Commit | Line | Data |
---|---|---|
8489eb67 | 1 | /* GNU Emacs routines to deal with syntax tables; also word and list parsing. |
ab422c4d PE |
2 | Copyright (C) 1985, 1987, 1993-1995, 1997-1999, 2001-2013 Free |
3 | Software Foundation, Inc. | |
8489eb67 RS |
4 | |
5 | This file is part of GNU Emacs. | |
6 | ||
9ec0b715 | 7 | GNU Emacs is free software: you can redistribute it and/or modify |
8489eb67 | 8 | it under the terms of the GNU General Public License as published by |
9ec0b715 GM |
9 | the Free Software Foundation, either version 3 of the License, or |
10 | (at your option) any later version. | |
8489eb67 RS |
11 | |
12 | GNU Emacs is distributed in the hope that it will be useful, | |
13 | but WITHOUT ANY WARRANTY; without even the implied warranty of | |
14 | MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the | |
15 | GNU General Public License for more details. | |
16 | ||
17 | You should have received a copy of the GNU General Public License | |
9ec0b715 | 18 | along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */ |
8489eb67 RS |
19 | |
20 | ||
18160b98 | 21 | #include <config.h> |
1b2de274 | 22 | |
1b2de274 | 23 | #include <sys/types.h> |
0328b6de | 24 | |
8489eb67 RS |
25 | #include "lisp.h" |
26 | #include "commands.h" | |
5c7b02ab | 27 | #include "character.h" |
e5560ff7 | 28 | #include "buffer.h" |
e35f6ff7 | 29 | #include "keymap.h" |
a1bc88d4 | 30 | #include "regex.h" |
195d1361 RS |
31 | |
32 | /* Make syntax table lookup grant data in gl_state. */ | |
33 | #define SYNTAX_ENTRY_VIA_PROPERTY | |
34 | ||
8489eb67 | 35 | #include "syntax.h" |
195d1361 | 36 | #include "intervals.h" |
c5683ceb SM |
37 | #include "category.h" |
38 | ||
39 | /* Then there are seven single-bit flags that have the following meanings: | |
40 | 1. This character is the first of a two-character comment-start sequence. | |
41 | 2. This character is the second of a two-character comment-start sequence. | |
42 | 3. This character is the first of a two-character comment-end sequence. | |
43 | 4. This character is the second of a two-character comment-end sequence. | |
44 | 5. This character is a prefix, for backward-prefix-chars. | |
45 | 6. The char is part of a delimiter for comments of style "b". | |
46 | 7. This character is part of a nestable comment sequence. | |
47 | 8. The char is part of a delimiter for comments of style "c". | |
48 | Note that any two-character sequence whose first character has flag 1 | |
49 | and whose second character has flag 2 will be interpreted as a comment start. | |
50 | ||
51 | bit 6 and 8 are used to discriminate between different comment styles. | |
52 | Languages such as C++ allow two orthogonal syntax start/end pairs | |
53 | and bit 6 is used to determine whether a comment-end or Scommentend | |
54 | ends style a or b. Comment markers can start style a, b, c, or bc. | |
55 | Style a is always the default. | |
56 | For 2-char comment markers, the style b flag is only looked up on the second | |
57 | char of the comment marker and on the first char of the comment ender. | |
58 | For style c (like to for the nested flag), the flag can be placed on any | |
59 | one of the chars. | |
60 | */ | |
61 | ||
62 | /* These macros extract specific flags from an integer | |
63 | that holds the syntax code and the flags. */ | |
64 | ||
65 | #define SYNTAX_FLAGS_COMSTART_FIRST(flags) (((flags) >> 16) & 1) | |
66 | ||
67 | #define SYNTAX_FLAGS_COMSTART_SECOND(flags) (((flags) >> 17) & 1) | |
68 | ||
69 | #define SYNTAX_FLAGS_COMEND_FIRST(flags) (((flags) >> 18) & 1) | |
70 | ||
71 | #define SYNTAX_FLAGS_COMEND_SECOND(flags) (((flags) >> 19) & 1) | |
72 | ||
73 | #define SYNTAX_FLAGS_PREFIX(flags) (((flags) >> 20) & 1) | |
74 | ||
75 | #define SYNTAX_FLAGS_COMMENT_STYLEB(flags) (((flags) >> 21) & 1) | |
76 | #define SYNTAX_FLAGS_COMMENT_STYLEC(flags) (((flags) >> 22) & 2) | |
77 | /* FLAGS should be the flags of the main char of the comment marker, e.g. | |
78 | the second for comstart and the first for comend. */ | |
79 | #define SYNTAX_FLAGS_COMMENT_STYLE(flags, other_flags) \ | |
80 | (SYNTAX_FLAGS_COMMENT_STYLEB (flags) \ | |
81 | | SYNTAX_FLAGS_COMMENT_STYLEC (flags) \ | |
82 | | SYNTAX_FLAGS_COMMENT_STYLEC (other_flags)) | |
83 | ||
84 | #define SYNTAX_FLAGS_COMMENT_NESTED(flags) (((flags) >> 22) & 1) | |
85 | ||
86 | /* These macros extract a particular flag for a given character. */ | |
87 | ||
88 | #define SYNTAX_COMEND_FIRST(c) \ | |
89 | (SYNTAX_FLAGS_COMEND_FIRST (SYNTAX_WITH_FLAGS (c))) | |
90 | #define SYNTAX_PREFIX(c) (SYNTAX_FLAGS_PREFIX (SYNTAX_WITH_FLAGS (c))) | |
195d1361 RS |
91 | |
92 | /* We use these constants in place for comment-style and | |
93 | string-ender-char to distinguish comments/strings started by | |
94 | comment_fence and string_fence codes. */ | |
95 | ||
eacd378d PE |
96 | #define ST_COMMENT_STYLE (256 + 1) |
97 | #define ST_STRING_STYLE (256 + 2) | |
8489eb67 | 98 | |
955cbe7b PE |
99 | static Lisp_Object Qsyntax_table_p; |
100 | static Lisp_Object Qsyntax_table, Qscan_error; | |
8489eb67 | 101 | |
d4b43b22 | 102 | #ifndef __GNUC__ |
8ea151b2 RS |
103 | /* Used as a temporary in SYNTAX_ENTRY and other macros in syntax.h, |
104 | if not compiled with GCC. No need to mark it, since it is used | |
105 | only very temporarily. */ | |
106 | Lisp_Object syntax_temp; | |
d4b43b22 | 107 | #endif |
8ea151b2 | 108 | |
e5d4f4dc RS |
109 | /* This is the internal form of the parse state used in parse-partial-sexp. */ |
110 | ||
111 | struct lisp_parse_state | |
112 | { | |
d311d28c | 113 | EMACS_INT depth; /* Depth at end of parsing. */ |
6449674e | 114 | int instring; /* -1 if not within string, else desired terminator. */ |
d311d28c | 115 | EMACS_INT incomment; /* -1 if in unnestable comment else comment nesting */ |
6449674e SM |
116 | int comstyle; /* comment style a=0, or b=1, or ST_COMMENT_STYLE. */ |
117 | int quoted; /* Nonzero if just after an escape char at end of parsing */ | |
d311d28c | 118 | EMACS_INT mindepth; /* Minimum depth seen while scanning. */ |
6449674e | 119 | /* Char number of most recent start-of-expression at current level */ |
d311d28c | 120 | ptrdiff_t thislevelstart; |
6449674e | 121 | /* Char number of start of containing expression */ |
d311d28c PE |
122 | ptrdiff_t prevlevelstart; |
123 | ptrdiff_t location; /* Char number at which parsing stopped. */ | |
124 | ptrdiff_t comstr_start; /* Position of last comment/string starter. */ | |
6449674e SM |
125 | Lisp_Object levelstarts; /* Char numbers of starts-of-expression |
126 | of levels (starting from outermost). */ | |
e5d4f4dc RS |
127 | }; |
128 | \f | |
37bef230 RS |
129 | /* These variables are a cache for finding the start of a defun. |
130 | find_start_pos is the place for which the defun start was found. | |
131 | find_start_value is the defun start position found for it. | |
6a140a74 | 132 | find_start_value_byte is the corresponding byte position. |
37bef230 RS |
133 | find_start_buffer is the buffer it was found in. |
134 | find_start_begv is the BEGV value when it was found. | |
135 | find_start_modiff is the value of MODIFF when it was found. */ | |
136 | ||
d311d28c PE |
137 | static ptrdiff_t find_start_pos; |
138 | static ptrdiff_t find_start_value; | |
139 | static ptrdiff_t find_start_value_byte; | |
37bef230 | 140 | static struct buffer *find_start_buffer; |
d311d28c PE |
141 | static ptrdiff_t find_start_begv; |
142 | static EMACS_INT find_start_modiff; | |
6a140a74 RS |
143 | |
144 | ||
f57e2426 J |
145 | static Lisp_Object skip_chars (int, Lisp_Object, Lisp_Object, int); |
146 | static Lisp_Object skip_syntaxes (int, Lisp_Object, Lisp_Object); | |
147 | static Lisp_Object scan_lists (EMACS_INT, EMACS_INT, EMACS_INT, int); | |
148 | static void scan_sexps_forward (struct lisp_parse_state *, | |
d311d28c | 149 | ptrdiff_t, ptrdiff_t, ptrdiff_t, EMACS_INT, |
f57e2426 J |
150 | int, Lisp_Object, int); |
151 | static int in_classes (int, Lisp_Object); | |
39eb03f1 PE |
152 | |
153 | /* This setter is used only in this file, so it can be private. */ | |
b0ab8123 | 154 | static void |
39eb03f1 PE |
155 | bset_syntax_table (struct buffer *b, Lisp_Object val) |
156 | { | |
157 | b->INTERNAL_FIELD (syntax_table) = val; | |
158 | } | |
195d1361 | 159 | \f |
c5683ceb SM |
160 | /* Whether the syntax of the character C has the prefix flag set. */ |
161 | int syntax_prefix_flag_p (int c) | |
162 | { | |
163 | return SYNTAX_PREFIX (c); | |
164 | } | |
195d1361 RS |
165 | |
166 | struct gl_state_s gl_state; /* Global state of syntax parser. */ | |
167 | ||
195d1361 RS |
168 | #define INTERVALS_AT_ONCE 10 /* 1 + max-number of intervals |
169 | to scan to property-change. */ | |
170 | ||
6a140a74 RS |
171 | /* Update gl_state to an appropriate interval which contains CHARPOS. The |
172 | sign of COUNT give the relative position of CHARPOS wrt the previously | |
195d1361 | 173 | valid interval. If INIT, only [be]_property fields of gl_state are |
6a140a74 | 174 | valid at start, the rest is filled basing on OBJECT. |
195d1361 | 175 | |
6a140a74 | 176 | `gl_state.*_i' are the intervals, and CHARPOS is further in the search |
195d1361 RS |
177 | direction than the intervals - or in an interval. We update the |
178 | current syntax-table basing on the property of this interval, and | |
6a140a74 | 179 | update the interval to start further than CHARPOS - or be |
77c7bcb1 | 180 | NULL. We also update lim_property to be the next value of |
6a140a74 | 181 | charpos to call this subroutine again - or be before/after the |
195d1361 RS |
182 | start/end of OBJECT. */ |
183 | ||
184 | void | |
d311d28c | 185 | update_syntax_table (ptrdiff_t charpos, EMACS_INT count, int init, |
4f3a2f8d | 186 | Lisp_Object object) |
195d1361 RS |
187 | { |
188 | Lisp_Object tmp_table; | |
12cbf13f PE |
189 | unsigned cnt = 0; |
190 | int invalidate = 1; | |
5a774522 | 191 | INTERVAL i; |
195d1361 RS |
192 | |
193 | if (init) | |
194 | { | |
bb0de084 | 195 | gl_state.old_prop = Qnil; |
195d1361 RS |
196 | gl_state.start = gl_state.b_property; |
197 | gl_state.stop = gl_state.e_property; | |
bb0de084 SM |
198 | i = interval_of (charpos, object); |
199 | gl_state.backward_i = gl_state.forward_i = i; | |
195d1361 | 200 | invalidate = 0; |
77c7bcb1 | 201 | if (!i) |
195d1361 | 202 | return; |
f902a008 | 203 | /* interval_of updates only ->position of the return value, so |
d80f4cc7 | 204 | update the parents manually to speed up update_interval. */ |
bb0de084 | 205 | while (!NULL_PARENT (i)) |
d80f4cc7 RS |
206 | { |
207 | if (AM_RIGHT_CHILD (i)) | |
439d5cb4 | 208 | INTERVAL_PARENT (i)->position = i->position |
d80f4cc7 | 209 | - LEFT_TOTAL_LENGTH (i) + TOTAL_LENGTH (i) /* right end */ |
439d5cb4 KR |
210 | - TOTAL_LENGTH (INTERVAL_PARENT (i)) |
211 | + LEFT_TOTAL_LENGTH (INTERVAL_PARENT (i)); | |
d80f4cc7 | 212 | else |
439d5cb4 | 213 | INTERVAL_PARENT (i)->position = i->position - LEFT_TOTAL_LENGTH (i) |
d80f4cc7 | 214 | + TOTAL_LENGTH (i); |
439d5cb4 | 215 | i = INTERVAL_PARENT (i); |
d80f4cc7 RS |
216 | } |
217 | i = gl_state.forward_i; | |
bb0de084 | 218 | gl_state.b_property = i->position - gl_state.offset; |
ee0cdb48 | 219 | gl_state.e_property = INTERVAL_LAST_POS (i) - gl_state.offset; |
195d1361 RS |
220 | goto update; |
221 | } | |
5a774522 | 222 | i = count > 0 ? gl_state.forward_i : gl_state.backward_i; |
195d1361 | 223 | |
423e705d | 224 | /* We are guaranteed to be called with CHARPOS either in i, |
6a140a74 | 225 | or further off. */ |
77c7bcb1 | 226 | if (!i) |
195d1361 | 227 | error ("Error in syntax_table logic for to-the-end intervals"); |
6a140a74 | 228 | else if (charpos < i->position) /* Move left. */ |
195d1361 RS |
229 | { |
230 | if (count > 0) | |
