428
|
1 /* XEmacs routines to deal with syntax tables; also word and list parsing.
|
|
2 Copyright (C) 1985-1994 Free Software Foundation, Inc.
|
|
3 Copyright (C) 1995 Sun Microsystems, Inc.
|
1296
|
4 Copyright (C) 2001, 2002, 2003 Ben Wing.
|
428
|
5
|
|
6 This file is part of XEmacs.
|
|
7
|
|
8 XEmacs is free software; you can redistribute it and/or modify it
|
|
9 under the terms of the GNU General Public License as published by the
|
|
10 Free Software Foundation; either version 2, or (at your option) any
|
|
11 later version.
|
|
12
|
|
13 XEmacs is distributed in the hope that it will be useful, but WITHOUT
|
|
14 ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or
|
|
15 FITNESS FOR A PARTICULAR PURPOSE. See the GNU General Public License
|
|
16 for more details.
|
|
17
|
|
18 You should have received a copy of the GNU General Public License
|
|
19 along with XEmacs; see the file COPYING. If not, write to
|
|
20 the Free Software Foundation, Inc., 59 Temple Place - Suite 330,
|
|
21 Boston, MA 02111-1307, USA. */
|
|
22
|
|
23 /* Synched up with: FSF 19.28. */
|
|
24
|
|
25 /* This file has been Mule-ized. */
|
|
26
|
|
27 #include <config.h>
|
|
28 #include "lisp.h"
|
|
29
|
|
30 #include "buffer.h"
|
|
31 #include "syntax.h"
|
460
|
32 #include "extents.h"
|
428
|
33
|
460
|
34 #define ST_COMMENT_STYLE 0x101
|
|
35 #define ST_STRING_STYLE 0x102
|
|
36
|
|
37 Lisp_Object Qsyntax_table;
|
|
38 int lookup_syntax_properties;
|
|
39
|
428
|
40 Lisp_Object Qsyntax_table_p;
|
|
41
|
|
42 int words_include_escapes;
|
|
43
|
|
44 int parse_sexp_ignore_comments;
|
|
45
|
|
46 /* The following two variables are provided to tell additional information
|
|
47 to the regex routines. We do it this way rather than change the
|
|
48 arguments to re_search_2() in an attempt to maintain some call
|
|
49 compatibility with other versions of the regex code. */
|
|
50
|
|
51 /* Tell the regex routines not to QUIT. Normally there is a QUIT
|
|
52 each iteration in re_search_2(). */
|
|
53 int no_quit_in_re_search;
|
|
54
|
826
|
55 /* The standard syntax table is stored where it will automatically
|
|
56 be used in all new buffers. */
|
428
|
57 Lisp_Object Vstandard_syntax_table;
|
|
58
|
|
59 Lisp_Object Vsyntax_designator_chars_string;
|
|
60
|
826
|
61 Lisp_Object Vtemp_table_for_use_updating_syntax_tables;
|
|
62
|
1296
|
63 /* A value that is guaranteed not be in a syntax table. */
|
|
64 Lisp_Object Vbogus_syntax_table_value;
|
|
65
|
826
|
66 static void syntax_cache_table_was_changed (struct buffer *buf);
|
|
67
|
428
|
68 /* This is the internal form of the parse state used in parse-partial-sexp. */
|
|
69
|
|
70 struct lisp_parse_state
|
|
71 {
|
|
72 int depth; /* Depth at end of parsing */
|
867
|
73 Ichar instring; /* -1 if not within string, else desired terminator */
|
428
|
74 int incomment; /* Nonzero if within a comment at end of parsing */
|
460
|
75 int comstyle; /* comment style a=0, or b=1, or ST_COMMENT_STYLE */
|
428
|
76 int quoted; /* Nonzero if just after an escape char at end of
|
|
77 parsing */
|
665
|
78 Charbpos thislevelstart;/* Char number of most recent start-of-expression
|
428
|
79 at current level */
|
665
|
80 Charbpos prevlevelstart;/* Char number of start of containing expression */
|
|
81 Charbpos location; /* Char number at which parsing stopped */
|
428
|
82 int mindepth; /* Minimum depth seen while scanning */
|
826
|
83 Charbpos comstr_start;/* Position just after last comment/string starter */
|
|
84 Lisp_Object levelstarts;/* Char numbers of starts-of-expression
|
|
85 of levels (starting from outermost). */
|
428
|
86 };
|
|
87
|
|
88 /* These variables are a cache for finding the start of a defun.
|
|
89 find_start_pos is the place for which the defun start was found.
|
|
90 find_start_value is the defun start position found for it.
|
|
91 find_start_buffer is the buffer it was found in.
|
|
92 find_start_begv is the BEGV value when it was found.
|
|
93 find_start_modiff is the value of MODIFF when it was found. */
|
|
94
|
665
|
95 static Charbpos find_start_pos;
|
|
96 static Charbpos find_start_value;
|
428
|
97 static struct buffer *find_start_buffer;
|
665
|
98 static Charbpos find_start_begv;
|
428
|
99 static int find_start_modiff;
|
|
100
|
|
101 /* Find a defun-start that is the last one before POS (or nearly the last).
|
|
102 We record what we find, so that another call in the same area
|
|
103 can return the same value right away. */
|
|
104
|
665
|
105 static Charbpos
|
|
106 find_defun_start (struct buffer *buf, Charbpos pos)
|
428
|
107 {
|
665
|
108 Charbpos tem;
|
826
|
109 struct syntax_cache *scache;
|
|
110
|
428
|
111 /* Use previous finding, if it's valid and applies to this inquiry. */
|
|
112 if (buf == find_start_buffer
|
|
113 /* Reuse the defun-start even if POS is a little farther on.
|
|
114 POS might be in the next defun, but that's ok.
|
|
115 Our value may not be the best possible, but will still be usable. */
|
|
116 && pos <= find_start_pos + 1000
|
|
117 && pos >= find_start_value
|
|
118 && BUF_BEGV (buf) == find_start_begv
|
|
119 && BUF_MODIFF (buf) == find_start_modiff)
|
|
120 return find_start_value;
|
|
121
|
|
122 /* Back up to start of line. */
|
|
123 tem = find_next_newline (buf, pos, -1);
|
|
124
|
826
|
125 scache = setup_buffer_syntax_cache (buf, tem, 1);
|
428
|
126 while (tem > BUF_BEGV (buf))
|
|
127 {
|
826
|
128 UPDATE_SYNTAX_CACHE_BACKWARD (scache, tem);
|
460
|
129
|
428
|
130 /* Open-paren at start of line means we found our defun-start. */
|
826
|
131 if (SYNTAX_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, tem)) == Sopen)
|
428
|
132 break;
|
|
133 /* Move to beg of previous line. */
|
|
134 tem = find_next_newline (buf, tem, -2);
|
|
135 }
|
|
136
|
|
137 /* Record what we found, for the next try. */
|
|
138 find_start_value = tem;
|
|
139 find_start_buffer = buf;
|
|
140 find_start_modiff = BUF_MODIFF (buf);
|
|
141 find_start_begv = BUF_BEGV (buf);
|
|
142 find_start_pos = pos;
|
|
143
|
|
144 return find_start_value;
|
|
145 }
|
|
146
|
|
147 DEFUN ("syntax-table-p", Fsyntax_table_p, 1, 1, 0, /*
|
444
|
148 Return t if OBJECT is a syntax table.
|
428
|
149 */
|
444
|
150 (object))
|
428
|
151 {
|
444
|
152 return (CHAR_TABLEP (object)
|
|
153 && XCHAR_TABLE_TYPE (object) == CHAR_TABLE_TYPE_SYNTAX)
|
428
|
154 ? Qt : Qnil;
|
|
155 }
|
|
156
|
|
157 static Lisp_Object
|
|
158 check_syntax_table (Lisp_Object obj, Lisp_Object default_)
|
|
159 {
|
|
160 if (NILP (obj))
|
|
161 obj = default_;
|
|
162 while (NILP (Fsyntax_table_p (obj)))
|
|
163 obj = wrong_type_argument (Qsyntax_table_p, obj);
|
|
164 return obj;
|
|
165 }
|
|
166
|
|
167 DEFUN ("syntax-table", Fsyntax_table, 0, 1, 0, /*
|
|
168 Return the current syntax table.
|
|
169 This is the one specified by the current buffer, or by BUFFER if it
|
|
170 is non-nil.
|
|
171 */
|
|
172 (buffer))
|
|
173 {
|
|
174 return decode_buffer (buffer, 0)->syntax_table;
|
|
175 }
|
|
176
|
826
|
177 #ifdef DEBUG_XEMACS
|
|
178
|
|
179 DEFUN ("mirror-syntax-table", Fmirror_syntax_table, 0, 1, 0, /*
|
|
180 Return the current mirror syntax table, for debugging purposes.
|
|
181 This is the one specified by the current buffer, or by BUFFER if it
|
|
182 is non-nil.
|
|
183 */
|
|
184 (buffer))
|
|
185 {
|
|
186 return decode_buffer (buffer, 0)->mirror_syntax_table;
|
|
187 }
|
|
188
|
|
189 DEFUN ("syntax-cache-info", Fsyntax_cache_info, 0, 1, 0, /*
|
|
190 Return info about the syntax cache in BUFFER.
|
|
191 BUFFER defaults to the current buffer if nil.
|
|
192 */
|
|
193 (buffer))
|
|
194 {
|
|
195 struct buffer *buf = decode_buffer (buffer, 0);
|
|
196 struct syntax_cache *cache = buf->syntax_cache;
|
|
197 return list4 (cache->start, cache->end, make_int (cache->prev_change),
|
|
198 make_int (cache->next_change));
|
|
199 }
|
|
200
|
|
201 #endif /* DEBUG_XEMACS */
|
|
202
|
428
|
203 DEFUN ("standard-syntax-table", Fstandard_syntax_table, 0, 0, 0, /*
|
|
204 Return the standard syntax table.
|
|
205 This is the one used for new buffers.
|
|
206 */
|
|
207 ())
|
|
208 {
|
|
209 return Vstandard_syntax_table;
|
|
210 }
|
|
211
|
|
212 DEFUN ("copy-syntax-table", Fcopy_syntax_table, 0, 1, 0, /*
|
444
|
213 Return a new syntax table which is a copy of SYNTAX-TABLE.
|
|
214 SYNTAX-TABLE defaults to the standard syntax table.
|
428
|
215 */
|
444
|
216 (syntax_table))
|
428
|
217 {
|
|
218 if (NILP (Vstandard_syntax_table))
|
|
219 return Fmake_char_table (Qsyntax);
|
|
220
|
444
|
221 syntax_table = check_syntax_table (syntax_table, Vstandard_syntax_table);
|
|
222 return Fcopy_char_table (syntax_table);
|
428
|
223 }
|
|
224
|
|
225 DEFUN ("set-syntax-table", Fset_syntax_table, 1, 2, 0, /*
|
444
|
226 Select SYNTAX-TABLE as the new syntax table for BUFFER.
|
428
|
227 BUFFER defaults to the current buffer if omitted.
|
|
228 */
|
444
|
229 (syntax_table, buffer))
|
428
|
230 {
|
|
231 struct buffer *buf = decode_buffer (buffer, 0);
|
444
|
232 syntax_table = check_syntax_table (syntax_table, Qnil);
|
|
233 buf->syntax_table = syntax_table;
|
|
234 buf->mirror_syntax_table = XCHAR_TABLE (syntax_table)->mirror_table;
|
826
|
235 syntax_cache_table_was_changed (buf);
|
428
|
236 /* Indicate that this buffer now has a specified syntax table. */
|
|
237 buf->local_var_flags |= XINT (buffer_local_flags.syntax_table);
|
444
|
238 return syntax_table;
|
428
|
239 }
|
3252
|
240
|
|
241
|
428
|
242
|
3252
|
243 /*
|
|
244 * Syntax caching
|
|
245 */
|
|
246
|
|
247 /* syntax_cache object implementation */
|
|
248
|
|
249 static const struct memory_description syntax_cache_description_1 [] = {
|
|
250 { XD_LISP_OBJECT, offsetof (struct syntax_cache, object) },
|
|
251 { XD_LISP_OBJECT, offsetof (struct syntax_cache, buffer) },
|
|
252 { XD_LISP_OBJECT, offsetof (struct syntax_cache, syntax_table) },
|
|
253 { XD_LISP_OBJECT, offsetof (struct syntax_cache, mirror_table) },
|
|
254 { XD_LISP_OBJECT, offsetof (struct syntax_cache, start) },
|
|
255 { XD_LISP_OBJECT, offsetof (struct syntax_cache, end) },
|
|
256 { XD_END }
|
|
257 };
|
|
258
|
|
259 #ifdef NEW_GC
|
|
260 DEFINE_LRECORD_IMPLEMENTATION ("syntax-cache", syntax_cache,
|
|
261 1, /*dumpable-flag*/
|
|
262 0, 0, 0, 0, 0,
|
|
263 syntax_cache_description_1,
|
|
264 Lisp_Syntax_Cache);
|
|
265 #else /* not NEW_GC */
|
|
266
|
|
267 const struct sized_memory_description syntax_cache_description = {
|
|
268 sizeof (struct syntax_cache),
|
|
269 syntax_cache_description_1
|
|
270 };
|
|
271 #endif /* not NEW_GC */
|
|
272
|
|
273 /* static syntax cache utilities */
|
|
274
|
|
275 static void
|
|
276 syntax_cache_table_was_changed (struct buffer *buf)
|
|
277 {
|
|
278 struct syntax_cache *cache = buf->syntax_cache;
|
|
279 if (cache->no_syntax_table_prop)
|
|
280 {
|
|
281 cache->syntax_table =
|
|
282 BUFFER_SYNTAX_TABLE (buf);
|
|
283 cache->mirror_table =
|
|
284 BUFFER_MIRROR_SYNTAX_TABLE (buf);
|
|
285 }
|
|
286 }
|
|
287
|
|
288 static void
|
|
289 reset_buffer_syntax_cache_range (struct syntax_cache *cache,
|
|
290 Lisp_Object buffer, int infinite)
|
|
291 {
|
|
292 Fset_marker (cache->start, make_int (1), buffer);
|
|
293 Fset_marker (cache->end, make_int (1), buffer);
|
|
294 Fset_marker_insertion_type (cache->start, Qt);
|
|
295 Fset_marker_insertion_type (cache->end, Qnil);
|
|
296 /* #### Should we "cache->no_syntax_table_prop = 1;" here? */
|
|
297 /* #### Cf comment on INFINITE in init_syntax_cache. -- sjt */
|
|
298 if (infinite)
|
|
299 {
|
|
300 cache->prev_change = EMACS_INT_MIN;
|
|
301 cache->next_change = EMACS_INT_MAX;
|
|
302 }
|
|
303 else
|
|
304 {
|
|
305 cache->prev_change = -1;
|
|
306 cache->next_change = -1;
|
|
307 }
|
|
308 }
|
826
|
309
|
|
310 static void
|
|
311 init_syntax_cache (struct syntax_cache *cache, Lisp_Object object,
|
|
312 struct buffer *buffer, int infinite)
|
|
313 {
|
|
314 xzero (*cache);
|
|
315 cache->object = object;
|
|
316 cache->buffer = buffer;
|
|
317 cache->no_syntax_table_prop = 1;
|
1296
|
318 cache->syntax_table =
|
|
319 BUFFER_SYNTAX_TABLE (cache->buffer);
|
|
320 cache->mirror_table =
|
826
|
321 BUFFER_MIRROR_SYNTAX_TABLE (cache->buffer);
|
|
322 cache->start = Qnil;
|
|
323 cache->end = Qnil;
|
3250
|
324 /* #### I'm not sure what INFINITE is for, but it's apparently needed by
|
|
325 setup_syntax_cache(). It looks like it's supposed to guarantee that
|
|
326 the test for POS outside of cache-valid range will never succeed, so
|
|
327 that update_syntax_cache won't get called, but it's hard to be sure.
