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. */
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 /* 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. */
843 syntax_parent_lookup (table
, character
)
851 table
= XCHAR_TABLE (table
)->parent
;
855 value
= XCHAR_TABLE (table
)->contents
[character
];
861 DEFUN ("char-syntax", Fchar_syntax
, Schar_syntax
, 1, 1, 0,
862 doc
: /* Return the syntax code of CHARACTER, described by a character.
863 For example, if CHARACTER is a word constituent,
864 the character `w' is returned.
865 The characters that correspond to various syntax codes
866 are listed in the documentation of `modify-syntax-entry'. */)
868 Lisp_Object character
;
871 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
873 gl_state
.use_global
= 0;
874 CHECK_NUMBER (character
);
875 char_int
= XINT (character
);
876 return make_number (syntax_code_spec
[(int) SYNTAX (char_int
)]);
879 DEFUN ("matching-paren", Fmatching_paren
, Smatching_paren
, 1, 1, 0,
880 doc
: /* Return the matching parenthesis of CHARACTER, or nil if none. */)
882 Lisp_Object character
;
885 gl_state
.current_syntax_table
= current_buffer
->syntax_table
;
886 gl_state
.use_global
= 0;
887 CHECK_NUMBER (character
);
888 char_int
= XINT (character
);
889 code
= SYNTAX (char_int
);
890 if (code
== Sopen
|| code
== Sclose
)
891 return SYNTAX_MATCH (char_int
);
895 DEFUN ("string-to-syntax", Fstring_to_syntax
, Sstring_to_syntax
, 1, 1, 0,
896 doc
: /* Convert a syntax specification STRING into syntax cell form.
897 STRING 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'
904 register unsigned char *p
;
905 register enum syntaxcode code
;
909 CHECK_STRING (string
);
911 p
= XSTRING (string
)->data
;
912 code
= (enum syntaxcode
) syntax_spec_code
[*p
++];
913 if (((int) code
& 0377) == 0377)
914 error ("invalid syntax description letter: %c", p
[-1]);
916 if (code
== Sinherit
)
922 int character
= (STRING_CHAR_AND_LENGTH
923 (p
, STRING_BYTES (XSTRING (string
)) - 1, len
));
924 XSETINT (match
, character
);
925 if (XFASTINT (match
) == ' ')
965 if (val
< XVECTOR (Vsyntax_code_object
)->size
&& NILP (match
))
966 return XVECTOR (Vsyntax_code_object
)->contents
[val
];
968 /* Since we can't use a shared object, let's make a new one. */
969 return Fcons (make_number (val
), match
);
972 /* I really don't know why this is interactive
973 help-form should at least be made useful whilst reading the second arg. */
974 DEFUN ("modify-syntax-entry", Fmodify_syntax_entry
, Smodify_syntax_entry
, 2, 3,
975 "cSet syntax for character: \nsSet syntax for %s to: ",
976 doc
: /* Set syntax for character CHAR according to string NEWENTRY.
977 The syntax is changed only for table SYNTAX_TABLE, which defaults to
978 the current buffer's syntax table.
979 The 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.
989 Only single-character comment start and end sequences are represented thus.
990 Two-character sequences are represented as described below.
991 The second character of NEWENTRY is the matching parenthesis,
992 used only if the first character is `(' or `)'.
993 Any additional characters are flags.
994 Defined flags are the characters 1, 2, 3, 4, b, p, and n.
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.
1000 There can be up to two orthogonal comment sequences. This is to support
1001 language modes such as C++. By default, all comment sequences are of style
1002 a, but you can set the comment sequence style to b (on the second character
1003 of a comment-start, or the first character of a comment-end sequence) using
1005 b means CHAR is part of comment sequence b.
1006 n means CHAR is part of a nestable comment sequence.
1008 p means CHAR is a prefix character for `backward-prefix-chars';
1009 such characters are treated as whitespace when they occur
1010 between expressions.
1011 usage: (modify-syntax-entry CHAR NEWENTRY &optional SYNTAX-TABLE) */)
1012 (c
, newentry
, syntax_table
)
1013 Lisp_Object c
, newentry
, syntax_table
;
1017 if (NILP (syntax_table
))
1018 syntax_table
= current_buffer
->syntax_table
;
1020 check_syntax_table (syntax_table
);
1022 SET_RAW_SYNTAX_ENTRY (syntax_table
, XINT (c
), Fstring_to_syntax (newentry
));
1026 /* Dump syntax table to buffer in human-readable format */
1028 DEFUN ("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. */)
1034 register enum syntaxcode code
;
1035 char desc
, start1
, start2
, end1
, end2
, prefix
, comstyle
, comnested
;
1037 Lisp_Object first
, match_lisp
, value
= syntax
;
1041 insert_string ("default");
1045 if (CHAR_TABLE_P (value
))
1047 insert_string ("deeper char-table ...");
1053 insert_string ("invalid");
1057 first
= XCAR (value
);
1058 match_lisp
= XCDR (value
);
1060 if (!INTEGERP (first
) || !(NILP (match_lisp
) || INTEGERP (match_lisp
)))
1062 insert_string ("invalid");
1066 code
= (enum syntaxcode
) (XINT (first
) & 0377);
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;
1073 comnested
= (XINT (first
) >> 22) & 1;
1075 if ((int) code
< 0 || (int) code
>= (int) Smax
)
1077 insert_string ("invalid");
1080 desc
= syntax_code_spec
[(int) code
];
1082 str
[0] = desc
, str
[1] = 0;
1085 if (NILP (match_lisp
))
1088 insert_char (XINT (match_lisp
));
1107 insert_string ("\twhich means: ");
1109 switch (SWITCH_ENUM_CAST (code
))
1112 insert_string ("whitespace"); break;
1114 insert_string ("punctuation"); break;