f902a008 | 231 | error ("Error in syntax_table logic for intervals <-"); |
195d1361 | 232 | /* Update the interval. */ |
6a140a74 | 233 | i = update_interval (i, charpos); |
bb0de084 | 234 | if (INTERVAL_LAST_POS (i) != gl_state.b_property) |
195d1361 RS |
235 | { |
236 | invalidate = 0; | |
195d1361 | 237 | gl_state.forward_i = i; |
ee0cdb48 | 238 | gl_state.e_property = INTERVAL_LAST_POS (i) - gl_state.offset; |
195d1361 | 239 | } |
423e705d | 240 | } |
6a140a74 | 241 | else if (charpos >= INTERVAL_LAST_POS (i)) /* Move right. */ |
195d1361 RS |
242 | { |
243 | if (count < 0) | |
f902a008 | 244 | error ("Error in syntax_table logic for intervals ->"); |
195d1361 | 245 | /* Update the interval. */ |
6a140a74 | 246 | i = update_interval (i, charpos); |
bb0de084 | 247 | if (i->position != gl_state.e_property) |
195d1361 RS |
248 | { |
249 | invalidate = 0; | |
195d1361 | 250 | gl_state.backward_i = i; |
bb0de084 | 251 | gl_state.b_property = i->position - gl_state.offset; |
195d1361 RS |
252 | } |
253 | } | |
195d1361 RS |
254 | |
255 | update: | |
256 | tmp_table = textget (i->plist, Qsyntax_table); | |
257 | ||
258 | if (invalidate) | |
259 | invalidate = !EQ (tmp_table, gl_state.old_prop); /* Need to invalidate? */ | |
7d0393cf | 260 | |
423e705d SM |
261 | if (invalidate) /* Did not get to adjacent interval. */ |
262 | { /* with the same table => */ | |
263 | /* invalidate the old range. */ | |
195d1361 RS |
264 | if (count > 0) |
265 | { | |
266 | gl_state.backward_i = i; | |
bb0de084 SM |
267 | gl_state.b_property = i->position - gl_state.offset; |
268 | } | |
269 | else | |
195d1361 | 270 | { |
bb0de084 | 271 | gl_state.forward_i = i; |
ee0cdb48 | 272 | gl_state.e_property = INTERVAL_LAST_POS (i) - gl_state.offset; |
195d1361 RS |
273 | } |
274 | } | |
37bef230 | 275 | |
bb0de084 | 276 | if (!EQ (tmp_table, gl_state.old_prop)) |
195d1361 | 277 | { |
bb0de084 SM |
278 | gl_state.current_syntax_table = tmp_table; |
279 | gl_state.old_prop = tmp_table; | |
280 | if (EQ (Fsyntax_table_p (tmp_table), Qt)) | |
281 | { | |
282 | gl_state.use_global = 0; | |
7d0393cf | 283 | } |
bb0de084 SM |
284 | else if (CONSP (tmp_table)) |
285 | { | |
286 | gl_state.use_global = 1; | |
287 | gl_state.global_code = tmp_table; | |
288 | } | |
7d0393cf | 289 | else |
bb0de084 SM |
290 | { |
291 | gl_state.use_global = 0; | |
4b4deea2 | 292 | gl_state.current_syntax_table = BVAR (current_buffer, syntax_table); |
bb0de084 | 293 | } |
195d1361 RS |
294 | } |
295 | ||
77c7bcb1 | 296 | while (i) |
195d1361 RS |
297 | { |
298 | if (cnt && !EQ (tmp_table, textget (i->plist, Qsyntax_table))) | |
299 | { | |
300 | if (count > 0) | |
bb0de084 SM |
301 | { |
302 | gl_state.e_property = i->position - gl_state.offset; | |
303 | gl_state.forward_i = i; | |
304 | } | |
305 | else | |
306 | { | |
5a774522 SM |
307 | gl_state.b_property |
308 | = i->position + LENGTH (i) - gl_state.offset; | |
bb0de084 SM |
309 | gl_state.backward_i = i; |
310 | } | |
311 | return; | |
195d1361 | 312 | } |
7d0393cf | 313 | else if (cnt == INTERVALS_AT_ONCE) |
195d1361 RS |
314 | { |
315 | if (count > 0) | |
bb0de084 | 316 | { |
5a774522 SM |
317 | gl_state.e_property |
318 | = i->position + LENGTH (i) - gl_state.offset | |
319 | /* e_property at EOB is not set to ZV but to ZV+1, so that | |
320 | we can do INC(from);UPDATE_SYNTAX_TABLE_FORWARD without | |
321 | having to check eob between the two. */ | |
77c7bcb1 | 322 | + (next_interval (i) ? 0 : 1); |
bb0de084 SM |
323 | gl_state.forward_i = i; |
324 | } | |
325 | else | |
326 | { | |
327 | gl_state.b_property = i->position - gl_state.offset; | |
328 | gl_state.backward_i = i; | |
329 | } | |
330 | return; | |
195d1361 RS |
331 | } |
332 | cnt++; | |
333 | i = count > 0 ? next_interval (i) : previous_interval (i); | |
334 | } | |
77c7bcb1 | 335 | eassert (i == NULL); /* This property goes to the end. */ |
bb0de084 SM |
336 | if (count > 0) |
337 | gl_state.e_property = gl_state.stop; | |
338 | else | |
339 | gl_state.b_property = gl_state.start; | |
195d1361 RS |
340 | } |
341 | \f | |
6a140a74 RS |
342 | /* Returns TRUE if char at CHARPOS is quoted. |
343 | Global syntax-table data should be set up already to be good at CHARPOS | |
344 | or after. On return global syntax data is good for lookup at CHARPOS. */ | |
195d1361 RS |
345 | |
346 | static int | |
d311d28c | 347 | char_quoted (ptrdiff_t charpos, ptrdiff_t bytepos) |
195d1361 RS |
348 | { |
349 | register enum syntaxcode code; | |
d311d28c | 350 | register ptrdiff_t beg = BEGV; |
195d1361 | 351 | register int quoted = 0; |
d311d28c | 352 | ptrdiff_t orig = charpos; |
6a140a74 | 353 | |
02d8b017 | 354 | while (charpos > beg) |
195d1361 | 355 | { |
ab229fdd | 356 | int c; |
02d8b017 | 357 | DEC_BOTH (charpos, bytepos); |
ab229fdd | 358 | |
6a140a74 | 359 | UPDATE_SYNTAX_TABLE_BACKWARD (charpos); |
327719ee | 360 | c = FETCH_CHAR_AS_MULTIBYTE (bytepos); |
ab229fdd | 361 | code = SYNTAX (c); |
6a140a74 RS |
362 | if (! (code == Scharquote || code == Sescape)) |
363 | break; | |
364 | ||
6a140a74 | 365 | quoted = !quoted; |
195d1361 | 366 | } |
6a140a74 RS |
367 | |
368 | UPDATE_SYNTAX_TABLE (orig); | |
195d1361 RS |
369 | return quoted; |
370 | } | |
6a140a74 | 371 | |
6a140a74 RS |
372 | /* Return the bytepos one character before BYTEPOS. |
373 | We assume that BYTEPOS is not at the start of the buffer. */ | |
374 | ||
b0ab8123 | 375 | static ptrdiff_t |
d311d28c | 376 | dec_bytepos (ptrdiff_t bytepos) |
6a140a74 | 377 | { |
4b4deea2 | 378 | if (NILP (BVAR (current_buffer, enable_multibyte_characters))) |
ef316cf0 RS |
379 | return bytepos - 1; |
380 | ||
6a140a74 RS |
381 | DEC_POS (bytepos); |
382 | return bytepos; | |
383 | } | |
195d1361 | 384 | \f |
2b34df4e | 385 | /* Return a defun-start position before POS and not too far before. |
7cc80f0a RS |
386 | It should be the last one before POS, or nearly the last. |
387 | ||
388 | When open_paren_in_column_0_is_defun_start is nonzero, | |
1fd1cc2f | 389 | only the beginning of the buffer is treated as a defun-start. |
7cc80f0a RS |
390 | |
391 | We record the information about where the scan started | |
392 | and what its result was, so that another call in the same area | |
393 | can return the same value very quickly. | |
195d1361 RS |
394 | |
395 | There is no promise at which position the global syntax data is | |
396 | valid on return from the subroutine, so the caller should explicitly | |
397 | update the global data. */ | |
37bef230 | 398 | |
d311d28c PE |
399 | static ptrdiff_t |
400 | find_defun_start (ptrdiff_t pos, ptrdiff_t pos_byte) | |
37bef230 | 401 | { |
d311d28c | 402 | ptrdiff_t opoint = PT, opoint_byte = PT_BYTE; |
37bef230 | 403 | |
1fd1cc2f RS |
404 | if (!open_paren_in_column_0_is_defun_start) |
405 | { | |
6e20a0d4 | 406 | find_start_value = BEGV; |
1fd1cc2f | 407 | find_start_value_byte = BEGV_BYTE; |
6e20a0d4 CY |
408 | find_start_buffer = current_buffer; |
409 | find_start_modiff = MODIFF; | |
410 | find_start_begv = BEGV; | |
411 | find_start_pos = pos; | |
1fd1cc2f RS |
412 | return BEGV; |
413 | } | |
414 | ||
37bef230 RS |
415 | /* Use previous finding, if it's valid and applies to this inquiry. */ |
416 | if (current_buffer == find_start_buffer | |
417 | /* Reuse the defun-start even if POS is a little farther on. | |
418 | POS might be in the next defun, but that's ok. | |
419 | Our value may not be the best possible, but will still be usable. */ | |
420 | && pos <= find_start_pos + 1000 | |
421 | && pos >= find_start_value | |
422 | && BEGV == find_start_begv | |
423 | && MODIFF == find_start_modiff) | |
424 | return find_start_value; | |
425 | ||
426 | /* Back up to start of line. */ | |
6a140a74 | 427 | scan_newline (pos, pos_byte, BEGV, BEGV_BYTE, -1, 1); |
37bef230 | 428 | |
195d1361 RS |
429 | /* We optimize syntax-table lookup for rare updates. Thus we accept |
430 | only those `^\s(' which are good in global _and_ text-property | |
431 | syntax-tables. */ | |
d48cd3f4 | 432 | SETUP_BUFFER_SYNTAX_TABLE (); |
1fd1cc2f | 433 | while (PT > BEGV) |
37bef230 | 434 | { |
ab229fdd AS |
435 | int c; |
436 | ||
1fd1cc2f RS |
437 | /* Open-paren at start of line means we may have found our |
438 | defun-start. */ | |
327719ee | 439 | c = FETCH_CHAR_AS_MULTIBYTE (PT_BYTE); |
ab229fdd | 440 | if (SYNTAX (c) == Sopen) |
195d1361 | 441 | { |
1fd1cc2f | 442 | SETUP_SYNTAX_TABLE (PT + 1, -1); /* Try again... */ |
327719ee | 443 | c = FETCH_CHAR_AS_MULTIBYTE (PT_BYTE); |
ab229fdd | 444 | if (SYNTAX (c) == Sopen) |
1fd1cc2f RS |
445 | break; |
446 | /* Now fallback to the default value. */ | |
d48cd3f4 | 447 | SETUP_BUFFER_SYNTAX_TABLE (); |
195d1361 | 448 | } |
1fd1cc2f RS |
449 | /* Move to beg of previous line. */ |
450 | scan_newline (PT, PT_BYTE, BEGV, BEGV_BYTE, -2, 1); | |
37bef230 RS |
451 | } |
452 | ||
453 | /* Record what we found, for the next try. */ | |
6a140a74 RS |
454 | find_start_value = PT; |
455 | find_start_value_byte = PT_BYTE; | |
37bef230 RS |
456 | find_start_buffer = current_buffer; |
457 | find_start_modiff = MODIFF; | |
458 | find_start_begv = BEGV; | |
459 | find_start_pos = pos; | |
460 | ||
6a140a74 RS |
461 | TEMP_SET_PT_BOTH (opoint, opoint_byte); |
462 | ||
37bef230 RS |
463 | return find_start_value; |
464 | } | |
465 | \f | |
f902a008 RS |
466 | /* Return the SYNTAX_COMEND_FIRST of the character before POS, POS_BYTE. */ |
467 | ||
468 | static int | |
d311d28c | 469 | prev_char_comend_first (ptrdiff_t pos, ptrdiff_t pos_byte) |
f902a008 RS |
470 | { |
471 | int c, val; | |
472 | ||
473 | DEC_BOTH (pos, pos_byte); | |
474 | UPDATE_SYNTAX_TABLE_BACKWARD (pos); | |
475 | c = FETCH_CHAR (pos_byte); | |
476 | val = SYNTAX_COMEND_FIRST (c); | |
477 | UPDATE_SYNTAX_TABLE_FORWARD (pos + 1); | |
478 | return val; | |
479 | } | |
480 | ||
481 | /* Return the SYNTAX_COMSTART_FIRST of the character before POS, POS_BYTE. */ | |
482 | ||
3f679f55 SM |
483 | /* static int |
484 | * prev_char_comstart_first (pos, pos_byte) | |
485 | * int pos, pos_byte; | |
486 | * { | |
487 | * int c, val; | |
7d0393cf | 488 | * |
3f679f55 SM |
489 | * DEC_BOTH (pos, pos_byte); |
490 | * UPDATE_SYNTAX_TABLE_BACKWARD (pos); | |
491 | * c = FETCH_CHAR (pos_byte); | |
492 | * val = SYNTAX_COMSTART_FIRST (c); | |
493 | * UPDATE_SYNTAX_TABLE_FORWARD (pos + 1); | |
494 | * return val; | |
495 | * } */ | |
f902a008 | 496 | |
6a140a74 RS |
497 | /* Checks whether charpos FROM is at the end of a comment. |
498 | FROM_BYTE is the bytepos corresponding to FROM. | |
499 | Do not move back before STOP. | |
500 | ||
501 | Return a positive value if we find a comment ending at FROM/FROM_BYTE; | |
502 | return -1 otherwise. | |
503 | ||
504 | If successful, store the charpos of the comment's beginning | |
505 | into *CHARPOS_PTR, and the bytepos into *BYTEPOS_PTR. | |
527a32d9 KH |
506 | |
507 | Global syntax data remains valid for backward search starting at | |
508 | the returned value (or at FROM, if the search was not successful). */ | |
195d1361 RS |
509 | |
510 | static int | |
d311d28c | 511 | back_comment (ptrdiff_t from, ptrdiff_t from_byte, ptrdiff_t stop, int comnested, int comstyle, ptrdiff_t *charpos_ptr, ptrdiff_t *bytepos_ptr) |
195d1361 RS |
512 | { |
513 | /* Look back, counting the parity of string-quotes, | |
514 | and recording the comment-starters seen. | |
515 | When we reach a safe place, assume that's not in a string; | |
516 | then step the main scan to the earliest comment-starter seen | |
517 | an even number of string quotes away from the safe place. | |
518 | ||
519 | OFROM[I] is position of the earliest comment-starter seen | |
520 | which is I+2X quotes from the comment-end. | |
521 | PARITY is current parity of quotes from the comment end. */ | |
