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 a 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 && ((com2end
&& 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 const unsigned char *p
;
905 register enum syntaxcode code
;
909 CHECK_STRING (string
);
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
, SBYTES (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.
1311 Note that char classes, e.g. `[:alpha:]', are not currently supported;
1312 they will be treated as literals. */)
1314 Lisp_Object string
, lim
;
1316 return skip_chars (1, 0, string
, lim
);
1319 DEFUN ("skip-chars-backward", Fskip_chars_backward
, Sskip_chars_backward
, 1, 2, 0,
1320 doc
: /* Move point backward, stopping after a char not in STRING, or at pos LIM.
1321 See `skip-chars-forward' for details.
1322 Returns the distance traveled, either zero or negative. */)
1324 Lisp_Object string
, lim
;
1326 return skip_chars (0, 0, string
, lim
);
1329 DEFUN ("skip-syntax-forward", Fskip_syntax_forward
, Sskip_syntax_forward
, 1, 2, 0,
1330 doc
: /* Move point forward across chars in specified syntax classes.
1331 SYNTAX is a string of syntax code characters.
1332 Stop before a char whose syntax is not in SYNTAX, or at position LIM.
1333 If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1334 This function returns the distance traveled, either zero or positive. */)
1336 Lisp_Object syntax
, lim
;
1338 return skip_chars (1, 1, syntax
, lim
);
1341 DEFUN ("skip-syntax-backward", Fskip_syntax_backward
, Sskip_syntax_backward
, 1, 2, 0,
1342 doc
: /* Move point backward across chars in specified syntax classes.
1343 SYNTAX is a string of syntax code characters.
1344 Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM.
1345 If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX.
1346 This function returns the distance traveled, either zero or negative. */)
1348 Lisp_Object syntax
, lim
;
1350 return skip_chars (0, 1, syntax
, lim
);
1354 skip_chars (forwardp
, syntaxp
, string
, lim
)
1355 int forwardp
, syntaxp
;
1356 Lisp_Object string
, lim
;
1358 register unsigned int c
;
1359 unsigned char fastmap
[0400];
1360 /* If SYNTAXP is 0, STRING may contain multi-byte form of characters
1361 of which codes don't fit in FASTMAP. In that case, set the
1362 ranges of characters in CHAR_RANGES. */
1364 int n_char_ranges
= 0;
1366 register int i
, i_byte
;
1367 int multibyte
= !NILP (current_buffer
->enable_multibyte_characters
);
1368 int string_multibyte
;
1370 const unsigned char *str
;
1373 CHECK_STRING (string
);
1374 char_ranges
= (int *) alloca (SCHARS (string
) * (sizeof (int)) * 2);
1375 string_multibyte
= STRING_MULTIBYTE (string
);
1376 str
= SDATA (string
);
1377 size_byte
= SBYTES (string
);
1379 /* Adjust the multibyteness of the string to that of the buffer. */
1380 if (multibyte
!= string_multibyte
)
1385 nbytes
= count_size_as_multibyte (SDATA (string
),
1388 nbytes
= SCHARS (string
);
1389 if (nbytes
!= size_byte
)
1391 unsigned char *tmp
= (unsigned char *) alloca (nbytes
);
1392 copy_text (SDATA (string
), tmp
, size_byte
,
1393 string_multibyte
, multibyte
);
1400 XSETINT (lim
, forwardp
? ZV
: BEGV
);
1402 CHECK_NUMBER_COERCE_MARKER (lim
);
1404 /* In any case, don't allow scan outside bounds of buffer. */
1405 if (XINT (lim
) > ZV
)
1406 XSETFASTINT (lim
, ZV
);
1407 if (XINT (lim
) < BEGV
)
1408 XSETFASTINT (lim
, BEGV
);
1410 bzero (fastmap
, sizeof fastmap
);
1414 if (i_byte
< size_byte
1415 && SREF (string
, 0) == '^')
1417 negate
= 1; i_byte
++;
1420 /* Find the characters specified and set their elements of fastmap.
1421 If syntaxp, each character counts as itself.
1422 Otherwise, handle backslashes and ranges specially. */
1424 while (i_byte
< size_byte
)
1426 c
= STRING_CHAR_AND_LENGTH (str
+ i_byte
, size_byte
- i_byte
, len
);
1430 fastmap
[syntax_spec_code
[c
& 0377]] = 1;
1435 if (i_byte
== size_byte
)
1438 c
= STRING_CHAR_AND_LENGTH (str
+i_byte
, size_byte
-i_byte
, len
);
1441 if (i_byte
< size_byte
1442 && str
[i_byte
] == '-')
1446 /* Skip over the dash. */
1449 if (i_byte
== size_byte
)
1452 /* Get the end of the range. */
1453 c2
=STRING_CHAR_AND_LENGTH (str
+i_byte
, size_byte
-i_byte
, len
);
1456 if (SINGLE_BYTE_CHAR_P (c
))
1458 if (! SINGLE_BYTE_CHAR_P (c2
))
1460 /* Handle a range starting with a character of
1461 less than 256, and ending with a character of
1462 not less than 256. Split that into two
1463 ranges, the low one ending at 0377, and the
1464 high one starting at the smallest character
1465 in the charset of C2 and ending at C2. */
1466 int charset
= CHAR_CHARSET (c2
);
1467 int c1
= MAKE_CHAR (charset
, 0, 0);
1469 char_ranges
[n_char_ranges
++] = c1
;
1470 char_ranges
[n_char_ranges
++] = c2
;
1479 else if (c
<= c2
) /* Both C and C2 are multibyte char. */
1481 char_ranges
[n_char_ranges
++] = c
;
1482 char_ranges
[n_char_ranges
++] = c2
;
1487 if (SINGLE_BYTE_CHAR_P (c
))
1491 char_ranges
[n_char_ranges
++] = c
;
1492 char_ranges
[n_char_ranges
++] = c
;
1498 /* If ^ was the first character, complement the fastmap. */
1500 for (i
= 0; i
< sizeof fastmap
; i
++)
1504 int start_point
= PT
;
1506 int pos_byte
= PT_BYTE
;
1511 SETUP_SYNTAX_TABLE (pos
, forwardp
? 1 : -1);
1516 if (pos
< XINT (lim
))
1517 while (fastmap
[(int) SYNTAX (FETCH_CHAR (pos_byte
))])
1519 /* Since we already checked for multibyteness,
1520 avoid using INC_BOTH which checks again. */
1523 if (pos
>= XINT (lim
))
1525 UPDATE_SYNTAX_TABLE_FORWARD (pos
);
1530 while (pos
< XINT (lim
)
1531 && fastmap
[(int) SYNTAX (FETCH_BYTE (pos
))])
1534 UPDATE_SYNTAX_TABLE_FORWARD (pos
);
1542 while (pos
> XINT (lim
))
1544 int savepos