1116 insert_string ("word"); break;
1118 insert_string ("symbol"); break;
1120 insert_string ("open"); break;
1122 insert_string ("close"); break;
1124 insert_string ("prefix"); break;
1126 insert_string ("string"); break;
1128 insert_string ("math"); break;
1130 insert_string ("escape"); break;
1132 insert_string ("charquote"); break;
1134 insert_string ("comment"); break;
1136 insert_string ("endcomment"); break;
1138 insert_string ("inherit"); break;
1139 case Scomment_fence
:
1140 insert_string ("comment fence"); break;
1142 insert_string ("string fence"); break;
1144 insert_string ("invalid");
1148 if (!NILP (match_lisp
))
1150 insert_string (", matches ");
1151 insert_char (XINT (match_lisp
));
1155 insert_string (",\n\t is the first character of a comment-start sequence");
1157 insert_string (",\n\t is the second character of a comment-start sequence");
1160 insert_string (",\n\t is the first character of a comment-end sequence");
1162 insert_string (",\n\t is the second character of a comment-end sequence");
1164 insert_string (" (comment style b)");
1166 insert_string (" (nestable)");
1169 insert_string (",\n\t is a prefix character for `backward-prefix-chars'");
1174 int parse_sexp_ignore_comments
;
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. */
1181 scan_words (from
, count
)
1182 register int from
, count
;
1184 register int beg
= BEGV
;
1185 register int end
= ZV
;
1186 register int from_byte
= CHAR_TO_BYTE (from
);
1187 register enum syntaxcode code
;
1193 SETUP_SYNTAX_TABLE (from
, count
);
1204 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1205 ch0
= FETCH_CHAR (from_byte
);
1206 code
= SYNTAX (ch0
);
1207 INC_BOTH (from
, from_byte
);
1208 if (words_include_escapes
1209 && (code
== Sescape
|| code
== Scharquote
))
1214 /* Now CH0 is a character which begins a word and FROM is the
1215 position of the next character. */
1218 if (from
== end
) break;
1219 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1220 ch1
= FETCH_CHAR (from_byte
);
1221 code
= SYNTAX (ch1
);
1222 if (!(words_include_escapes
1223 && (code
== Sescape
|| code
== Scharquote
)))
1224 if (code
!= Sword
|| WORD_BOUNDARY_P (ch0
, ch1
))
1226 INC_BOTH (from
, from_byte
);
1240 DEC_BOTH (from
, from_byte
);
1241 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1242 ch1
= FETCH_CHAR (from_byte
);
1243 code
= SYNTAX (ch1
);
1244 if (words_include_escapes
1245 && (code
== Sescape
|| code
== Scharquote
))
1250 /* Now CH1 is a character which ends a word and FROM is the
1258 temp_byte
= dec_bytepos (from_byte
);
1259 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1260 ch0
= FETCH_CHAR (temp_byte
);
1261 code
= SYNTAX (ch0
);
1262 if (!(words_include_escapes
1263 && (code
== Sescape
|| code
== Scharquote
)))
1264 if (code
!= Sword
|| WORD_BOUNDARY_P (ch0
, ch1
))
1266 DEC_BOTH (from
, from_byte
);
1277 DEFUN ("forward-word", Fforward_word
, Sforward_word
, 1, 1, "p",
1278 doc
: /* Move point forward ARG words (backward if ARG is negative).
1280 If an edge of the buffer or a field boundary is reached, point is left there
1281 and the function returns nil. Field boundaries are not noticed if
1282 `inhibit-field-text-motion' is non-nil. */)
1287 CHECK_NUMBER (count
);
1289 val
= orig_val
= scan_words (PT
, XINT (count
));
1291 val
= XINT (count
) > 0 ? ZV
: BEGV
;
1293 /* Avoid jumping out of an input field. */
1294 val
= XFASTINT (Fconstrain_to_field (make_number (val
), make_number (PT
),
1298 return val
== orig_val
? Qt
: Qnil
;
1301 Lisp_Object
skip_chars ();
1303 DEFUN ("skip-chars-forward", Fskip_chars_forward
, Sskip_chars_forward
, 1, 2, 0,
1304 doc
: /* Move point forward, stopping before a char not in STRING, or at pos LIM.
1305 STRING is like the inside of a `[...]' in a regular expression
1306 except that `]' is never special and `\\' quotes `^', `-' or `\\'
1307 (but not as the end of a range; quoting is never needed there).
1308 Thus, with arg "a-zA-Z", this skips letters stopping before first nonletter.
1309 With arg "^a-zA-Z", skips nonletters stopping before first letter.
1310 Returns the distance traveled, either zero or positive. */)
1312 Lisp_Object string
, lim
;
1314 return skip_chars (1, 0, string
, lim
);
1317 DEFUN ("skip-chars-backward", Fskip_chars_backward
, Sskip_chars_backward
, 1, 2, 0,
1318 doc
: /* Move point backward, stopping after a char not in STRING, or at pos LIM.
1319 See `skip-chars-forward' for details.
1320 Returns the distance traveled, either zero or negative. */)
1322 Lisp_Object string
, lim
;
1324 return skip_chars (0, 0, string
, lim
);
1327 DEFUN ("skip-syntax-forward", Fskip_syntax_forward
, Sskip_syntax_forward
, 1, 2, 0,
1328 doc
: /* Move point forward across chars in specified syntax classes.
1329 SYNTAX is a string of syntax code characters.
1330 Stop before a char whose syntax is not in SYNTAX, or at position LIM.
1331 If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1332 This function returns the distance traveled, either zero or positive. */)
1334 Lisp_Object syntax
, lim
;
1336 return skip_chars (1, 1, syntax
, lim
);
1339 DEFUN ("skip-syntax-backward", Fskip_syntax_backward
, Sskip_syntax_backward
, 1, 2, 0,
1340 doc
: /* Move point backward across chars in specified syntax classes.
1341 SYNTAX is a string of syntax code characters.
1342 Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.