3ee5041c | 522 | int string_style = -1; /* Presumed outside of any string. */ |
195d1361 | 523 | int string_lossage = 0; |
f7c436c1 | 524 | /* Not a real lossage: indicates that we have passed a matching comment |
7d0393cf | 525 | starter plus a non-matching comment-ender, meaning that any matching |
f7c436c1 SM |
526 | comment-starter we might see later could be a false positive (hidden |
527 | inside another comment). | |
528 | Test case: { a (* b } c (* d *) */ | |
529 | int comment_lossage = 0; | |
d311d28c PE |
530 | ptrdiff_t comment_end = from; |
531 | ptrdiff_t comment_end_byte = from_byte; | |
532 | ptrdiff_t comstart_pos = 0; | |
533 | ptrdiff_t comstart_byte IF_LINT (= 0); | |
eb35b628 RS |
534 | /* Place where the containing defun starts, |
535 | or 0 if we didn't come across it yet. */ | |
d311d28c PE |
536 | ptrdiff_t defun_start = 0; |
537 | ptrdiff_t defun_start_byte = 0; | |
195d1361 | 538 | register enum syntaxcode code; |
95ff8dfc | 539 | int nesting = 1; /* current comment nesting */ |
ea315ed6 | 540 | int c; |
3f679f55 SM |
541 | int syntax = 0; |
542 | ||
543 | /* FIXME: A }} comment-ender style leads to incorrect behavior | |
544 | in the case of {{ c }}} because we ignore the last two chars which are | |
545 | assumed to be comment-enders although they aren't. */ | |
195d1361 RS |
546 | |
547 | /* At beginning of range to scan, we're outside of strings; | |
548 | that determines quote parity to the comment-end. */ | |
549 | while (from != stop) | |
550 | { | |
d311d28c | 551 | ptrdiff_t temp_byte; |
4f3a2f8d | 552 | int prev_syntax, com2start, com2end; |
fbb3da77 | 553 | int comstart; |
6a140a74 | 554 | |
195d1361 | 555 | /* Move back and examine a character. */ |
6a140a74 | 556 | DEC_BOTH (from, from_byte); |
195d1361 RS |
557 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
558 | ||
3f679f55 | 559 | prev_syntax = syntax; |
b7dbcc19 | 560 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
3f679f55 | 561 | syntax = SYNTAX_WITH_FLAGS (c); |
195d1361 RS |
562 | code = SYNTAX (c); |
563 | ||
3f679f55 SM |
564 | /* Check for 2-char comment markers. */ |
565 | com2start = (SYNTAX_FLAGS_COMSTART_FIRST (syntax) | |
566 | && SYNTAX_FLAGS_COMSTART_SECOND (prev_syntax) | |
c5683ceb SM |
567 | && (comstyle |
568 | == SYNTAX_FLAGS_COMMENT_STYLE (prev_syntax, syntax)) | |
3f679f55 SM |
569 | && (SYNTAX_FLAGS_COMMENT_NESTED (prev_syntax) |
570 | || SYNTAX_FLAGS_COMMENT_NESTED (syntax)) == comnested); | |
571 | com2end = (SYNTAX_FLAGS_COMEND_FIRST (syntax) | |
572 | && SYNTAX_FLAGS_COMEND_SECOND (prev_syntax)); | |
fbb3da77 | 573 | comstart = (com2start || code == Scomment); |
01f44d5a | 574 | |
3f679f55 SM |
575 | /* Nasty cases with overlapping 2-char comment markers: |
576 | - snmp-mode: -- c -- foo -- c -- | |
577 | --- c -- | |
578 | ------ c -- | |
579 | - c-mode: *||* | |
580 | |* *|* *| | |
581 | |*| |* |*| | |
582 | /// */ | |
583 | ||
584 | /* If a 2-char comment sequence partly overlaps with another, | |
fbb3da77 SM |
585 | we don't try to be clever. E.g. |*| in C, or }% in modes that |
586 | have %..\n and %{..}%. */ | |
587 | if (from > stop && (com2end || comstart)) | |
195d1361 | 588 | { |
d311d28c | 589 | ptrdiff_t next = from, next_byte = from_byte; |
4f3a2f8d | 590 | int next_c, next_syntax; |
3f679f55 SM |
591 | DEC_BOTH (next, next_byte); |
592 | UPDATE_SYNTAX_TABLE_BACKWARD (next); | |
b7dbcc19 | 593 | next_c = FETCH_CHAR_AS_MULTIBYTE (next_byte); |
3f679f55 | 594 | next_syntax = SYNTAX_WITH_FLAGS (next_c); |
fbb3da77 | 595 | if (((comstart || comnested) |
3f679f55 SM |
596 | && SYNTAX_FLAGS_COMEND_SECOND (syntax) |
597 | && SYNTAX_FLAGS_COMEND_FIRST (next_syntax)) | |
598 | || ((com2end || comnested) | |
599 | && SYNTAX_FLAGS_COMSTART_SECOND (syntax) | |
c5683ceb SM |
600 | && (comstyle |
601 | == SYNTAX_FLAGS_COMMENT_STYLE (syntax, prev_syntax)) | |
3f679f55 SM |
602 | && SYNTAX_FLAGS_COMSTART_FIRST (next_syntax))) |
603 | goto lossage; | |
604 | /* UPDATE_SYNTAX_TABLE_FORWARD (next + 1); */ | |
f902a008 | 605 | } |
3f679f55 SM |
606 | |
607 | if (com2start && comstart_pos == 0) | |
608 | /* We're looking at a comment starter. But it might be a comment | |
609 | ender as well (see snmp-mode). The first time we see one, we | |
610 | need to consider it as a comment starter, | |
611 | and the subsequent times as a comment ender. */ | |
612 | com2end = 0; | |
613 | ||
614 | /* Turn a 2-char comment sequences into the appropriate syntax. */ | |
615 | if (com2end) | |
616 | code = Sendcomment; | |
617 | else if (com2start) | |
618 | code = Scomment; | |
619 | /* Ignore comment starters of a different style. */ | |
620 | else if (code == Scomment | |
c5683ceb | 621 | && (comstyle != SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0) |
3f679f55 | 622 | || SYNTAX_FLAGS_COMMENT_NESTED (syntax) != comnested)) |
1aa963c8 | 623 | continue; |
195d1361 | 624 | |
9828a477 KH |
625 | /* Ignore escaped characters, except comment-enders. */ |
626 | if (code != Sendcomment && char_quoted (from, from_byte)) | |
195d1361 RS |
627 | continue; |
628 | ||
3ee5041c | 629 | switch (code) |
195d1361 | 630 | { |
3ee5041c SM |
631 | case Sstring_fence: |
632 | case Scomment_fence: | |
633 | c = (code == Sstring_fence ? ST_STRING_STYLE : ST_COMMENT_STYLE); | |
634 | case Sstring: | |
635 | /* Track parity of quotes. */ | |
636 | if (string_style == -1) | |
637 | /* Entering a string. */ | |
638 | string_style = c; | |
639 | else if (string_style == c) | |
640 | /* Leaving the string. */ | |
641 | string_style = -1; | |
642 | else | |
643 | /* If we have two kinds of string delimiters. | |
644 | There's no way to grok this scanning backwards. */ | |
195d1361 | 645 | string_lossage = 1; |
3ee5041c | 646 | break; |
7d0393cf | 647 | |
3ee5041c SM |
648 | case Scomment: |
649 | /* We've already checked that it is the relevant comstyle. */ | |
f7c436c1 | 650 | if (string_style != -1 || comment_lossage || string_lossage) |
3ee5041c SM |
651 | /* There are odd string quotes involved, so let's be careful. |
652 | Test case in Pascal: " { " a { " } */ | |
653 | goto lossage; | |
654 | ||
f7c436c1 SM |
655 | if (!comnested) |
656 | { | |
657 | /* Record best comment-starter so far. */ | |
658 | comstart_pos = from; | |
659 | comstart_byte = from_byte; | |
660 | } | |
661 | else if (--nesting <= 0) | |
95ff8dfc RS |
662 | /* nested comments have to be balanced, so we don't need to |
663 | keep looking for earlier ones. We use here the same (slightly | |
664 | incorrect) reasoning as below: since it is followed by uniform | |
665 | paired string quotes, this comment-start has to be outside of | |
666 | strings, else the comment-end itself would be inside a string. */ | |
667 | goto done; | |
3ee5041c SM |
668 | break; |
669 | ||
02010917 | 670 | case Sendcomment: |
c5683ceb | 671 | if (SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0) == comstyle |
d070eb22 | 672 | && ((com2end && SYNTAX_FLAGS_COMMENT_NESTED (prev_syntax)) |
3f679f55 | 673 | || SYNTAX_FLAGS_COMMENT_NESTED (syntax)) == comnested) |
02010917 SM |
674 | /* This is the same style of comment ender as ours. */ |
675 | { | |
676 | if (comnested) | |
677 | nesting++; | |
678 | else | |
679 | /* Anything before that can't count because it would match | |
680 | this comment-ender rather than ours. */ | |
681 | from = stop; /* Break out of the loop. */ | |
682 | } | |
f7c436c1 SM |
683 | else if (comstart_pos != 0 || c != '\n') |
684 | /* We're mixing comment styles here, so we'd better be careful. | |
685 | The (comstart_pos != 0 || c != '\n') check is not quite correct | |
686 | (we should just always set comment_lossage), but removing it | |
687 | would imply that any multiline comment in C would go through | |
688 | lossage, which seems overkill. | |
689 | The failure should only happen in the rare cases such as | |
690 | { (* } *) */ | |
691 | comment_lossage = 1; | |
02010917 | 692 | break; |
195d1361 | 693 | |
02010917 SM |
694 | case Sopen: |
695 | /* Assume a defun-start point is outside of strings. */ | |
696 | if (open_paren_in_column_0_is_defun_start | |
697 | && (from == stop | |
698 | || (temp_byte = dec_bytepos (from_byte), | |
699 | FETCH_CHAR (temp_byte) == '\n'))) | |
700 | { | |
701 | defun_start = from; | |
702 | defun_start_byte = from_byte; | |
703 | from = stop; /* Break out of the loop. */ | |
704 | } | |
eb35b628 | 705 | break; |
02010917 SM |
706 | |
707 | default: | |
f7c436c1 | 708 | break; |
eb35b628 | 709 | } |
195d1361 RS |
710 | } |
711 | ||
712 | if (comstart_pos == 0) | |
713 | { | |
714 | from = comment_end; | |
6a140a74 | 715 | from_byte = comment_end_byte; |
195d1361 RS |
716 | UPDATE_SYNTAX_TABLE_FORWARD (comment_end - 1); |
717 | } | |
f7c436c1 SM |
718 | /* If comstart_pos is set and we get here (ie. didn't jump to `lossage' |
719 | or `done'), then we've found the beginning of the non-nested comment. */ | |
720 | else if (1) /* !comnested */ | |
195d1361 RS |
721 | { |
722 | from = comstart_pos; | |
6a140a74 | 723 | from_byte = comstart_byte; |
789f3320 | 724 | UPDATE_SYNTAX_TABLE_FORWARD (from - 1); |
195d1361 RS |
725 | } |
726 | else | |
727 | { | |
3ee5041c SM |
728 | struct lisp_parse_state state; |
729 | lossage: | |
195d1361 RS |
730 | /* We had two kinds of string delimiters mixed up |
731 | together. Decode this going forwards. | |
02010917 | 732 | Scan fwd from a known safe place (beginning-of-defun) |
195d1361 RS |
733 | to the one in question; this records where we |
734 | last passed a comment starter. */ | |
eb35b628 RS |
735 | /* If we did not already find the defun start, find it now. */ |
736 | if (defun_start == 0) | |
737 | { | |
738 | defun_start = find_defun_start (comment_end, comment_end_byte); | |
739 | defun_start_byte = find_start_value_byte; | |
740 | } | |
02010917 | 741 | do |
195d1361 | 742 | { |
02010917 SM |
743 | scan_sexps_forward (&state, |
744 | defun_start, defun_start_byte, | |
d311d28c PE |
745 | comment_end, TYPE_MINIMUM (EMACS_INT), |
746 | 0, Qnil, 0); | |
02010917 SM |
747 | defun_start = comment_end; |
748 | if (state.incomment == (comnested ? 1 : -1) | |
749 | && state.comstyle == comstyle) | |
750 | from = state.comstr_start; | |
751 | else | |
752 | { | |
753 | from = comment_end; | |
754 | if (state.incomment) | |
755 | /* If comment_end is inside some other comment, maybe ours | |
756 | is nested, so we need to try again from within the | |
757 | surrounding comment. Example: { a (* " *) */ | |
758 | { | |
759 | /* FIXME: We should advance by one or two chars. */ | |
760 | defun_start = state.comstr_start + 2; | |
761 | defun_start_byte = CHAR_TO_BYTE (defun_start); | |
762 | } | |
763 | } | |
764 | } while (defun_start < comment_end); | |
765 | ||
6a140a74 | 766 | from_byte = CHAR_TO_BYTE (from); |
195d1361 RS |
767 | UPDATE_SYNTAX_TABLE_FORWARD (from - 1); |
768 | } | |
7d0393cf | 769 | |
95ff8dfc | 770 | done: |
6a140a74 RS |
771 | *charpos_ptr = from; |
772 | *bytepos_ptr = from_byte; | |
773 | ||
1aa963c8 | 774 | return (from == comment_end) ? -1 : from; |
195d1361 RS |
775 | } |
776 | \f | |
8489eb67 | 777 | DEFUN ("syntax-table-p", Fsyntax_table_p, Ssyntax_table_p, 1, 1, 0, |
fdb82f93 PJ |
778 | doc: /* Return t if OBJECT is a syntax table. |
779 | Currently, any char-table counts as a syntax table. */) | |
5842a27b | 780 | (Lisp_Object object) |
8489eb67 | 781 | { |
2203e1e8 | 782 | if (CHAR_TABLE_P (object) |
e704cb4b | 783 | && EQ (XCHAR_TABLE (object)->purpose, Qsyntax_table)) |
8489eb67 RS |
784 | return Qt; |
785 | return Qnil; | |
786 | } | |
787 | ||
8ea151b2 | 788 | static void |
971de7fb | 789 | check_syntax_table (Lisp_Object obj) |
8489eb67 | 790 | { |
47f5f6ae KS |
791 | CHECK_TYPE (CHAR_TABLE_P (obj) && EQ (XCHAR_TABLE (obj)->purpose, Qsyntax_table), |
