITCH_ENUM_CAST (code
))
2639 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2640 curlevel
->last
= prev_from
;
2642 if (from
== end
) goto endquoted
;
2645 /* treat following character as a word constituent */
2648 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2649 curlevel
->last
= prev_from
;
2653 /* Some compilers can't handle this inside the switch. */
2654 temp
= SYNTAX (FETCH_CHAR (from_byte
));
2660 if (from
== end
) goto endquoted
;
2672 curlevel
->prev
= curlevel
->last
;
2676 if (commentstop
|| boundary_stop
) goto done
;
2678 /* The (from == BEGV) test was to enter the loop in the middle so
2679 that we find a 2-char comment ender even if we start in the
2680 middle of it. We don't want to do that if we're just at the
2681 beginning of the comment (think of (*) ... (*)). */
2682 found
= forw_comment (from
, from_byte
, end
,
2683 state
.incomment
, state
.comstyle
,
2684 (from
== BEGV
|| from
< state
.comstr_start
+ 3)
2685 ? 0 : prev_from_syntax
,
2686 &out_charpos
, &out_bytepos
, &state
.incomment
);
2687 from
= out_charpos
; from_byte
= out_bytepos
;
2688 /* Beware! prev_from and friends are invalid now.
2689 Luckily, the `done' doesn't use them and the INC_FROM
2690 sets them to a sane value without looking at them. */
2691 if (!found
) goto done
;
2693 state
.incomment
= 0;
2694 state
.comstyle
= 0; /* reset the comment style */
2695 if (boundary_stop
) goto done
;
2699 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2701 /* curlevel++->last ran into compiler bug on Apollo */
2702 curlevel
->last
= prev_from
;
2703 if (++curlevel
== endlevel
)
2704 curlevel
--; /* error ("Nesting too deep for parser"); */
2705 curlevel
->prev
= -1;
2706 curlevel
->last
= -1;
2707 if (targetdepth
== depth
) goto done
;
2712 if (depth
< mindepth
)
2714 if (curlevel
!= levelstart
)
2716 curlevel
->prev
= curlevel
->last
;
2717 if (targetdepth
== depth
) goto done
;
2722 state
.comstr_start
= from
- 1;
2723 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2724 curlevel
->last
= prev_from
;
2725 state
.instring
= (code
== Sstring
2726 ? (FETCH_CHAR (prev_from_byte
))
2728 if (boundary_stop
) goto done
;
2731 nofence
= state
.instring
!= ST_STRING_STYLE
;
2737 if (from
>= end
) goto done
;
2738 c
= FETCH_CHAR (from_byte
);
2739 /* Some compilers can't handle this inside the switch. */
2742 /* Check TEMP here so that if the char has
2743 a syntax-table property which says it is NOT
2744 a string character, it does not end the string. */
2745 if (nofence
&& c
== state
.instring
&& temp
== Sstring
)
2751 if (!nofence
) goto string_end
;
2756 startquotedinstring
:
2757 if (from
>= end
) goto endquoted
;
2763 state
.instring
= -1;
2764 curlevel
->prev
= curlevel
->last
;
2766 if (boundary_stop
) goto done
;
2775 stop
: /* Here if stopping before start of sexp. */
2776 from
= prev_from
; /* We have just fetched the char that starts it; */
2777 goto done
; /* but return the position before it. */
2782 state
.depth
= depth
;
2783 state
.mindepth
= mindepth
;
2784 state
.thislevelstart
= curlevel
->prev
;
2785 state
.prevlevelstart
2786 = (curlevel
== levelstart
) ? -1 : (curlevel
- 1)->last
;
2787 state
.location
= from
;
2788 state
.levelstarts
= Qnil
;
2789 while (--curlevel
>= levelstart
)
2790 state
.levelstarts
= Fcons (make_number (curlevel
->last
),
2797 DEFUN ("parse-partial-sexp", Fparse_partial_sexp
, Sparse_partial_sexp
, 2, 6, 0,
2798 doc
: /* Parse Lisp syntax starting at FROM until TO; return status of parse at TO.
2799 Parsing stops at TO or when certain criteria are met;
2800 point is set to where parsing stops.
2801 If fifth arg OLDSTATE is omitted or nil,
2802 parsing assumes that FROM is the beginning of a function.
2803 Value is a list of ten elements describing final state of parsing:
2805 1. character address of start of innermost containing list; nil if none.