1343 If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1344 This function returns the distance traveled, either zero or negative. */)
1346 Lisp_Object syntax
, lim
;
1348 return skip_chars (0, 1, syntax
, lim
);
1352 skip_chars (forwardp
, syntaxp
, string
, lim
)
1353 int forwardp
, syntaxp
;
1354 Lisp_Object string
, lim
;
1356 register unsigned int c
;
1357 unsigned char fastmap
[0400];
1358 /* If SYNTAXP is 0, STRING may contain multi-byte form of characters
1359 of which codes don't fit in FASTMAP. In that case, set the
1360 ranges of characters in CHAR_RANGES. */
1362 int n_char_ranges
= 0;
1364 register int i
, i_byte
;
1365 int multibyte
= !NILP (current_buffer
->enable_multibyte_characters
);
1366 int string_multibyte
;
1371 CHECK_STRING (string
);
1372 char_ranges
= (int *) alloca (XSTRING (string
)->size
* (sizeof (int)) * 2);
1373 string_multibyte
= STRING_MULTIBYTE (string
);
1374 str
= XSTRING (string
)->data
;
1375 size_byte
= STRING_BYTES (XSTRING (string
));
1377 /* Adjust the multibyteness of the string to that of the buffer. */
1378 if (multibyte
!= string_multibyte
)
1383 nbytes
= count_size_as_multibyte (XSTRING (string
)->data
,
1384 XSTRING (string
)->size
);
1386 nbytes
= XSTRING (string
)->size
;
1387 if (nbytes
!= size_byte
)
1389 str
= (unsigned char *) alloca (nbytes
);
1390 copy_text (XSTRING (string
)->data
, str
, size_byte
,
1391 string_multibyte
, multibyte
);
1397 XSETINT (lim
, forwardp
? ZV
: BEGV
);
1399 CHECK_NUMBER_COERCE_MARKER (lim
);
1401 /* In any case, don't allow scan outside bounds of buffer. */
1402 if (XINT (lim
) > ZV
)
1403 XSETFASTINT (lim
, ZV
);
1404 if (XINT (lim
) < BEGV
)
1405 XSETFASTINT (lim
, BEGV
);
1407 bzero (fastmap
, sizeof fastmap
);
1411 if (i_byte
< size_byte
1412 && XSTRING (string
)->data
[0] == '^')
1414 negate
= 1; i_byte
++;
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. */
1421 while (i_byte
< size_byte
)
1423 c
= STRING_CHAR_AND_LENGTH (str
+ i_byte
, size_byte
- i_byte
, len
);
1427 fastmap
[syntax_spec_code
[c
& 0377]] = 1;
1432 if (i_byte
== size_byte
)
1435 c
= STRING_CHAR_AND_LENGTH (str
+i_byte
, size_byte
-i_byte
, len
);
1438 if (i_byte
< size_byte
1439 && str
[i_byte
] == '-')
1443 /* Skip over the dash. */
1446 if (i_byte
== size_byte
)
1449 /* Get the end of the range. */
1450 c2
=STRING_CHAR_AND_LENGTH (str
+i_byte
, size_byte
-i_byte
, len
);
1453 if (SINGLE_BYTE_CHAR_P (c
))
1455 if (! SINGLE_BYTE_CHAR_P (c2
))
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. */
1463 int charset
= CHAR_CHARSET (c2
);
1464 int c1
= MAKE_CHAR (charset
, 0, 0);
1466 char_ranges
[n_char_ranges
++] = c1
;
1467 char_ranges
[n_char_ranges
++] = c2
;
1476 else if (c
<= c2
) /* Both C and C2 are multibyte char. */
1478 char_ranges
[n_char_ranges
++] = c
;
1479 char_ranges
[n_char_ranges
++] = c2
;
1484 if (SINGLE_BYTE_CHAR_P (c
))
1488 char_ranges
[n_char_ranges
++] = c
;
1489 char_ranges
[n_char_ranges
++] = c
;
1495 /* If ^ was the first character, complement the fastmap. */
1497 for (i
= 0; i
< sizeof fastmap
; i
++)
1501 int start_point
= PT
;
1503 int pos_byte
= PT_BYTE
;
1508 SETUP_SYNTAX_TABLE (pos
, forwardp
? 1 : -1);
1513 if (pos
< XINT (lim
))
1514 while (fastmap
[(int) SYNTAX (FETCH_CHAR (pos_byte
))])
1516 /* Since we already checked for multibyteness,
1517 avoid using INC_BOTH which checks again. */
1520 if (pos
>= XINT (lim
))
1522 UPDATE_SYNTAX_TABLE_FORWARD (pos
);
1527 while (pos
< XINT (lim
)
1528 && fastmap
[(int) SYNTAX (FETCH_BYTE (pos
))])
1531 UPDATE_SYNTAX_TABLE_FORWARD (pos
);
1539 while (pos
> XINT (lim
))
1541 int savepos
= pos_byte
;
1542 /* Since we already checked for multibyteness,
1543 avoid using DEC_BOTH which checks again. */
1546 UPDATE_SYNTAX_TABLE_BACKWARD (pos
);
1547 if (!fastmap
[(int) SYNTAX (FETCH_CHAR (pos_byte
))])
1557 if (pos
> XINT (lim
))
1558 while (fastmap
[(int) SYNTAX (FETCH_BYTE (pos
- 1))])
1561 if (pos
<= XINT (lim
))
1563 UPDATE_SYNTAX_TABLE_BACKWARD (pos
- 1);
1573 while (pos
< XINT (lim
))
1575 c
= FETCH_MULTIBYTE_CHAR (pos_byte
);
1576 if (SINGLE_BYTE_CHAR_P (c
))
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
1589 /* The following code do the right thing even if
1590 n_char_ranges is zero (i.e. no data in
1592 for (i
= 0; i
< n_char_ranges
; i
+= 2)
1593 if (c
>= char_ranges
[i
] && c
<= char_ranges
[i
+ 1])
1595 if (!(negate
^ (i
< n_char_ranges
)))
1598 INC_BOTH (pos
, pos_byte
);
1601 while (pos
< XINT (lim
) && fastmap
[FETCH_BYTE (pos
)])
1607 while (pos
> XINT (lim
))
1609 int prev_pos_byte
= pos_byte
;
1611 DEC_POS (prev_pos_byte
);
1612 c
= FETCH_MULTIBYTE_CHAR (prev_pos_byte
);
1613 if (SINGLE_BYTE_CHAR_P (c
))
1620 /* See the comment in the previous similar code. */
1621 for (i
= 0; i
< n_char_ranges
; i
+= 2)
1622 if (c
>= char_ranges
[i
] && c
<= char_ranges
[i
+ 1])
1624 if (!(negate
^ (i
< n_char_ranges
)))
1628 pos_byte
= prev_pos_byte
;
1631 while (pos
> XINT (lim
) && fastmap
[FETCH_BYTE (pos
- 1)])
1636 #if 0 /* Not needed now that a position in mid-character
1637 cannot be specified in Lisp. */
1639 /* INC_POS or DEC_POS might have moved POS over LIM. */
1640 && (forwardp
? (pos
> XINT (lim
)) : (pos
< XINT (lim
))))
1647 SET_PT_BOTH (pos
, pos_byte
);
1650 return make_number (PT
- start_point
);
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).
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.
1672 The comment end is the last character of the comment rather than the
1673 character just after the comment.