792 | Qsyntax_table_p, obj); | |
7d0393cf | 793 | } |
8489eb67 | 794 | |
8489eb67 | 795 | DEFUN ("syntax-table", Fsyntax_table, Ssyntax_table, 0, 0, 0, |
fdb82f93 PJ |
796 | doc: /* Return the current syntax table. |
797 | This is the one specified by the current buffer. */) | |
5842a27b | 798 | (void) |
8489eb67 | 799 | { |
4b4deea2 | 800 | return BVAR (current_buffer, syntax_table); |
8489eb67 RS |
801 | } |
802 | ||
803 | DEFUN ("standard-syntax-table", Fstandard_syntax_table, | |
804 | Sstandard_syntax_table, 0, 0, 0, | |
fdb82f93 PJ |
805 | doc: /* Return the standard syntax table. |
806 | This is the one used for new buffers. */) | |
5842a27b | 807 | (void) |
8489eb67 RS |
808 | { |
809 | return Vstandard_syntax_table; | |
810 | } | |
811 | ||
812 | DEFUN ("copy-syntax-table", Fcopy_syntax_table, Scopy_syntax_table, 0, 1, 0, | |
fdb82f93 PJ |
813 | doc: /* Construct a new syntax table and return it. |
814 | It is a copy of the TABLE, which defaults to the standard syntax table. */) | |
5842a27b | 815 | (Lisp_Object table) |
8489eb67 | 816 | { |
8ea151b2 RS |
817 | Lisp_Object copy; |
818 | ||
265a9e55 | 819 | if (!NILP (table)) |
8ea151b2 RS |
820 | check_syntax_table (table); |
821 | else | |
822 | table = Vstandard_syntax_table; | |
823 | ||
824 | copy = Fcopy_sequence (table); | |
0f867324 RS |
825 | |
826 | /* Only the standard syntax table should have a default element. | |
827 | Other syntax tables should inherit from parents instead. */ | |
742af32f | 828 | set_char_table_defalt (copy, Qnil); |
0f867324 RS |
829 | |
830 | /* Copied syntax tables should all have parents. | |
831 | If we copied one with no parent, such as the standard syntax table, | |
832 | use the standard syntax table as the copy's parent. */ | |
833 | if (NILP (XCHAR_TABLE (copy)->parent)) | |
834 | Fset_char_table_parent (copy, Vstandard_syntax_table); | |
8ea151b2 | 835 | return copy; |
8489eb67 RS |
836 | } |
837 | ||
838 | DEFUN ("set-syntax-table", Fset_syntax_table, Sset_syntax_table, 1, 1, 0, | |
fdb82f93 PJ |
839 | doc: /* Select a new syntax table for the current buffer. |
840 | One argument, a syntax table. */) | |
5842a27b | 841 | (Lisp_Object table) |
8489eb67 | 842 | { |
4e4dfa78 | 843 | int idx; |
8ea151b2 | 844 | check_syntax_table (table); |
39eb03f1 | 845 | bset_syntax_table (current_buffer, table); |
8489eb67 | 846 | /* Indicate that this buffer now has a specified syntax table. */ |
f6cd0527 GM |
847 | idx = PER_BUFFER_VAR_IDX (syntax_table); |
848 | SET_PER_BUFFER_VALUE_P (current_buffer, idx, 1); | |
8489eb67 RS |
849 | return table; |
850 | } | |
851 | \f | |
852 | /* Convert a letter which signifies a syntax code | |
853 | into the code it signifies. | |
195d1361 | 854 | This is used by modify-syntax-entry, and other things. */ |
8489eb67 RS |
855 | |
856 | unsigned char syntax_spec_code[0400] = | |
857 | { 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, | |
858 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, | |
859 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, | |
860 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, | |
195d1361 | 861 | (char) Swhitespace, (char) Scomment_fence, (char) Sstring, 0377, |
8489eb67 RS |
862 | (char) Smath, 0377, 0377, (char) Squote, |
863 | (char) Sopen, (char) Sclose, 0377, 0377, | |
864 | 0377, (char) Swhitespace, (char) Spunct, (char) Scharquote, | |
865 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, | |
866 | 0377, 0377, 0377, 0377, | |
867 | (char) Scomment, 0377, (char) Sendcomment, 0377, | |
6cb71bf6 | 868 | (char) Sinherit, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* @, A ... */ |
8489eb67 RS |
869 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, |
870 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword, | |
871 | 0377, 0377, 0377, 0377, (char) Sescape, 0377, 0377, (char) Ssymbol, | |
872 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, /* `, a, ... */ | |
873 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, 0377, | |
874 | 0377, 0377, 0377, 0377, 0377, 0377, 0377, (char) Sword, | |
195d1361 | 875 | 0377, 0377, 0377, 0377, (char) Sstring_fence, 0377, 0377, 0377 |
8489eb67 RS |
876 | }; |
877 | ||
195d1361 | 878 | /* Indexed by syntax code, give the letter that describes it. */ |
8489eb67 | 879 | |
195d1361 | 880 | char syntax_code_spec[16] = |
8489eb67 | 881 | { |
195d1361 RS |
882 | ' ', '.', 'w', '_', '(', ')', '\'', '\"', '$', '\\', '/', '<', '>', '@', |
883 | '!', '|' | |
8489eb67 | 884 | }; |
93da5fff KH |
885 | |
886 | /* Indexed by syntax code, give the object (cons of syntax code and | |
887 | nil) to be stored in syntax table. Since these objects can be | |
888 | shared among syntax tables, we generate them in advance. By | |
889 | sharing objects, the function `describe-syntax' can give a more | |
890 | compact listing. */ | |
891 | static Lisp_Object Vsyntax_code_object; | |
892 | ||
8489eb67 RS |
893 | \f |
894 | DEFUN ("char-syntax", Fchar_syntax, Schar_syntax, 1, 1, 0, | |
fdb82f93 | 895 | doc: /* Return the syntax code of CHARACTER, described by a character. |
e2b6daf4 CY |
896 | For example, if CHARACTER is a word constituent, the |
897 | character `w' (119) is returned. | |
fdb82f93 PJ |
898 | The characters that correspond to various syntax codes |
899 | are listed in the documentation of `modify-syntax-entry'. */) | |
5842a27b | 900 | (Lisp_Object character) |
8489eb67 | 901 | { |
8ea151b2 | 902 | int char_int; |
774b9a60 | 903 | CHECK_CHARACTER (character); |
2203e1e8 | 904 | char_int = XINT (character); |
d48cd3f4 | 905 | SETUP_BUFFER_SYNTAX_TABLE (); |
8ea151b2 | 906 | return make_number (syntax_code_spec[(int) SYNTAX (char_int)]); |
beefa22e RS |
907 | } |
908 | ||
909 | DEFUN ("matching-paren", Fmatching_paren, Smatching_paren, 1, 1, 0, | |
fdb82f93 | 910 | doc: /* Return the matching parenthesis of CHARACTER, or nil if none. */) |
5842a27b | 911 | (Lisp_Object character) |
beefa22e | 912 | { |
8ea151b2 | 913 | int char_int, code; |
b7826503 | 914 | CHECK_NUMBER (character); |
2203e1e8 | 915 | char_int = XINT (character); |
d48cd3f4 | 916 | SETUP_BUFFER_SYNTAX_TABLE (); |
8ea151b2 | 917 | code = SYNTAX (char_int); |
a8bd7cd8 | 918 | if (code == Sopen || code == Sclose) |
2e34157c | 919 | return SYNTAX_MATCH (char_int); |
beefa22e | 920 | return Qnil; |
8489eb67 RS |
921 | } |
922 | ||
c65adb44 | 923 | DEFUN ("string-to-syntax", Fstring_to_syntax, Sstring_to_syntax, 1, 1, 0, |
6dcef6ec CY |
924 | doc: /* Convert a syntax descriptor STRING into a raw syntax descriptor. |
925 | STRING should be a string of the form allowed as argument of | |
926 | `modify-syntax-entry'. The return value is a raw syntax descriptor: a | |
927 | cons cell \(CODE . MATCHING-CHAR) which can be used, for example, as | |
928 | the value of a `syntax-table' text property. */) | |
5842a27b | 929 | (Lisp_Object string) |
8489eb67 | 930 | { |
2e567bd3 | 931 | register const unsigned char *p; |
8489eb67 | 932 | register enum syntaxcode code; |
247e20a8 | 933 | int val; |
8ea151b2 | 934 | Lisp_Object match; |
8489eb67 | 935 | |
b7826503 | 936 | CHECK_STRING (string); |
8489eb67 | 937 | |
d5db4077 | 938 | p = SDATA (string); |
8489eb67 RS |
939 | code = (enum syntaxcode) syntax_spec_code[*p++]; |
940 | if (((int) code & 0377) == 0377) | |
b764a653 | 941 | error ("Invalid syntax description letter: %c", p[-1]); |
8489eb67 | 942 | |
8ea151b2 | 943 | if (code == Sinherit) |
c65adb44 | 944 | return Qnil; |
8ea151b2 RS |
945 | |
946 | if (*p) | |
d1be9f0f | 947 | { |
93da5fff | 948 | int len; |
62a6e103 | 949 | int character = STRING_CHAR_AND_LENGTH (p, len); |
93da5fff | 950 | XSETINT (match, character); |
d1be9f0f RS |
951 | if (XFASTINT (match) == ' ') |
952 | match = Qnil; | |
93da5fff | 953 | p += len; |
d1be9f0f RS |
954 | } |
955 | else | |
8ea151b2 | 956 | match = Qnil; |
8489eb67 | 957 | |
8ea151b2 | 958 | val = (int) code; |
8489eb67 RS |
959 | while (*p) |
960 | switch (*p++) | |
961 | { | |
962 | case '1': | |
247e20a8 | 963 | val |= 1 << 16; |
8489eb67 RS |
964 | break; |
965 | ||
966 | case '2': | |
247e20a8 | 967 | val |= 1 << 17; |
8489eb67 RS |
968 | break; |
969 | ||
970 | case '3': | |
247e20a8 | 971 | val |= 1 << 18; |
8489eb67 RS |
972 | break; |
973 | ||
974 | case '4': | |
247e20a8 | 975 | val |= 1 << 19; |
8489eb67 RS |
976 | break; |
977 | ||
978 | case 'p': | |
247e20a8 | 979 | val |= 1 << 20; |
8489eb67 | 980 | break; |
e5d4f4dc RS |
981 | |
982 | case 'b': | |
247e20a8 | 983 | val |= 1 << 21; |
e5d4f4dc | 984 | break; |
95ff8dfc RS |
985 | |
986 | case 'n': | |
987 | val |= 1 << 22; | |
988 | break; | |
c5683ceb SM |
989 | |
990 | case 'c': | |
991 | val |= 1 << 23; | |
992 | break; | |
8489eb67 | 993 | } |
7d0393cf | 994 | |
77b37c05 | 995 | if (val < ASIZE (Vsyntax_code_object) && NILP (match)) |
28be1ada | 996 | return AREF (Vsyntax_code_object, val); |
93da5fff KH |
997 | else |
998 | /* Since we can't use a shared object, let's make a new one. */ | |
c65adb44 SM |
999 | return Fcons (make_number (val), match); |
1000 | } | |
1001 | ||
fdb82f93 | 1002 | /* I really don't know why this is interactive |
2c9e1900 | 1003 | help-form should at least be made useful whilst reading the second arg. */ |
7d0393cf | 1004 | DEFUN ("modify-syntax-entry", Fmodify_syntax_entry, Smodify_syntax_entry, 2, 3, |
c65adb44 | 1005 | "cSet syntax for character: \nsSet syntax for %s to: ", |
a50a10a0 | 1006 | doc: /* Set syntax for character CHAR according to string NEWENTRY. |
d7ee9fab | 1007 | The syntax is changed only for table SYNTAX-TABLE, which defaults to |
fdb82f93 | 1008 | the current buffer's syntax table. |
5c7b02ab | 1009 | CHAR may be a cons (MIN . MAX), in which case, syntaxes of all characters |
fedc6ab5 | 1010 | in the range MIN to MAX are changed. |
fdb82f93 PJ |
1011 | The first character of NEWENTRY should be one of the following: |
1012 | Space or - whitespace syntax. w word constituent. | |
1013 | _ symbol constituent. . punctuation. | |
1014 | ( open-parenthesis. ) close-parenthesis. | |
1015 | " string quote. \\ escape. | |
1016 | $ paired delimiter. ' expression quote or prefix operator. | |
1017 | < comment starter. > comment ender. | |
02676e5d | 1018 | / character-quote. @ inherit from parent table. |
fdb82f93 PJ |
1019 | | generic string fence. ! generic comment fence. |
1020 | ||
1021 | Only single-character comment start and end sequences are represented thus. | |
1022 | Two-character sequences are represented as described below. | |
1023 | The second character of NEWENTRY is the matching parenthesis, | |
1024 | used only if the first character is `(' or `)'. | |
1025 | Any additional characters are flags. | |
1026 | Defined flags are the characters 1, 2, 3, 4, b, p, and n. | |
a50a10a0 PJ |
1027 | 1 means CHAR is the start of a two-char comment start sequence. |
1028 | 2 means CHAR is the second character of such a sequence. | |
1029 | 3 means CHAR is the start of a two-char comment end sequence. | |
1030 | 4 means CHAR is the second character of such a sequence. | |
fdb82f93 | 1031 | |
c5683ceb | 1032 | There can be several orthogonal comment sequences. This is to support |
fdb82f93 PJ |
1033 | language modes such as C++. By default, all comment sequences are of style |
1034 | a, but you can set the comment sequence style to b (on the second character | |
c5683ceb SM |
1035 | of a comment-start, and the first character of a comment-end sequence) and/or |
1036 | c (on any of its chars) using this flag: | |
a50a10a0 | 1037 | b means CHAR is part of comment sequence b. |
c5683ceb | 1038 | c means CHAR is part of comment sequence c. |
a50a10a0 | 1039 | n means CHAR is part of a nestable comment sequence. |
fdb82f93 | 1040 | |