|
|
328 Cf reset_buffer_syntax_cache_range. -- sjt */
|
826
|
329 if (infinite)
|
|
330 {
|
|
331 cache->prev_change = EMACS_INT_MIN;
|
|
332 cache->next_change = EMACS_INT_MAX;
|
|
333 }
|
|
334 else
|
|
335 {
|
|
336 cache->prev_change = -1;
|
|
337 cache->next_change = -1;
|
|
338 }
|
|
339 }
|
|
340
|
3252
|
341 /* external syntax cache API */
|
|
342
|
3250
|
343 /* #### This function and associated logic still needs work, and especially
|
|
344 documentation. */
|
|
345 struct syntax_cache * /* return CACHE or the cache of OBJECT */
|
|
346 setup_syntax_cache (struct syntax_cache *cache, /* syntax cache, may be NULL
|
|
347 if OBJECT is a buffer */
|
|
348 Lisp_Object object, /* the object (if any) cache
|
|
349 is associated with */
|
|
350 struct buffer *buffer, /* the buffer to use as source
|
|
351 of the syntax table */
|
|
352 Charxpos from, /* initial position of cache */
|
|
353 int count) /* direction? see code */
|
826
|
354 {
|
3250
|
355 /* If OBJECT is a buffer, use its cache. Initialize cache. Make it valid
|
|
356 for the whole buffer if the syntax-table property is not being respected.
|
|
357 Else if OBJECT is not a buffer, initialize the cache passed in CACHE.
|
|
358 If the syntax-table property is being respected, update the cache. */
|
826
|
359 if (BUFFERP (object))
|
3250
|
360 {
|
|
361 cache = XBUFFER (object)->syntax_cache;
|
|
362 if (!lookup_syntax_properties)
|
|
363 reset_buffer_syntax_cache_range (cache, object, 1);
|
|
364 }
|
|
365 else
|
826
|
366 init_syntax_cache (cache, object, buffer, 0);
|
|
367 if (lookup_syntax_properties)
|
|
368 {
|
|
369 if (count <= 0)
|
|
370 {
|
|
371 from--;
|
2167
|
372 from = buffer_or_string_clip_to_accessible_char (cache->object,
|
826
|
373 from);
|
|
374 }
|
|
375 if (!(from >= cache->prev_change && from < cache->next_change))
|
|
376 update_syntax_cache (cache, from, count);
|
|
377 }
|
1296
|
378 #ifdef NOT_WORTH_THE_EFFORT
|
|
379 update_mirror_syntax_if_dirty (cache->mirror_table);
|
|
380 #endif /* NOT_WORTH_THE_EFFORT */
|
826
|
381 return cache;
|
|
382 }
|
|
383
|
|
384 struct syntax_cache *
|
|
385 setup_buffer_syntax_cache (struct buffer *buffer, Charxpos from, int count)
|
|
386 {
|
|
387 return setup_syntax_cache (NULL, wrap_buffer (buffer), buffer, from, count);
|
|
388 }
|
|
389
|
460
|
390 /*
|
|
391 Update syntax_cache to an appropriate setting for position POS
|
|
392
|
|
393 The sign of COUNT gives the relative position of POS wrt the
|
|
394 previously valid interval. (not currently used)
|
|
395
|
|
396 `syntax_cache.*_change' are the next and previous positions at
|
|
397 which syntax_code and c_s_t will need to be recalculated.
|
|
398
|
3025
|
399 #### Currently this code uses `get-char-property', which will
|
460
|
400 return the "last smallest" extent at a given position. In cases
|
|
401 where overlapping extents are defined, this code will simply use
|
|
402 whatever is returned by get-char-property.
|
|
403
|
|
404 It might be worth it at some point to merge provided syntax tables
|
826
|
405 outward to the current buffer (#### rewrite in English please?!). */
|
460
|
406
|
|
407 void
|
2286
|
408 update_syntax_cache (struct syntax_cache *cache, Charxpos cpos,
|
|
409 int UNUSED (count))
|
460
|
410 {
|
|
411 Lisp_Object tmp_table;
|
826
|
412 Bytexpos pos;
|
|
413 Bytexpos lim;
|
|
414 Bytexpos next, prev;
|
|
415 int at_begin = 0, at_end = 0;
|
460
|
416
|
826
|
417 if (NILP (cache->object))
|
|
418 return;
|
|
419
|
|
420 pos = buffer_or_string_charxpos_to_bytexpos (cache->object, cpos);
|
|
421
|
|
422 tmp_table = get_char_property (pos, Qsyntax_table, cache->object,
|
|
423 EXTENT_AT_AFTER, 0);
|
2506
|
424 lim = next_previous_single_property_change (pos, Qsyntax_table,
|
|
425 cache->object, -1, 1, 0);
|
826
|
426 if (lim < 0)
|
460
|
427 {
|
826
|
428 next = buffer_or_string_absolute_end_byte (cache->object);
|
|
429 at_begin = 1;
|
460
|
430 }
|
826
|
431 else
|
|
432 next = lim;
|
460
|
433
|
826
|
434 if (pos < buffer_or_string_absolute_end_byte (cache->object))
|
|
435 pos = next_bytexpos (cache->object, pos);
|
2506
|
436 lim = next_previous_single_property_change (pos, Qsyntax_table,
|
|
437 cache->object, -1, 0, 0);
|
826
|
438 if (lim < 0)
|
460
|
439 {
|
826
|
440 prev = buffer_or_string_absolute_begin_byte (cache->object);
|
|
441 at_end = 1;
|
460
|
442 }
|
|
443 else
|
826
|
444 prev = lim;
|
460
|
445
|
826
|
446 cache->prev_change =
|
|
447 buffer_or_string_bytexpos_to_charxpos (cache->object, prev);
|
|
448 cache->next_change =
|
|
449 buffer_or_string_bytexpos_to_charxpos (cache->object, next);
|
460
|
450
|
826
|
451 if (BUFFERP (cache->object))
|
|
452 {
|
|
453 /* If we are at the beginning or end of buffer, check to see if there's
|
|
454 a zero-length `syntax-table' extent there (highly unlikely); if not,
|
|
455 then we can safely make the end closed, so it will take in newly
|
|
456 inserted text. (If such an extent is inserted, we will be informed
|
3250
|
457 through signal_syntax_cache_extent_changed().) */
|
826
|
458 Fset_marker (cache->start, make_int (cache->prev_change), cache->object);
|
|
459 Fset_marker_insertion_type
|
|
460 (cache->start,
|
|
461 at_begin && NILP (extent_at (prev, cache->object, Qsyntax_table,
|
|
462 NULL, EXTENT_AT_AT, 0))
|
|
463 ? Qnil : Qt);
|
|
464 Fset_marker (cache->end, make_int (cache->next_change), cache->object);
|
|
465 Fset_marker_insertion_type
|
|
466 (cache->end,
|
|
467 at_end && NILP (extent_at (next, cache->object, Qsyntax_table,
|
|
468 NULL, EXTENT_AT_AT, 0))
|
|
469 ? Qt : Qnil);
|
|
470 }
|
|
471
|
|
472 if (!NILP (Fsyntax_table_p (tmp_table)))
|
|
473 {
|
|
474 cache->use_code = 0;
|
1296
|
475 cache->syntax_table = tmp_table;
|
|
476 cache->mirror_table = XCHAR_TABLE (tmp_table)->mirror_table;
|
826
|
477 cache->no_syntax_table_prop = 0;
|
1296
|
478 #ifdef NOT_WORTH_THE_EFFORT
|
|
479 update_mirror_syntax_if_dirty (cache->mirror_table);
|
|
480 #endif /* NOT_WORTH_THE_EFFORT */
|
826
|
481 }
|
|
482 else if (CONSP (tmp_table) && INTP (XCAR (tmp_table)))
|
|
483 {
|
|
484 cache->use_code = 1;
|
|
485 cache->syntax_code = XINT (XCAR (tmp_table));
|
|
486 cache->no_syntax_table_prop = 0;
|
|
487 }
|
|
488 else
|
|
489 {
|
|
490 cache->use_code = 0;
|
|
491 cache->no_syntax_table_prop = 1;
|
1296
|
492 cache->syntax_table = BUFFER_SYNTAX_TABLE (cache->buffer);
|
|
493 cache->mirror_table = BUFFER_MIRROR_SYNTAX_TABLE (cache->buffer);
|
|
494 #ifdef NOT_WORTH_THE_EFFORT
|
|
495 update_mirror_syntax_if_dirty (cache->mirror_table);
|
|
496 #endif /* NOT_WORTH_THE_EFFORT */
|
460
|
497 }
|
|
498 }
|
3252
|
499
|
|
500 /* buffer-specific APIs used in buffer.c
|
|
501 #### This is really unclean;
|
|
502 the syntax cache should just be a LISP object */
|
|
503
|
|
504 void
|
|
505 mark_buffer_syntax_cache (struct buffer *buf)
|
|
506 {
|
|
507 struct syntax_cache *cache = buf->syntax_cache;
|
|
508 if (!cache) /* Vbuffer_defaults and such don't have caches */
|
|
509 return;
|
|
510 mark_object (cache->object);
|
|
511 if (cache->buffer)
|
|
512 mark_object (wrap_buffer (cache->buffer));
|
|
513 mark_object (cache->syntax_table);
|
|
514 mark_object (cache->mirror_table);
|
|
515 mark_object (cache->start);
|
|
516 mark_object (cache->end);
|
|
517 }
|
|
518
|
|
519 void
|
|
520 init_buffer_syntax_cache (struct buffer *buf)
|
|
521 {
|
|
522 struct syntax_cache *cache;
|
|
523 #ifdef NEW_GC
|
|
524 buf->syntax_cache = alloc_lrecord_type (struct syntax_cache,
|
|
525 &lrecord_syntax_cache);
|
|
526 #else /* not NEW_GC */
|
|
527 buf->syntax_cache = xnew_and_zero (struct syntax_cache);
|
|
528 #endif /* not NEW_GC */
|
|
529 cache = buf->syntax_cache;
|
|
530 cache->object = wrap_buffer (buf);
|
|
531 cache->buffer = buf;
|
|
532 cache->no_syntax_table_prop = 1;
|
|
533 cache->syntax_table = BUFFER_SYNTAX_TABLE (cache->buffer);
|
|
534 cache->mirror_table = BUFFER_MIRROR_SYNTAX_TABLE (cache->buffer);
|
|
535 cache->start = Fmake_marker ();
|
|
536 cache->end = Fmake_marker ();
|
|
537 reset_buffer_syntax_cache_range (cache, cache->object, 0);
|
|
538 }
|
|
539
|
|
540 /* finalize the syntax cache for BUF */
|
|
541
|
|
542 void
|
|
543 uninit_buffer_syntax_cache (struct buffer *buf)
|
|
544 {
|
|
545 #ifdef NEW_GC
|
|
546 mc_free (buf->syntax_cache);
|
|
547 #else /* not NEW_GC */
|
|
548 xfree (buf->syntax_cache, struct syntax_cache *);
|
|
549 #endif /* not NEW_GC */
|
|
550 buf->syntax_cache = 0;
|
|
551 }
|
|
552
|
|
553 /* extent-specific APIs used in extents.c and insdel.c */
|
|
554
|
|
555 /* The syntax-table property on the range covered by EXTENT may be changing,
|
|
556 either because EXTENT has a syntax-table property and is being attached
|
|
557 or detached (this includes having its endpoints changed), or because
|
|
558 the value of EXTENT's syntax-table property is changing. */
|
|
559
|
|
560 void
|
|
561 signal_syntax_cache_extent_changed (EXTENT extent)
|
|
562 {
|
|
563 Lisp_Object buffer = Fextent_object (wrap_extent (extent));
|
|
564 if (BUFFERP (buffer))
|
|
565 {
|
|
566 /* This was getting called with the buffer's start and end null, eg in
|
|
567 cperl mode, which triggers an assert in byte_marker_position. Cf
|
|
568 thread rooted at <yxz7j7xzk97.fsf@gimli.holgi.priv> on xemacs-beta.
|
|
569 <yxzfymklb6p.fsf@gimli.holgi.priv> has a recipe, but you also need
|
|
570 to delete or type SPC to get the crash.
|
|
571 #### Delete this comment when setup_syntax_cache is made sane. */
|
|
572 struct syntax_cache *cache = XBUFFER (buffer)->syntax_cache;
|
|
573 /* #### would this be slower or less accurate in character terms? */
|
|
574 Bytexpos start = extent_endpoint_byte (extent, 0);
|
|
575 Bytexpos end = extent_endpoint_byte (extent, 1);
|
|
576 Bytexpos start2 = byte_marker_position (cache->start);
|
|
577 Bytexpos end2 = byte_marker_position (cache->end);
|
|
578 /* If the extent is entirely before or entirely after the cache
|
|
579 range, it doesn't overlap. Otherwise, invalidate the range. */
|
|
580 if (!(end < start2 || start > end2))
|
|
581 reset_buffer_syntax_cache_range (cache, buffer, 0);
|
|
582 }
|
|
583 }
|
|
584
|
|
585 /* Extents have been adjusted for insertion or deletion, so we need to
|
|
586 refetch the start and end position of the extent */
|
|
587 void
|
|
588 signal_syntax_cache_extent_adjust (struct buffer *buf)
|
|
589 {
|
|
590 struct syntax_cache *cache = buf->syntax_cache;
|
|
591 /* If the cache was invalid before, leave it that way. We only want
|
|
592 to update the limits of validity when they were actually valid. */
|
|
593 if (cache->prev_change < 0)
|
|
594 return;
|
|
595 cache->prev_change = marker_position (cache->start);
|
|
596 cache->next_change = marker_position (cache->end);
|
|
597 }
|
|
598
|
|
599
|
460
|
600
|
428
|
601 /* Convert a letter which signifies a syntax code
|
|
602 into the code it signifies.