1675 Global syntax data is assumed to initially be valid for FROM and
1676 remains valid for forward search starting at the returned position. */
1679 forw_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
;
1686 register enum syntaxcode code
;
1687 register int syntax
;
1689 if (nesting
<= 0) nesting
= -1;
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
;
1700 *incomment_ptr
= nesting
;
1701 *charpos_ptr
= from
;
1702 *bytepos_ptr
= from_byte
;
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
1710 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ?
1711 (nesting
> 0 && --nesting
== 0) : nesting
< 0))
1712 /* we have encountered a comment end of the same style
1713 as the comment sequence which began this comment
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
1724 && SYNTAX_FLAGS_COMMENT_NESTED (syntax
)
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 */
1729 INC_BOTH (from
, from_byte
);
1730 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1733 if (from
< stop
&& SYNTAX_FLAGS_COMEND_FIRST (syntax
)
1734 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
1735 && (c1
= FETCH_CHAR (from_byte
),
1736 SYNTAX_COMEND_SECOND (c1
))
1737 && ((SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ||
1738 SYNTAX_COMMENT_NESTED (c1
)) ? nesting
> 0 : nesting
< 0))
1741 /* we have encountered a comment end of the same style
1742 as the comment sequence which began this comment
1747 INC_BOTH (from
, from_byte
);
1748 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1753 && SYNTAX_FLAGS_COMSTART_FIRST (syntax
)
1754 && (c1
= FETCH_CHAR (from_byte
),
1755 SYNTAX_COMMENT_STYLE (c1
) == style
1756 && SYNTAX_COMSTART_SECOND (c1
))
1757 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ||
1758 SYNTAX_COMMENT_NESTED (c1
)))
1759 /* we have encountered a nested comment of the same style
1760 as the comment sequence which began this comment
1763 INC_BOTH (from
, from_byte
);
1764 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1768 *charpos_ptr
= from
;
1769 *bytepos_ptr
= from_byte
;
1773 DEFUN ("forward-comment", Fforward_comment
, Sforward_comment
, 1, 1, 0,
1774 doc
: /* Move forward across up to N comments. If N is negative, move backward.
1775 Stop scanning if we find something other than a comment or whitespace.
1776 Set point to where scanning stops.
1777 If N comments are found as expected, with nothing except whitespace
1778 between them, return t; otherwise return nil. */)
1786 register enum syntaxcode code
;
1787 int comstyle
= 0; /* style of comment encountered */
1788 int comnested
= 0; /* whether the comment is nestable or not */
1791 int out_charpos
, out_bytepos
;
1794 CHECK_NUMBER (count
);
1795 count1
= XINT (count
);
1796 stop
= count1
> 0 ? ZV
: BEGV
;
1802 from_byte
= PT_BYTE
;
1804 SETUP_SYNTAX_TABLE (from
, count1
);
1813 SET_PT_BOTH (from
, from_byte
);
1817 c
= FETCH_CHAR (from_byte
);
1819 comstart_first
= SYNTAX_COMSTART_FIRST (c
);
1820 comnested
= SYNTAX_COMMENT_NESTED (c
);
1821 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1822 INC_BOTH (from
, from_byte
);
1823 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1824 if (from
< stop
&& comstart_first
1825 && (c1
= FETCH_CHAR (from_byte
),
1826 SYNTAX_COMSTART_SECOND (c1
)))
1828 /* We have encountered a comment start sequence and we
1829 are ignoring all text inside comments. We must record
1830 the comment style this sequence begins so that later,
1831 only a comment end of the same style actually ends
1832 the comment section. */
1834 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
1835 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
1836 INC_BOTH (from
, from_byte
);
1837 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1840 while (code
== Swhitespace
|| (code
== Sendcomment
&& c
== '\n'));
1842 if (code
== Scomment_fence
)
1843 comstyle
= ST_COMMENT_STYLE
;
1844 else if (code
!= Scomment
)
1847 DEC_BOTH (from
, from_byte
);
1848 SET_PT_BOTH (from
, from_byte
);
1851 /* We're at the start of a comment. */
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
;
1858 SET_PT_BOTH (from
, from_byte
);
1861 INC_BOTH (from
, from_byte
);
1862 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1863 /* We have skipped one comment. */
1875 SET_PT_BOTH (BEGV
, BEGV_BYTE
);
1880 DEC_BOTH (from
, from_byte
);
1881 /* char_quoted does UPDATE_SYNTAX_TABLE_BACKWARD (from). */
1882 quoted
= char_quoted (from
, from_byte
);
1883 c
= FETCH_CHAR (from_byte
);
1886 comnested
= SYNTAX_COMMENT_NESTED (c
);
1887 if (code
== Sendcomment
)
1888 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1889 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
1890 && prev_char_comend_first (from
, from_byte
)
1891 && !char_quoted (from
- 1, dec_bytepos (from_byte
)))
1893 /* We must record the comment style encountered so that
1894 later, we can match only the proper comment begin
1895 sequence of the same style. */
1896 DEC_BOTH (from
, from_byte
);
1898 /* Calling char_quoted, above, set up global syntax position
1899 at the new value of FROM. */
1900 c1
= FETCH_CHAR (from_byte
);
1901 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
1902 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
1905 if (code
== Scomment_fence
)
1907 /* Skip until first preceding unquoted comment_fence. */
1908 int found
= 0, ini
= from
, ini_byte
= from_byte
;
1912 DEC_BOTH (from
, from_byte
);
1915 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1916 c
= FETCH_CHAR (from_byte
);
1917 if (SYNTAX (c
) == Scomment_fence