a50a10a0 | 1041 | p means CHAR is a prefix character for `backward-prefix-chars'; |
fdb82f93 | 1042 | such characters are treated as whitespace when they occur |
a50a10a0 | 1043 | between expressions. |
fedc6ab5 | 1044 | usage: (modify-syntax-entry CHAR NEWENTRY &optional SYNTAX-TABLE) */) |
5842a27b | 1045 | (Lisp_Object c, Lisp_Object newentry, Lisp_Object syntax_table) |
c65adb44 | 1046 | { |
5c7b02ab KH |
1047 | if (CONSP (c)) |
1048 | { | |
8f924df7 KH |
1049 | CHECK_CHARACTER_CAR (c); |
1050 | CHECK_CHARACTER_CDR (c); | |
5c7b02ab KH |
1051 | } |
1052 | else | |
1053 | CHECK_CHARACTER (c); | |
c65adb44 SM |
1054 | |
1055 | if (NILP (syntax_table)) | |
4b4deea2 | 1056 | syntax_table = BVAR (current_buffer, syntax_table); |
c65adb44 SM |
1057 | else |
1058 | check_syntax_table (syntax_table); | |
8489eb67 | 1059 | |
5c7b02ab KH |
1060 | newentry = Fstring_to_syntax (newentry); |
1061 | if (CONSP (c)) | |
1062 | SET_RAW_SYNTAX_ENTRY_RANGE (syntax_table, c, newentry); | |
1063 | else | |
1064 | SET_RAW_SYNTAX_ENTRY (syntax_table, XINT (c), newentry); | |
e5b94d44 CY |
1065 | |
1066 | /* We clear the regexp cache, since character classes can now have | |
1067 | different values from those in the compiled regexps.*/ | |
1068 | clear_regexp_cache (); | |
1069 | ||
8489eb67 RS |
1070 | return Qnil; |
1071 | } | |
1072 | \f | |
1073 | /* Dump syntax table to buffer in human-readable format */ | |
1074 | ||
62abe9cb SM |
1075 | DEFUN ("internal-describe-syntax-value", Finternal_describe_syntax_value, |
1076 | Sinternal_describe_syntax_value, 1, 1, 0, | |
1077 | doc: /* Insert a description of the internal syntax description SYNTAX at point. */) | |
5842a27b | 1078 | (Lisp_Object syntax) |
8489eb67 RS |
1079 | { |
1080 | register enum syntaxcode code; | |
c5683ceb SM |
1081 | int syntax_code; |
1082 | char desc, start1, start2, end1, end2, prefix, | |
1083 | comstyleb, comstylec, comnested; | |
8489eb67 | 1084 | char str[2]; |
62abe9cb | 1085 | Lisp_Object first, match_lisp, value = syntax; |
8489eb67 | 1086 | |
8ea151b2 RS |
1087 | if (NILP (value)) |
1088 | { | |
62abe9cb SM |
1089 | insert_string ("default"); |
1090 | return syntax; | |
8ea151b2 RS |
1091 | } |
1092 | ||
908b7fea KH |
1093 | if (CHAR_TABLE_P (value)) |
1094 | { | |
62abe9cb SM |
1095 | insert_string ("deeper char-table ..."); |
1096 | return syntax; | |
908b7fea KH |
1097 | } |
1098 | ||
8ea151b2 RS |
1099 | if (!CONSP (value)) |
1100 | { | |
62abe9cb SM |
1101 | insert_string ("invalid"); |
1102 | return syntax; | |
8ea151b2 RS |
1103 | } |
1104 | ||
c1d497be KR |
1105 | first = XCAR (value); |
1106 | match_lisp = XCDR (value); | |
8ea151b2 | 1107 | |
d311d28c | 1108 | if (!INTEGERP (first) || !(NILP (match_lisp) || CHARACTERP (match_lisp))) |
8489eb67 | 1109 | { |
62abe9cb SM |
1110 | insert_string ("invalid"); |
1111 | return syntax; | |
8489eb67 RS |
1112 | } |
1113 | ||
d311d28c | 1114 | syntax_code = XINT (first) & INT_MAX; |
c5683ceb SM |
1115 | code = (enum syntaxcode) (syntax_code & 0377); |
1116 | start1 = SYNTAX_FLAGS_COMSTART_FIRST (syntax_code); | |
1117 | start2 = SYNTAX_FLAGS_COMSTART_SECOND (syntax_code);; | |
1118 | end1 = SYNTAX_FLAGS_COMEND_FIRST (syntax_code); | |
1119 | end2 = SYNTAX_FLAGS_COMEND_SECOND (syntax_code); | |
1120 | prefix = SYNTAX_FLAGS_PREFIX (syntax_code); | |
1121 | comstyleb = SYNTAX_FLAGS_COMMENT_STYLEB (syntax_code); | |
1122 | comstylec = SYNTAX_FLAGS_COMMENT_STYLEC (syntax_code); | |
1123 | comnested = SYNTAX_FLAGS_COMMENT_NESTED (syntax_code); | |
8489eb67 RS |
1124 | |
1125 | if ((int) code < 0 || (int) code >= (int) Smax) | |
1126 | { | |
1127 | insert_string ("invalid"); | |
62abe9cb | 1128 | return syntax; |
8489eb67 RS |
1129 | } |
1130 | desc = syntax_code_spec[(int) code]; | |
1131 | ||
1132 | str[0] = desc, str[1] = 0; | |
1133 | insert (str, 1); | |
1134 | ||
93da5fff KH |
1135 | if (NILP (match_lisp)) |
1136 | insert (" ", 1); | |
1137 | else | |
1138 | insert_char (XINT (match_lisp)); | |
8489eb67 | 1139 | |
8489eb67 RS |
1140 | if (start1) |
1141 | insert ("1", 1); | |
1142 | if (start2) | |
1143 | insert ("2", 1); | |
1144 | ||
1145 | if (end1) | |
1146 | insert ("3", 1); | |
1147 | if (end2) | |
1148 | insert ("4", 1); | |
1149 | ||
1150 | if (prefix) | |
1151 | insert ("p", 1); | |
c5683ceb | 1152 | if (comstyleb) |
e5d4f4dc | 1153 | insert ("b", 1); |
c5683ceb SM |
1154 | if (comstylec) |
1155 | insert ("c", 1); | |
e6365855 SM |
1156 | if (comnested) |
1157 | insert ("n", 1); | |
8489eb67 RS |
1158 | |
1159 | insert_string ("\twhich means: "); | |
1160 | ||
7393bcbb | 1161 | switch (code) |
8489eb67 RS |
1162 | { |
1163 | case Swhitespace: | |
1164 | insert_string ("whitespace"); break; | |
1165 | case Spunct: | |
1166 | insert_string ("punctuation"); break; | |
1167 | case Sword: | |
1168 | insert_string ("word"); break; | |
1169 | case Ssymbol: | |
1170 | insert_string ("symbol"); break; | |
1171 | case Sopen: | |
1172 | insert_string ("open"); break; | |
1173 | case Sclose: | |
1174 | insert_string ("close"); break; | |
1175 | case Squote: | |
d671382e | 1176 | insert_string ("prefix"); break; |
8489eb67 RS |
1177 | case Sstring: |
1178 | insert_string ("string"); break; | |
1179 | case Smath: | |
1180 | insert_string ("math"); break; | |
1181 | case Sescape: | |
1182 | insert_string ("escape"); break; | |
1183 | case Scharquote: | |
1184 | insert_string ("charquote"); break; | |
1185 | case Scomment: | |
1186 | insert_string ("comment"); break; | |
1187 | case Sendcomment: | |
1188 | insert_string ("endcomment"); break; | |
d671382e SM |
1189 | case Sinherit: |
1190 | insert_string ("inherit"); break; | |
1191 | case Scomment_fence: | |
1192 | insert_string ("comment fence"); break; | |
1193 | case Sstring_fence: | |
1194 | insert_string ("string fence"); break; | |
8489eb67 RS |
1195 | default: |
1196 | insert_string ("invalid"); | |
62abe9cb | 1197 | return syntax; |
8489eb67 RS |
1198 | } |
1199 | ||
8ea151b2 | 1200 | if (!NILP (match_lisp)) |
8489eb67 RS |
1201 | { |
1202 | insert_string (", matches "); | |
8ea151b2 | 1203 | insert_char (XINT (match_lisp)); |
8489eb67 RS |
1204 | } |
1205 | ||
1206 | if (start1) | |
1207 | insert_string (",\n\t is the first character of a comment-start sequence"); | |
1208 | if (start2) | |
1209 | insert_string (",\n\t is the second character of a comment-start sequence"); | |
1210 | ||
1211 | if (end1) | |
1212 | insert_string (",\n\t is the first character of a comment-end sequence"); | |
1213 | if (end2) | |
1214 | insert_string (",\n\t is the second character of a comment-end sequence"); | |
c5683ceb | 1215 | if (comstyleb) |
e5d4f4dc | 1216 | insert_string (" (comment style b)"); |
c5683ceb SM |
1217 | if (comstylec) |
1218 | insert_string (" (comment style c)"); | |
e6365855 SM |
1219 | if (comnested) |
1220 | insert_string (" (nestable)"); | |
e5d4f4dc | 1221 | |
8489eb67 RS |
1222 | if (prefix) |
1223 | insert_string (",\n\t is a prefix character for `backward-prefix-chars'"); | |
1224 | ||
62abe9cb SM |
1225 | return syntax; |
1226 | } | |
8489eb67 RS |
1227 | \f |
1228 | /* Return the position across COUNT words from FROM. | |
1229 | If that many words cannot be found before the end of the buffer, return 0. | |
1230 | COUNT negative means scan backward and stop at word beginning. */ | |
1231 | ||
d311d28c PE |
1232 | ptrdiff_t |
1233 | scan_words (register ptrdiff_t from, register EMACS_INT count) | |
8489eb67 | 1234 | { |
d311d28c PE |
1235 | register ptrdiff_t beg = BEGV; |
1236 | register ptrdiff_t end = ZV; | |
1237 | register ptrdiff_t from_byte = CHAR_TO_BYTE (from); | |
93da5fff KH |
1238 | register enum syntaxcode code; |
1239 | int ch0, ch1; | |
fca8fe46 | 1240 | Lisp_Object func, pos; |
8489eb67 RS |
1241 | |
1242 | immediate_quit = 1; | |
1243 | QUIT; | |
1244 | ||
195d1361 RS |
1245 | SETUP_SYNTAX_TABLE (from, count); |
1246 | ||
8489eb67 RS |
1247 | while (count > 0) |
1248 | { | |
1249 | while (1) | |
1250 | { | |
1251 | if (from == end) | |
1252 | { | |
1253 | immediate_quit = 0; | |
1254 | return 0; | |
1255 | } | |
195d1361 | 1256 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
b7dbcc19 | 1257 | ch0 = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
93da5fff | 1258 | code = SYNTAX (ch0); |
6a140a74 | 1259 | INC_BOTH (from, from_byte); |
8489eb67 RS |
1260 | if (words_include_escapes |
1261 | && (code == Sescape || code == Scharquote)) | |
1262 | break; | |
1263 | if (code == Sword) | |
1264 | break; | |
8489eb67 | 1265 | } |
93da5fff KH |
1266 | /* Now CH0 is a character which begins a word and FROM is the |
1267 | position of the next character. */ | |
8f924df7 | 1268 | func = CHAR_TABLE_REF (Vfind_word_boundary_function_table, ch0); |
869bb237 | 1269 | if (! NILP (Ffboundp (func))) |
8489eb67 | 1270 | { |
869bb237 | 1271 | pos = call2 (func, make_number (from - 1), make_number (end)); |
17f85215 | 1272 | if (INTEGERP (pos) && from < XINT (pos) && XINT (pos) <= ZV) |
8f924df7 KH |
1273 | { |
1274 | from = XINT (pos); | |
1275 | from_byte = CHAR_TO_BYTE (from); | |
1276 | } | |
8489eb67 | 1277 | } |
869bb237 | 1278 | else |
8cb8232a | 1279 | { |
8cb8232a KH |
1280 | while (1) |
1281 | { | |
1282 | if (from == end) break; | |
1283 | UPDATE_SYNTAX_TABLE_FORWARD (from); | |
b7dbcc19 | 1284 | ch1 = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
8cb8232a KH |
1285 | code = SYNTAX (ch1); |
1286 | if ((code != Sword | |
1287 | && (! words_include_escapes | |
1288 | || (code != Sescape && code != Scharquote))) | |
2b70a11c | 1289 | || word_boundary_p (ch0, ch1)) |
869bb237 | 1290 | break; |
8cb8232a KH |
1291 | INC_BOTH (from, from_byte); |
1292 | ch0 = ch1; | |
1293 | } | |
8489eb67 RS |
1294 | } |
1295 | count--; | |
1296 | } | |
1297 | while (count < 0) | |
1298 | { | |
1299 | while (1) | |
1300 | { | |
1301 | if (from == beg) | |
1302 | { | |
1303 | immediate_quit = 0; | |
1304 | return 0; | |
1305 | } | |
6a140a74 | 1306 | DEC_BOTH (from, from_byte); |
195d1361 | 1307 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
b7dbcc19 | 1308 | ch1 = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
93da5fff | 1309 | code = SYNTAX (ch1); |
8489eb67 RS |
1310 | if (words_include_escapes |
1311 | && (code == Sescape || code == Scharquote)) | |
1312 | break; | |
1313 | if (code == Sword) | |
1314 | break; | |
8489eb67 | 1315 | } |
93da5fff KH |
1316 | /* Now CH1 is a character which ends a word and FROM is the |
1317 | position of it. */ | |
8f924df7 | 1318 | func = CHAR_TABLE_REF (Vfind_word_boundary_function_table, ch1); |
869bb237 | 1319 | if (! NILP (Ffboundp (func))) |
8f924df7 | 1320 | { |
869bb237 | 1321 | pos = call2 (func, make_number (from), make_number (beg)); |
17f85215 | 1322 | if (INTEGERP (pos) && BEGV <= XINT (pos) && XINT (pos) < from) |
8f924df7 KH |
1323 | { |
1324 | from = XINT (pos); | |
1325 | from_byte = CHAR_TO_BYTE (from); | |
1326 | } | |
8489eb67 | 1327 | } |
869bb237 | 1328 | else |
8489eb67 | 1329 | { |
8cb8232a KH |
1330 | while (1) |
1331 | { | |
8cb8232a KH |
1332 | if (from == beg) |
1333 | break; | |
262be72a | 1334 | DEC_BOTH (from, from_byte); |
8cb8232a | 1335 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
262be72a | 1336 | ch0 = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
8cb8232a KH |
1337 | code = SYNTAX (ch0); |
1338 | if ((code != Sword | |
1339 | && (! words_include_escapes | |
1340 | || (code != Sescape && code != Scharquote))) | |
2b70a11c | 1341 | || word_boundary_p (ch0, ch1)) |
262be72a MB |
1342 | { |
1343 | INC_BOTH (from, from_byte); | |
1344 | break; | |
1345 | } | |
8cb8232a KH |
1346 | ch1 = ch0; |
1347 | } | |
8489eb67 RS |
1348 | } |
1349 | count++; | |
1350 | } | |
1351 | ||
1352 | immediate_quit = 0; | |
1353 | ||
1354 | return from; | |
1355 | } | |