2806 2. character address of start of last complete sexp terminated.
2807 3. non-nil if inside a string.
2808 (it is the character that will terminate the string,
2809 or t if the string should be terminated by a generic string delimiter.)
2810 4. nil if outside a comment, t if inside a non-nestable comment,
2811 else an integer (the current comment nesting).
2812 5. t if following a quote character.
2813 6. the minimum paren-depth encountered during this scan.
2814 7. t if in a comment of style b; symbol `syntax-table' if the comment
2815 should be terminated by a generic comment delimiter.
2816 8. character address of start of comment or string; nil if not in one.
2817 9. Intermediate data for continuation of parsing (subject to change).
2818 If third arg TARGETDEPTH is non-nil, parsing stops if the depth
2819 in parentheses becomes equal to TARGETDEPTH.
2820 Fourth arg STOPBEFORE non-nil means stop when come to
2821 any character that starts a sexp.
2822 Fifth arg OLDSTATE is a nine-element list like what this function returns.
2823 It is used to initialize the state of the parse. Elements number 1, 2, 6
2824 and 8 are ignored; you can leave off element 8 (the last) entirely.
2825 Sixth arg COMMENTSTOP non-nil means stop at the start of a comment.
2826 If it is symbol `syntax-table', stop after the start of a comment or a
2827 string, or after end of a comment or a string. */)
2828 (from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
)
2829 Lisp_Object from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
;
2831 struct lisp_parse_state state
;
2834 if (!NILP (targetdepth
))
2836 CHECK_NUMBER (targetdepth
);
2837 target
= XINT (targetdepth
);
2840 target
= -100000; /* We won't reach this depth */
2842 validate_region (&from
, &to
);
2843 scan_sexps_forward (&state
, XINT (from
), CHAR_TO_BYTE (XINT (from
)),
2845 target
, !NILP (stopbefore
), oldstate
,
2847 ? 0 : (EQ (commentstop
, Qsyntax_table
) ? -1 : 1)));
2849 SET_PT (state
.location
);
2851 return Fcons (make_number (state
.depth
),
2852 Fcons (state
.prevlevelstart
< 0 ? Qnil
: make_number (state
.prevlevelstart
),
2853 Fcons (state
.thislevelstart
< 0 ? Qnil
: make_number (state
.thislevelstart
),
2854 Fcons (state
.instring
>= 0
2855 ? (state
.instring
== ST_STRING_STYLE
2856 ? Qt
: make_number (state
.instring
)) : Qnil
,
2857 Fcons (state
.incomment
< 0 ? Qt
:
2858 (state
.incomment
== 0 ? Qnil
:
2859 make_number (state
.incomment
)),
2860 Fcons (state
.quoted
? Qt
: Qnil
,
2861 Fcons (make_number (state
.mindepth
),
2862 Fcons ((state
.comstyle
2863 ? (state
.comstyle
== ST_COMMENT_STYLE
2864 ? Qsyntax_table
: Qt
) :
2866 Fcons (((state
.incomment
2867 || (state
.instring
>= 0))
2868 ? make_number (state
.comstr_start
)
2870 Fcons (state
.levelstarts
, Qnil
))))))))));
2879 /* This has to be done here, before we call Fmake_char_table. */
2880 Qsyntax_table
= intern ("syntax-table");
2881 staticpro (&Qsyntax_table
);
2883 /* Intern this now in case it isn't already done.