= pos_byte
;
1545 /* Since we already checked for multibyteness,
1546 avoid using DEC_BOTH which checks again. */
1549 UPDATE_SYNTAX_TABLE_BACKWARD (pos
);
1550 if (!fastmap
[(int) SYNTAX (FETCH_CHAR (pos_byte
))])
1560 if (pos
> XINT (lim
))
1561 while (fastmap
[(int) SYNTAX (FETCH_BYTE (pos
- 1))])
1564 if (pos
<= XINT (lim
))
1566 UPDATE_SYNTAX_TABLE_BACKWARD (pos
- 1);
1576 while (pos
< XINT (lim
))
1578 c
= FETCH_MULTIBYTE_CHAR (pos_byte
);
1579 if (SINGLE_BYTE_CHAR_P (c
))
1586 /* If we are looking at a multibyte character,
1587 we must look up the character in the table
1588 CHAR_RANGES. If there's no data in the
1589 table, that character is not what we want to
1592 /* The following code do the right thing even if
1593 n_char_ranges is zero (i.e. no data in
1595 for (i
= 0; i
< n_char_ranges
; i
+= 2)
1596 if (c
>= char_ranges
[i
] && c
<= char_ranges
[i
+ 1])
1598 if (!(negate
^ (i
< n_char_ranges
)))
1601 INC_BOTH (pos
, pos_byte
);
1604 while (pos
< XINT (lim
) && fastmap
[FETCH_BYTE (pos
)])
1610 while (pos
> XINT (lim
))
1612 int prev_pos_byte
= pos_byte
;
1614 DEC_POS (prev_pos_byte
);
1615 c
= FETCH_MULTIBYTE_CHAR (prev_pos_byte
);
1616 if (SINGLE_BYTE_CHAR_P (c
))
1623 /* See the comment in the previous similar code. */
1624 for (i
= 0; i
< n_char_ranges
; i
+= 2)
1625 if (c
>= char_ranges
[i
] && c
<= char_ranges
[i
+ 1])
1627 if (!(negate
^ (i
< n_char_ranges
)))
1631 pos_byte
= prev_pos_byte
;
1634 while (pos
> XINT (lim
) && fastmap
[FETCH_BYTE (pos
- 1)])
1639 #if 0 /* Not needed now that a position in mid-character
1640 cannot be specified in Lisp. */
1642 /* INC_POS or DEC_POS might have moved POS over LIM. */
1643 && (forwardp
? (pos
> XINT (lim
)) : (pos
< XINT (lim
))))
1650 SET_PT_BOTH (pos
, pos_byte
);
1653 return make_number (PT
- start_point
);
1657 /* Jump over a comment, assuming we are at the beginning of one.
1658 FROM is the current position.
1659 FROM_BYTE is the bytepos corresponding to FROM.
1660 Do not move past STOP (a charpos).
1661 The comment over which we have to jump is of style STYLE
1662 (either SYNTAX_COMMENT_STYLE(foo) or ST_COMMENT_STYLE).
1663 NESTING should be positive to indicate the nesting at the beginning
1664 for nested comments and should be zero or negative else.
1665 ST_COMMENT_STYLE cannot be nested.
1666 PREV_SYNTAX is the SYNTAX_WITH_FLAGS of the previous character
1667 (or 0 If the search cannot start in the middle of a two-character).
1669 If successful, return 1 and store the charpos of the comment's end
1670 into *CHARPOS_PTR and the corresponding bytepos into *BYTEPOS_PTR.
1671 Else, return 0 and store the charpos STOP into *CHARPOS_PTR, the
1672 corresponding bytepos into *BYTEPOS_PTR and the current nesting
1673 (as defined for state.incomment) in *INCOMMENT_PTR.
1675 The comment end is the last character of the comment rather than the
1676 character just after the comment.
1678 Global syntax data is assumed to initially be valid for FROM and
1679 remains valid for forward search starting at the returned position. */
1682 forw_comment (from
, from_byte
, stop
, nesting
, style
, prev_syntax
,
1683 charpos_ptr
, bytepos_ptr
, incomment_ptr
)
1684 int from
, from_byte
, stop
;
1685 int nesting
, style
, prev_syntax
;
1686 int *charpos_ptr
, *bytepos_ptr
, *incomment_ptr
;
1689 register enum syntaxcode code
;
1690 register int syntax
;
1692 if (nesting
<= 0) nesting
= -1;
1694 /* Enter the loop in the middle so that we find
1695 a 2-char comment ender if we start in the middle of it. */
1696 syntax
= prev_syntax
;
1697 if (syntax
!= 0) goto forw_incomment
;
1703 *incomment_ptr
= nesting
;
1704 *charpos_ptr
= from
;
1705 *bytepos_ptr
= from_byte
;
1708 c
= FETCH_CHAR (from_byte
);
1709 syntax
= SYNTAX_WITH_FLAGS (c
);
1710 code
= syntax
& 0xff;
1711 if (code
== Sendcomment
1712 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
1713 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ?
1714 (nesting
> 0 && --nesting
== 0) : nesting
< 0))
1715 /* we have encountered a comment end of the same style
1716 as the comment sequence which began this comment
1719 if (code
== Scomment_fence
1720 && style
== ST_COMMENT_STYLE
)
1721 /* we have encountered a comment end of the same style
1722 as the comment sequence which began this comment
1727 && SYNTAX_FLAGS_COMMENT_NESTED (syntax
)
1728 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
)
1729 /* we have encountered a nested comment of the same style
1730 as the comment sequence which began this comment section */
1732 INC_BOTH (from
, from_byte
);
1733 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1736 if (from
< stop
&& SYNTAX_FLAGS_COMEND_FIRST (syntax
)
1737 && SYNTAX_FLAGS_COMMENT_STYLE (syntax
) == style
1738 && (c1
= FETCH_CHAR (from_byte
),
1739 SYNTAX_COMEND_SECOND (c1
))
1740 && ((SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ||
1741 SYNTAX_COMMENT_NESTED (c1
)) ? nesting
> 0 : nesting
< 0))
1744 /* we have encountered a comment end of the same style
1745 as the comment sequence which began this comment
1750 INC_BOTH (from
, from_byte
);
1751 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1756 && SYNTAX_FLAGS_COMSTART_FIRST (syntax
)
1757 && (c1
= FETCH_CHAR (from_byte
),
1758 SYNTAX_COMMENT_STYLE (c1
) == style
1759 && SYNTAX_COMSTART_SECOND (c1
))
1760 && (SYNTAX_FLAGS_COMMENT_NESTED (syntax
) ||
1761 SYNTAX_COMMENT_NESTED (c1
)))
1762 /* we have encountered a nested comment of the same style
1763 as the comment sequence which began this comment
1766 INC_BOTH (from
, from_byte
);
1767 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1771 *charpos_ptr
= from
;
1772 *bytepos_ptr
= from_byte
;