1356 | ||
a7ca3326 | 1357 | DEFUN ("forward-word", Fforward_word, Sforward_word, 0, 1, "^p", |
fdb82f93 PJ |
1358 | doc: /* Move point forward ARG words (backward if ARG is negative). |
1359 | Normally returns t. | |
1360 | If an edge of the buffer or a field boundary is reached, point is left there | |
1361 | and the function returns nil. Field boundaries are not noticed if | |
1362 | `inhibit-field-text-motion' is non-nil. */) | |
5842a27b | 1363 | (Lisp_Object arg) |
8489eb67 | 1364 | { |
c096ae4d | 1365 | Lisp_Object tmp; |
d311d28c | 1366 | ptrdiff_t orig_val, val; |
8489eb67 | 1367 | |
839966f3 KH |
1368 | if (NILP (arg)) |
1369 | XSETFASTINT (arg, 1); | |
1370 | else | |
1371 | CHECK_NUMBER (arg); | |
1372 | ||
1373 | val = orig_val = scan_words (PT, XINT (arg)); | |
2c6ea900 | 1374 | if (! orig_val) |
839966f3 | 1375 | val = XINT (arg) > 0 ? ZV : BEGV; |
5878ee6f | 1376 | |
b8855607 | 1377 | /* Avoid jumping out of an input field. */ |
c096ae4d SM |
1378 | tmp = Fconstrain_to_field (make_number (val), make_number (PT), |
1379 | Qt, Qnil, Qnil); | |
1380 | val = XFASTINT (tmp); | |
7d0393cf | 1381 | |
8489eb67 | 1382 | SET_PT (val); |
e6d8341f | 1383 | return val == orig_val ? Qt : Qnil; |
8489eb67 RS |
1384 | } |
1385 | \f | |
a7ca3326 | 1386 | DEFUN ("skip-chars-forward", Fskip_chars_forward, Sskip_chars_forward, 1, 2, 0, |
fdb82f93 PJ |
1387 | doc: /* Move point forward, stopping before a char not in STRING, or at pos LIM. |
1388 | STRING is like the inside of a `[...]' in a regular expression | |
1389 | except that `]' is never special and `\\' quotes `^', `-' or `\\' | |
7087d5e9 | 1390 | (but not at the end of a range; quoting is never needed there). |
fdb82f93 PJ |
1391 | Thus, with arg "a-zA-Z", this skips letters stopping before first nonletter. |
1392 | With arg "^a-zA-Z", skips nonletters stopping before first letter. | |
a1bc88d4 RS |
1393 | Char classes, e.g. `[:alpha:]', are supported. |
1394 | ||
1395 | Returns the distance traveled, either zero or positive. */) | |
5842a27b | 1396 | (Lisp_Object string, Lisp_Object lim) |
195d1361 | 1397 | { |
327719ee | 1398 | return skip_chars (1, string, lim, 1); |
195d1361 RS |
1399 | } |
1400 | ||
a7ca3326 | 1401 | DEFUN ("skip-chars-backward", Fskip_chars_backward, Sskip_chars_backward, 1, 2, 0, |
fdb82f93 PJ |
1402 | doc: /* Move point backward, stopping after a char not in STRING, or at pos LIM. |
1403 | See `skip-chars-forward' for details. | |
1404 | Returns the distance traveled, either zero or negative. */) | |
5842a27b | 1405 | (Lisp_Object string, Lisp_Object lim) |
195d1361 | 1406 | { |
327719ee | 1407 | return skip_chars (0, string, lim, 1); |
195d1361 RS |
1408 | } |
1409 | ||
1410 | DEFUN ("skip-syntax-forward", Fskip_syntax_forward, Sskip_syntax_forward, 1, 2, 0, | |
fdb82f93 PJ |
1411 | doc: /* Move point forward across chars in specified syntax classes. |
1412 | SYNTAX is a string of syntax code characters. | |
1413 | Stop before a char whose syntax is not in SYNTAX, or at position LIM. | |
1414 | If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX. | |
1415 | This function returns the distance traveled, either zero or positive. */) | |
5842a27b | 1416 | (Lisp_Object syntax, Lisp_Object lim) |
195d1361 | 1417 | { |
b7dbcc19 | 1418 | return skip_syntaxes (1, syntax, lim); |
195d1361 RS |
1419 | } |
1420 | ||
1421 | DEFUN ("skip-syntax-backward", Fskip_syntax_backward, Sskip_syntax_backward, 1, 2, 0, | |
fdb82f93 PJ |
1422 | doc: /* Move point backward across chars in specified syntax classes. |
1423 | SYNTAX is a string of syntax code characters. | |
1424 | Stop on reaching a char whose syntax is not in SYNTAX, or at position LIM. | |
1425 | If SYNTAX starts with ^, skip characters whose syntax is NOT in SYNTAX. | |
1426 | This function returns the distance traveled, either zero or negative. */) | |
5842a27b | 1427 | (Lisp_Object syntax, Lisp_Object lim) |
195d1361 | 1428 | { |
b7dbcc19 | 1429 | return skip_syntaxes (0, syntax, lim); |
195d1361 RS |
1430 | } |
1431 | ||
6a140a74 | 1432 | static Lisp_Object |
971de7fb | 1433 | skip_chars (int forwardp, Lisp_Object string, Lisp_Object lim, int handle_iso_classes) |
195d1361 | 1434 | { |
195d1361 RS |
1435 | register unsigned int c; |
1436 | unsigned char fastmap[0400]; | |
b7dbcc19 | 1437 | /* Store the ranges of non-ASCII characters. */ |
4f63c6bb | 1438 | int *char_ranges IF_LINT (= NULL); |
7e68b0ea | 1439 | int n_char_ranges = 0; |
195d1361 | 1440 | int negate = 0; |
d311d28c | 1441 | register ptrdiff_t i, i_byte; |
b7dbcc19 KH |
1442 | /* Set to 1 if the current buffer is multibyte and the region |
1443 | contains non-ASCII chars. */ | |
1444 | int multibyte; | |
1445 | /* Set to 1 if STRING is multibyte and it contains non-ASCII | |
1446 | chars. */ | |
1674d9a2 | 1447 | int string_multibyte; |
d311d28c | 1448 | ptrdiff_t size_byte; |
2e567bd3 | 1449 | const unsigned char *str; |
82d497fc | 1450 | int len; |
a1bc88d4 | 1451 | Lisp_Object iso_classes; |
195d1361 | 1452 | |
b7826503 | 1453 | CHECK_STRING (string); |
a1bc88d4 | 1454 | iso_classes = Qnil; |
82d497fc | 1455 | |
195d1361 RS |
1456 | if (NILP (lim)) |
1457 | XSETINT (lim, forwardp ? ZV : BEGV); | |
1458 | else | |
b7826503 | 1459 | CHECK_NUMBER_COERCE_MARKER (lim); |
195d1361 RS |
1460 | |
1461 | /* In any case, don't allow scan outside bounds of buffer. */ | |
195d1361 RS |
1462 | if (XINT (lim) > ZV) |
1463 | XSETFASTINT (lim, ZV); | |
1464 | if (XINT (lim) < BEGV) | |
1465 | XSETFASTINT (lim, BEGV); | |
1466 | ||
4b4deea2 | 1467 | multibyte = (!NILP (BVAR (current_buffer, enable_multibyte_characters)) |
92eaa22e | 1468 | && (XINT (lim) - PT != CHAR_TO_BYTE (XINT (lim)) - PT_BYTE)); |
8f924df7 | 1469 | string_multibyte = SBYTES (string) > SCHARS (string); |
b7dbcc19 | 1470 | |
72af86bd | 1471 | memset (fastmap, 0, sizeof fastmap); |
195d1361 | 1472 | |
8f924df7 KH |
1473 | str = SDATA (string); |
1474 | size_byte = SBYTES (string); | |
4101e6fe | 1475 | |
82d497fc | 1476 | i_byte = 0; |
13090112 | 1477 | if (i_byte < size_byte |
d5db4077 | 1478 | && SREF (string, 0) == '^') |
195d1361 | 1479 | { |
82d497fc | 1480 | negate = 1; i_byte++; |
195d1361 RS |
1481 | } |
1482 | ||
1483 | /* Find the characters specified and set their elements of fastmap. | |
b7dbcc19 | 1484 | Handle backslashes and ranges specially. |
195d1361 | 1485 | |
b7dbcc19 KH |
1486 | If STRING contains non-ASCII characters, setup char_ranges for |
1487 | them and use fastmap only for their leading codes. */ | |
4101e6fe | 1488 | |
b7dbcc19 | 1489 | if (! string_multibyte) |
195d1361 | 1490 | { |
b7dbcc19 | 1491 | int string_has_eight_bit = 0; |
4101e6fe | 1492 | |
b7dbcc19 KH |
1493 | /* At first setup fastmap. */ |
1494 | while (i_byte < size_byte) | |
1495 | { | |
1496 | c = str[i_byte++]; | |
1497 | ||
a1bc88d4 RS |
1498 | if (handle_iso_classes && c == '[' |
1499 | && i_byte < size_byte | |
327719ee | 1500 | && str[i_byte] == ':') |
a1bc88d4 RS |
1501 | { |
1502 | const unsigned char *class_beg = str + i_byte + 1; | |
1503 | const unsigned char *class_end = class_beg; | |
b3bda4fd | 1504 | const unsigned char *class_limit = str + size_byte - 2; |
db9cd97a | 1505 | /* Leave room for the null. */ |
a1bc88d4 RS |
1506 | unsigned char class_name[CHAR_CLASS_MAX_LENGTH + 1]; |
1507 | re_wctype_t cc; | |
1508 | ||
1509 | if (class_limit - class_beg > CHAR_CLASS_MAX_LENGTH) | |
1510 | class_limit = class_beg + CHAR_CLASS_MAX_LENGTH; | |
1511 | ||
b3bda4fd KS |
1512 | while (class_end < class_limit |
1513 | && *class_end >= 'a' && *class_end <= 'z') | |
a1bc88d4 RS |
1514 | class_end++; |
1515 | ||
b3bda4fd KS |
1516 | if (class_end == class_beg |
1517 | || *class_end != ':' || class_end[1] != ']') | |
1518 | goto not_a_class_name; | |
a1bc88d4 | 1519 | |
72af86bd | 1520 | memcpy (class_name, class_beg, class_end - class_beg); |
a1bc88d4 RS |
1521 | class_name[class_end - class_beg] = 0; |
1522 | ||
1523 | cc = re_wctype (class_name); | |
1524 | if (cc == 0) | |
1525 | error ("Invalid ISO C character class"); | |
1526 | ||
1527 | iso_classes = Fcons (make_number (cc), iso_classes); | |
1528 | ||
1529 | i_byte = class_end + 2 - str; | |
1530 | continue; | |
1531 | } | |
1532 | ||
b3bda4fd | 1533 | not_a_class_name: |
b7dbcc19 KH |
1534 | if (c == '\\') |
1535 | { | |
13090112 | 1536 | if (i_byte == size_byte) |
4101e6fe RS |
1537 | break; |
1538 | ||
5c7b02ab | 1539 | c = str[i_byte++]; |
195d1361 | 1540 | } |
839966f3 KH |
1541 | /* Treat `-' as range character only if another character |
1542 | follows. */ | |
1543 | if (i_byte + 1 < size_byte | |
82d497fc | 1544 | && str[i_byte] == '-') |
195d1361 | 1545 | { |
9690d026 | 1546 | unsigned int c2; |
4101e6fe RS |
1547 | |
1548 | /* Skip over the dash. */ | |
82d497fc | 1549 | i_byte++; |
4101e6fe | 1550 | |
4101e6fe | 1551 | /* Get the end of the range. */ |
5c7b02ab | 1552 | c2 = str[i_byte++]; |
b7dbcc19 KH |
1553 | if (c2 == '\\' |
1554 | && i_byte < size_byte) | |
1555 | c2 = str[i_byte++]; | |
1556 | ||
8f924df7 KH |
1557 | if (c <= c2) |
1558 | { | |
1e69125e PE |
1559 | unsigned lim2 = c2 + 1; |
1560 | while (c < lim2) | |
8f924df7 KH |
1561 | fastmap[c++] = 1; |
1562 | if (! ASCII_CHAR_P (c2)) | |
1563 | string_has_eight_bit = 1; | |
1564 | } | |
195d1361 RS |
1565 | } |
1566 | else | |
b7dbcc19 KH |
1567 | { |
1568 | fastmap[c] = 1; | |
1569 | if (! ASCII_CHAR_P (c)) | |
1570 | string_has_eight_bit = 1; | |
1571 | } | |
1572 | } | |
1573 | ||
1574 | /* If the current range is multibyte and STRING contains | |
1575 | eight-bit chars, arrange fastmap and setup char_ranges for | |
1576 | the corresponding multibyte chars. */ | |
1577 | if (multibyte && string_has_eight_bit) | |
1578 | { | |
1579 | unsigned char fastmap2[0400]; | |
1580 | int range_start_byte, range_start_char; | |
1581 | ||
72af86bd AS |
1582 | memcpy (fastmap + 0200, fastmap2 + 0200, 0200); |
1583 | memset (fastmap + 0200, 0, 0200); | |
b7dbcc19 KH |
1584 | /* We are sure that this loop stops. */ |
1585 | for (i = 0200; ! fastmap2[i]; i++); | |
4c0354d7 | 1586 | c = BYTE8_TO_CHAR (i); |
b7dbcc19 KH |
1587 | fastmap[CHAR_LEADING_CODE (c)] = 1; |
1588 | range_start_byte = i; | |
1589 | range_start_char = c; | |
38182d90 | 1590 | char_ranges = alloca (sizeof *char_ranges * 128 * 2); |
b7dbcc19 KH |
1591 | for (i = 129; i < 0400; i++) |
1592 | { | |
4c0354d7 | 1593 | c = BYTE8_TO_CHAR (i); |
b7dbcc19 KH |
1594 | fastmap[CHAR_LEADING_CODE (c)] = 1; |
1595 | if (i - range_start_byte != c - range_start_char) | |
1596 | { | |
1597 | char_ranges[n_char_ranges++] = range_start_char; | |
1598 | char_ranges[n_char_ranges++] = ((i - 1 - range_start_byte) | |
1599 | + range_start_char); | |
1600 | range_start_byte = i; | |
1601 | range_start_char = c; | |
8f924df7 | 1602 | } |
b7dbcc19 KH |
1603 | } |
1604 | char_ranges[n_char_ranges++] = range_start_char; | |
1605 | char_ranges[n_char_ranges++] = ((i - 1 - range_start_byte) | |
1606 | + range_start_char); | |
195d1361 RS |
1607 | } |
1608 | } | |
f003ca54 | 1609 | else /* STRING is multibyte */ |
b7dbcc19 | 1610 | { |
38182d90 | 1611 | char_ranges = alloca (sizeof *char_ranges * SCHARS (string) * 2); |
f003ca54 | 1612 | |
b7dbcc19 | 1613 | while (i_byte < size_byte) |
195d1361 | 1614 | { |
b7dbcc19 | 1615 | unsigned char leading_code; |
5c7b02ab | 1616 | |
b7dbcc19 | 1617 | leading_code = str[i_byte]; |
62a6e103 | 1618 | c = STRING_CHAR_AND_LENGTH (str + i_byte, len); |
b7dbcc19 | 1619 | i_byte += len; |
5c7b02ab | 1620 | |
327719ee MB |
1621 | if (handle_iso_classes && c == '[' |
1622 | && i_byte < size_byte | |
62a6e103 | 1623 | && STRING_CHAR (str + i_byte) == ':') |
327719ee MB |
1624 | { |
1625 | const unsigned char *class_beg = str + i_byte + 1; | |