2884 Setting this variable twice is harmless.
2885 But don't staticpro it here--that is done in alloc.c. */
2886 Qchar_table_extra_slots
= intern ("char-table-extra-slots");
2888 /* Create objects which can be shared among syntax tables. */
2889 Vsyntax_code_object
= Fmake_vector (make_number (Smax
), Qnil
);
2890 for (i
= 0; i
< XVECTOR (Vsyntax_code_object
)->size
; i
++)
2891 XVECTOR (Vsyntax_code_object
)->contents
[i
]
2892 = Fcons (make_number (i
), Qnil
);
2894 /* Now we are ready to set up this property, so we can
2895 create syntax tables. */
2896 Fput (Qsyntax_table
, Qchar_table_extra_slots
, make_number (0));
2898 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Swhitespace
];
2900 Vstandard_syntax_table
= Fmake_char_table (Qsyntax_table
, temp
);
2902 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Sword
];
2903 for (i
= 'a'; i
<= 'z'; i
++)
2904 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2905 for (i
= 'A'; i
<= 'Z'; i
++)
2906 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2907 for (i
= '0'; i
<= '9'; i
++)
2908 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2910 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '$', temp
);
2911 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '%', temp
);
2913 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '(',
2914 Fcons (make_number (Sopen
), make_number (')')));
2915 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, ')',
2916 Fcons (make_number (Sclose
), make_number ('(')));
2917 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '[',
2918 Fcons (make_number (Sopen
), make_number (']')));
2919 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, ']',
2920 Fcons (make_number (Sclose
), make_number ('[')));
2921 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '{',
2922 Fcons (make_number (Sopen
), make_number ('}')));
2923 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '}',
2924 Fcons (make_number (Sclose
), make_number ('{')));
2925 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '"',
2926 Fcons (make_number ((int) Sstring
), Qnil
));
2927 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '\\',
2928 Fcons (make_number ((int) Sescape
), Qnil
));
2930 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Ssymbol
];
2931 for (i
= 0; i
< 10; i
++)
2933 c
= "_-+*/&|<>="[i
];
2934 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, c
, temp
);
2937 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Spunct
];
2938 for (i
= 0; i
< 12; i
++)
2940 c
= ".,;:?!#@~^'`"[i
];
2941 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, c
, temp
);
2944 /* All multibyte characters have syntax `word' by default. */
2945 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Sword
];
2946 char_table_set_range (Vstandard_syntax_table
, 0x80, MAX_CHAR
, temp
);
2952 Qsyntax_table_p
= intern ("syntax-table-p");
2953 staticpro (&Qsyntax_table_p
);
2955 staticpro (&Vsyntax_code_object
);
2957 Qscan_error
= intern ("scan-error");
2958 staticpro (&Qscan_error
);
2959 Fput (Qscan_error
, Qerror_conditions
,
2960 Fcons (Qscan_error
, Fcons (Qerror
, Qnil
)));
2961 Fput (Qscan_error
, Qerror_message
,
2962 build_string ("Scan error"));
2964 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments
,
2965 doc
: /* Non-nil means `forward-sexp', etc., should treat comments as whitespace. */);
2967 DEFVAR_BOOL ("parse-sexp-lookup-properties", &parse_sexp_lookup_properties
,
2968 doc
: /* Non-nil means `forward-sexp', etc., obey `syntax-table' property.
2969 Otherwise, that text property is simply ignored.
2970 See the info node `(elisp)Syntax Properties' for a description of the
2971 `syntax-table' property. */);
2973 words_include_escapes
= 0;
2974 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes
,
2975 doc
: /* Non-nil means `forward-word', etc., should treat escape chars part of words. */);
2977 DEFVAR_BOOL ("multibyte-syntax-as-symbol", &multibyte_syntax_as_symbol
,
2978 doc
: /* Non-nil means `scan-sexps' treats all multibyte characters as symbol. */);
2979 multibyte_syntax_as_symbol
= 0;
2981 DEFVAR_BOOL ("open-paren-in-column-0-is-defun-start",
2982 &open_paren_in_column_0_is_defun_start
,
2983 doc
: /* Non-nil means an open paren in column 0 denotes the start of a defun. */);
2984 open_paren_in_column_0_is_defun_start
= 1;
2986 defsubr (&Ssyntax_table_p
);
2987 defsubr (&Ssyntax_table
);
2988 defsubr (&Sstandard_syntax_table
);
2989 defsubr (&Scopy_syntax_table
);
2990 defsubr (&Sset_syntax_table
);
2991 defsubr (&Schar_syntax
);
2992 defsubr (&Smatching_paren
);
2993 defsubr (&Sstring_to_syntax
);
2994 defsubr (&Smodify_syntax_entry
);
2995 defsubr (&Sinternal_describe_syntax_value
);
2997 defsubr (&Sforward_word
);
2999 defsubr (&Sskip_chars_forward
);
3000 defsubr (&Sskip_chars_backward
);
3001 defsubr (&Sskip_syntax_forward
);
3002 defsubr (&Sskip_syntax_backward
);
3004 defsubr (&Sforward_comment
);
3005 defsubr (&Sscan_lists
);
3006 defsubr (&Sscan_sexps
);
3007 defsubr (&Sbackward_prefix_chars
);
3008 defsubr (&Sparse_partial_sexp
);