1776 DEFUN ("forward-comment", Fforward_comment
, Sforward_comment
, 1, 1, 0,
1778 Move forward across up to COUNT comments. If COUNT is negative, move backward.
1779 Stop scanning if we find something other than a comment or whitespace.
1780 Set point to where scanning stops.
1781 If COUNT comments are found as expected, with nothing except whitespace
1782 between them, return t; otherwise return nil. */)
1790 register enum syntaxcode code
;
1791 int comstyle
= 0; /* style of comment encountered */
1792 int comnested
= 0; /* whether the comment is nestable or not */
1795 int out_charpos
, out_bytepos
;
1798 CHECK_NUMBER (count
);
1799 count1
= XINT (count
);
1800 stop
= count1
> 0 ? ZV
: BEGV
;
1806 from_byte
= PT_BYTE
;
1808 SETUP_SYNTAX_TABLE (from
, count1
);
1817 SET_PT_BOTH (from
, from_byte
);
1821 c
= FETCH_CHAR (from_byte
);
1823 comstart_first
= SYNTAX_COMSTART_FIRST (c
);
1824 comnested
= SYNTAX_COMMENT_NESTED (c
);
1825 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1826 INC_BOTH (from
, from_byte
);
1827 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1828 if (from
< stop
&& comstart_first
1829 && (c1
= FETCH_CHAR (from_byte
),
1830 SYNTAX_COMSTART_SECOND (c1
)))
1832 /* We have encountered a comment start sequence and we
1833 are ignoring all text inside comments. We must record
1834 the comment style this sequence begins so that later,
1835 only a comment end of the same style actually ends
1836 the comment section. */
1838 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
1839 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
1840 INC_BOTH (from
, from_byte
);
1841 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1844 while (code
== Swhitespace
|| (code
== Sendcomment
&& c
== '\n'));
1846 if (code
== Scomment_fence
)
1847 comstyle
= ST_COMMENT_STYLE
;
1848 else if (code
!= Scomment
)
1851 DEC_BOTH (from
, from_byte
);
1852 SET_PT_BOTH (from
, from_byte
);
1855 /* We're at the start of a comment. */
1856 found
= forw_comment (from
, from_byte
, stop
, comnested
, comstyle
, 0,
1857 &out_charpos
, &out_bytepos
, &dummy
);
1858 from
= out_charpos
; from_byte
= out_bytepos
;
1862 SET_PT_BOTH (from
, from_byte
);
1865 INC_BOTH (from
, from_byte
);
1866 UPDATE_SYNTAX_TABLE_FORWARD (from
);
1867 /* We have skipped one comment. */
1879 SET_PT_BOTH (BEGV
, BEGV_BYTE
);
1884 DEC_BOTH (from
, from_byte
);
1885 /* char_quoted does UPDATE_SYNTAX_TABLE_BACKWARD (from). */
1886 quoted
= char_quoted (from
, from_byte
);
1887 c
= FETCH_CHAR (from_byte
);
1890 comnested
= SYNTAX_COMMENT_NESTED (c
);
1891 if (code
== Sendcomment
)
1892 comstyle
= SYNTAX_COMMENT_STYLE (c
);
1893 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
1894 && prev_char_comend_first (from
, from_byte
)
1895 && !char_quoted (from
- 1, dec_bytepos (from_byte
)))
1897 /* We must record the comment style encountered so that
1898 later, we can match only the proper comment begin
1899 sequence of the same style. */
1900 DEC_BOTH (from
, from_byte
);
1902 /* Calling char_quoted, above, set up global syntax position
1903 at the new value of FROM. */
1904 c1
= FETCH_CHAR (from_byte
);
1905 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
1906 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
1909 if (code
== Scomment_fence
)
1911 /* Skip until first preceding unquoted comment_fence. */
1912 int found
= 0, ini
= from
, ini_byte
= from_byte
;
1916 DEC_BOTH (from
, from_byte
);
1919 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
1920 c
= FETCH_CHAR (from_byte
);
1921 if (SYNTAX (c
) == Scomment_fence
1922 && !char_quoted (from
, from_byte
))
1930 from
= ini
; /* Set point to ini + 1. */
1931 from_byte
= ini_byte
;
1935 else if (code
== Sendcomment
)
1937 found
= back_comment (from
, from_byte
, stop
, comnested
, comstyle
,
1938 &out_charpos
, &out_bytepos
);
1942 /* This end-of-line is not an end-of-comment.