|
|
603 This is used by modify-syntax-entry, and other things. */
|
|
604
|
442
|
605 const unsigned char syntax_spec_code[0400] =
|
428
|
606 { 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
607 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
608 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
609 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
610 (char) Swhitespace, 0377, (char) Sstring, 0377,
|
|
611 (char) Smath, 0377, 0377, (char) Squote,
|
|
612 (char) Sopen, (char) Sclose, 0377, 0377,
|
|
613 0377, (char) Swhitespace, (char) Spunct, (char) Scharquote,
|
|
614 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
615 0377, 0377, 0377, 0377,
|
|
616 (char) Scomment, 0377, (char) Sendcomment, 0377,
|
|
617 (char) Sinherit, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */
|
|
618 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
619 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
|
|
620 0377, 0377, 0377, 0377, (char) Sescape, 0377, 0377, (char) Ssymbol,
|
|
621 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* `, a, ... */
|
|
622 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377,
|
|
623 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword,
|
460
|
624 0377, 0377, 0377, 0377, (char) Sstring_fence, 0377, 0377, 0377
|
428
|
625 };
|
|
626
|
460
|
627 const unsigned char syntax_code_spec[] = " .w_()'\"$\\/<>@!|";
|
428
|
628
|
|
629 DEFUN ("syntax-designator-chars", Fsyntax_designator_chars, 0, 0, 0, /*
|
|
630 Return a string of the recognized syntax designator chars.
|
|
631 The chars are ordered by their internal syntax codes, which are
|
|
632 numbered starting at 0.
|
|
633 */
|
|
634 ())
|
|
635 {
|
|
636 return Vsyntax_designator_chars_string;
|
|
637 }
|
|
638
|
|
639 DEFUN ("char-syntax", Fchar_syntax, 1, 2, 0, /*
|
444
|
640 Return the syntax code of CHARACTER, described by a character.
|
|
641 For example, if CHARACTER is a word constituent,
|
|
642 the character `?w' is returned.
|
428
|
643 The characters that correspond to various syntax codes
|
|
644 are listed in the documentation of `modify-syntax-entry'.
|
444
|
645 Optional second argument SYNTAX-TABLE defaults to the current buffer's
|
428
|
646 syntax table.
|
|
647 */
|
444
|
648 (character, syntax_table))
|
428
|
649 {
|
826
|
650 Lisp_Object mirrortab;
|
428
|
651
|
444
|
652 if (NILP (character))
|
428
|
653 {
|
444
|
654 character = make_char ('\000');
|
428
|
655 }
|
444
|
656 CHECK_CHAR_COERCE_INT (character);
|
826
|
657 syntax_table = check_syntax_table (syntax_table,
|
|
658 current_buffer->syntax_table);
|
|
659 mirrortab = XCHAR_TABLE (syntax_table)->mirror_table;
|
|
660 return make_char (syntax_code_spec[(int) SYNTAX (mirrortab,
|
|
661 XCHAR (character))]);
|
428
|
662 }
|
|
663
|
|
664 #ifdef MULE
|
|
665
|
|
666 enum syntaxcode
|
2286
|
667 charset_syntax (struct buffer *UNUSED (buf), Lisp_Object UNUSED (charset),
|
|
668 int *multi_p_out)
|
428
|
669 {
|
|
670 *multi_p_out = 1;
|
826
|
671 /* !!#### get this right */
|
3152
|
672 return Sword;
|
428
|
673 }
|
|
674
|
|
675 #endif
|
|
676
|
|
677 Lisp_Object
|
867
|
678 syntax_match (Lisp_Object syntax_table, Ichar ch)
|
428
|
679 {
|
826
|
680 Lisp_Object code = get_char_table (ch, syntax_table);
|
428
|
681 Lisp_Object code2 = code;
|
|
682
|
|
683 if (CONSP (code))
|
|
684 code2 = XCAR (code);
|
|
685 if (SYNTAX_FROM_CODE (XINT (code2)) == Sinherit)
|
826
|
686 code = get_char_table (ch, Vstandard_syntax_table);
|
428
|
687
|
|
688 return CONSP (code) ? XCDR (code) : Qnil;
|
|
689 }
|
|
690
|
|
691 DEFUN ("matching-paren", Fmatching_paren, 1, 2, 0, /*
|
444
|
692 Return the matching parenthesis of CHARACTER, or nil if none.
|
|
693 Optional second argument SYNTAX-TABLE defaults to the current buffer's
|
428
|
694 syntax table.
|
|
695 */
|
444
|
696 (character, syntax_table))
|
428
|
697 {
|
826
|
698 Lisp_Object mirrortab;
|
1315
|
699 enum syntaxcode code;
|
428
|
700
|
444
|
701 CHECK_CHAR_COERCE_INT (character);
|
826
|
702 syntax_table = check_syntax_table (syntax_table,
|
|
703 current_buffer->syntax_table);
|
|
704 mirrortab = XCHAR_TABLE (syntax_table)->mirror_table;
|
444
|
705 code = SYNTAX (mirrortab, XCHAR (character));
|
428
|
706 if (code == Sopen || code == Sclose || code == Sstring)
|
444
|
707 return syntax_match (syntax_table, XCHAR (character));
|
428
|
708 return Qnil;
|
|
709 }
|
|
710
|
|
711
|
|
712
|
|
713 #ifdef MULE
|
|
714 /* Return 1 if there is a word boundary between two word-constituent
|
|
715 characters C1 and C2 if they appear in this order, else return 0.
|
|
716 There is no word boundary between two word-constituent ASCII
|
|
717 characters. */
|
|
718 #define WORD_BOUNDARY_P(c1, c2) \
|
867
|
719 (!(ichar_ascii_p (c1) && ichar_ascii_p (c2)) \
|
428
|
720 && word_boundary_p (c1, c2))
|
|
721 #endif
|
|
722
|
|
723 /* Return the position across COUNT words from FROM.
|
|
724 If that many words cannot be found before the end of the buffer, return 0.
|
|
725 COUNT negative means scan backward and stop at word beginning. */
|
|
726
|
665
|
727 Charbpos
|
|
728 scan_words (struct buffer *buf, Charbpos from, int count)
|
428
|
729 {
|
665
|
730 Charbpos limit = count > 0 ? BUF_ZV (buf) : BUF_BEGV (buf);
|
867
|
731 Ichar ch0, ch1;
|
428
|
732 enum syntaxcode code;
|
826
|
733 struct syntax_cache *scache = setup_buffer_syntax_cache (buf, from, count);
|
460
|
734
|
428
|
735 /* #### is it really worth it to hand expand both cases? JV */
|
|
736 while (count > 0)
|
|
737 {
|
|
738 QUIT;
|
|
739
|
|
740 while (1)
|
|
741 {
|
|
742 if (from == limit)
|
|
743 return 0;
|
|
744
|
826
|
745 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
746 ch0 = BUF_FETCH_CHAR (buf, from);
|
826
|
747 code = SYNTAX_FROM_CACHE (scache, ch0);
|
428
|
748
|
442
|
749 from++;
|
428
|
750 if (words_include_escapes
|
|
751 && (code == Sescape || code == Scharquote))
|
|
752 break;
|
|
753 if (code == Sword)
|
|
754 break;
|
|
755 }
|
|
756
|
|
757 QUIT;
|
|
758
|
|
759 while (from != limit)
|
|
760 {
|
826
|
761 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
762 ch1 = BUF_FETCH_CHAR (buf, from);
|
826
|
763 code = SYNTAX_FROM_CACHE (scache, ch1);
|
428
|
764 if (!(words_include_escapes
|
|
765 && (code == Sescape || code == Scharquote)))
|
|
766 if (code != Sword
|
|
767 #ifdef MULE
|
|
768 || WORD_BOUNDARY_P (ch0, ch1)
|
434
|
769 #endif
|
428
|
770 )
|
|
771 break;
|
|
772 #ifdef MULE
|
|
773 ch0 = ch1;
|
434
|
774 #endif
|
428
|
775 from++;
|
|
776 }
|
|
777 count--;
|
|
778 }
|
|
779
|
|
780 while (count < 0)
|
|
781 {
|
|
782 QUIT;
|
|
783
|
|
784 while (1)
|
|
785 {
|
|
786 if (from == limit)
|
|
787 return 0;
|
|
788
|
826
|
789 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
428
|
790 ch1 = BUF_FETCH_CHAR (buf, from - 1);
|
826
|
791 code = SYNTAX_FROM_CACHE (scache, ch1);
|
460
|
792 from--;
|
442
|
793
|
428
|
794 if (words_include_escapes
|
|
795 && (code == Sescape || code == Scharquote))
|
|
796 break;
|
|
797 if (code == Sword)
|
|
798 break;
|
|
799 }
|
|
800
|
|
801 QUIT;
|
|
802
|
|
803 while (from != limit)
|
|
804 {
|
826
|
805 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
428
|
806 ch0 = BUF_FETCH_CHAR (buf, from - 1);
|
826
|
807 code = SYNTAX_FROM_CACHE (scache, ch0);
|
460
|
808
|
428
|
809 if (!(words_include_escapes
|
|
810 && (code == Sescape || code == Scharquote)))
|
|
811 if (code != Sword
|
|
812 #ifdef MULE
|
|
813 || WORD_BOUNDARY_P (ch0, ch1)
|
|
814 #endif
|
|
815 )
|
|
816 break;
|
|
817 #ifdef MULE
|
|
818 ch1 = ch0;
|
|
819 #endif
|
|
820 from--;
|
|
821 }
|
|
822 count++;
|
|
823 }
|
|
824
|
|
825 return from;
|
|
826 }
|
|
827
|
446
|
828 DEFUN ("forward-word", Fforward_word, 0, 2, "_p", /*
|
428
|
829 Move point forward COUNT words (backward if COUNT is negative).
|
446
|
830 Normally t is returned, but if an edge of the buffer is reached,
|
|
831 point is left there and nil is returned.
|
428
|
832
|
462
|
833 The characters that are moved over may be added to the current selection
|
|
834 \(i.e. active region) if the Shift key is held down, a motion key is used
|
|
835 to invoke this command, and `shifted-motion-keys-select-region' is t; see
|
|
836 the documentation for this variable for more details.
|
|
837
|
446
|
838 COUNT defaults to 1, and BUFFER defaults to the current buffer.
|
428
|
839 */
|
|
840 (count, buffer))
|
|
841 {
|
665
|
842 Charbpos val;
|
428
|
843 struct buffer *buf = decode_buffer (buffer, 0);
|
446
|
844 EMACS_INT n;
|
|
845
|
|
846 if (NILP (count))
|
|
847 n = 1;
|
|
848 else
|
|
849 {
|
|
850 CHECK_INT (count);
|
|
851 n = XINT (count);
|
|
852 }
|
428
|
853
|
446
|
854 val = scan_words (buf, BUF_PT (buf), n);
|
|
855 if (val)
|
428
|
856 {
|
446
|
857 BUF_SET_PT (buf, val);
|
|
858 return Qt;
|
|
859 }
|
|
860 else
|
|
861 {
|
|
862 BUF_SET_PT (buf, n > 0 ? BUF_ZV (buf) : BUF_BEGV (buf));
|
428
|
863 return Qnil;
|
|
864 }
|
|
865 }
|
|
866
|
|
867 static void scan_sexps_forward (struct buffer *buf,
|
|
868 struct lisp_parse_state *,
|
665
|
869 Charbpos from, Charbpos end,
|
428
|
870 int targetdepth, int stopbefore,
|
|
871 Lisp_Object oldstate,
|
|
872 int commentstop);
|
|
873
|
|
874 static int
|
665
|
875 find_start_of_comment (struct buffer *buf, Charbpos from, Charbpos stop,
|
460
|
876 int comstyle)
|
428
|
877 {
|
867
|
878 Ichar c;
|
428
|
879 enum syntaxcode code;
|
|
880
|
|
881 /* Look back, counting the parity of string-quotes,
|
|
882 and recording the comment-starters seen.
|
|
883 When we reach a safe place, assume that's not in a string;
|
|
884 then step the main scan to the earliest comment-starter seen
|
|
885 an even number of string quotes away from the safe place.
|
|
886
|
|
887 OFROM[I] is position of the earliest comment-starter seen
|
|
888 which is I+2X quotes from the comment-end.