1918 && !char_quoted (from
, from_byte
))
1926 from
= ini
; /* Set point to ini + 1. */
1927 from_byte
= ini_byte
;
1931 else if (code
== Sendcomment
)
1933 found
= back_comment (from
, from_byte
, stop
, comnested
, comstyle
,
1934 &out_charpos
, &out_bytepos
);
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. */
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
);
1954 /* We have skipped one comment. */
1955 from
= out_charpos
, from_byte
= out_bytepos
;
1959 else if (code
!= Swhitespace
|| quoted
)
1963 INC_BOTH (from
, from_byte
);
1964 SET_PT_BOTH (from
, from_byte
);
1972 SET_PT_BOTH (from
, from_byte
);
1977 /* Return syntax code of character C if C is a single byte character
1978 or `multibyte_symbol_p' is zero. Otherwise, return Ssymbol. */
1980 #define SYNTAX_WITH_MULTIBYTE_CHECK(c) \
1981 ((SINGLE_BYTE_CHAR_P (c) || !multibyte_symbol_p) \
1982 ? SYNTAX (c) : Ssymbol)
1985 scan_lists (from
, count
, depth
, sexpflag
)
1987 int count
, depth
, sexpflag
;
1990 register int stop
= count
> 0 ? ZV
: BEGV
;
1995 register enum syntaxcode code
, temp_code
;
1996 int min_depth
= depth
; /* Err out if depth gets less than this. */
1997 int comstyle
= 0; /* style of comment encountered */
1998 int comnested
= 0; /* whether the comment is nestable or not */
2000 int last_good
= from
;
2003 int out_bytepos
, out_charpos
;
2005 int multibyte_symbol_p
= sexpflag
&& multibyte_syntax_as_symbol
;
2007 if (depth
> 0) min_depth
= 0;
2009 if (from
> ZV
) from
= ZV
;
2010 if (from
< BEGV
) from
= BEGV
;
2012 from_byte
= CHAR_TO_BYTE (from
);
2017 SETUP_SYNTAX_TABLE (from
, count
);
2022 int comstart_first
, prefix
;
2023 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2024 c
= FETCH_CHAR (from_byte
);
2025 code
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2026 comstart_first
= SYNTAX_COMSTART_FIRST (c
);
2027 comnested
= SYNTAX_COMMENT_NESTED (c
);
2028 comstyle
= SYNTAX_COMMENT_STYLE (c
);
2029 prefix
= SYNTAX_PREFIX (c
);
2030 if (depth
== min_depth
)
2032 INC_BOTH (from
, from_byte
);
2033 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2034 if (from
< stop
&& comstart_first
2035 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from_byte
))
2036 && parse_sexp_ignore_comments
)
2038 /* we have encountered a comment start sequence and we
2039 are ignoring all text inside comments. We must record
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 */
2044 c1
= FETCH_CHAR (from_byte
);
2045 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2046 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2047 INC_BOTH (from
, from_byte
);
2048 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2054 switch (SWITCH_ENUM_CAST (code
))
2058 if (from
== stop
) goto lose
;
2059 INC_BOTH (from
, from_byte
);
2060 /* treat following character as a word constituent */
2063 if (depth
|| !sexpflag
) break;
2064 /* This word counts as a sexp; return at end of it. */
2067 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2069 /* Some compilers can't handle this inside the switch. */
2070 c
= FETCH_CHAR (from_byte
);
2071 temp
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2076 INC_BOTH (from
, from_byte
);
2077 if (from
== stop
) goto lose
;
2086 INC_BOTH (from
, from_byte
);
2090 case Scomment_fence
:
2091 comstyle
= ST_COMMENT_STYLE
;
2094 if (!parse_sexp_ignore_comments
) break;
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
;
2106 INC_BOTH (from
, from_byte
);
2107 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2113 if (from
!= stop
&& c
== FETCH_CHAR (from_byte
))
2115 INC_BOTH (from
, from_byte
);
2125 if (!++depth
) goto done
;
2130 if (!--depth
) goto done
;
2131 if (depth
< min_depth
)
2132 Fsignal (Qscan_error
,
2133 Fcons (build_string ("Containing expression ends prematurely"),
2134 Fcons (make_number (last_good
),
2135 Fcons (make_number (from
), Qnil
))));
2140 temp_pos
= dec_bytepos (from_byte
);
2141 stringterm
= FETCH_CHAR (temp_pos
);
2144 if (from
>= stop
) goto lose
;
2145 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2146 c
= FETCH_CHAR (from_byte
);
2149 && SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring
)
2150 : SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring_fence
)
2153 /* Some compilers can't handle this inside the switch. */
2154 temp
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2159 INC_BOTH (from
, from_byte
);
2161 INC_BOTH (from
, from_byte
);
2163 INC_BOTH (from
, from_byte
);
2164 if (!depth
&& sexpflag
) goto done
;
2169 /* Reached end of buffer. Error if within object, return nil if between */
2170 if (depth
) goto lose
;
2175 /* End of object reached */
2184 DEC_BOTH (from
, from_byte
);
2185 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2186 c
= FETCH_CHAR (from_byte
);
2187 code
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2188 if (depth
== min_depth
)
2191 comnested
= SYNTAX_COMMENT_NESTED (c
);
2192 if (code
== Sendcomment
)
2193 comstyle
= SYNTAX_COMMENT_STYLE (c
);
2194 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
2195 && prev_char_comend_first (from
, from_byte
)
2196 && parse_sexp_ignore_comments
)
2198 /* We must record the comment style encountered so that
2199 later, we can match only the proper comment begin
2200 sequence of the same style. */
2201 DEC_BOTH (from
, from_byte
);
2202 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2204 c1
= FETCH_CHAR (from_byte
);
2205 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2206 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2209 /* Quoting turns anything except a comment-ender