1626 | const unsigned char *class_end = class_beg; | |
1627 | const unsigned char *class_limit = str + size_byte - 2; | |
1628 | /* Leave room for the null. */ | |
1629 | unsigned char class_name[CHAR_CLASS_MAX_LENGTH + 1]; | |
1630 | re_wctype_t cc; | |
1631 | ||
1632 | if (class_limit - class_beg > CHAR_CLASS_MAX_LENGTH) | |
1633 | class_limit = class_beg + CHAR_CLASS_MAX_LENGTH; | |
1634 | ||
1635 | while (class_end < class_limit | |
1636 | && *class_end >= 'a' && *class_end <= 'z') | |
1637 | class_end++; | |
1638 | ||
1639 | if (class_end == class_beg | |
1640 | || *class_end != ':' || class_end[1] != ']') | |
1641 | goto not_a_class_name_multibyte; | |
1642 | ||
72af86bd | 1643 | memcpy (class_name, class_beg, class_end - class_beg); |
327719ee MB |
1644 | class_name[class_end - class_beg] = 0; |
1645 | ||
1646 | cc = re_wctype (class_name); | |
1647 | if (cc == 0) | |
1648 | error ("Invalid ISO C character class"); | |
1649 | ||
1650 | iso_classes = Fcons (make_number (cc), iso_classes); | |
1651 | ||
1652 | i_byte = class_end + 2 - str; | |
1653 | continue; | |
1654 | } | |
1655 | ||
1656 | not_a_class_name_multibyte: | |
195d1361 RS |
1657 | if (c == '\\') |
1658 | { | |
13090112 | 1659 | if (i_byte == size_byte) |
4101e6fe RS |
1660 | break; |
1661 | ||
b7dbcc19 | 1662 | leading_code = str[i_byte]; |
62a6e103 | 1663 | c = STRING_CHAR_AND_LENGTH (str + i_byte, len); |
82d497fc | 1664 | i_byte += len; |
195d1361 | 1665 | } |
839966f3 KH |
1666 | /* Treat `-' as range character only if another character |
1667 | follows. */ | |
1668 | if (i_byte + 1 < size_byte | |
82d497fc | 1669 | && str[i_byte] == '-') |
195d1361 | 1670 | { |
9690d026 | 1671 | unsigned int c2; |
b7dbcc19 | 1672 | unsigned char leading_code2; |
4101e6fe RS |
1673 | |
1674 | /* Skip over the dash. */ | |
82d497fc | 1675 | i_byte++; |
4101e6fe | 1676 | |
4101e6fe | 1677 | /* Get the end of the range. */ |
b7dbcc19 | 1678 | leading_code2 = str[i_byte]; |
62a6e103 | 1679 | c2 = STRING_CHAR_AND_LENGTH (str + i_byte, len); |
82d497fc | 1680 | i_byte += len; |
7e68b0ea | 1681 | |
b7dbcc19 KH |
1682 | if (c2 == '\\' |
1683 | && i_byte < size_byte) | |
1684 | { | |
1685 | leading_code2 = str[i_byte]; | |
62a6e103 | 1686 | c2 =STRING_CHAR_AND_LENGTH (str + i_byte, len); |
b7dbcc19 KH |
1687 | i_byte += len; |
1688 | } | |
1689 | ||
f003ca54 KH |
1690 | if (c > c2) |
1691 | continue; | |
b7dbcc19 | 1692 | if (ASCII_CHAR_P (c)) |
06274af5 | 1693 | { |
5c7b02ab KH |
1694 | while (c <= c2 && c < 0x80) |
1695 | fastmap[c++] = 1; | |
b7dbcc19 KH |
1696 | leading_code = CHAR_LEADING_CODE (c); |
1697 | } | |
1698 | if (! ASCII_CHAR_P (c)) | |
1699 | { | |
1e69125e PE |
1700 | unsigned lim2 = leading_code2 + 1; |
1701 | while (leading_code < lim2) | |
b7dbcc19 KH |
1702 | fastmap[leading_code++] = 1; |
1703 | if (c <= c2) | |
e39091b6 | 1704 | { |
b7dbcc19 | 1705 | char_ranges[n_char_ranges++] = c; |
e39091b6 | 1706 | char_ranges[n_char_ranges++] = c2; |
06274af5 KH |
1707 | } |
1708 | } | |
195d1361 RS |
1709 | } |
1710 | else | |
7e68b0ea | 1711 | { |
b7dbcc19 | 1712 | if (ASCII_CHAR_P (c)) |
e39091b6 KH |
1713 | fastmap[c] = 1; |
1714 | else | |
7e68b0ea | 1715 | { |
b7dbcc19 | 1716 | fastmap[leading_code] = 1; |
4101e6fe RS |
1717 | char_ranges[n_char_ranges++] = c; |
1718 | char_ranges[n_char_ranges++] = c; | |
7e68b0ea RS |
1719 | } |
1720 | } | |
195d1361 | 1721 | } |
b7dbcc19 KH |
1722 | |
1723 | /* If the current range is unibyte and STRING contains non-ASCII | |
1724 | chars, arrange fastmap for the corresponding unibyte | |
1725 | chars. */ | |
1726 | ||
1727 | if (! multibyte && n_char_ranges > 0) | |
1728 | { | |
72af86bd | 1729 | memset (fastmap + 0200, 0, 0200); |
b7dbcc19 KH |
1730 | for (i = 0; i < n_char_ranges; i += 2) |
1731 | { | |
1732 | int c1 = char_ranges[i]; | |
1e69125e | 1733 | unsigned lim2 = char_ranges[i + 1] + 1; |
b7dbcc19 | 1734 | |
1e69125e | 1735 | for (; c1 < lim2; c1++) |
2afc21f5 SM |
1736 | { |
1737 | int b = CHAR_TO_BYTE_SAFE (c1); | |
1738 | if (b >= 0) | |
1739 | fastmap[b] = 1; | |
1740 | } | |
b7dbcc19 KH |
1741 | } |
1742 | } | |
195d1361 RS |
1743 | } |
1744 | ||
9690d026 | 1745 | /* If ^ was the first character, complement the fastmap. */ |
195d1361 | 1746 | if (negate) |
b7dbcc19 KH |
1747 | { |
1748 | if (! multibyte) | |
1749 | for (i = 0; i < sizeof fastmap; i++) | |
1750 | fastmap[i] ^= 1; | |
1751 | else | |
1752 | { | |
1753 | for (i = 0; i < 0200; i++) | |
1754 | fastmap[i] ^= 1; | |
1755 | /* All non-ASCII chars possibly match. */ | |
1756 | for (; i < sizeof fastmap; i++) | |
1757 | fastmap[i] = 1; | |
1758 | } | |
1759 | } | |
195d1361 RS |
1760 | |
1761 | { | |
d311d28c PE |
1762 | ptrdiff_t start_point = PT; |
1763 | ptrdiff_t pos = PT; | |
1764 | ptrdiff_t pos_byte = PT_BYTE; | |
9af7511a KH |
1765 | unsigned char *p = PT_ADDR, *endp, *stop; |
1766 | ||
1767 | if (forwardp) | |
1768 | { | |
1769 | endp = (XINT (lim) == GPT) ? GPT_ADDR : CHAR_POS_ADDR (XINT (lim)); | |
d7ee9fab | 1770 | stop = (pos < GPT && GPT < XINT (lim)) ? GPT_ADDR : endp; |
9af7511a KH |
1771 | } |
1772 | else | |
1773 | { | |
1774 | endp = CHAR_POS_ADDR (XINT (lim)); | |
d7ee9fab | 1775 | stop = (pos >= GPT && GPT > XINT (lim)) ? GAP_END_ADDR : endp; |
9af7511a | 1776 | } |
195d1361 RS |
1777 | |
1778 | immediate_quit = 1; | |
618db430 SM |
1779 | /* This code may look up syntax tables using macros that rely on the |
1780 | gl_state object. To make sure this object is not out of date, | |
1781 | let's initialize it manually. | |
1782 | We ignore syntax-table text-properties for now, since that's | |
1783 | what we've done in the past. */ | |
d48cd3f4 | 1784 | SETUP_BUFFER_SYNTAX_TABLE (); |
b7dbcc19 | 1785 | if (forwardp) |
195d1361 | 1786 | { |
b7dbcc19 | 1787 | if (multibyte) |
8f924df7 | 1788 | while (1) |
b7dbcc19 | 1789 | { |
8f924df7 | 1790 | int nbytes; |
9af7511a | 1791 | |
8f924df7 | 1792 | if (p >= stop) |
9af7511a | 1793 | { |
8f924df7 | 1794 | if (p >= endp) |
9af7511a | 1795 | break; |
8f924df7 KH |
1796 | p = GAP_END_ADDR; |
1797 | stop = endp; | |
9af7511a | 1798 | } |
62a6e103 | 1799 | c = STRING_CHAR_AND_LENGTH (p, nbytes); |
327719ee | 1800 | if (! NILP (iso_classes) && in_classes (c, iso_classes)) |
9af7511a | 1801 | { |
327719ee | 1802 | if (negate) |
9af7511a | 1803 | break; |
327719ee MB |
1804 | else |
1805 | goto fwd_ok; | |
9af7511a | 1806 | } |
9af7511a | 1807 | |
8f924df7 | 1808 | if (! fastmap[*p]) |
b7dbcc19 KH |
1809 | break; |
1810 | if (! ASCII_CHAR_P (c)) | |
9af7511a | 1811 | { |
b7dbcc19 KH |
1812 | /* As we are looking at a multibyte character, we |
1813 | must look up the character in the table | |
1814 | CHAR_RANGES. If there's no data in the table, | |
1815 | that character is not what we want to skip. */ | |
1816 | ||
1817 | /* The following code do the right thing even if | |
1818 | n_char_ranges is zero (i.e. no data in | |
1819 | CHAR_RANGES). */ | |
1820 | for (i = 0; i < n_char_ranges; i += 2) | |
1821 | if (c >= char_ranges[i] && c <= char_ranges[i + 1]) | |
9af7511a | 1822 | break; |
b7dbcc19 KH |
1823 | if (!(negate ^ (i < n_char_ranges))) |
1824 | break; | |
9af7511a | 1825 | } |
327719ee | 1826 | fwd_ok: |
8f924df7 | 1827 | p += nbytes, pos++, pos_byte += nbytes; |
b7dbcc19 KH |
1828 | } |
1829 | else | |
8f924df7 KH |
1830 | while (1) |
1831 | { | |
1832 | if (p >= stop) | |
9af7511a | 1833 | { |
8f924df7 | 1834 | if (p >= endp) |
9af7511a | 1835 | break; |
8f924df7 KH |
1836 | p = GAP_END_ADDR; |
1837 | stop = endp; | |
9af7511a | 1838 | } |
327719ee MB |
1839 | |
1840 | if (!NILP (iso_classes) && in_classes (*p, iso_classes)) | |
1841 | { | |
1842 | if (negate) | |
1843 | break; | |
1844 | else | |
1845 | goto fwd_unibyte_ok; | |
1846 | } | |
1847 | ||
8f924df7 KH |
1848 | if (!fastmap[*p]) |
1849 | break; | |
327719ee | 1850 | fwd_unibyte_ok: |
8f924df7 KH |
1851 | p++, pos++, pos_byte++; |
1852 | } | |
195d1361 RS |
1853 | } |
1854 | else | |
1855 | { | |
b7dbcc19 | 1856 | if (multibyte) |
8f924df7 | 1857 | while (1) |
b7dbcc19 | 1858 | { |
8f924df7 | 1859 | unsigned char *prev_p; |
9af7511a | 1860 | |
8f924df7 | 1861 | if (p <= stop) |
9af7511a | 1862 | { |
8f924df7 | 1863 | if (p <= endp) |
9af7511a | 1864 | break; |
8f924df7 KH |
1865 | p = GPT_ADDR; |
1866 | stop = endp; | |
9af7511a | 1867 | } |
8f924df7 KH |
1868 | prev_p = p; |
1869 | while (--p >= stop && ! CHAR_HEAD_P (*p)); | |
62a6e103 | 1870 | c = STRING_CHAR (p); |
a1bc88d4 | 1871 | |
327719ee MB |
1872 | if (! NILP (iso_classes) && in_classes (c, iso_classes)) |
1873 | { | |
1874 | if (negate) | |
1875 | break; | |
9690d026 | 1876 | else |
327719ee | 1877 | goto back_ok; |
7e68b0ea | 1878 | } |
a1bc88d4 | 1879 | |
8f924df7 | 1880 | if (! fastmap[*p]) |
b7dbcc19 KH |
1881 | break; |
1882 | if (! ASCII_CHAR_P (c)) | |
7e68b0ea | 1883 | { |
b7dbcc19 KH |
1884 | /* See the comment in the previous similar code. */ |
1885 | for (i = 0; i < n_char_ranges; i += 2) | |
1886 | if (c >= char_ranges[i] && c <= char_ranges[i + 1]) | |
1887 | break; | |
1888 | if (!(negate ^ (i < n_char_ranges))) | |
1889 | break; | |
7e68b0ea | 1890 | } |
327719ee | 1891 | back_ok: |
8f924df7 | 1892 | pos--, pos_byte -= prev_p - p; |
b7dbcc19 KH |
1893 | } |
1894 | else | |
8f924df7 KH |
1895 | while (1) |
1896 | { | |
1897 | if (p <= stop) | |
9af7511a | 1898 | { |
8f924df7 | 1899 | if (p <= endp) |
9af7511a | 1900 | break; |
8f924df7 KH |
1901 | p = GPT_ADDR; |
1902 | stop = endp; | |
9af7511a | 1903 | } |
e39091b6 | 1904 | |
327719ee MB |
1905 | if (! NILP (iso_classes) && in_classes (p[-1], iso_classes)) |
1906 | { | |
1907 | if (negate) | |
1908 | break; | |
9af7511a | 1909 | else |
327719ee MB |
1910 | goto back_unibyte_ok; |
1911 | } | |
a1bc88d4 | 1912 | |
8f924df7 KH |
1913 | if (!fastmap[p[-1]]) |
1914 | break; | |
327719ee | 1915 | back_unibyte_ok: |
8f924df7 KH |
1916 | p--, pos--, pos_byte--; |
1917 | } | |
195d1361 | 1918 | } |
7e68b0ea | 1919 | |
b7dbcc19 KH |
1920 | SET_PT_BOTH (pos, pos_byte); |
1921 | immediate_quit = 0; | |
1922 | ||
1923 | return make_number (PT - start_point); | |
1924 | } | |
1925 | } | |
7e68b0ea | 1926 | |
b7dbcc19 KH |
1927 | |
1928 | static Lisp_Object | |
971de7fb | 1929 | skip_syntaxes (int forwardp, Lisp_Object string, Lisp_Object lim) |
b7dbcc19 KH |
1930 | { |
1931 | register unsigned int c; | |
1932 | unsigned char fastmap[0400]; | |
1933 | int negate = 0; | |
d311d28c | 1934 | register ptrdiff_t i, i_byte; |
b7dbcc19 | 1935 | int multibyte; |
d311d28c | 1936 | ptrdiff_t size_byte; |
b7dbcc19 KH |
1937 | unsigned char *str; |
1938 | ||
1939 | CHECK_STRING (string); | |
1940 | ||
1941 | if (NILP (lim)) | |
1942 | XSETINT (lim, forwardp ? ZV : BEGV); | |
1943 | else | |
1944 | CHECK_NUMBER_COERCE_MARKER (lim); | |
1945 | ||
1946 | /* In any case, don't allow scan outside bounds of buffer. */ | |
1947 | if (XINT (lim) > ZV) | |
1948 | XSETFASTINT (lim, ZV); | |
1949 | if (XINT (lim) < BEGV) | |
1950 | XSETFASTINT (lim, BEGV); | |
1951 | ||
ade8ee9e | 1952 | if (forwardp ? (PT >= XFASTINT (lim)) : (PT <= XFASTINT (lim))) |
49feb1cd | 1953 | return make_number (0); |
8c6e735b | 1954 | |
4b4deea2 | 1955 | multibyte = (!NILP (BVAR (current_buffer, enable_multibyte_characters)) |
92eaa22e | 1956 | && (XINT (lim) - PT != CHAR_TO_BYTE (XINT (lim)) - PT_BYTE)); |
b7dbcc19 | 1957 | |
72af86bd | 1958 | memset (fastmap, 0, sizeof fastmap); |
b7dbcc19 | 1959 | |
8f924df7 KH |
1960 | if (SBYTES (string) > SCHARS (string)) |
1961 | /* As this is very rare case (syntax spec is ASCII only), don't | |
1962 | consider efficiency. */ | |
bc796a59 KH |
1963 | string = string_make_unibyte (string); |
1964 | ||
8f924df7 KH |
1965 | str = SDATA (string); |
1966 | size_byte = SBYTES (string); | |
bc796a59 | 1967 | |
b7dbcc19 KH |