|
|
889 PARITY is current parity of quotes from the comment end. */
|
|
890 int parity = 0;
|
867
|
891 Ichar my_stringend = 0;
|
428
|
892 int string_lossage = 0;
|
665
|
893 Charbpos comment_end = from;
|
|
894 Charbpos comstart_pos = 0;
|
428
|
895 int comstart_parity = 0;
|
|
896 int styles_match_p = 0;
|
460
|
897 /* mask to match comment styles against; for ST_COMMENT_STYLE, this
|
|
898 will get set to SYNTAX_COMMENT_STYLE_B, but never get checked */
|
|
899 int mask = comstyle ? SYNTAX_COMMENT_STYLE_B : SYNTAX_COMMENT_STYLE_A;
|
826
|
900 struct syntax_cache *scache = buf->syntax_cache;
|
428
|
901
|
|
902 /* At beginning of range to scan, we're outside of strings;
|
|
903 that determines quote parity to the comment-end. */
|
|
904 while (from != stop)
|
|
905 {
|
460
|
906 int syncode;
|
|
907
|
428
|
908 /* Move back and examine a character. */
|
|
909 from--;
|
826
|
910 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from);
|
428
|
911
|
|
912 c = BUF_FETCH_CHAR (buf, from);
|
826
|
913 syncode = SYNTAX_CODE_FROM_CACHE (scache, c);
|
|
914 code = SYNTAX_FROM_CODE (syncode);
|
428
|
915
|
|
916 /* is this a 1-char comment end sequence? if so, try
|
|
917 to see if style matches previously extracted mask */
|
|
918 if (code == Sendcomment)
|
|
919 {
|
|
920 styles_match_p =
|
460
|
921 SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode) & mask;
|
428
|
922 }
|
|
923
|
|
924 /* or are we looking at a 1-char comment start sequence
|
|
925 of the style matching mask? */
|
460
|
926 else if (code == Scomment)
|
428
|
927 {
|
460
|
928 styles_match_p =
|
|
929 SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode) & mask;
|
428
|
930 }
|
|
931
|
460
|
932 /* otherwise, is this a 2-char comment end or start sequence? */
|
|
933 else if (from > stop)
|
|
934 do
|
|
935 {
|
|
936 /* 2-char comment end sequence? */
|
|
937 if (SYNTAX_CODE_END_SECOND_P (syncode))
|
|
938 {
|
|
939 int prev_syncode;
|
826
|
940 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
460
|
941 prev_syncode =
|
1315
|
942 SYNTAX_CODE_FROM_CACHE (scache,
|
|
943 BUF_FETCH_CHAR (buf, from - 1));
|
460
|
944
|
|
945 if (SYNTAX_CODES_END_P (prev_syncode, syncode))
|
|
946 {
|
|
947 code = Sendcomment;
|
|
948 styles_match_p =
|
826
|
949 SYNTAX_CODES_COMMENT_MASK_END (prev_syncode,
|
|
950 syncode) & mask;
|
460
|
951 from--;
|
826
|
952 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from);
|
460
|
953 c = BUF_FETCH_CHAR (buf, from);
|
|
954
|
|
955 /* Found a comment-end sequence, so skip past the
|
|
956 check for a comment-start */
|
|
957 break;
|
|
958 }
|
|
959 }
|
|
960
|
|
961 /* 2-char comment start sequence? */
|
|
962 if (SYNTAX_CODE_START_SECOND_P (syncode))
|
|
963 {
|
|
964 int prev_syncode;
|
826
|
965 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
460
|
966 prev_syncode =
|
1315
|
967 SYNTAX_CODE_FROM_CACHE (scache,
|
|
968 BUF_FETCH_CHAR (buf, from - 1));
|
460
|
969
|
|
970 if (SYNTAX_CODES_START_P (prev_syncode, syncode))
|
|
971 {
|
|
972 code = Scomment;
|
|
973 styles_match_p =
|
826
|
974 SYNTAX_CODES_COMMENT_MASK_START (prev_syncode,
|
|
975 syncode) & mask;
|
460
|
976 from--;
|
826
|
977 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from);
|
460
|
978 c = BUF_FETCH_CHAR (buf, from);
|
|
979 }
|
|
980 }
|
|
981 } while (0);
|
428
|
982
|
|
983 /* Ignore escaped characters. */
|
|
984 if (char_quoted (buf, from))
|
|
985 continue;
|
|
986
|
|
987 /* Track parity of quotes. */
|
|
988 if (code == Sstring)
|
|
989 {
|
|
990 parity ^= 1;
|
|
991 if (my_stringend == 0)
|
|
992 my_stringend = c;
|
|
993 /* If we have two kinds of string delimiters.
|
|
994 There's no way to grok this scanning backwards. */
|
|
995 else if (my_stringend != c)
|
|
996 string_lossage = 1;
|
|
997 }
|
|
998
|
460
|
999 if (code == Sstring_fence || code == Scomment_fence)
|
|
1000 {
|
|
1001 parity ^= 1;
|
|
1002 if (my_stringend == 0)
|
|
1003 my_stringend =
|
|
1004 code == Sstring_fence ? ST_STRING_STYLE : ST_COMMENT_STYLE;
|
|
1005 /* If we have two kinds of string delimiters.
|
|
1006 There's no way to grok this scanning backwards. */
|
|
1007 else if (my_stringend != (code == Sstring_fence
|
|
1008 ? ST_STRING_STYLE : ST_COMMENT_STYLE))
|
|
1009 string_lossage = 1;
|
|
1010 }
|
|
1011
|
428
|
1012 /* Record comment-starters according to that
|
|
1013 quote-parity to the comment-end. */
|
|
1014 if (code == Scomment && styles_match_p)
|
|
1015 {
|
|
1016 comstart_parity = parity;
|
|
1017 comstart_pos = from;
|
|
1018 }
|
|
1019
|
|
1020 /* If we find another earlier comment-ender,
|
|
1021 any comment-starts earlier than that don't count
|
|
1022 (because they go with the earlier comment-ender). */
|
|
1023 if (code == Sendcomment && styles_match_p)
|
|
1024 break;
|
|
1025
|
|
1026 /* Assume a defun-start point is outside of strings. */
|
|
1027 if (code == Sopen
|
|
1028 && (from == stop || BUF_FETCH_CHAR (buf, from - 1) == '\n'))
|
|
1029 break;
|
|
1030 }
|
|
1031
|
|
1032 if (comstart_pos == 0)
|
|
1033 from = comment_end;
|
|
1034 /* If the earliest comment starter
|
|
1035 is followed by uniform paired string quotes or none,
|
|
1036 we know it can't be inside a string
|
|
1037 since if it were then the comment ender would be inside one.
|
|
1038 So it does start a comment. Skip back to it. */
|
|
1039 else if (comstart_parity == 0 && !string_lossage)
|
|
1040 from = comstart_pos;
|
|
1041 else
|
|
1042 {
|
|
1043 /* We had two kinds of string delimiters mixed up
|
|
1044 together. Decode this going forwards.
|
|
1045 Scan fwd from the previous comment ender
|
|
1046 to the one in question; this records where we
|
|
1047 last passed a comment starter. */
|
|
1048
|
|
1049 struct lisp_parse_state state;
|
|
1050 scan_sexps_forward (buf, &state, find_defun_start (buf, comment_end),
|
|
1051 comment_end - 1, -10000, 0, Qnil, 0);
|
|
1052 if (state.incomment)
|
460
|
1053 from = state.comstr_start;
|
428
|
1054 else
|
|
1055 /* We can't grok this as a comment; scan it normally. */
|
|
1056 from = comment_end;
|
826
|
1057 UPDATE_SYNTAX_CACHE_FORWARD (scache, from - 1);
|
428
|
1058 }
|
|
1059 return from;
|
|
1060 }
|
|
1061
|
665
|
1062 static Charbpos
|
826
|
1063 find_end_of_comment (struct buffer *buf, Charbpos from, Charbpos stop,
|
|
1064 int comstyle)
|
428
|
1065 {
|
|
1066 int c;
|
460
|
1067 int prev_code;
|
|
1068 /* mask to match comment styles against; for ST_COMMENT_STYLE, this
|
|
1069 will get set to SYNTAX_COMMENT_STYLE_B, but never get checked */
|
|
1070 int mask = comstyle ? SYNTAX_COMMENT_STYLE_B : SYNTAX_COMMENT_STYLE_A;
|
826
|
1071 struct syntax_cache *scache = buf->syntax_cache;
|
428
|
1072
|
460
|
1073 /* This is only called by functions which have already set up the
|
|
1074 syntax_cache and are keeping it up-to-date */
|
428
|
1075 while (1)
|
|
1076 {
|
|
1077 if (from == stop)
|
|
1078 {
|
|
1079 return -1;
|
|
1080 }
|
460
|
1081
|
826
|
1082 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
1083 c = BUF_FETCH_CHAR (buf, from);
|
460
|
1084
|
|
1085 /* Test for generic comments */
|
|
1086 if (comstyle == ST_COMMENT_STYLE)
|
|
1087 {
|
826
|
1088 if (SYNTAX_FROM_CACHE (scache, c) == Scomment_fence)
|
460
|
1089 {
|
|
1090 from++;
|
826
|
1091 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1092 break;
|
|
1093 }
|
|
1094 from++;
|
|
1095 continue; /* No need to test other comment styles in a
|
|
1096 generic comment */
|
|
1097 }
|
|
1098 else
|
|
1099
|
826
|
1100 if (SYNTAX_FROM_CACHE (scache, c) == Sendcomment
|
460
|
1101 && SYNTAX_CODE_MATCHES_1CHAR_P
|
826
|
1102 (SYNTAX_CODE_FROM_CACHE (scache, c), mask))
|
428
|
1103 /* we have encountered a comment end of the same style
|
|
1104 as the comment sequence which began this comment
|
|
1105 section */
|
460
|
1106 {
|
|
1107 from++;
|
826
|
1108 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1109 break;
|
|
1110 }
|
428
|
1111
|
826
|
1112 prev_code = SYNTAX_CODE_FROM_CACHE (scache, c);
|
428
|
1113 from++;
|
826
|
1114 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
1115 if (from < stop
|
460
|
1116 && SYNTAX_CODES_MATCH_END_P
|
|
1117 (prev_code,
|
826
|
1118 SYNTAX_CODE_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from)),
|
460
|
1119 mask)
|
|
1120
|
|
1121 )
|
428
|
1122 /* we have encountered a comment end of the same style
|
|
1123 as the comment sequence which began this comment
|
|
1124 section */
|
460
|
1125 {
|
|
1126 from++;
|
826
|
1127 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1128 break;
|
|
1129 }
|
428
|
1130 }
|
|
1131 return from;
|
|
1132 }
|
|
1133
|
|
1134
|
|
1135 /* #### between FSF 19.23 and 19.28 there are some changes to the logic
|
|
1136 in this function (and minor changes to find_start_of_comment(),
|
|
1137 above, which is part of Fforward_comment() in FSF). Attempts to port
|
|
1138 that logic made this function break, so I'm leaving it out. If anyone
|
|
1139 ever complains about this function not working properly, take a look
|
|
1140 at those changes. --ben */
|
|
1141
|
446
|
1142 DEFUN ("forward-comment", Fforward_comment, 0, 2, 0, /*
|
444
|
1143 Move forward across up to COUNT comments, or backwards if COUNT is negative.
|
428
|
1144 Stop scanning if we find something other than a comment or whitespace.
|
|
1145 Set point to where scanning stops.
|
444
|
1146 If COUNT comments are found as expected, with nothing except whitespace
|
428
|
1147 between them, return t; otherwise return nil.
|
|
1148 Point is set in either case.
|
446
|
1149 COUNT defaults to 1, and BUFFER defaults to the current buffer.
|
428
|
1150 */
|
444
|
1151 (count, buffer))
|
428
|
1152 {
|
665
|
1153 Charbpos from;
|
|
1154 Charbpos stop;
|
867
|
1155 Ichar c;
|
428
|
1156 enum syntaxcode code;
|
460
|
1157 int syncode;
|
444
|
1158 EMACS_INT n;
|
428
|
1159 struct buffer *buf = decode_buffer (buffer, 0);
|
826
|
1160 struct syntax_cache *scache;
|
|
1161
|
446
|
1162 if (NILP (count))
|
|
1163 n = 1;
|
|
1164 else
|
|
1165 {
|
|
1166 CHECK_INT (count);
|
|
1167 n = XINT (count);
|
|
1168 }
|
428
|
1169
|
|
1170 from = BUF_PT (buf);
|
|
1171
|
826
|
1172 scache = setup_buffer_syntax_cache (buf, from, n);
|
444
|
1173 while (n > 0)
|
428
|
1174 {
|
|
1175 QUIT;
|
|
1176
|
|
1177 stop = BUF_ZV (buf);
|
|
1178 while (from < stop)
|
|
1179 {
|
460
|
1180 int comstyle = 0; /* mask for finding matching comment style */
|
428
|
1181
|
|
1182 if (char_quoted (buf, from))
|
|
1183 {
|
|
1184 from++;
|
|
1185 continue;
|
|
1186 }
|
|
1187
|
826
|
1188 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
1189 c = BUF_FETCH_CHAR (buf, from);
|
826
|
1190 syncode = SYNTAX_CODE_FROM_CACHE (scache, c);
|
|
1191 code = SYNTAX_FROM_CODE (syncode);
|
428
|
1192
|
|
1193 if (code == Scomment)
|
|
1194 {
|
|
1195 /* we have encountered a single character comment start
|
|
1196 sequence, and we are ignoring all text inside comments.