2210 into a word character. Note that this cannot be true
2211 if we decremented FROM in the if-statement above. */
2212 if (code
!= Sendcomment
&& char_quoted (from
, from_byte
))
2214 else if (SYNTAX_PREFIX (c
))
2217 switch (SWITCH_ENUM_CAST (code
))
2223 if (depth
|| !sexpflag
) break;
2224 /* This word counts as a sexp; count object finished
2225 after passing it. */
2228 temp_pos
= from_byte
;
2229 if (! NILP (current_buffer
->enable_multibyte_characters
))
2233 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2234 c1
= FETCH_CHAR (temp_pos
);
2235 temp_code
= SYNTAX_WITH_MULTIBYTE_CHECK (c1
);
2236 /* Don't allow comment-end to be quoted. */
2237 if (temp_code
== Sendcomment
)
2239 quoted
= char_quoted (from
- 1, temp_pos
);
2242 DEC_BOTH (from
, from_byte
);
2243 temp_pos
= dec_bytepos (temp_pos
);
2244 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2246 c1
= FETCH_CHAR (temp_pos
);
2247 temp_code
= SYNTAX_WITH_MULTIBYTE_CHECK (c1
);
2248 if (! (quoted
|| temp_code
== Sword
2249 || temp_code
== Ssymbol
2250 || temp_code
== Squote
))
2252 DEC_BOTH (from
, from_byte
);
2259 temp_pos
= dec_bytepos (from_byte
);
2260 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2261 if (from
!= stop
&& c
== FETCH_CHAR (temp_pos
))
2262 DEC_BOTH (from
, from_byte
);
2271 if (!++depth
) goto done2
;
2276 if (!--depth
) goto done2
;
2277 if (depth
< min_depth
)
2278 Fsignal (Qscan_error
,
2279 Fcons (build_string ("Containing expression ends prematurely"),
2280 Fcons (make_number (last_good
),
2281 Fcons (make_number (from
), Qnil
))));
2285 if (!parse_sexp_ignore_comments
)
2287 found
= back_comment (from
, from_byte
, stop
, comnested
, comstyle
,
2288 &out_charpos
, &out_bytepos
);
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 */
2296 from
= out_charpos
, from_byte
= out_bytepos
;
2299 case Scomment_fence
:
2303 DEC_BOTH (from
, from_byte
);
2304 if (from
== stop
) goto lose
;
2305 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2306 if (!char_quoted (from
, from_byte
)
2307 && (c
= FETCH_CHAR (from_byte
),
2308 SYNTAX_WITH_MULTIBYTE_CHECK (c
) == code
))
2311 if (code
== Sstring_fence
&& !depth
&& sexpflag
) goto done2
;
2315 stringterm
= FETCH_CHAR (from_byte
);
2318 if (from
== stop
) goto lose
;
2319 temp_pos
= from_byte
;
2320 if (! NILP (current_buffer
->enable_multibyte_characters
))
2324 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2325 if (!char_quoted (from
- 1, temp_pos
)
2326 && stringterm
== (c
= FETCH_CHAR (temp_pos
))
2327 && SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring
)
2329 DEC_BOTH (from
, from_byte
);
2331 DEC_BOTH (from
, from_byte
);
2332 if (!depth
&& sexpflag
) goto done2
;
2337 /* Reached start of buffer. Error if within object, return nil if between */
2338 if (depth
) goto lose
;
2349 XSETFASTINT (val
, from
);
2353 Fsignal (Qscan_error
,
2354 Fcons (build_string ("Unbalanced parentheses"),
2355 Fcons (make_number (last_good
),
2356 Fcons (make_number (from
), Qnil
))));
2361 DEFUN ("scan-lists", Fscan_lists
, Sscan_lists
, 3, 3, 0,
2362 doc
: /* Scan from character number FROM by COUNT lists.
2363 Returns the character number of the position thus found.
2365 If DEPTH is nonzero, paren depth begins counting from that value,
2366 only places where the depth in parentheses becomes zero
2367 are candidates for stopping; COUNT such places are counted.
2368 Thus, a positive value for DEPTH means go out levels.
2370 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2372 If the beginning or end of (the accessible part of) the buffer is reached
2373 and the depth is wrong, an error is signaled.
2374 If the depth is right but the count is not used up, nil is returned. */)
2375 (from
, count
, depth
)
2376 Lisp_Object from
, count
, depth
;
2378 CHECK_NUMBER (from
);
2379 CHECK_NUMBER (count
);
2380 CHECK_NUMBER (depth
);
2382 return scan_lists (XINT (from
), XINT (count
), XINT (depth
), 0);
2385 DEFUN ("scan-sexps", Fscan_sexps
, Sscan_sexps
, 2, 2, 0,
2386 doc
: /* Scan from character number FROM by COUNT balanced expressions.
2387 If COUNT is negative, scan backwards.
2388 Returns the character number of the position thus found.
2390 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2392 If the beginning or end of (the accessible part of) the buffer is reached
2393 in the middle of a parenthetical grouping, an error is signaled.
2394 If the beginning or end is reached between groupings
2395 but before count is used up, nil is returned. */)
2397 Lisp_Object from
, count
;
2399 CHECK_NUMBER (from
);
2400 CHECK_NUMBER (count
);
2402 return scan_lists (XINT (from
), XINT (count
), 0, 1);
2405 DEFUN ("backward-prefix-chars", Fbackward_prefix_chars
, Sbackward_prefix_chars
,
2407 doc
: /* Move point backward over any number of chars with prefix syntax.
2408 This includes chars with "quote" or "prefix" syntax (' or p). */)
2413 int opoint_byte
= PT_BYTE
;
2415 int pos_byte
= PT_BYTE
;
2420 SET_PT_BOTH (opoint
, opoint_byte
);
2425 SETUP_SYNTAX_TABLE (pos
, -1);
2427 DEC_BOTH (pos
, pos_byte
);
2429 while (!char_quoted (pos
, pos_byte
)
2430 /* Previous statement updates syntax table. */
2431 && ((c
= FETCH_CHAR (pos_byte
), SYNTAX (c
) == Squote
)
2432 || SYNTAX_PREFIX (c
)))
2435 opoint_byte
= pos_byte
;
2438 DEC_BOTH (pos
, pos_byte
);
2441 SET_PT_BOTH (opoint
, opoint_byte
);
2446 /* Parse forward from FROM / FROM_BYTE to END,
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.