1943 Treat it like a whitespace.
1944 CC-mode (and maybe others) relies on this behavior. */
1948 /* Failure: we should go back to the end of this
1949 not-quite-endcomment. */
1950 if (SYNTAX(c
) != code
)
1951 /* It was a two-char Sendcomment. */
1952 INC_BOTH (from
, from_byte
);
1958 /* We have skipped one comment. */
1959 from
= out_charpos
, from_byte
= out_bytepos
;
1963 else if (code
!= Swhitespace
|| quoted
)
1967 INC_BOTH (from
, from_byte
);
1968 SET_PT_BOTH (from
, from_byte
);
1976 SET_PT_BOTH (from
, from_byte
);
1981 /* Return syntax code of character C if C is a single byte character
1982 or `multibyte_symbol_p' is zero. Otherwise, return Ssymbol. */
1984 #define SYNTAX_WITH_MULTIBYTE_CHECK(c) \
1985 ((SINGLE_BYTE_CHAR_P (c) || !multibyte_symbol_p) \
1986 ? SYNTAX (c) : Ssymbol)
1989 scan_lists (from
, count
, depth
, sexpflag
)
1991 int count
, depth
, sexpflag
;
1994 register int stop
= count
> 0 ? ZV
: BEGV
;
1999 register enum syntaxcode code
, temp_code
;
2000 int min_depth
= depth
; /* Err out if depth gets less than this. */
2001 int comstyle
= 0; /* style of comment encountered */
2002 int comnested
= 0; /* whether the comment is nestable or not */
2004 int last_good
= from
;
2007 int out_bytepos
, out_charpos
;
2009 int multibyte_symbol_p
= sexpflag
&& multibyte_syntax_as_symbol
;
2011 if (depth
> 0) min_depth
= 0;
2013 if (from
> ZV
) from
= ZV
;
2014 if (from
< BEGV
) from
= BEGV
;
2016 from_byte
= CHAR_TO_BYTE (from
);
2021 SETUP_SYNTAX_TABLE (from
, count
);
2026 int comstart_first
, prefix
;
2027 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2028 c
= FETCH_CHAR (from_byte
);
2029 code
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2030 comstart_first
= SYNTAX_COMSTART_FIRST (c
);
2031 comnested
= SYNTAX_COMMENT_NESTED (c
);
2032 comstyle
= SYNTAX_COMMENT_STYLE (c
);
2033 prefix
= SYNTAX_PREFIX (c
);
2034 if (depth
== min_depth
)
2036 INC_BOTH (from
, from_byte
);
2037 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2038 if (from
< stop
&& comstart_first
2039 && SYNTAX_COMSTART_SECOND (FETCH_CHAR (from_byte
))
2040 && parse_sexp_ignore_comments
)
2042 /* we have encountered a comment start sequence and we
2043 are ignoring all text inside comments. We must record
2044 the comment style this sequence begins so that later,
2045 only a comment end of the same style actually ends
2046 the comment section */
2048 c1
= FETCH_CHAR (from_byte
);
2049 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2050 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2051 INC_BOTH (from
, from_byte
);
2052 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2058 switch (SWITCH_ENUM_CAST (code
))
2062 if (from
== stop
) goto lose
;
2063 INC_BOTH (from
, from_byte
);
2064 /* treat following character as a word constituent */
2067 if (depth
|| !sexpflag
) break;
2068 /* This word counts as a sexp; return at end of it. */
2071 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2073 /* Some compilers can't handle this inside the switch. */
2074 c
= FETCH_CHAR (from_byte
);
2075 temp
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2080 INC_BOTH (from
, from_byte
);
2081 if (from
== stop
) goto lose
;
2090 INC_BOTH (from
, from_byte
);
2094 case Scomment_fence
:
2095 comstyle
= ST_COMMENT_STYLE
;
2098 if (!parse_sexp_ignore_comments
) break;
2099 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2100 found
= forw_comment (from
, from_byte
, stop
,
2101 comnested
, comstyle
, 0,
2102 &out_charpos
, &out_bytepos
, &dummy
);
2103 from
= out_charpos
, from_byte
= out_bytepos
;
2110 INC_BOTH (from
, from_byte
);
2111 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2117 if (from
!= stop
&& c
== FETCH_CHAR (from_byte
))
2119 INC_BOTH (from
, from_byte
);
2129 if (!++depth
) goto done
;
2134 if (!--depth
) goto done
;
2135 if (depth
< min_depth
)
2136 Fsignal (Qscan_error
,
2137 Fcons (build_string ("Containing expression ends prematurely"),
2138 Fcons (make_number (last_good
),
2139 Fcons (make_number (from
), Qnil
))));
2144 temp_pos
= dec_bytepos (from_byte
);
2145 stringterm
= FETCH_CHAR (temp_pos
);
2148 if (from
>= stop
) goto lose
;
2149 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2150 c
= FETCH_CHAR (from_byte
);
2153 && SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring
)
2154 : SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring_fence
)
2157 /* Some compilers can't handle this inside the switch. */
2158 temp
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2163 INC_BOTH (from
, from_byte
);
2165 INC_BOTH (from
, from_byte
);
2167 INC_BOTH (from
, from_byte
);
2168 if (!depth
&& sexpflag
) goto done
;
2171 /* Ignore whitespace, punctuation, quote, endcomment. */
2176 /* Reached end of buffer. Error if within object, return nil if between */
2177 if (depth
) goto lose
;
2182 /* End of object reached */
2191 DEC_BOTH (from
, from_byte
);
2192 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2193 c
= FETCH_CHAR (from_byte
);
2194 code
= SYNTAX_WITH_MULTIBYTE_CHECK (c
);
2195 if (depth
== min_depth
)
2198 comnested
= SYNTAX_COMMENT_NESTED (c
);
2199 if (code
== Sendcomment
)
2200 comstyle
= SYNTAX_COMMENT_STYLE (c
);
2201 if (from
> stop
&& SYNTAX_COMEND_SECOND (c
)
2202 && prev_char_comend_first (from
, from_byte
)
2203 && parse_sexp_ignore_comments
)
2205 /* We must record the comment style encountered so that
2206 later, we can match only the proper comment begin
2207 sequence of the same style. */
2208 DEC_BOTH (from
, from_byte
);
2209 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2211 c1
= FETCH_CHAR (from_byte
);
2212 comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2213 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2216 /* Quoting turns anything except a comment-ender
2217 into a word character. Note that this cannot be true
2218 if we decremented FROM in the if-statement above. */
2219 if (code
!= Sendcomment
&& char_quoted (from
, from_byte
))
2221 DEC_BOTH (from
, from_byte
);
2224 else if (SYNTAX_PREFIX (c
))
2227 switch (SWITCH_ENUM_CAST (code
))
2233 if (depth
|| !sexpflag
) break;
2234 /* This word counts as a sexp; count object finished
2235 after passing it. */
2238 temp_pos
= from_byte
;
2239 if (! NILP (current_buffer
->enable_multibyte_characters
))
2243 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2244 c1
= FETCH_CHAR (temp_pos
);
2245 temp_code
= SYNTAX_WITH_MULTIBYTE_CHECK (c1
);
2246 /* Don't allow comment-end to be quoted. */
2247 if (temp_code
== Sendcomment
)
2249 quoted
= char_quoted (from
- 1, temp_pos
);
2252 DEC_BOTH (from
, from_byte
);
2253 temp_pos
= dec_bytepos (temp_pos
);
2254 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2256 c1
= FETCH_CHAR (temp_pos
);
2257 temp_code
= SYNTAX_WITH_MULTIBYTE_CHECK (c1
);
2258 if (! (quoted
|| temp_code
== Sword
2259 || temp_code
== Ssymbol
2260 || temp_code
== Squote
))
2262 DEC_BOTH (from
, from_byte
);
2269 temp_pos
= dec_bytepos (from_byte
);
2270 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2271 if (from
!= stop
&& c
== FETCH_CHAR (temp_pos
))
2272 DEC_BOTH (from
, from_byte
);
2281 if (!++depth
) goto done2
;
2286 if (!--depth
) goto done2
;
2287 if (depth
< min_depth
)
2288 Fsignal (Qscan_error
,
2289 Fcons (build_string ("Containing expression ends prematurely"),
2290 Fcons (make_number (last_good
),
2291 Fcons (make_number (from
), Qnil
))));
2295 if (!parse_sexp_ignore_comments
)
2297 found
= back_comment (from
, from_byte
, stop
, comnested
, comstyle
,
2298 &out_charpos
, &out_bytepos
);
2299 /* FIXME: if found == -1, then it really wasn't a comment-end.