|
|
1197 we must record the comment style this character begins
|
|
1198 so that later, only a comment end of the same style actually
|
|
1199 ends the comment section */
|
460
|
1200 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
|
|
1201 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
428
|
1202 }
|
|
1203
|
460
|
1204 else if (code == Scomment_fence)
|
|
1205 {
|
|
1206 from++;
|
|
1207 code = Scomment;
|
|
1208 comstyle = ST_COMMENT_STYLE;
|
|
1209 }
|
|
1210
|
428
|
1211 else if (from < stop
|
460
|
1212 && SYNTAX_CODE_START_FIRST_P (syncode))
|
428
|
1213 {
|
460
|
1214 int next_syncode;
|
826
|
1215 UPDATE_SYNTAX_CACHE_FORWARD (scache, from + 1);
|
460
|
1216 next_syncode =
|
826
|
1217 SYNTAX_CODE_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from + 1));
|
460
|
1218
|
|
1219 if (SYNTAX_CODES_START_P (syncode, next_syncode))
|
|
1220 {
|
|
1221 /* we have encountered a 2char comment start sequence and we
|
|
1222 are ignoring all text inside comments. we must record
|
|
1223 the comment style this sequence begins so that later,
|
|
1224 only a comment end of the same style actually ends
|
|
1225 the comment section */
|
|
1226 code = Scomment;
|
|
1227 comstyle =
|
|
1228 SYNTAX_CODES_COMMENT_MASK_START (syncode, next_syncode)
|
|
1229 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
|
1230 from++;
|
|
1231 }
|
428
|
1232 }
|
|
1233
|
|
1234 if (code == Scomment)
|
|
1235 {
|
826
|
1236 Charbpos newfrom = find_end_of_comment (buf, from, stop,
|
|
1237 comstyle);
|
428
|
1238 if (newfrom < 0)
|
|
1239 {
|
|
1240 /* we stopped because from==stop */
|
|
1241 BUF_SET_PT (buf, stop);
|
|
1242 return Qnil;
|
|
1243 }
|
|
1244 from = newfrom;
|
|
1245
|
|
1246 /* We have skipped one comment. */
|
|
1247 break;
|
|
1248 }
|
|
1249 else if (code != Swhitespace
|
|
1250 && code != Sendcomment
|
|
1251 && code != Scomment )
|
|
1252 {
|
|
1253 BUF_SET_PT (buf, from);
|
|
1254 return Qnil;
|
|
1255 }
|
|
1256 from++;
|
|
1257 }
|
|
1258
|
|
1259 /* End of comment reached */
|
444
|
1260 n--;
|
428
|
1261 }
|
|
1262
|
444
|
1263 while (n < 0)
|
428
|
1264 {
|
|
1265 QUIT;
|
|
1266
|
|
1267 stop = BUF_BEGV (buf);
|
|
1268 while (from > stop)
|
|
1269 {
|
460
|
1270 int comstyle = 0; /* mask for finding matching comment style */
|
428
|
1271
|
|
1272 from--;
|
|
1273 if (char_quoted (buf, from))
|
|
1274 {
|
|
1275 from--;
|
|
1276 continue;
|
|
1277 }
|
|
1278
|
|
1279 c = BUF_FETCH_CHAR (buf, from);
|
826
|
1280 syncode = SYNTAX_CODE_FROM_CACHE (scache, c);
|
|
1281 code = SYNTAX_FROM_CODE (syncode);
|
428
|
1282
|
|
1283 if (code == Sendcomment)
|
|
1284 {
|
|
1285 /* we have found a single char end comment. we must record
|
|
1286 the comment style encountered so that later, we can match
|
|
1287 only the proper comment begin sequence of the same style */
|
460
|
1288 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
|
|
1289 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
|
1290 }
|
|
1291
|
|
1292 else if (code == Scomment_fence)
|
|
1293 {
|
|
1294 code = Sendcomment;
|
|
1295 comstyle = ST_COMMENT_STYLE;
|
428
|
1296 }
|
|
1297
|
|
1298 else if (from > stop
|
460
|
1299 && SYNTAX_CODE_END_SECOND_P (syncode))
|
428
|
1300 {
|
460
|
1301 int prev_syncode;
|
826
|
1302 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
460
|
1303 prev_syncode =
|
826
|
1304 SYNTAX_CODE_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from - 1));
|
460
|
1305 if (SYNTAX_CODES_END_P (prev_syncode, syncode))
|
|
1306 {
|
|
1307 /* We must record the comment style encountered so that
|
|
1308 later, we can match only the proper comment begin
|
|
1309 sequence of the same style. */
|
|
1310 code = Sendcomment;
|
|
1311 comstyle = SYNTAX_CODES_COMMENT_MASK_END
|
|
1312 (prev_syncode, syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
|
1313 from--;
|
|
1314 }
|
428
|
1315 }
|
|
1316
|
|
1317 if (code == Sendcomment)
|
|
1318 {
|
460
|
1319 from = find_start_of_comment (buf, from, stop, comstyle);
|
428
|
1320 break;
|
|
1321 }
|
|
1322
|
|
1323 else if (code != Swhitespace
|
460
|
1324 && code != Scomment
|
|
1325 && code != Sendcomment)
|
428
|
1326 {
|
|
1327 BUF_SET_PT (buf, from + 1);
|
|
1328 return Qnil;
|
|
1329 }
|
|
1330 }
|
|
1331
|
444
|
1332 n++;
|
428
|
1333 }
|
|
1334
|
|
1335 BUF_SET_PT (buf, from);
|
|
1336 return Qt;
|
|
1337 }
|
|
1338
|
|
1339
|
|
1340 Lisp_Object
|
665
|
1341 scan_lists (struct buffer *buf, Charbpos from, int count, int depth,
|
444
|
1342 int sexpflag, int noerror)
|
428
|
1343 {
|
665
|
1344 Charbpos stop;
|
867
|
1345 Ichar c;
|
428
|
1346 int quoted;
|
|
1347 int mathexit = 0;
|
|
1348 enum syntaxcode code;
|
460
|
1349 int syncode;
|
428
|
1350 int min_depth = depth; /* Err out if depth gets less than this. */
|
826
|
1351 struct syntax_cache *scache;
|
|
1352
|
428
|
1353 if (depth > 0) min_depth = 0;
|
|
1354
|
826
|
1355 scache = setup_buffer_syntax_cache (buf, from, count);
|
428
|
1356 while (count > 0)
|
|
1357 {
|
|
1358 QUIT;
|
|
1359
|
|
1360 stop = BUF_ZV (buf);
|
|
1361 while (from < stop)
|
|
1362 {
|
460
|
1363 int comstyle = 0; /* mask for finding matching comment style */
|
428
|
1364
|
826
|
1365 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
1366 c = BUF_FETCH_CHAR (buf, from);
|
826
|
1367 syncode = SYNTAX_CODE_FROM_CACHE (scache, c);
|
|
1368 code = SYNTAX_FROM_CODE (syncode);
|
428
|
1369 from++;
|
|
1370
|
|
1371 /* a 1-char comment start sequence */
|
|
1372 if (code == Scomment && parse_sexp_ignore_comments)
|
|
1373 {
|
460
|
1374 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode) ==
|
|
1375 SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
428
|
1376 }
|
|
1377
|
|
1378 /* else, a 2-char comment start sequence? */
|
|
1379 else if (from < stop
|
460
|
1380 && SYNTAX_CODE_START_FIRST_P (syncode)
|
428
|
1381 && parse_sexp_ignore_comments)
|
|
1382 {
|
460
|
1383 int next_syncode;
|
826
|
1384 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1385 next_syncode =
|
826
|
1386 SYNTAX_CODE_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from));
|
460
|
1387
|
|
1388 if (SYNTAX_CODES_START_P (syncode, next_syncode))
|
|
1389 {
|
826
|
1390 /* we have encountered a comment start sequence and we
|
|
1391 are ignoring all text inside comments. we must record
|
|
1392 the comment style this sequence begins so that later,
|
|
1393 only a comment end of the same style actually ends
|
|
1394 the comment section */
|
|
1395 code = Scomment;
|
460
|
1396 comstyle = SYNTAX_CODES_COMMENT_MASK_START
|
|
1397 (syncode, next_syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
826
|
1398 from++;
|
|
1399 }
|
428
|
1400 }
|
826
|
1401 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
1402
|
460
|
1403 if (SYNTAX_CODE_PREFIX (syncode))
|
428
|
1404 continue;
|
|
1405
|
|
1406 switch (code)
|
|
1407 {
|
|
1408 case Sescape:
|
|
1409 case Scharquote:
|
|
1410 if (from == stop) goto lose;
|
|
1411 from++;
|
|
1412 /* treat following character as a word constituent */
|
|
1413 case Sword:
|
|
1414 case Ssymbol:
|
|
1415 if (depth || !sexpflag) break;
|
|
1416 /* This word counts as a sexp; return at end of it. */
|
|
1417 while (from < stop)
|
|
1418 {
|
826
|
1419 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
|
1420 switch (SYNTAX_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from)))
|
428
|
1421 {
|
|
1422 case Scharquote:
|
|
1423 case Sescape:
|
|
1424 from++;
|
|
1425 if (from == stop) goto lose;
|
|
1426 break;
|
|
1427 case Sword:
|
|
1428 case Ssymbol:
|
|
1429 case Squote:
|
|
1430 break;
|
|
1431 default:
|
|
1432 goto done;
|
|
1433 }
|
|
1434 from++;
|
|
1435 }
|
|
1436 goto done;
|
|
1437
|
460
|
1438 case Scomment_fence:
|
|
1439 comstyle = ST_COMMENT_STYLE;
|
428
|
1440 case Scomment:
|
|
1441 if (!parse_sexp_ignore_comments)
|
|
1442 break;
|
826
|
1443 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
1444 {
|
665
|
1445 Charbpos newfrom =
|
460
|
1446 find_end_of_comment (buf, from, stop, comstyle);
|
428
|
1447 if (newfrom < 0)
|
|
1448 {
|
|
1449 /* we stopped because from == stop in search forward */
|
|
1450 from = stop;
|
|
1451 if (depth == 0)
|
|
1452 goto done;
|
|
1453 goto lose;
|
|
1454 }
|
|
1455 from = newfrom;
|
|
1456 }
|
|
1457 break;
|
|
1458
|
|
1459 case Smath:
|
|
1460 if (!sexpflag)
|
|
1461 break;
|
|
1462 if (from != stop && c == BUF_FETCH_CHAR (buf, from))
|
|
1463 from++;
|
|
1464 if (mathexit)
|
|
1465 {
|
|
1466 mathexit = 0;
|
|
1467 goto close1;
|
|
1468 }
|
|
1469 mathexit = 1;
|
|
1470
|
|
1471 case Sopen:
|
|
1472 if (!++depth) goto done;
|
|
1473 break;
|
|
1474
|
|
1475 case Sclose:
|
|
1476 close1:
|
|
1477 if (!--depth) goto done;
|
|
1478 if (depth < min_depth)
|
|
1479 {
|
444
|
1480 if (noerror)
|
428
|
1481 return Qnil;
|
826
|
1482 syntax_error ("Containing expression ends prematurely",
|
|
1483 Qunbound);
|
428
|
1484 }
|
|
1485 break;
|
|
1486
|
460
|
1487 case Sstring_fence:
|
428
|
1488 case Sstring:
|
|
1489 {
|
867
|
1490 Ichar stringterm;
|
460
|
1491
|
|
1492 if (code != Sstring_fence)
|
|
1493 {
|
826
|
1494 /* XEmacs change: call syntax_match on character */
|
867
|
1495 Ichar ch = BUF_FETCH_CHAR (buf, from - 1);
|
460
|
1496 Lisp_Object stermobj =
|
1296
|
1497 syntax_match (scache->syntax_table, ch);
|
428
|
1498
|
|
1499 if (CHARP (stermobj))
|
|
1500 stringterm = XCHAR (stermobj);
|
|
1501 else
|
|
1502 stringterm = ch;
|
460
|
1503 }
|
|
1504 else
|
|
1505 stringterm = '\0'; /* avoid compiler warnings */
|
428
|
1506
|
|
1507 while (1)
|
|
1508 {
|
|
1509 if (from >= stop)
|
|
1510 goto lose;
|
826
|
1511 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1512 c = BUF_FETCH_CHAR (buf, from);
|
|
1513 if (code == Sstring
|
|
1514 ? c == stringterm
|
826
|
1515 : SYNTAX_FROM_CACHE (scache, c) == Sstring_fence)
|
428
|
1516 break;
|
460
|
1517
|
826
|
1518 switch (SYNTAX_FROM_CACHE (scache, c))
|
428
|
1519 {
|
|
1520 case Scharquote:
|
|
1521 case Sescape:
|
|
1522 from++;
|
|
1523 break;
|
|
1524 default:
|
|
1525 break;
|
|
1526 }
|
|
1527 from++;
|
|
1528 }
|
|
1529 from++;
|
|
1530 if (!depth && sexpflag) goto done;
|
|
1531 break;
|
|
1532 }
|
|
1533
|
|
1534 default:
|
|
1535 break;
|
|
1536 }
|
|
1537 }
|
|
1538
|
|
1539 /* Reached end of buffer. Error if within object,
|
|
1540 return nil if between */
|
|
1541 if (depth) goto lose;
|
|
1542
|
|
1543 return Qnil;
|
|
1544
|
|
1545 /* End of object reached */
|
|
1546 done:
|
|
1547 count--;
|
|
1548 }
|
|
1549
|
|
1550 while (count < 0)
|
|
1551 {
|
|