2449 If STOPBEFORE is nonzero, stop at the start of an atom.
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. */
2455 scan_sexps_forward (stateptr
, from
, from_byte
, end
, targetdepth
,
2456 stopbefore
, oldstate
, commentstop
)
2457 struct lisp_parse_state
*stateptr
;
2459 int end
, targetdepth
, stopbefore
;
2460 Lisp_Object oldstate
;
2463 struct lisp_parse_state state
;
2465 register enum syntaxcode code
;
2468 struct level
{ int last
, prev
; };
2469 struct level levelstart
[100];
2470 register struct level
*curlevel
= levelstart
;
2471 struct level
*endlevel
= levelstart
+ 100;
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 */
2478 int prev_from
; /* Keep one character before FROM. */
2480 int prev_from_syntax
;
2481 int boundary_stop
= commentstop
== -1;
2484 int out_bytepos
, out_charpos
;
2488 prev_from_byte
= from_byte
;
2490 DEC_BOTH (prev_from
, prev_from_byte
);
2492 /* Use this macro instead of `from++'. */
2494 do { prev_from = from; \
2495 prev_from_byte = from_byte; \
2497 = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte)); \
2498 INC_BOTH (from, from_byte); \
2499 UPDATE_SYNTAX_TABLE_FORWARD (from); \
2505 if (NILP (oldstate
))
2508 state
.instring
= -1;
2509 state
.incomment
= 0;
2510 state
.comstyle
= 0; /* comment style a by default. */
2511 state
.comstr_start
= -1; /* no comment/string seen. */
2515 tem
= Fcar (oldstate
);
2521 oldstate
= Fcdr (oldstate
);
2522 oldstate
= Fcdr (oldstate
);
2523 oldstate
= Fcdr (oldstate
);
2524 tem
= Fcar (oldstate
);
2525 /* Check whether we are inside string_fence-style string: */
2526 state
.instring
= (!NILP (tem
)
2527 ? (INTEGERP (tem
) ? XINT (tem
) : ST_STRING_STYLE
)
2530 oldstate
= Fcdr (oldstate
);
2531 tem
= Fcar (oldstate
);
2532 state
.incomment
= (!NILP (tem
)
2533 ? (INTEGERP (tem
) ? XINT (tem
) : -1)
2536 oldstate
= Fcdr (oldstate
);
2537 tem
= Fcar (oldstate
);
2538 start_quoted
= !NILP (tem
);
2540 /* if the eighth element of the list is nil, we are in comment
2541 style a. If it is non-nil, we are in comment style b */
2542 oldstate
= Fcdr (oldstate
);
2543 oldstate
= Fcdr (oldstate
);
2544 tem
= Fcar (oldstate
);
2545 state
.comstyle
= NILP (tem
) ? 0 : (EQ (tem
, Qsyntax_table
)
2546 ? ST_COMMENT_STYLE
: 1);
2548 oldstate
= Fcdr (oldstate
);
2549 tem
= Fcar (oldstate
);
2550 state
.comstr_start
= NILP (tem
) ? -1 : XINT (tem
) ;
2551 oldstate
= Fcdr (oldstate
);
2552 tem
= Fcar (oldstate
);
2553 while (!NILP (tem
)) /* >= second enclosing sexps. */
2555 /* curlevel++->last ran into compiler bug on Apollo */
2556 curlevel
->last
= XINT (Fcar (tem
));
2557 if (++curlevel
== endlevel
)
2558 curlevel
--; /* error ("Nesting too deep for parser"); */
2559 curlevel
->prev
= -1;
2560 curlevel
->last
= -1;
2567 curlevel
->prev
= -1;
2568 curlevel
->last
= -1;
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
);
2574 /* Enter the loop at a place appropriate for initial state. */
2576 if (state
.incomment
)
2577 goto startincomment
;
2578 if (state
.instring
>= 0)
2580 nofence
= state
.instring
!= ST_STRING_STYLE
;
2582 goto startquotedinstring
;
2585 else if (start_quoted
)
2588 #if 0 /* This seems to be redundant with the identical code above. */
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
);
2597 code
= prev_from_syntax
& 0xff;
2599 if (code
== Scomment
)
2601 state
.comstyle
= SYNTAX_FLAGS_COMMENT_STYLE (prev_from_syntax
);
2602 state
.incomment
= (SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax
) ?
2604 state
.comstr_start
= prev_from
;
2606 else if (code
== Scomment_fence
)
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. */
2611 state
.comstyle
= ST_COMMENT_STYLE
;
2612 state
.incomment
= -1;
2613 state
.comstr_start
= prev_from
;
2616 else if (from
< end
)
2617 if (SYNTAX_FLAGS_COMSTART_FIRST (prev_from_syntax
))
2618 if (c1
= FETCH_CHAR (from_byte
),
2619 SYNTAX_COMSTART_SECOND (c1
))
2620 /* Duplicate code to avoid a complex if-expression
2621 which causes trouble for the SGI compiler. */
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. */
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;
2630 state
.comstr_start
= prev_from
;
2635 if (SYNTAX_FLAGS_PREFIX (prev_from_syntax
))
2637 switch (SWITCH_ENUM_CAST (code
))
2641 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2642 curlevel
->last
= prev_from
;
2644 if (from
== end
) goto endquoted
;
2647 /* treat following character as a word constituent */
2650 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2651 curlevel
->last
= prev_from
;
2655 /* Some compilers can't handle this inside the switch. */
2656 temp
= SYNTAX (FETCH_CHAR (from_byte
));
2662 if (from
== end
) goto endquoted
;
2674 curlevel
->prev
= curlevel
->last
;
2678 if (commentstop
|| boundary_stop
) goto done
;
2680 /* The (from == BEGV) test was to enter the loop in the middle so
2681 that we find a 2-char comment ender even if we start in the
2682 middle of it. We don't want to do that if we're just at the
2683 beginning of the comment (think of (*) ... (*)). */
2684 found
= forw_comment (from
, from_byte
, end
,
2685 state
.incomment
, state
.comstyle
,
2686 (from
== BEGV
|| from
< state
.comstr_start
+ 3)
2687 ? 0 : prev_from_syntax
,
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
;
2695 state
.incomment
= 0;
2696 state
.comstyle
= 0; /* reset the comment style */
2697 if (boundary_stop
) goto done
;
2701 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2703 /* curlevel++->last ran into compiler bug on Apollo */
2704 curlevel
->last
= prev_from
;
2705 if (++curlevel
== endlevel
)
2706 curlevel
--; /* error ("Nesting too deep for parser"); */
2707 curlevel
->prev
= -1;
2708 curlevel
->last
= -1;
2709 if (targetdepth
== depth
) goto done
;
2714 if (depth
< mindepth
)
2716 if (curlevel
!= levelstart
)
2718 curlevel
->prev
= curlevel
->last
;
2719 if (targetdepth
== depth
) goto done
;
2724 state
.comstr_start
= from
- 1;
2725 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2726 curlevel
->last
= prev_from
;
2727 state
.instring
= (code
== Sstring
2728 ? (FETCH_CHAR (prev_from_byte
))
2730 if (boundary_stop
) goto done
;
2733 nofence
= state
.instring
!= ST_STRING_STYLE
;
2739 if (from
>= end
) goto done
;
2740 c
= FETCH_CHAR (from_byte
);
2741 /* Some compilers can't handle this inside the switch. */
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
)
2753 if (!nofence
) goto string_end
;
2758 startquotedinstring
:
2759 if (from
>= end
) goto endquoted
;
2765 state
.instring
= -1;
2766 curlevel
->prev
= curlevel
->last
;
2768 if (boundary_stop
) goto done
;
2777 stop
: /* Here if stopping before start of sexp. */
2778 from
= prev_from
; /* We have just fetched the char that starts it; */
2779 goto done
; /* but return the position before it. */
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
;
2790 state
.levelstarts
= Qnil
;
2791 while (--curlevel
>= levelstart
)
2792 state
.levelstarts
= Fcons (make_number (curlevel
->last
),
2799 DEFUN ("parse-partial-sexp", Fparse_partial_sexp
, Sparse_partial_sexp
, 2, 6, 0,
2800 doc
: /* Parse Lisp syntax starting at FROM until TO; return status of parse at TO.