2300 For single-char Sendcomment, we can't do much about it apart
2301 from skipping the char.
2302 For 2-char endcomments, we could try again, taking both
2303 chars as separate entities, but it's a lot of trouble
2304 for very little gain, so we don't bother either. -sm */
2306 from
= out_charpos
, from_byte
= out_bytepos
;
2309 case Scomment_fence
:
2313 DEC_BOTH (from
, from_byte
);
2314 if (from
== stop
) goto lose
;
2315 UPDATE_SYNTAX_TABLE_BACKWARD (from
);
2316 if (!char_quoted (from
, from_byte
)
2317 && (c
= FETCH_CHAR (from_byte
),
2318 SYNTAX_WITH_MULTIBYTE_CHECK (c
) == code
))
2321 if (code
== Sstring_fence
&& !depth
&& sexpflag
) goto done2
;
2325 stringterm
= FETCH_CHAR (from_byte
);
2328 if (from
== stop
) goto lose
;
2329 temp_pos
= from_byte
;
2330 if (! NILP (current_buffer
->enable_multibyte_characters
))
2334 UPDATE_SYNTAX_TABLE_BACKWARD (from
- 1);
2335 if (!char_quoted (from
- 1, temp_pos
)
2336 && stringterm
== (c
= FETCH_CHAR (temp_pos
))
2337 && SYNTAX_WITH_MULTIBYTE_CHECK (c
) == Sstring
)
2339 DEC_BOTH (from
, from_byte
);
2341 DEC_BOTH (from
, from_byte
);
2342 if (!depth
&& sexpflag
) goto done2
;
2345 /* Ignore whitespace, punctuation, quote, endcomment. */
2350 /* Reached start of buffer. Error if within object, return nil if between */
2351 if (depth
) goto lose
;
2362 XSETFASTINT (val
, from
);
2366 Fsignal (Qscan_error
,
2367 Fcons (build_string ("Unbalanced parentheses"),
2368 Fcons (make_number (last_good
),
2369 Fcons (make_number (from
), Qnil
))));
2374 DEFUN ("scan-lists", Fscan_lists
, Sscan_lists
, 3, 3, 0,
2375 doc
: /* Scan from character number FROM by COUNT lists.
2376 Returns the character number of the position thus found.
2378 If DEPTH is nonzero, paren depth begins counting from that value,
2379 only places where the depth in parentheses becomes zero
2380 are candidates for stopping; COUNT such places are counted.
2381 Thus, a positive value for DEPTH means go out levels.
2383 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2385 If the beginning or end of (the accessible part of) the buffer is reached
2386 and the depth is wrong, an error is signaled.
2387 If the depth is right but the count is not used up, nil is returned. */)
2388 (from
, count
, depth
)
2389 Lisp_Object from
, count
, depth
;
2391 CHECK_NUMBER (from
);
2392 CHECK_NUMBER (count
);
2393 CHECK_NUMBER (depth
);
2395 return scan_lists (XINT (from
), XINT (count
), XINT (depth
), 0);
2398 DEFUN ("scan-sexps", Fscan_sexps
, Sscan_sexps
, 2, 2, 0,
2399 doc
: /* Scan from character number FROM by COUNT balanced expressions.
2400 If COUNT is negative, scan backwards.
2401 Returns the character number of the position thus found.
2403 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
2405 If the beginning or end of (the accessible part of) the buffer is reached
2406 in the middle of a parenthetical grouping, an error is signaled.
2407 If the beginning or end is reached between groupings
2408 but before count is used up, nil is returned. */)
2410 Lisp_Object from
, count
;
2412 CHECK_NUMBER (from
);
2413 CHECK_NUMBER (count
);
2415 return scan_lists (XINT (from
), XINT (count
), 0, 1);
2418 DEFUN ("backward-prefix-chars", Fbackward_prefix_chars
, Sbackward_prefix_chars
,
2420 doc
: /* Move point backward over any number of chars with prefix syntax.
2421 This includes chars with "quote" or "prefix" syntax (' or p). */)
2426 int opoint_byte
= PT_BYTE
;
2428 int pos_byte
= PT_BYTE
;
2433 SET_PT_BOTH (opoint
, opoint_byte
);
2438 SETUP_SYNTAX_TABLE (pos
, -1);
2440 DEC_BOTH (pos
, pos_byte
);
2442 while (!char_quoted (pos
, pos_byte
)
2443 /* Previous statement updates syntax table. */
2444 && ((c
= FETCH_CHAR (pos_byte
), SYNTAX (c
) == Squote
)
2445 || SYNTAX_PREFIX (c
)))
2448 opoint_byte
= pos_byte
;
2451 DEC_BOTH (pos
, pos_byte
);
2454 SET_PT_BOTH (opoint
, opoint_byte
);
2459 /* Parse forward from FROM / FROM_BYTE to END,
2460 assuming that FROM has state OLDSTATE (nil means FROM is start of function),
2461 and return a description of the state of the parse at END.
2462 If STOPBEFORE is nonzero, stop at the start of an atom.
2463 If COMMENTSTOP is 1, stop at the start of a comment.
2464 If COMMENTSTOP is -1, stop at the start or end of a comment,
2465 after the beginning of a string, or after the end of a string. */
2468 scan_sexps_forward (stateptr
, from
, from_byte
, end
, targetdepth
,
2469 stopbefore
, oldstate
, commentstop
)
2470 struct lisp_parse_state
*stateptr
;
2472 int end
, targetdepth
, stopbefore
, from_byte
;
2473 Lisp_Object oldstate
;
2476 struct lisp_parse_state state
;
2478 register enum syntaxcode code
;
2481 struct level
{ int last
, prev
; };
2482 struct level levelstart
[100];
2483 register struct level
*curlevel
= levelstart
;
2484 struct level
*endlevel
= levelstart
+ 100;
2485 register int depth
; /* Paren depth of current scanning location.