1552 QUIT;
|
|
1553
|
|
1554 stop = BUF_BEGV (buf);
|
|
1555 while (from > stop)
|
|
1556 {
|
460
|
1557 int comstyle = 0; /* mask for finding matching comment style */
|
428
|
1558
|
|
1559 from--;
|
826
|
1560 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from);
|
428
|
1561 quoted = char_quoted (buf, from);
|
|
1562 if (quoted)
|
460
|
1563 {
|
428
|
1564 from--;
|
826
|
1565 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from);
|
460
|
1566 }
|
428
|
1567
|
|
1568 c = BUF_FETCH_CHAR (buf, from);
|
826
|
1569 syncode = SYNTAX_CODE_FROM_CACHE (scache, c);
|
|
1570 code = SYNTAX_FROM_CODE (syncode);
|
428
|
1571
|
|
1572 if (code == Sendcomment && parse_sexp_ignore_comments)
|
|
1573 {
|
|
1574 /* we have found a single char end comment. we must record
|
|
1575 the comment style encountered so that later, we can match
|
|
1576 only the proper comment begin sequence of the same style */
|
460
|
1577 comstyle = SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
|
|
1578 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
428
|
1579 }
|
|
1580
|
|
1581 else if (from > stop
|
460
|
1582 && SYNTAX_CODE_END_SECOND_P (syncode)
|
428
|
1583 && !char_quoted (buf, from - 1)
|
|
1584 && parse_sexp_ignore_comments)
|
|
1585 {
|
460
|
1586 int prev_syncode;
|
826
|
1587 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
|
1588 prev_syncode =
|
|
1589 SYNTAX_CODE_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from - 1));
|
460
|
1590
|
|
1591 if (SYNTAX_CODES_END_P (prev_syncode, syncode))
|
|
1592 {
|
428
|
1593 /* we must record the comment style encountered so that
|
|
1594 later, we can match only the proper comment begin
|
|
1595 sequence of the same style */
|
|
1596 code = Sendcomment;
|
460
|
1597 comstyle = SYNTAX_CODES_COMMENT_MASK_END
|
|
1598 (prev_syncode, syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
428
|
1599 from--;
|
|
1600 }
|
460
|
1601 }
|
428
|
1602
|
460
|
1603 if (SYNTAX_CODE_PREFIX (syncode))
|
428
|
1604 continue;
|
|
1605
|
434
|
1606 switch (quoted ? Sword : code)
|
428
|
1607 {
|
|
1608 case Sword:
|
|
1609 case Ssymbol:
|
|
1610 if (depth || !sexpflag) break;
|
|
1611 /* This word counts as a sexp; count object finished after
|
|
1612 passing it. */
|
|
1613 while (from > stop)
|
|
1614 {
|
826
|
1615 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from);
|
428
|
1616 quoted = char_quoted (buf, from - 1);
|
|
1617
|
|
1618 if (quoted)
|
|
1619 from--;
|
|
1620 if (! (quoted
|
|
1621 || (syncode =
|
826
|
1622 SYNTAX_FROM_CACHE (scache, BUF_FETCH_CHAR (buf,
|
|
1623 from - 1)))
|
428
|
1624 == Sword
|
|
1625 || syncode == Ssymbol
|
|
1626 || syncode == Squote))
|
|
1627 goto done2;
|
|
1628 from--;
|
|
1629 }
|
|
1630 goto done2;
|
|
1631
|
|
1632 case Smath:
|
|
1633 if (!sexpflag)
|
|
1634 break;
|
|
1635 if (from != stop && c == BUF_FETCH_CHAR (buf, from - 1))
|
|
1636 from--;
|
|
1637 if (mathexit)
|
|
1638 {
|
|
1639 mathexit = 0;
|
|
1640 goto open2;
|
|
1641 }
|
|
1642 mathexit = 1;
|
|
1643
|
|
1644 case Sclose:
|
|
1645 if (!++depth) goto done2;
|
|
1646 break;
|
|
1647
|
|
1648 case Sopen:
|
|
1649 open2:
|
|
1650 if (!--depth) goto done2;
|
|
1651 if (depth < min_depth)
|
|
1652 {
|
444
|
1653 if (noerror)
|
428
|
1654 return Qnil;
|
826
|
1655 syntax_error ("Containing expression ends prematurely",
|
|
1656 Qunbound);
|
428
|
1657 }
|
|
1658 break;
|
|
1659
|
460
|
1660 case Scomment_fence:
|
|
1661 comstyle = ST_COMMENT_STYLE;
|
428
|
1662 case Sendcomment:
|
|
1663 if (parse_sexp_ignore_comments)
|
460
|
1664 from = find_start_of_comment (buf, from, stop, comstyle);
|
428
|
1665 break;
|
|
1666
|
460
|
1667 case Sstring_fence:
|
428
|
1668 case Sstring:
|
|
1669 {
|
867
|
1670 Ichar stringterm;
|
460
|
1671
|
|
1672 if (code != Sstring_fence)
|
|
1673 {
|
428
|
1674 /* XEmacs change: call syntax_match() on character */
|
867
|
1675 Ichar ch = BUF_FETCH_CHAR (buf, from);
|
460
|
1676 Lisp_Object stermobj =
|
1296
|
1677 syntax_match (scache->syntax_table, ch);
|
428
|
1678
|
|
1679 if (CHARP (stermobj))
|
|
1680 stringterm = XCHAR (stermobj);
|
|
1681 else
|
|
1682 stringterm = ch;
|
460
|
1683 }
|
|
1684 else
|
|
1685 stringterm = '\0'; /* avoid compiler warnings */
|
428
|
1686
|
|
1687 while (1)
|
|
1688 {
|
|
1689 if (from == stop) goto lose;
|
460
|
1690
|
826
|
1691 UPDATE_SYNTAX_CACHE_BACKWARD (scache, from - 1);
|
460
|
1692 c = BUF_FETCH_CHAR (buf, from - 1);
|
|
1693
|
|
1694 if ((code == Sstring
|
|
1695 ? c == stringterm
|
826
|
1696 : SYNTAX_FROM_CACHE (scache, c) == Sstring_fence)
|
460
|
1697 && !char_quoted (buf, from - 1))
|
|
1698 {
|
428
|
1699 break;
|
460
|
1700 }
|
|
1701
|
428
|
1702 from--;
|
|
1703 }
|
|
1704 from--;
|
|
1705 if (!depth && sexpflag) goto done2;
|
|
1706 break;
|
|
1707 }
|
|
1708 }
|
|
1709 }
|
|
1710
|
|
1711 /* Reached start of buffer. Error if within object,
|
|
1712 return nil if between */
|
|
1713 if (depth) goto lose;
|
|
1714
|
|
1715 return Qnil;
|
|
1716
|
|
1717 done2:
|
|
1718 count++;
|
|
1719 }
|
|
1720
|
|
1721
|
|
1722 return (make_int (from));
|
|
1723
|
|
1724 lose:
|
444
|
1725 if (!noerror)
|
826
|
1726 syntax_error ("Unbalanced parentheses", Qunbound);
|
428
|
1727 return Qnil;
|
|
1728 }
|
|
1729
|
|
1730 int
|
665
|
1731 char_quoted (struct buffer *buf, Charbpos pos)
|
428
|
1732 {
|
|
1733 enum syntaxcode code;
|
665
|
1734 Charbpos beg = BUF_BEGV (buf);
|
428
|
1735 int quoted = 0;
|
665
|
1736 Charbpos startpos = pos;
|
826
|
1737 struct syntax_cache *scache = buf->syntax_cache;
|
460
|
1738
|
|
1739 while (pos > beg)
|
|
1740 {
|
826
|
1741 UPDATE_SYNTAX_CACHE_BACKWARD (scache, pos - 1);
|
|
1742 code = SYNTAX_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, pos - 1));
|
428
|
1743
|
460
|
1744 if (code != Scharquote && code != Sescape)
|
|
1745 break;
|
|
1746 pos--;
|
|
1747 quoted = !quoted;
|
|
1748 }
|
|
1749
|
826
|
1750 UPDATE_SYNTAX_CACHE (scache, startpos);
|
428
|
1751 return quoted;
|
|
1752 }
|
|
1753
|
|
1754 DEFUN ("scan-lists", Fscan_lists, 3, 5, 0, /*
|
|
1755 Scan from character number FROM by COUNT lists.
|
|
1756 Returns the character number of the position thus found.
|
|
1757
|
|
1758 If DEPTH is nonzero, paren depth begins counting from that value,
|
|
1759 only places where the depth in parentheses becomes zero
|
|
1760 are candidates for stopping; COUNT such places are counted.
|
|
1761 Thus, a positive value for DEPTH means go out levels.
|
|
1762
|
|
1763 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
|
|
1764
|
|
1765 If the beginning or end of (the accessible part of) the buffer is reached
|
|
1766 and the depth is wrong, an error is signaled.
|
|
1767 If the depth is right but the count is not used up, nil is returned.
|
|
1768
|
|
1769 If optional arg BUFFER is non-nil, scanning occurs in that buffer instead
|
|
1770 of in the current buffer.
|
|
1771
|
|
1772 If optional arg NOERROR is non-nil, scan-lists will return nil instead of
|
|
1773 signalling an error.
|
|
1774 */
|
444
|
1775 (from, count, depth, buffer, noerror))
|
428
|
1776 {
|
|
1777 struct buffer *buf;
|
|
1778
|
|
1779 CHECK_INT (from);
|
|
1780 CHECK_INT (count);
|
|
1781 CHECK_INT (depth);
|
|
1782 buf = decode_buffer (buffer, 0);
|
|
1783
|
|
1784 return scan_lists (buf, XINT (from), XINT (count), XINT (depth), 0,
|
444
|
1785 !NILP (noerror));
|
428
|
1786 }
|
|
1787
|
|
1788 DEFUN ("scan-sexps", Fscan_sexps, 2, 4, 0, /*
|
|
1789 Scan from character number FROM by COUNT balanced expressions.
|
|
1790 If COUNT is negative, scan backwards.
|
|
1791 Returns the character number of the position thus found.
|
|
1792
|
|
1793 Comments are ignored if `parse-sexp-ignore-comments' is non-nil.
|
|
1794
|
|
1795 If the beginning or end of (the accessible part of) the buffer is reached
|
|
1796 in the middle of a parenthetical grouping, an error is signaled.
|
|
1797 If the beginning or end is reached between groupings
|
|
1798 but before count is used up, nil is returned.
|
|
1799
|
|
1800 If optional arg BUFFER is non-nil, scanning occurs in that buffer instead
|
|
1801 of in the current buffer.
|
|
1802
|
|
1803 If optional arg NOERROR is non-nil, scan-sexps will return nil instead of
|
|
1804 signalling an error.
|
|
1805 */
|
444
|
1806 (from, count, buffer, noerror))
|
428
|
1807 {
|
|
1808 struct buffer *buf = decode_buffer (buffer, 0);
|
|
1809 CHECK_INT (from);
|
|
1810 CHECK_INT (count);
|
|
1811
|
444
|
1812 return scan_lists (buf, XINT (from), XINT (count), 0, 1, !NILP (noerror));
|
428
|
1813 }
|
|
1814
|
|
1815 DEFUN ("backward-prefix-chars", Fbackward_prefix_chars, 0, 1, 0, /*
|
|
1816 Move point backward over any number of chars with prefix syntax.
|
|
1817 This includes chars with "quote" or "prefix" syntax (' or p).
|
|
1818
|
|
1819 Optional arg BUFFER defaults to the current buffer.
|
|
1820 */
|
|
1821 (buffer))
|
|
1822 {
|
|
1823 struct buffer *buf = decode_buffer (buffer, 0);
|
665
|
1824 Charbpos beg = BUF_BEGV (buf);
|
|
1825 Charbpos pos = BUF_PT (buf);
|
867
|
1826 Ichar c = '\0'; /* initialize to avoid compiler warnings */
|
826
|
1827 struct syntax_cache *scache;
|
|
1828
|
|
1829 scache = setup_buffer_syntax_cache (buf, pos, -1);
|
428
|
1830
|
|
1831 while (pos > beg && !char_quoted (buf, pos - 1)
|
460
|
1832 /* Previous statement updates syntax table. */
|
826
|
1833 && (SYNTAX_FROM_CACHE (scache, c = BUF_FETCH_CHAR (buf, pos - 1)) == Squote
|
|
1834 || SYNTAX_CODE_PREFIX (SYNTAX_CODE_FROM_CACHE (scache, c))))
|
428
|
1835 pos--;
|
|
1836
|
|
1837 BUF_SET_PT (buf, pos);
|
|
1838
|
|
1839 return Qnil;
|
|
1840 }
|
|
1841
|
|
1842 /* Parse forward from FROM to END,
|
|
1843 assuming that FROM has state OLDSTATE (nil means FROM is start of function),
|
|
1844 and return a description of the state of the parse at END.
|
|
1845 If STOPBEFORE is nonzero, stop at the start of an atom.
|
|
1846 If COMMENTSTOP is nonzero, stop at the start of a comment. */
|
|
1847
|
|
1848 static void
|
|
1849 scan_sexps_forward (struct buffer *buf, struct lisp_parse_state *stateptr,
|
665
|
1850 Charbpos from, Charbpos end,
|
428
|
1851 int targetdepth, int stopbefore,
|
|
1852 Lisp_Object oldstate,
|
|
1853 int commentstop)
|
|
1854 {
|
|
1855 struct lisp_parse_state state;
|
|
1856
|
|
1857 enum syntaxcode code;
|
|
1858 struct level { int last, prev; };
|
|
1859 struct level levelstart[100];
|
|
1860 struct level *curlevel = levelstart;
|
|
1861 struct level *endlevel = levelstart + 100;
|
|
1862 int depth; /* Paren depth of current scanning location.