2801 Parsing stops at TO or when certain criteria are met;
2802 point is set to where parsing stops.
2803 If fifth arg OLDSTATE is omitted or nil,
2804 parsing assumes that FROM is the beginning of a function.
2805 Value is a list of ten elements describing final state of parsing:
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).
2820 If third arg TARGETDEPTH is non-nil, parsing stops if the depth
2821 in parentheses becomes equal to TARGETDEPTH.
2822 Fourth arg STOPBEFORE non-nil means stop when come to
2823 any character that starts a sexp.
2824 Fifth 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.
2827 Sixth 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
)
2831 Lisp_Object from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
;
2833 struct lisp_parse_state state
;
2836 if (!NILP (targetdepth
))
2838 CHECK_NUMBER (targetdepth
);
2839 target
= XINT (targetdepth
);
2842 target
= -100000; /* We won't reach this depth */
2844 validate_region (&from
, &to
);
2845 scan_sexps_forward (&state
, XINT (from
), CHAR_TO_BYTE (XINT (from
)),
2847 target
, !NILP (stopbefore
), oldstate
,
2849 ? 0 : (EQ (commentstop
, Qsyntax_table
) ? -1 : 1)));
2851 SET_PT (state
.location
);
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
),
2856 Fcons (state
.instring
>= 0
2857 ? (state
.instring
== ST_STRING_STYLE
2858 ? Qt
: make_number (state
.instring
)) : Qnil
,
2859 Fcons (state
.incomment
< 0 ? Qt
:
2860 (state
.incomment
== 0 ? Qnil
:
2861 make_number (state
.incomment
)),
2862 Fcons (state
.quoted
? Qt
: Qnil
,
2863 Fcons (make_number (state
.mindepth
),
2864 Fcons ((state
.comstyle
2865 ? (state
.comstyle
== ST_COMMENT_STYLE
2866 ? Qsyntax_table
: Qt
) :
2868 Fcons (((state
.incomment
2869 || (state
.instring
>= 0))
2870 ? make_number (state
.comstr_start
)
2872 Fcons (state
.levelstarts
, Qnil
))))))))));
2881 /* This has to be done here, before we call Fmake_char_table. */
2882 Qsyntax_table
= intern ("syntax-table");
2883 staticpro (&Qsyntax_table
);
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");
2890 /* Create objects which can be shared among syntax tables. */
2891 Vsyntax_code_object
= Fmake_vector (make_number (Smax
), Qnil
);
2892 for (i
= 0; i
< XVECTOR (Vsyntax_code_object
)->size
; i
++)
2893 XVECTOR (Vsyntax_code_object
)->contents
[i
]
2894 = Fcons (make_number (i
), Qnil
);
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));
2900 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Swhitespace
];
2902 Vstandard_syntax_table
= Fmake_char_table (Qsyntax_table
, temp
);
2904 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Sword
];
2905 for (i
= 'a'; i
<= 'z'; i
++)
2906 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2907 for (i
= 'A'; i
<= 'Z'; i
++)
2908 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2909 for (i
= '0'; i
<= '9'; i
++)
2910 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2912 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '$', temp
);
2913 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '%', temp
);
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
));
2932 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Ssymbol
];
2933 for (i
= 0; i
< 10; i
++)
2935 c
= "_-+*/&|<>="[i
];
2936 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, c
, temp
);
2939 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Spunct
];
2940 for (i
= 0; i
< 12; i
++)
2942 c
= ".,;:?!#@~^'`"[i
];
2943 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, c
, temp
);
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
;
2955 Qsyntax_table_p
= intern ("syntax-table-p");
2956 staticpro (&Qsyntax_table_p
);
2958 staticpro (&Vsyntax_code_object
);
2960 Qscan_error
= intern ("scan-error");
2961 staticpro (&Qscan_error
);
2962 Fput (Qscan_error
, Qerror_conditions
,
2963 Fcons (Qscan_error
, Fcons (Qerror
, Qnil
)));
2964 Fput (Qscan_error
, Qerror_message
,
2965 build_string ("Scan error"));
2967 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments
,
2968 doc
: /* Non-nil means `forward-sexp', etc., should treat comments as whitespace. */);
2970 DEFVAR_BOOL ("parse-sexp-lookup-properties", &parse_sexp_lookup_properties
,
2971 doc
: /* Non-nil means `forward-sexp', etc., obey `syntax-table' property.
2972 Otherwise, that text property is simply ignored.
2973 See the info node `(elisp)Syntax Properties' for a description of the
2974 `syntax-table' property. */);
2976 words_include_escapes
= 0;
2977 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes
,
2978 doc
: /* Non-nil means `forward-word', etc., should treat escape chars part of words. */);
2980 DEFVAR_BOOL ("multibyte-syntax-as-symbol", &multibyte_syntax_as_symbol
,
2981 doc
: /* Non-nil means `scan-sexps' treats all multibyte characters as symbol. */);
2982 multibyte_syntax_as_symbol
= 0;
2984 DEFVAR_BOOL ("open-paren-in-column-0-is-defun-start",
2985 &open_paren_in_column_0_is_defun_start
,
2986 doc
: /* Non-nil means an open paren in column 0 denotes the start of a defun. */);
2987 open_paren_in_column_0_is_defun_start
= 1;
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
);
2995 defsubr (&Smatching_paren
);
2996 defsubr (&Sstring_to_syntax
);
2997 defsubr (&Smodify_syntax_entry
);
2998 defsubr (&Sinternal_describe_syntax_value
);
3000 defsubr (&Sforward_word
);
3002 defsubr (&Sskip_chars_forward
);
3003 defsubr (&Sskip_chars_backward
);
3004 defsubr (&Sskip_syntax_forward
);
3005 defsubr (&Sskip_syntax_backward
);
3007 defsubr (&Sforward_comment
);
3008 defsubr (&Sscan_lists
);
3009 defsubr (&Sscan_sexps
);
3010 defsubr (&Sbackward_prefix_chars
);
3011 defsubr (&Sparse_partial_sexp
);