2486 level - levelstart equals this except
2487 when the depth becomes negative. */
2488 int mindepth
; /* Lowest DEPTH value seen. */
2489 int start_quoted
= 0; /* Nonzero means starting after a char quote */
2491 int prev_from
; /* Keep one character before FROM. */
2493 int prev_from_syntax
;
2494 int boundary_stop
= commentstop
== -1;
2497 int out_bytepos
, out_charpos
;
2501 prev_from_byte
= from_byte
;
2503 DEC_BOTH (prev_from
, prev_from_byte
);
2505 /* Use this macro instead of `from++'. */
2507 do { prev_from = from; \
2508 prev_from_byte = from_byte; \
2510 = SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte)); \
2511 INC_BOTH (from, from_byte); \
2513 UPDATE_SYNTAX_TABLE_FORWARD (from); \
2519 if (NILP (oldstate
))
2522 state
.instring
= -1;
2523 state
.incomment
= 0;
2524 state
.comstyle
= 0; /* comment style a by default. */
2525 state
.comstr_start
= -1; /* no comment/string seen. */
2529 tem
= Fcar (oldstate
);
2535 oldstate
= Fcdr (oldstate
);
2536 oldstate
= Fcdr (oldstate
);
2537 oldstate
= Fcdr (oldstate
);
2538 tem
= Fcar (oldstate
);
2539 /* Check whether we are inside string_fence-style string: */
2540 state
.instring
= (!NILP (tem
)
2541 ? (INTEGERP (tem
) ? XINT (tem
) : ST_STRING_STYLE
)
2544 oldstate
= Fcdr (oldstate
);
2545 tem
= Fcar (oldstate
);
2546 state
.incomment
= (!NILP (tem
)
2547 ? (INTEGERP (tem
) ? XINT (tem
) : -1)
2550 oldstate
= Fcdr (oldstate
);
2551 tem
= Fcar (oldstate
);
2552 start_quoted
= !NILP (tem
);
2554 /* if the eighth element of the list is nil, we are in comment
2555 style a. If it is non-nil, we are in comment style b */
2556 oldstate
= Fcdr (oldstate
);
2557 oldstate
= Fcdr (oldstate
);
2558 tem
= Fcar (oldstate
);
2559 state
.comstyle
= NILP (tem
) ? 0 : (EQ (tem
, Qsyntax_table
)
2560 ? ST_COMMENT_STYLE
: 1);
2562 oldstate
= Fcdr (oldstate
);
2563 tem
= Fcar (oldstate
);
2564 state
.comstr_start
= NILP (tem
) ? -1 : XINT (tem
) ;
2565 oldstate
= Fcdr (oldstate
);
2566 tem
= Fcar (oldstate
);
2567 while (!NILP (tem
)) /* >= second enclosing sexps. */
2569 /* curlevel++->last ran into compiler bug on Apollo */
2570 curlevel
->last
= XINT (Fcar (tem
));
2571 if (++curlevel
== endlevel
)
2572 curlevel
--; /* error ("Nesting too deep for parser"); */
2573 curlevel
->prev
= -1;
2574 curlevel
->last
= -1;
2581 curlevel
->prev
= -1;
2582 curlevel
->last
= -1;
2584 SETUP_SYNTAX_TABLE (prev_from
, 1);
2585 prev_from_syntax
= SYNTAX_WITH_FLAGS (FETCH_CHAR (prev_from_byte
));
2586 UPDATE_SYNTAX_TABLE_FORWARD (from
);
2588 /* Enter the loop at a place appropriate for initial state. */
2590 if (state
.incomment
)
2591 goto startincomment
;
2592 if (state
.instring
>= 0)
2594 nofence
= state
.instring
!= ST_STRING_STYLE
;
2596 goto startquotedinstring
;
2599 else if (start_quoted
)
2605 code
= prev_from_syntax
& 0xff;
2607 if (code
== Scomment
)
2609 state
.comstyle
= SYNTAX_FLAGS_COMMENT_STYLE (prev_from_syntax
);
2610 state
.incomment
= (SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax
) ?
2612 state
.comstr_start
= prev_from
;
2614 else if (code
== Scomment_fence
)
2616 /* Record the comment style we have entered so that only
2617 the comment-end sequence of the same style actually
2618 terminates the comment section. */
2619 state
.comstyle
= ST_COMMENT_STYLE
;
2620 state
.incomment
= -1;
2621 state
.comstr_start
= prev_from
;
2624 else if (from
< end
)
2625 if (SYNTAX_FLAGS_COMSTART_FIRST (prev_from_syntax
))
2626 if (c1
= FETCH_CHAR (from_byte
),
2627 SYNTAX_COMSTART_SECOND (c1
))
2628 /* Duplicate code to avoid a complex if-expression
2629 which causes trouble for the SGI compiler. */
2631 /* Record the comment style we have entered so that only
2632 the comment-end sequence of the same style actually
2633 terminates the comment section. */
2634 state
.comstyle
= SYNTAX_COMMENT_STYLE (c1
);
2635 comnested
= SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax
);
2636 comnested
= comnested
|| SYNTAX_COMMENT_NESTED (c1
);
2637 state
.incomment
= comnested
? 1 : -1;
2638 state
.comstr_start
= prev_from
;
2643 if (SYNTAX_FLAGS_PREFIX (prev_from_syntax
))
2645 switch (SWITCH_ENUM_CAST (code
))
2649 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2650 curlevel
->last
= prev_from
;
2652 if (from
== end
) goto endquoted
;
2655 /* treat following character as a word constituent */
2658 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2659 curlevel
->last
= prev_from
;
2663 /* Some compilers can't handle this inside the switch. */
2664 temp
= SYNTAX (FETCH_CHAR (from_byte
));
2670 if (from
== end
) goto endquoted
;
2682 curlevel
->prev
= curlevel
->last
;
2685 case Scomment_fence
: /* Can't happen because it's handled above. */
2687 if (commentstop
|| boundary_stop
) goto done
;
2689 /* The (from == BEGV) test was to enter the loop in the middle so
2690 that we find a 2-char comment ender even if we start in the
2691 middle of it. We don't want to do that if we're just at the
2692 beginning of the comment (think of (*) ... (*)). */
2693 found
= forw_comment (from
, from_byte
, end
,
2694 state
.incomment
, state
.comstyle
,
2695 (from
== BEGV
|| from
< state
.comstr_start
+ 3)
2696 ? 0 : prev_from_syntax
,
2697 &out_charpos
, &out_bytepos
, &state
.incomment
);
2698 from
= out_charpos
; from_byte
= out_bytepos
;
2699 /* Beware! prev_from and friends are invalid now.