|
|
1863 level - levelstart equals this except
|
|
1864 when the depth becomes negative. */
|
|
1865 int mindepth; /* Lowest DEPTH value seen. */
|
|
1866 int start_quoted = 0; /* Nonzero means starting after a char quote */
|
460
|
1867 int boundary_stop = commentstop == -1;
|
428
|
1868 Lisp_Object tem;
|
826
|
1869 struct syntax_cache *scache;
|
|
1870
|
|
1871 scache = setup_buffer_syntax_cache (buf, from, 1);
|
428
|
1872 if (NILP (oldstate))
|
|
1873 {
|
|
1874 depth = 0;
|
|
1875 state.instring = -1;
|
|
1876 state.incomment = 0;
|
|
1877 state.comstyle = 0; /* comment style a by default */
|
460
|
1878 state.comstr_start = -1; /* no comment/string seen. */
|
428
|
1879 }
|
|
1880 else
|
|
1881 {
|
|
1882 tem = Fcar (oldstate); /* elt 0, depth */
|
|
1883 if (!NILP (tem))
|
|
1884 depth = XINT (tem);
|
|
1885 else
|
|
1886 depth = 0;
|
|
1887
|
|
1888 oldstate = Fcdr (oldstate);
|
|
1889 oldstate = Fcdr (oldstate);
|
|
1890 oldstate = Fcdr (oldstate);
|
|
1891 tem = Fcar (oldstate); /* elt 3, instring */
|
460
|
1892 state.instring = ( !NILP (tem)
|
|
1893 ? ( INTP (tem) ? XINT (tem) : ST_STRING_STYLE)
|
|
1894 : -1);
|
428
|
1895
|
460
|
1896 oldstate = Fcdr (oldstate);
|
|
1897 tem = Fcar (oldstate); /* elt 4, incomment */
|
428
|
1898 state.incomment = !NILP (tem);
|
|
1899
|
|
1900 oldstate = Fcdr (oldstate);
|
|
1901 tem = Fcar (oldstate); /* elt 5, follows-quote */
|
|
1902 start_quoted = !NILP (tem);
|
|
1903
|
|
1904 /* if the eighth element of the list is nil, we are in comment style
|
3025
|
1905 a; if it is t, we are in comment style b; if it is `syntax-table',
|
460
|
1906 we are in a generic comment */
|
428
|
1907 oldstate = Fcdr (oldstate);
|
|
1908 oldstate = Fcdr (oldstate);
|
460
|
1909 tem = Fcar (oldstate); /* elt 7, comment style a/b/fence */
|
|
1910 state.comstyle = NILP (tem) ? 0 : ( EQ (tem, Qsyntax_table)
|
|
1911 ? ST_COMMENT_STYLE : 1 );
|
|
1912
|
|
1913 oldstate = Fcdr (oldstate); /* elt 8, start of last comment/string */
|
|
1914 tem = Fcar (oldstate);
|
|
1915 state.comstr_start = NILP (tem) ? -1 : XINT (tem);
|
|
1916
|
|
1917 /* elt 9, char numbers of starts-of-expression of levels
|
|
1918 (starting from outermost). */
|
|
1919 oldstate = Fcdr (oldstate);
|
|
1920 tem = Fcar (oldstate); /* elt 9, intermediate data for
|
|
1921 continuation of parsing (subject
|
|
1922 to change). */
|
|
1923 while (!NILP (tem)) /* >= second enclosing sexps. */
|
|
1924 {
|
|
1925 curlevel->last = XINT (Fcar (tem));
|
|
1926 if (++curlevel == endlevel)
|
826
|
1927 stack_overflow ("Nesting too deep for parser",
|
|
1928 make_int (curlevel - levelstart));
|
460
|
1929 curlevel->prev = -1;
|
|
1930 curlevel->last = -1;
|
|
1931 tem = Fcdr (tem);
|
|
1932 }
|
428
|
1933 }
|
|
1934 state.quoted = 0;
|
|
1935 mindepth = depth;
|
|
1936
|
|
1937 curlevel->prev = -1;
|
|
1938 curlevel->last = -1;
|
|
1939
|
|
1940 /* Enter the loop at a place appropriate for initial state. */
|
|
1941
|
|
1942 if (state.incomment) goto startincomment;
|
|
1943 if (state.instring >= 0)
|
|
1944 {
|
|
1945 if (start_quoted) goto startquotedinstring;
|
|
1946 goto startinstring;
|
|
1947 }
|
|
1948 if (start_quoted) goto startquoted;
|
|
1949
|
|
1950 while (from < end)
|
|
1951 {
|
867
|
1952 Ichar c;
|
460
|
1953 int syncode;
|
|
1954
|
428
|
1955 QUIT;
|
|
1956
|
826
|
1957 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1958 c = BUF_FETCH_CHAR (buf, from);
|
826
|
1959 syncode = SYNTAX_CODE_FROM_CACHE (scache, c);
|
|
1960 code = SYNTAX_FROM_CODE (syncode);
|
428
|
1961 from++;
|
|
1962
|
|
1963 /* record the comment style we have entered so that only the
|
|
1964 comment-ender sequence (or single char) of the same style
|
|
1965 actually terminates the comment section. */
|
460
|
1966 if (code == Scomment)
|
|
1967 {
|
|
1968 state.comstyle =
|
|
1969 SYNTAX_CODE_COMMENT_1CHAR_MASK (syncode)
|
|
1970 == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
|
1971 state.comstr_start = from - 1;
|
|
1972 }
|
|
1973
|
|
1974 /* a generic comment delimiter? */
|
|
1975 else if (code == Scomment_fence)
|
|
1976 {
|
|
1977 state.comstyle = ST_COMMENT_STYLE;
|
|
1978 state.comstr_start = from - 1;
|
|
1979 code = Scomment;
|
428
|
1980 }
|
|
1981
|
|
1982 else if (from < end &&
|
460
|
1983 SYNTAX_CODE_START_FIRST_P (syncode))
|
428
|
1984 {
|
460
|
1985 int next_syncode;
|
826
|
1986 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1987 next_syncode =
|
826
|
1988 SYNTAX_CODE_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from));
|
460
|
1989
|
|
1990 if (SYNTAX_CODES_START_P (syncode, next_syncode))
|
|
1991 {
|
428
|
1992 code = Scomment;
|
460
|
1993 state.comstyle = SYNTAX_CODES_COMMENT_MASK_START
|
|
1994 (syncode, next_syncode) == SYNTAX_COMMENT_STYLE_A ? 0 : 1;
|
|
1995 state.comstr_start = from - 1;
|
428
|
1996 from++;
|
826
|
1997 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
1998 }
|
428
|
1999 }
|
|
2000
|
460
|
2001 if (SYNTAX_CODE_PREFIX (syncode))
|
428
|
2002 continue;
|
|
2003 switch (code)
|
|
2004 {
|
|
2005 case Sescape:
|
|
2006 case Scharquote:
|
|
2007 if (stopbefore) goto stop; /* this arg means stop at sexp start */
|
|
2008 curlevel->last = from - 1;
|
|
2009 startquoted:
|
|
2010 if (from == end) goto endquoted;
|
|
2011 from++;
|
|
2012 goto symstarted;
|
|
2013 /* treat following character as a word constituent */
|
|
2014 case Sword:
|
|
2015 case Ssymbol:
|
|
2016 if (stopbefore) goto stop; /* this arg means stop at sexp start */
|
|
2017 curlevel->last = from - 1;
|
|
2018 symstarted:
|
|
2019 while (from < end)
|
|
2020 {
|
826
|
2021 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
|
2022 switch (SYNTAX_FROM_CACHE (scache, BUF_FETCH_CHAR (buf, from)))
|
428
|
2023 {
|
|
2024 case Scharquote:
|
|
2025 case Sescape:
|
|
2026 from++;
|
|
2027 if (from == end) goto endquoted;
|
|
2028 break;
|
|
2029 case Sword:
|
|
2030 case Ssymbol:
|
|
2031 case Squote:
|
|
2032 break;
|
|
2033 default:
|
|
2034 goto symdone;
|
|
2035 }
|
|
2036 from++;
|
|
2037 }
|
|
2038 symdone:
|
|
2039 curlevel->prev = curlevel->last;
|
|
2040 break;
|
|
2041
|
|
2042 case Scomment:
|
|
2043 state.incomment = 1;
|
460
|
2044 if (commentstop || boundary_stop) goto done;
|
428
|
2045 startincomment:
|
460
|
2046 if (commentstop == 1)
|
428
|
2047 goto done;
|
826
|
2048 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
428
|
2049 {
|
826
|
2050 Charbpos newfrom = find_end_of_comment (buf, from, end,
|
|
2051 state.comstyle);
|
428
|
2052 if (newfrom < 0)
|
|
2053 {
|
|
2054 /* we terminated search because from == end */
|
|
2055 from = end;
|
|
2056 goto done;
|
|
2057 }
|
|
2058 from = newfrom;
|
|
2059 }
|
|
2060 state.incomment = 0;
|
|
2061 state.comstyle = 0; /* reset the comment style */
|
460
|
2062 if (boundary_stop) goto done;
|
428
|
2063 break;
|
|
2064
|
|
2065 case Sopen:
|
|
2066 if (stopbefore) goto stop; /* this arg means stop at sexp start */
|
|
2067 depth++;
|
|
2068 /* curlevel++->last ran into compiler bug on Apollo */
|
|
2069 curlevel->last = from - 1;
|
|
2070 if (++curlevel == endlevel)
|
826
|
2071 stack_overflow ("Nesting too deep for parser",
|
|
2072 make_int (curlevel - levelstart));
|
428
|
2073 curlevel->prev = -1;
|
|
2074 curlevel->last = -1;
|
|
2075 if (targetdepth == depth) goto done;
|
|
2076 break;
|
|
2077
|
|
2078 case Sclose:
|
|
2079 depth--;
|
|
2080 if (depth < mindepth)
|
|
2081 mindepth = depth;
|
|
2082 if (curlevel != levelstart)
|
|
2083 curlevel--;
|
|
2084 curlevel->prev = curlevel->last;
|
|
2085 if (targetdepth == depth) goto done;
|
|
2086 break;
|
|
2087
|
|
2088 case Sstring:
|
460
|
2089 case Sstring_fence:
|
|
2090 state.comstr_start = from - 1;
|
428
|
2091 if (stopbefore) goto stop; /* this arg means stop at sexp start */
|
|
2092 curlevel->last = from - 1;
|
460
|
2093 if (code == Sstring_fence)
|
428
|
2094 {
|
460
|
2095 state.instring = ST_STRING_STYLE;
|
|
2096 }
|
|
2097 else
|
|
2098 {
|
|
2099 /* XEmacs change: call syntax_match() on character */
|
867
|
2100 Ichar ch = BUF_FETCH_CHAR (buf, from - 1);
|
460
|
2101 Lisp_Object stermobj =
|
1296
|
2102 syntax_match (scache->syntax_table, ch);
|
428
|
2103
|
|
2104 if (CHARP (stermobj))
|
|
2105 state.instring = XCHAR (stermobj);
|
|
2106 else
|
|
2107 state.instring = ch;
|
|
2108 }
|
460
|
2109 if (boundary_stop) goto done;
|
428
|
2110 startinstring:
|
|
2111 while (1)
|
|
2112 {
|
460
|
2113 enum syntaxcode temp_code;
|
|
2114
|
428
|
2115 if (from >= end) goto done;
|
460
|
2116
|
826
|
2117 UPDATE_SYNTAX_CACHE_FORWARD (scache, from);
|
460
|
2118 c = BUF_FETCH_CHAR (buf, from);
|
826
|
2119 temp_code = SYNTAX_FROM_CACHE (scache, c);
|
460
|
2120
|
|
2121 if (
|
|
2122 state.instring != ST_STRING_STYLE &&
|
|
2123 temp_code == Sstring &&
|
|
2124 c == state.instring) break;
|
|
2125
|
|
2126 switch (temp_code)
|
428
|
2127 {
|
460
|
2128 case Sstring_fence:
|
|
2129 if (state.instring == ST_STRING_STYLE)
|
|
2130 goto string_end;
|
|
2131 break;
|
428
|
2132 case Scharquote:
|
|
2133 case Sescape:
|
|
2134 {
|
|
2135 from++;
|
|
2136 startquotedinstring:
|
|
2137 if (from >= end) goto endquoted;
|
|
2138 break;
|
|
2139 }
|
|
2140 default:
|
|
2141 break;
|
|
2142 }
|
|
2143 from++;
|
|
2144 }
|
460
|
2145 string_end:
|
428
|
2146 state.instring = -1;
|
|
2147 curlevel->prev = curlevel->last;
|
|
2148 from++;
|
460
|
2149 if (boundary_stop) goto done;
|
428
|
2150 break;
|
|
2151
|
|
2152 case Smath:
|
|
2153 break;
|
|
2154
|
|
2155 case Swhitespace:
|
|
2156 case Spunct:
|
|
2157 case Squote:
|
|
2158 case Sendcomment:
|
460
|
2159 case Scomment_fence:
|
428
|
2160 case Sinherit:
|
|
2161 case Smax:
|
|
2162 break;
|
|
2163 }
|
|
2164 }
|
|
2165 goto done;
|
|
2166
|
|
2167 stop: /* Here if stopping before start of sexp. */
|
|
2168 from--; /* We have just fetched the char that starts it; */
|
|
2169 goto done; /* but return the position before it. */
|
|
2170
|
|
2171 endquoted:
|
|
2172 state.quoted = 1;
|
|
2173 done:
|
|
2174 state.depth = depth;
|
|
2175 state.mindepth = mindepth;
|
|
2176 state.thislevelstart = curlevel->prev;
|
|
2177 state.prevlevelstart
|
|
2178 = (curlevel == levelstart) ? -1 : (curlevel - 1)->last;
|
|
2179 state.location = from;
|
460
|
2180 state.levelstarts = Qnil;
|
|
2181 while (--curlevel >= levelstart)
|
|
2182 state.levelstarts = Fcons (make_int (curlevel->last),
|
|
2183 state.levelstarts);
|
428
|
2184
|
|
2185 *stateptr = state;
|
|
2186 }
|
|
2187
|
|
2188 DEFUN ("parse-partial-sexp", Fparse_partial_sexp, 2, 7, 0, /*
|
|
2189 Parse Lisp syntax starting at FROM until TO; return status of parse at TO.
|
|
2190 Parsing stops at TO or when certain criteria are met;
|
|
2191 point is set to where parsing stops.
|
444
|
2192 If fifth arg OLDSTATE is omitted or nil,
|
428
|
2193 parsing assumes that FROM is the beginning of a function.
|
460
|
2194 Value is a list of nine elements describing final state of parsing:
|
428
|
2195 0. depth in parens.
|
|
2196 1. character address of start of innermost containing list; nil if none.
|
|
2197 2. character address of start of last complete sexp terminated.
|
|
2198 3. non-nil if inside a string.
|
460
|
2199 (It is the character that will terminate the string,
|
|
2200 or t if the string should be terminated by an explicit
|
|
2201 `syntax-table' property.)
|
428
|
2202 4. t if inside a comment.
|
|
2203 5. t if following a quote character.
|
|
2204 6. the minimum paren-depth encountered during this scan.
|
460
|
2205 7. nil if in comment style a, or not in a comment; t if in comment style b;
|
|
2206 `syntax-table' if given by an explicit `syntax-table' property.
|
|
2207 8. character address of start of last comment or string; nil if none.
|
|
2208 9. Intermediate data for continuation of parsing (subject to change).
|
428
|
2209 If third arg TARGETDEPTH is non-nil, parsing stops if the depth
|
|
2210 in parentheses becomes equal to TARGETDEPTH.
|
|
2211 Fourth arg STOPBEFORE non-nil means stop when come to
|
|
2212 any character that starts a sexp.
|
460
|
2213 Fifth arg OLDSTATE is a nine-element list like what this function returns.
|
428
|
2214 It is used to initialize the state of the parse. Its second and third
|
|
2215 elements are ignored.
|
460
|
2216 Sixth arg COMMENTSTOP non-nil means stop at the start of a comment. If it
|
|
2217 is `syntax-table', stop after the start of a comment or a string, or after
|
|
2218 the end of a comment or string.
|
826
|
2219 Seventh arg BUFFER specifies the buffer to do the parsing in, and defaults
|
|
2220 to the current buffer.
|
428
|
2221 */
|
|
2222 (from, to, targetdepth, stopbefore, oldstate, commentstop, buffer))
|
|
2223 {
|
|
2224 struct lisp_parse_state state;
|
|
2225 int target;
|
665
|
2226 Charbpos start, end;
|
428
|
2227 struct buffer *buf = decode_buffer (buffer, 0);
|
|
2228 Lisp_Object val;
|
|
2229
|
|
2230 if (!NILP (targetdepth))
|
|
2231 {
|
|
2232 CHECK_INT (targetdepth);
|
|
2233 target = XINT (targetdepth);
|
|
2234 }
|
|
2235 else
|
|
2236 target = -100000; /* We won't reach this depth */
|
|
2237
|
|
2238 get_buffer_range_char (buf, from, to, &start, &end, 0);
|
|
2239 scan_sexps_forward (buf, &state, start, end,
|
|
2240 target, !NILP (stopbefore), oldstate,
|
460
|
2241 (NILP (commentstop)
|
|
2242 ? 0 : (EQ (commentstop, Qsyntax_table) ? -1 : 1)));
|
428
|
2243 BUF_SET_PT (buf, state.location);
|
|
2244
|
|
2245 /* reverse order */
|
|
2246 val = Qnil;
|
460
|
2247 val = Fcons (state.levelstarts, val);
|
|
2248 val = Fcons ((state.incomment || (state.instring >= 0))
|
|
2249 ? make_int (state.comstr_start) : Qnil, val);
|
|
2250 val = Fcons (state.comstyle ? (state.comstyle == ST_COMMENT_STYLE
|
|
2251 ? Qsyntax_table : Qt) : Qnil, val);
|
428
|
2252 val = Fcons (make_int (state.mindepth), val);
|
|
2253 val = Fcons (state.quoted ? Qt : Qnil, val);
|
|
2254 val = Fcons (state.incomment ? Qt : Qnil, val);
|
460
|
2255 val = Fcons (state.instring < 0
|
|
2256 ? Qnil
|
|
2257 : (state.instring == ST_STRING_STYLE
|
|
2258 ? Qt : make_int (state.instring)), val);
|
826
|
2259 val = Fcons (state.thislevelstart < 0 ? Qnil :
|
|
2260 make_int (state.thislevelstart), val);
|
|
2261 val = Fcons (state.prevlevelstart < 0 ? Qnil :
|
|
2262 make_int (state.prevlevelstart), val);
|
428
|
2263 val = Fcons (make_int (state.depth), val);
|
|
2264
|
|
2265 return val;
|
|
2266 }
|
|
2267
|
|
2268
|
|
2269 /* Updating of the mirror syntax table.