2700 Luckily, the `done' doesn't use them and the INC_FROM
2701 sets them to a sane value without looking at them. */
2702 if (!found
) goto done
;
2704 state
.incomment
= 0;
2705 state
.comstyle
= 0; /* reset the comment style */
2706 if (boundary_stop
) goto done
;
2710 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2712 /* curlevel++->last ran into compiler bug on Apollo */
2713 curlevel
->last
= prev_from
;
2714 if (++curlevel
== endlevel
)
2715 curlevel
--; /* error ("Nesting too deep for parser"); */
2716 curlevel
->prev
= -1;
2717 curlevel
->last
= -1;
2718 if (targetdepth
== depth
) goto done
;
2723 if (depth
< mindepth
)
2725 if (curlevel
!= levelstart
)
2727 curlevel
->prev
= curlevel
->last
;
2728 if (targetdepth
== depth
) goto done
;
2733 state
.comstr_start
= from
- 1;
2734 if (stopbefore
) goto stop
; /* this arg means stop at sexp start */
2735 curlevel
->last
= prev_from
;
2736 state
.instring
= (code
== Sstring
2737 ? (FETCH_CHAR (prev_from_byte
))
2739 if (boundary_stop
) goto done
;
2742 nofence
= state
.instring
!= ST_STRING_STYLE
;
2748 if (from
>= end
) goto done
;
2749 c
= FETCH_CHAR (from_byte
);
2750 /* Some compilers can't handle this inside the switch. */
2753 /* Check TEMP here so that if the char has
2754 a syntax-table property which says it is NOT
2755 a string character, it does not end the string. */
2756 if (nofence
&& c
== state
.instring
&& temp
== Sstring
)
2762 if (!nofence
) goto string_end
;
2767 startquotedinstring
:
2768 if (from
>= end
) goto endquoted
;
2774 state
.instring
= -1;
2775 curlevel
->prev
= curlevel
->last
;
2777 if (boundary_stop
) goto done
;
2781 /* FIXME: We should do something with it. */
2784 /* Ignore whitespace, punctuation, quote, endcomment. */
2790 stop
: /* Here if stopping before start of sexp. */
2791 from
= prev_from
; /* We have just fetched the char that starts it; */
2792 goto done
; /* but return the position before it. */
2797 state
.depth
= depth
;
2798 state
.mindepth
= mindepth
;
2799 state
.thislevelstart
= curlevel
->prev
;
2800 state
.prevlevelstart
2801 = (curlevel
== levelstart
) ? -1 : (curlevel
- 1)->last
;
2802 state
.location
= from
;
2803 state
.levelstarts
= Qnil
;
2804 while (--curlevel
>= levelstart
)
2805 state
.levelstarts
= Fcons (make_number (curlevel
->last
),
2812 DEFUN ("parse-partial-sexp", Fparse_partial_sexp
, Sparse_partial_sexp
, 2, 6, 0,
2813 doc
: /* Parse Lisp syntax starting at FROM until TO; return status of parse at TO.
2814 Parsing stops at TO or when certain criteria are met;
2815 point is set to where parsing stops.
2816 If fifth arg OLDSTATE is omitted or nil,
2817 parsing assumes that FROM is the beginning of a function.
2818 Value is a list of ten elements describing final state of parsing:
2820 1. character address of start of innermost containing list; nil if none.
2821 2. character address of start of last complete sexp terminated.
2822 3. non-nil if inside a string.
2823 (it is the character that will terminate the string,
2824 or t if the string should be terminated by a generic string delimiter.)
2825 4. nil if outside a comment, t if inside a non-nestable comment,
2826 else an integer (the current comment nesting).
2827 5. t if following a quote character.
2828 6. the minimum paren-depth encountered during this scan.
2829 7. t if in a comment of style b; symbol `syntax-table' if the comment
2830 should be terminated by a generic comment delimiter.
2831 8. character address of start of comment or string; nil if not in one.
2832 9. Intermediate data for continuation of parsing (subject to change).
2833 If third arg TARGETDEPTH is non-nil, parsing stops if the depth
2834 in parentheses becomes equal to TARGETDEPTH.
2835 Fourth arg STOPBEFORE non-nil means stop when come to
2836 any character that starts a sexp.
2837 Fifth arg OLDSTATE is a nine-element list like what this function returns.
2838 It is used to initialize the state of the parse. Elements number 1, 2, 6
2839 and 8 are ignored; you can leave off element 8 (the last) entirely.
2840 Sixth arg COMMENTSTOP non-nil means stop at the start of a comment.
2841 If it is symbol `syntax-table', stop after the start of a comment or a
2842 string, or after end of a comment or a string. */)
2843 (from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
)
2844 Lisp_Object from
, to
, targetdepth
, stopbefore
, oldstate
, commentstop
;
2846 struct lisp_parse_state state
;
2849 if (!NILP (targetdepth
))
2851 CHECK_NUMBER (targetdepth
);
2852 target
= XINT (targetdepth
);
2855 target
= -100000; /* We won't reach this depth */
2857 validate_region (&from
, &to
);
2858 scan_sexps_forward (&state
, XINT (from
), CHAR_TO_BYTE (XINT (from
)),
2860 target
, !NILP (stopbefore
), oldstate
,
2862 ? 0 : (EQ (commentstop
, Qsyntax_table
) ? -1 : 1)));
2864 SET_PT (state
.location
);
2866 return Fcons (make_number (state
.depth
),
2867 Fcons (state
.prevlevelstart
< 0 ? Qnil
: make_number (state
.prevlevelstart
),
2868 Fcons (state
.thislevelstart
< 0 ? Qnil
: make_number (state
.thislevelstart
),
2869 Fcons (state
.instring
>= 0
2870 ? (state
.instring
== ST_STRING_STYLE
2871 ? Qt
: make_number (state
.instring
)) : Qnil
,
2872 Fcons (state
.incomment
< 0 ? Qt
:
2873 (state
.incomment
== 0 ? Qnil
:
2874 make_number (state
.incomment
)),
2875 Fcons (state
.quoted
? Qt
: Qnil
,
2876 Fcons (make_number (state
.mindepth
),
2877 Fcons ((state
.comstyle
2878 ? (state
.comstyle
== ST_COMMENT_STYLE
2879 ? Qsyntax_table
: Qt
) :
2881 Fcons (((state
.incomment
2882 || (state
.instring
>= 0))
2883 ? make_number (state
.comstr_start
)
2885 Fcons (state
.levelstarts
, Qnil
))))))))));
2894 /* This has to be done here, before we call Fmake_char_table. */
2895 Qsyntax_table
= intern ("syntax-table");
2896 staticpro (&Qsyntax_table
);
2898 /* Intern this now in case it isn't already done.