|
|
2270
|
1296
|
2271 Each syntax table has a corresponding mirror table in it. Whenever we
|
|
2272 make a change to a syntax table, we set a dirty flag. When accessing a
|
|
2273 value from the mirror table and the table is dirty, we call
|
|
2274 update_syntax_table() to clean it up.
|
428
|
2275
|
|
2276 #### We really only need to map over the changed range.
|
|
2277
|
|
2278 If we change the standard syntax table, we need to map over
|
|
2279 all tables because any of them could be inheriting from the
|
|
2280 standard syntax table.
|
|
2281
|
|
2282 When `set-syntax-table' is called, we set the buffer's mirror
|
|
2283 syntax table as well.
|
|
2284 */
|
|
2285
|
826
|
2286 static int
|
2286
|
2287 copy_to_mirrortab (struct chartab_range *range, Lisp_Object UNUSED (table),
|
826
|
2288 Lisp_Object val, void *arg)
|
428
|
2289 {
|
826
|
2290 Lisp_Object mirrortab = VOID_TO_LISP (arg);
|
428
|
2291
|
|
2292 if (CONSP (val))
|
|
2293 val = XCAR (val);
|
826
|
2294 if (SYNTAX_FROM_CODE (XINT (val)) != Sinherit)
|
|
2295 put_char_table (mirrortab, range, val);
|
|
2296 return 0;
|
|
2297 }
|
|
2298
|
|
2299 static int
|
2286
|
2300 copy_if_not_already_present (struct chartab_range *range,
|
|
2301 Lisp_Object UNUSED (table),
|
826
|
2302 Lisp_Object val, void *arg)
|
|
2303 {
|
1296
|
2304 Lisp_Object mirrortab = VOID_TO_LISP (arg);
|
826
|
2305 if (CONSP (val))
|
|
2306 val = XCAR (val);
|
|
2307 if (SYNTAX_FROM_CODE (XINT (val)) != Sinherit)
|
|
2308 {
|
|
2309 Lisp_Object existing =
|
1296
|
2310 updating_mirror_get_range_char_table (range, mirrortab,
|
|
2311 Vbogus_syntax_table_value);
|
826
|
2312 if (NILP (existing))
|
|
2313 /* nothing at all */
|
1296
|
2314 put_char_table (mirrortab, range, val);
|
|
2315 else if (!EQ (existing, Vbogus_syntax_table_value))
|
826
|
2316 /* full */
|
|
2317 ;
|
|
2318 else
|
|
2319 {
|
|
2320 Freset_char_table (Vtemp_table_for_use_updating_syntax_tables);
|
|
2321 copy_char_table_range
|
1296
|
2322 (mirrortab, Vtemp_table_for_use_updating_syntax_tables, range);
|
|
2323 put_char_table (mirrortab, range, val);
|
826
|
2324 copy_char_table_range
|
1296
|
2325 (Vtemp_table_for_use_updating_syntax_tables, mirrortab, range);
|
826
|
2326 }
|
428
|
2327 }
|
826
|
2328
|
428
|
2329 return 0;
|
|
2330 }
|
|
2331
|
|
2332 static void
|
826
|
2333 update_just_this_syntax_table (Lisp_Object table)
|
428
|
2334 {
|
|
2335 struct chartab_range range;
|
826
|
2336 Lisp_Object mirrortab = XCHAR_TABLE (table)->mirror_table;
|
|
2337
|
1296
|
2338 assert (!XCHAR_TABLE (table)->mirror_table_p);
|
826
|
2339 range.type = CHARTAB_RANGE_ALL;
|
|
2340 Freset_char_table (mirrortab);
|
1296
|
2341
|
826
|
2342 /* First, copy the tables values other than inherit into the mirror
|
|
2343 table. Then, for tables other than the standard syntax table, map
|
|
2344 over the standard table, copying values into the mirror table only if
|
|
2345 entries don't already exist in that table. (The copying step requires
|
|
2346 another mapping.)
|
|
2347 */
|
428
|
2348
|
826
|
2349 map_char_table (table, &range, copy_to_mirrortab, LISP_TO_VOID (mirrortab));
|
|
2350 /* second clause catches bootstrapping problems when initializing the
|
|
2351 standard syntax table */
|
|
2352 if (!EQ (table, Vstandard_syntax_table) && !NILP (Vstandard_syntax_table))
|
1296
|
2353 map_char_table (Vstandard_syntax_table, &range,
|
|
2354 copy_if_not_already_present, LISP_TO_VOID (mirrortab));
|
3152
|
2355 /* The resetting made the default be Qnil. Put it back to Sword. */
|
|
2356 set_char_table_default (mirrortab, make_int (Sword));
|
1296
|
2357 XCHAR_TABLE (mirrortab)->dirty = 0;
|
428
|
2358 }
|
|
2359
|
|
2360 /* Called from chartab.c when a change is made to a syntax table.
|
|
2361 If this is the standard syntax table, we need to recompute
|
|
2362 *all* syntax tables (yuck). Otherwise we just recompute this
|
|
2363 one. */
|
|
2364
|
|
2365 void
|
826
|
2366 update_syntax_table (Lisp_Object table)
|
428
|
2367 {
|
1296
|
2368 Lisp_Object nonmirror = XCHAR_TABLE (table)->mirror_table;
|
|
2369 assert (XCHAR_TABLE (table)->mirror_table_p);
|
|
2370 if (EQ (nonmirror, Vstandard_syntax_table))
|
428
|
2371 {
|
|
2372 Lisp_Object syntab;
|
|
2373
|
|
2374 for (syntab = Vall_syntax_tables; !NILP (syntab);
|
|
2375 syntab = XCHAR_TABLE (syntab)->next_table)
|
826
|
2376 update_just_this_syntax_table (syntab);
|
428
|
2377 }
|
|
2378 else
|
1296
|
2379 update_just_this_syntax_table (nonmirror);
|
428
|
2380 }
|
|
2381
|
|
2382
|
|
2383 /************************************************************************/
|
|
2384 /* initialization */
|
|
2385 /************************************************************************/
|
|
2386
|
|
2387 void
|
|
2388 syms_of_syntax (void)
|
|
2389 {
|
3092
|
2390 #ifdef NEW_GC
|
|
2391 INIT_LRECORD_IMPLEMENTATION (syntax_cache);
|
|
2392 #endif /* NEW_GC */
|
563
|
2393 DEFSYMBOL (Qsyntax_table_p);
|
|
2394 DEFSYMBOL (Qsyntax_table);
|
428
|
2395
|
|
2396 DEFSUBR (Fsyntax_table_p);
|
|
2397 DEFSUBR (Fsyntax_table);
|
826
|
2398 #ifdef DEBUG_XEMACS
|
|
2399 DEFSUBR (Fmirror_syntax_table);
|
|
2400 DEFSUBR (Fsyntax_cache_info);
|
|
2401 #endif /* DEBUG_XEMACS */
|
428
|
2402 DEFSUBR (Fstandard_syntax_table);
|
|
2403 DEFSUBR (Fcopy_syntax_table);
|
|
2404 DEFSUBR (Fset_syntax_table);
|
|
2405 DEFSUBR (Fsyntax_designator_chars);
|
|
2406 DEFSUBR (Fchar_syntax);
|
|
2407 DEFSUBR (Fmatching_paren);
|
|
2408 /* DEFSUBR (Fmodify_syntax_entry); now in Lisp. */
|
|
2409 /* DEFSUBR (Fdescribe_syntax); now in Lisp. */
|
|
2410
|
|
2411 DEFSUBR (Fforward_word);
|
|
2412
|
|
2413 DEFSUBR (Fforward_comment);
|
|
2414 DEFSUBR (Fscan_lists);
|
|
2415 DEFSUBR (Fscan_sexps);
|
|
2416 DEFSUBR (Fbackward_prefix_chars);
|
|
2417 DEFSUBR (Fparse_partial_sexp);
|
|
2418 }
|
|
2419
|
|
2420 void
|
|
2421 vars_of_syntax (void)
|
|
2422 {
|
|
2423 DEFVAR_BOOL ("parse-sexp-ignore-comments", &parse_sexp_ignore_comments /*
|
|
2424 Non-nil means `forward-sexp', etc., should treat comments as whitespace.
|
|
2425 */ );
|
434
|
2426 parse_sexp_ignore_comments = 0;
|
428
|
2427
|
460
|
2428 DEFVAR_BOOL ("lookup-syntax-properties", &lookup_syntax_properties /*
|
826
|
2429 Non-nil means `forward-sexp', etc., respect the `syntax-table' property.
|
|
2430 This property can be placed on buffers or strings and can be used to explicitly
|
|
2431 specify the syntax table to be used for looking up the syntax of the chars
|
|
2432 having this property, or to directly specify the syntax of the chars.
|
|
2433
|
460
|
2434 The value of this property should be either a syntax table, or a cons
|
|
2435 of the form (SYNTAXCODE . MATCHCHAR), SYNTAXCODE being the numeric
|
|
2436 syntax code, MATCHCHAR being nil or the character to match (which is
|
826
|
2437 relevant only when the syntax code is open/close-type).
|
460
|
2438 */ );
|
|
2439 lookup_syntax_properties = 1;
|
|
2440
|
428
|
2441 DEFVAR_BOOL ("words-include-escapes", &words_include_escapes /*
|
|
2442 Non-nil means `forward-word', etc., should treat escape chars part of words.
|
|
2443 */ );
|
434
|
2444 words_include_escapes = 0;
|
428
|
2445
|
|
2446 no_quit_in_re_search = 0;
|
1296
|
2447
|
|
2448 Vbogus_syntax_table_value = make_float (0.0);
|
|
2449 staticpro (&Vbogus_syntax_table_value);
|
428
|
2450 }
|
|
2451
|
|
2452 static void
|
3540
|
2453 define_standard_syntax (const UExtbyte *p, enum syntaxcode syn)
|
428
|
2454 {
|
|
2455 for (; *p; p++)
|
|
2456 Fput_char_table (make_char (*p), make_int (syn), Vstandard_syntax_table);
|
|
2457 }
|
|
2458
|
|
2459 void
|
|
2460 complex_vars_of_syntax (void)
|
|
2461 {
|
867
|
2462 Ichar i;
|
3540
|
2463 const UExtbyte *p; /* Latin-1, not internal format. */
|
|
2464
|
|
2465 #define SET_RANGE_SYNTAX(start, end, syntax) \
|
|
2466 do { \
|
|
2467 for (i = start; i <= end; i++) \
|
|
2468 Fput_char_table(make_char(i), make_int(syntax), \
|
|
2469 Vstandard_syntax_table); \
|
|
2470 } while (0)
|
|
2471
|
|
2472 /* Set this now, so first buffer creation can refer to it.
|
|
2473
|
|
2474 Make it nil before calling copy-syntax-table so that copy-syntax-table
|
|
2475 will know not to try to copy from garbage */
|
428
|
2476 Vstandard_syntax_table = Qnil;
|
|
2477 Vstandard_syntax_table = Fcopy_syntax_table (Qnil);
|
|
2478 staticpro (&Vstandard_syntax_table);
|
|
2479
|
826
|
2480 Vtemp_table_for_use_updating_syntax_tables = Fmake_char_table (Qgeneric);
|
|
2481 staticpro (&Vtemp_table_for_use_updating_syntax_tables);
|
|
2482
|
428
|
2483 Vsyntax_designator_chars_string = make_string_nocopy (syntax_code_spec,
|
|
2484 Smax);
|
|
2485 staticpro (&Vsyntax_designator_chars_string);
|
|
2486
|
3540
|
2487 /* Default character syntax is word. */
|
3152
|
2488 set_char_table_default (Vstandard_syntax_table, make_int (Sword));
|
428
|
2489
|
3540
|
2490 /* Control 0; treat as punctuation */
|
|
2491 SET_RANGE_SYNTAX(0, 32, Spunct);
|
428
|
2492
|
3544
|
2493 /* The whitespace--overwriting some of the above changes.
|
|
2494
|
|
2495 String literals are const char *s, not const unsigned char *s. */
|
|
2496 define_standard_syntax((const UExtbyte *)" \t\015\014", Swhitespace);
|
3540
|
2497
|
|
2498 /* DEL plus Control 1 */
|
|
2499 SET_RANGE_SYNTAX(127, 159, Spunct);
|
|
2500
|
3544
|
2501 define_standard_syntax ((const UExtbyte *)"\"", Sstring);
|
|
2502 define_standard_syntax ((const UExtbyte *)"\\", Sescape);
|
|
2503 define_standard_syntax ((const UExtbyte *)"_-+*/&|<>=", Ssymbol);
|
|
2504 define_standard_syntax ((const UExtbyte *)".,;:?!#@~^'`", Spunct);
|
428
|
2505
|
3544
|
2506 for (p = (const UExtbyte *)"()[]{}"; *p; p+=2)
|
428
|
2507 {
|
|
2508 Fput_char_table (make_char (p[0]),
|
|
2509 Fcons (make_int (Sopen), make_char (p[1])),
|
|
2510 Vstandard_syntax_table);
|
|
2511 Fput_char_table (make_char (p[1]),
|
|
2512 Fcons (make_int (Sclose), make_char (p[0])),
|
|
2513 Vstandard_syntax_table);
|
|
2514 }
|
3540
|
2515
|
|
2516 /* Latin 1 "symbols." This contrasts with the FSF, where they're word
|
|
2517 constituents. */
|
|
2518 SET_RANGE_SYNTAX(0240, 0277, Ssymbol);
|
|
2519
|
|
2520 /* The guillemets. These are not parentheses, in contrast to what the old
|
|
2521 code did. */
|
3569
|
2522 define_standard_syntax((const UExtbyte *)"\253\273", Spunct);
|
3540
|
2523
|
|
2524 /* The inverted exclamation mark, and the multiplication and division
|
|
2525 signs. */
|
3544
|
2526 define_standard_syntax((const UExtbyte *)"\241\327\367", Spunct);
|
3540
|
2527
|
|
2528 #undef SET_RANGE_SYNTAX
|
428
|
2529 }
|