2899 Setting this variable twice is harmless.
2900 But don't staticpro it here--that is done in alloc.c. */
2901 Qchar_table_extra_slots
= intern ("char-table-extra-slots");
2903 /* Create objects which can be shared among syntax tables. */
2904 Vsyntax_code_object
= Fmake_vector (make_number (Smax
), Qnil
);
2905 for (i
= 0; i
< XVECTOR (Vsyntax_code_object
)->size
; i
++)
2906 XVECTOR (Vsyntax_code_object
)->contents
[i
]
2907 = Fcons (make_number (i
), Qnil
);
2909 /* Now we are ready to set up this property, so we can
2910 create syntax tables. */
2911 Fput (Qsyntax_table
, Qchar_table_extra_slots
, make_number (0));
2913 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Swhitespace
];
2915 Vstandard_syntax_table
= Fmake_char_table (Qsyntax_table
, temp
);
2917 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Sword
];
2918 for (i
= 'a'; i
<= 'z'; i
++)
2919 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2920 for (i
= 'A'; i
<= 'Z'; i
++)
2921 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2922 for (i
= '0'; i
<= '9'; i
++)
2923 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, i
, temp
);
2925 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '$', temp
);
2926 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '%', temp
);
2928 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '(',
2929 Fcons (make_number (Sopen
), make_number (')')));
2930 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, ')',
2931 Fcons (make_number (Sclose
), make_number ('(')));
2932 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '[',
2933 Fcons (make_number (Sopen
), make_number (']')));
2934 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, ']',
2935 Fcons (make_number (Sclose
), make_number ('[')));
2936 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '{',
2937 Fcons (make_number (Sopen
), make_number ('}')));
2938 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '}',
2939 Fcons (make_number (Sclose
), make_number ('{')));
2940 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '"',
2941 Fcons (make_number ((int) Sstring
), Qnil
));
2942 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, '\\',
2943 Fcons (make_number ((int) Sescape
), Qnil
));
2945 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Ssymbol
];
2946 for (i
= 0; i
< 10; i
++)
2948 c
= "_-+*/&|<>="[i
];
2949 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, c
, temp
);
2952 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Spunct
];
2953 for (i
= 0; i
< 12; i
++)
2955 c
= ".,;:?!#@~^'`"[i
];
2956 SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table
, c
, temp
);
2959 /* All multibyte characters have syntax `word' by default. */
2960 temp
= XVECTOR (Vsyntax_code_object
)->contents
[(int) Sword
];
2961 for (i
= CHAR_TABLE_SINGLE_BYTE_SLOTS
; i
< CHAR_TABLE_ORDINARY_SLOTS
; i
++)
2962 XCHAR_TABLE (Vstandard_syntax_table
)->contents
[i
] = temp
;
2968 Qsyntax_table_p
= intern ("syntax-table-p");
2969 staticpro (&Qsyntax_table_p
);
2971 staticpro (&Vsyntax_code_object
);
2973 Qscan_error
= intern ("scan-error");
2974 staticpro (&Qscan_error
);
2975 Fput (Qscan_error
, Qerror_conditions
,
2976 Fcons (Qscan_error
, Fcons (Qerror
, Qnil
)));
2977 Fput (Qscan_error
, Qerror_message
,
2978 build_string ("Scan error"));
2980 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments
,
2981 doc
: /* Non-nil means `forward-sexp', etc., should treat comments as whitespace. */);
2983 DEFVAR_BOOL ("parse-sexp-lookup-properties", &parse_sexp_lookup_properties
,
2984 doc
: /* Non-nil means `forward-sexp', etc., obey `syntax-table' property.
2985 Otherwise, that text property is simply ignored.
2986 See the info node `(elisp)Syntax Properties' for a description of the
2987 `syntax-table' property. */);
2989 words_include_escapes
= 0;
2990 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes
,
2991 doc
: /* Non-nil means `forward-word', etc., should treat escape chars part of words. */);
2993 DEFVAR_BOOL ("multibyte-syntax-as-symbol", &multibyte_syntax_as_symbol
,
2994 doc
: /* Non-nil means `scan-sexps' treats all multibyte characters as symbol. */);
2995 multibyte_syntax_as_symbol
= 0;
2997 DEFVAR_BOOL ("open-paren-in-column-0-is-defun-start",
2998 &open_paren_in_column_0_is_defun_start
,
2999 doc
: /* *Non-nil means an open paren in column 0 denotes the start of a defun. */);
3000 open_paren_in_column_0_is_defun_start
= 1;
3002 defsubr (&Ssyntax_table_p
);
3003 defsubr (&Ssyntax_table
);
3004 defsubr (&Sstandard_syntax_table
);
3005 defsubr (&Scopy_syntax_table
);
3006 defsubr (&Sset_syntax_table
);
3007 defsubr (&Schar_syntax
);
3008 defsubr (&Smatching_paren
);
3009 defsubr (&Sstring_to_syntax
);
3010 defsubr (&Smodify_syntax_entry
);
3011 defsubr (&Sinternal_describe_syntax_value
);
3013 defsubr (&Sforward_word
);
3015 defsubr (&Sskip_chars_forward
);
3016 defsubr (&Sskip_chars_backward
);
3017 defsubr (&Sskip_syntax_forward
);
3018 defsubr (&Sskip_syntax_backward
);
3020 defsubr (&Sforward_comment
);
3021 defsubr (&Sscan_lists
);
3022 defsubr (&Sscan_sexps
);
3023 defsubr (&Sbackward_prefix_chars
);
3024 defsubr (&Sparse_partial_sexp
);