1968 | i_byte = 0; |
1969 | if (i_byte < size_byte | |
8f924df7 | 1970 | && SREF (string, 0) == '^') |
b7dbcc19 KH |
1971 | { |
1972 | negate = 1; i_byte++; | |
1973 | } | |
1974 | ||
b7dbcc19 KH |
1975 | /* Find the syntaxes specified and set their elements of fastmap. */ |
1976 | ||
1977 | while (i_byte < size_byte) | |
1978 | { | |
1979 | c = str[i_byte++]; | |
8f924df7 | 1980 | fastmap[syntax_spec_code[c]] = 1; |
b7dbcc19 | 1981 | } |
195d1361 | 1982 | |
9690d026 | 1983 | /* If ^ was the first character, complement the fastmap. */ |
195d1361 RS |
1984 | if (negate) |
1985 | for (i = 0; i < sizeof fastmap; i++) | |
9690d026 | 1986 | fastmap[i] ^= 1; |
195d1361 RS |
1987 | |
1988 | { | |
d311d28c PE |
1989 | ptrdiff_t start_point = PT; |
1990 | ptrdiff_t pos = PT; | |
1991 | ptrdiff_t pos_byte = PT_BYTE; | |
8f924df7 KH |
1992 | unsigned char *p = PT_ADDR, *endp, *stop; |
1993 | ||
1994 | if (forwardp) | |
1995 | { | |
1996 | endp = (XINT (lim) == GPT) ? GPT_ADDR : CHAR_POS_ADDR (XINT (lim)); | |
1997 | stop = (pos < GPT && GPT < XINT (lim)) ? GPT_ADDR : endp; | |
1998 | } | |
1999 | else | |
2000 | { | |
2001 | endp = CHAR_POS_ADDR (XINT (lim)); | |
2002 | stop = (pos >= GPT && GPT > XINT (lim)) ? GAP_END_ADDR : endp; | |
2003 | } | |
195d1361 RS |
2004 | |
2005 | immediate_quit = 1; | |
b7dbcc19 KH |
2006 | SETUP_SYNTAX_TABLE (pos, forwardp ? 1 : -1); |
2007 | if (forwardp) | |
195d1361 | 2008 | { |
b7dbcc19 | 2009 | if (multibyte) |
195d1361 | 2010 | { |
8f924df7 | 2011 | while (1) |
8c6e735b | 2012 | { |
8f924df7 KH |
2013 | int nbytes; |
2014 | ||
2015 | if (p >= stop) | |
2016 | { | |
2017 | if (p >= endp) | |
2018 | break; | |
2019 | p = GAP_END_ADDR; | |
2020 | stop = endp; | |
2021 | } | |
62a6e103 | 2022 | c = STRING_CHAR_AND_LENGTH (p, nbytes); |
8f924df7 | 2023 | if (! fastmap[(int) SYNTAX (c)]) |
8c6e735b | 2024 | break; |
8f924df7 | 2025 | p += nbytes, pos++, pos_byte += nbytes; |
8c6e735b KH |
2026 | UPDATE_SYNTAX_TABLE_FORWARD (pos); |
2027 | } | |
195d1361 RS |
2028 | } |
2029 | else | |
2030 | { | |
8c6e735b | 2031 | while (1) |
7e68b0ea | 2032 | { |
8f924df7 KH |
2033 | if (p >= stop) |
2034 | { | |
2035 | if (p >= endp) | |
2036 | break; | |
2037 | p = GAP_END_ADDR; | |
2038 | stop = endp; | |
2039 | } | |
2040 | if (! fastmap[(int) SYNTAX (*p)]) | |
8c6e735b | 2041 | break; |
8f924df7 | 2042 | p++, pos++, pos_byte++; |
b7dbcc19 | 2043 | UPDATE_SYNTAX_TABLE_FORWARD (pos); |
195d1361 RS |
2044 | } |
2045 | } | |
2046 | } | |
2047 | else | |
2048 | { | |
b7dbcc19 | 2049 | if (multibyte) |
195d1361 | 2050 | { |
8c6e735b | 2051 | while (1) |
b7dbcc19 | 2052 | { |
8f924df7 KH |
2053 | unsigned char *prev_p; |
2054 | ||
2055 | if (p <= stop) | |
b7dbcc19 | 2056 | { |
8f924df7 KH |
2057 | if (p <= endp) |
2058 | break; | |
2059 | p = GPT_ADDR; | |
2060 | stop = endp; | |
b7dbcc19 | 2061 | } |
262be72a | 2062 | UPDATE_SYNTAX_TABLE_BACKWARD (pos - 1); |
8f924df7 KH |
2063 | prev_p = p; |
2064 | while (--p >= stop && ! CHAR_HEAD_P (*p)); | |
62a6e103 | 2065 | c = STRING_CHAR (p); |
8f924df7 KH |
2066 | if (! fastmap[(int) SYNTAX (c)]) |
2067 | break; | |
2068 | pos--, pos_byte -= prev_p - p; | |
b7dbcc19 | 2069 | } |
195d1361 RS |
2070 | } |
2071 | else | |
2072 | { | |
8c6e735b KH |
2073 | while (1) |
2074 | { | |
8f924df7 KH |
2075 | if (p <= stop) |
2076 | { | |
2077 | if (p <= endp) | |
2078 | break; | |
2079 | p = GPT_ADDR; | |
2080 | stop = endp; | |
2081 | } | |
102f6132 | 2082 | UPDATE_SYNTAX_TABLE_BACKWARD (pos - 1); |
8f924df7 | 2083 | if (! fastmap[(int) SYNTAX (p[-1])]) |
8c6e735b | 2084 | break; |
8f924df7 | 2085 | p--, pos--, pos_byte--; |
8c6e735b | 2086 | } |
195d1361 RS |
2087 | } |
2088 | } | |
6a140a74 RS |
2089 | |
2090 | SET_PT_BOTH (pos, pos_byte); | |
195d1361 RS |
2091 | immediate_quit = 0; |
2092 | ||
2093 | return make_number (PT - start_point); | |
2094 | } | |
2095 | } | |
a1bc88d4 RS |
2096 | |
2097 | /* Return 1 if character C belongs to one of the ISO classes | |
2098 | in the list ISO_CLASSES. Each class is represented by an | |
2099 | integer which is its type according to re_wctype. */ | |
2100 | ||
2101 | static int | |
971de7fb | 2102 | in_classes (int c, Lisp_Object iso_classes) |
a1bc88d4 RS |
2103 | { |
2104 | int fits_class = 0; | |
2105 | ||
618db430 | 2106 | while (CONSP (iso_classes)) |
a1bc88d4 RS |
2107 | { |
2108 | Lisp_Object elt; | |
2109 | elt = XCAR (iso_classes); | |
2110 | iso_classes = XCDR (iso_classes); | |
2111 | ||
2112 | if (re_iswctype (c, XFASTINT (elt))) | |
2113 | fits_class = 1; | |
2114 | } | |
2115 | ||
2116 | return fits_class; | |
2117 | } | |
195d1361 | 2118 | \f |
95ff8dfc RS |
2119 | /* Jump over a comment, assuming we are at the beginning of one. |
2120 | FROM is the current position. | |
2121 | FROM_BYTE is the bytepos corresponding to FROM. | |
2122 | Do not move past STOP (a charpos). | |
2123 | The comment over which we have to jump is of style STYLE | |
c5683ceb | 2124 | (either SYNTAX_FLAGS_COMMENT_STYLE(foo) or ST_COMMENT_STYLE). |
95ff8dfc RS |
2125 | NESTING should be positive to indicate the nesting at the beginning |
2126 | for nested comments and should be zero or negative else. | |
2127 | ST_COMMENT_STYLE cannot be nested. | |
2128 | PREV_SYNTAX is the SYNTAX_WITH_FLAGS of the previous character | |
2129 | (or 0 If the search cannot start in the middle of a two-character). | |
2130 | ||
2131 | If successful, return 1 and store the charpos of the comment's end | |
2132 | into *CHARPOS_PTR and the corresponding bytepos into *BYTEPOS_PTR. | |
2133 | Else, return 0 and store the charpos STOP into *CHARPOS_PTR, the | |
2134 | corresponding bytepos into *BYTEPOS_PTR and the current nesting | |
2135 | (as defined for state.incomment) in *INCOMMENT_PTR. | |
2136 | ||
2137 | The comment end is the last character of the comment rather than the | |
2138 | character just after the comment. | |
2139 | ||
2140 | Global syntax data is assumed to initially be valid for FROM and | |
2141 | remains valid for forward search starting at the returned position. */ | |
2142 | ||
2143 | static int | |
d311d28c PE |
2144 | forw_comment (ptrdiff_t from, ptrdiff_t from_byte, ptrdiff_t stop, |
2145 | EMACS_INT nesting, int style, int prev_syntax, | |
2146 | ptrdiff_t *charpos_ptr, ptrdiff_t *bytepos_ptr, | |
2147 | EMACS_INT *incomment_ptr) | |
95ff8dfc RS |
2148 | { |
2149 | register int c, c1; | |
2150 | register enum syntaxcode code; | |
c5683ceb | 2151 | register int syntax, other_syntax; |
95ff8dfc RS |
2152 | |
2153 | if (nesting <= 0) nesting = -1; | |
2154 | ||
2155 | /* Enter the loop in the middle so that we find | |
2156 | a 2-char comment ender if we start in the middle of it. */ | |
2157 | syntax = prev_syntax; | |
2158 | if (syntax != 0) goto forw_incomment; | |
2159 | ||
2160 | while (1) | |
2161 | { | |
2162 | if (from == stop) | |
2163 | { | |
2164 | *incomment_ptr = nesting; | |
2165 | *charpos_ptr = from; | |
2166 | *bytepos_ptr = from_byte; | |
2167 | return 0; | |
2168 | } | |
8f924df7 | 2169 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
95ff8dfc RS |
2170 | syntax = SYNTAX_WITH_FLAGS (c); |
2171 | code = syntax & 0xff; | |
2172 | if (code == Sendcomment | |
c5683ceb | 2173 | && SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0) == style |
abf8a9ff SM |
2174 | && (SYNTAX_FLAGS_COMMENT_NESTED (syntax) ? |
2175 | (nesting > 0 && --nesting == 0) : nesting < 0)) | |
95ff8dfc RS |
2176 | /* we have encountered a comment end of the same style |
2177 | as the comment sequence which began this comment | |
2178 | section */ | |
2179 | break; | |
2180 | if (code == Scomment_fence | |
2181 | && style == ST_COMMENT_STYLE) | |
2182 | /* we have encountered a comment end of the same style | |
2183 | as the comment sequence which began this comment | |
2184 | section. */ | |
2185 | break; | |
2186 | if (nesting > 0 | |
2187 | && code == Scomment | |
abf8a9ff | 2188 | && SYNTAX_FLAGS_COMMENT_NESTED (syntax) |
c5683ceb | 2189 | && SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0) == style) |
95ff8dfc RS |
2190 | /* we have encountered a nested comment of the same style |
2191 | as the comment sequence which began this comment section */ | |
2192 | nesting++; | |
2193 | INC_BOTH (from, from_byte); | |
2194 | UPDATE_SYNTAX_TABLE_FORWARD (from); | |
7d0393cf | 2195 | |
95ff8dfc RS |
2196 | forw_incomment: |
2197 | if (from < stop && SYNTAX_FLAGS_COMEND_FIRST (syntax) | |
8f924df7 | 2198 | && (c1 = FETCH_CHAR_AS_MULTIBYTE (from_byte), |
c5683ceb SM |
2199 | other_syntax = SYNTAX_WITH_FLAGS (c1), |
2200 | SYNTAX_FLAGS_COMEND_SECOND (other_syntax)) | |
2201 | && SYNTAX_FLAGS_COMMENT_STYLE (syntax, other_syntax) == style | |
abf8a9ff | 2202 | && ((SYNTAX_FLAGS_COMMENT_NESTED (syntax) || |
c5683ceb SM |
2203 | SYNTAX_FLAGS_COMMENT_NESTED (other_syntax)) |
2204 | ? nesting > 0 : nesting < 0)) | |
4ffe723b GM |
2205 | { |
2206 | if (--nesting <= 0) | |
2207 | /* we have encountered a comment end of the same style | |
2208 | as the comment sequence which began this comment | |
2209 | section */ | |
2210 | break; | |
2211 | else | |
2212 | { | |
2213 | INC_BOTH (from, from_byte); | |
2214 | UPDATE_SYNTAX_TABLE_FORWARD (from); | |
2215 | } | |
2216 | } | |
95ff8dfc RS |
2217 | if (nesting > 0 |
2218 | && from < stop | |
2219 | && SYNTAX_FLAGS_COMSTART_FIRST (syntax) | |
8f924df7 | 2220 | && (c1 = FETCH_CHAR_AS_MULTIBYTE (from_byte), |
c5683ceb SM |
2221 | other_syntax = SYNTAX_WITH_FLAGS (c1), |
2222 | SYNTAX_FLAGS_COMMENT_STYLE (other_syntax, syntax) == style | |
2223 | && SYNTAX_FLAGS_COMSTART_SECOND (other_syntax)) | |
abf8a9ff | 2224 | && (SYNTAX_FLAGS_COMMENT_NESTED (syntax) || |
c5683ceb | 2225 | SYNTAX_FLAGS_COMMENT_NESTED (other_syntax))) |
95ff8dfc RS |
2226 | /* we have encountered a nested comment of the same style |
2227 | as the comment sequence which began this comment | |
2228 | section */ | |
2229 | { | |
2230 | INC_BOTH (from, from_byte); | |
2231 | UPDATE_SYNTAX_TABLE_FORWARD (from); | |
2232 | nesting++; | |
2233 | } | |
2234 | } | |
2235 | *charpos_ptr = from; | |
2236 | *bytepos_ptr = from_byte; | |
2237 | return 1; | |
2238 | } | |
2239 | ||
b3cfe0c8 | 2240 | DEFUN ("forward-comment", Fforward_comment, Sforward_comment, 1, 1, 0, |
177c0ea7 | 2241 | doc: /* |
ee648c11 | 2242 | Move forward across up to COUNT comments. If COUNT is negative, move backward. |
fdb82f93 PJ |
2243 | Stop scanning if we find something other than a comment or whitespace. |
2244 | Set point to where scanning stops. | |
ee648c11 | 2245 | If COUNT comments are found as expected, with nothing except whitespace |
fdb82f93 | 2246 | between them, return t; otherwise return nil. */) |
5842a27b | 2247 | (Lisp_Object count) |
b3cfe0c8 | 2248 | { |
d311d28c PE |
2249 | register ptrdiff_t from; |
2250 | ptrdiff_t from_byte; | |
2251 | register ptrdiff_t stop; | |
8ea151b2 | 2252 | register int c, c1; |
b3cfe0c8 RS |
2253 | register enum syntaxcode code; |
2254 | int comstyle = 0; /* style of comment encountered */ | |
95ff8dfc | 2255 | int comnested = 0; /* whether the comment is nestable or not */ |
be720845 | 2256 | int found; |
2312c580 | 2257 | EMACS_INT count1; |
d311d28c | 2258 | ptrdiff_t out_charpos, out_bytepos; |
7c85f529 | 2259 | EMACS_INT dummy; |
840f481c | 2260 | |
b7826503 | 2261 | CHECK_NUMBER (count); |
840f481c | 2262 | count1 = XINT (count); |
195d1361 | 2263 | stop = count1 > 0 ? ZV : BEGV; |
b3cfe0c8 RS |
2264 | |
2265 | immediate_quit = 1; | |
2266 | QUIT; | |
2267 | ||
2268 | from = PT; | |
6a140a74 | 2269 | from_byte = PT_BYTE; |
b3cfe0c8 | 2270 | |
195d1361 | 2271 | SETUP_SYNTAX_TABLE (from, count1); |
840f481c | 2272 | while (count1 > 0) |
b3cfe0c8 | 2273 | { |
04882296 | 2274 | do |
b3cfe0c8 | 2275 | { |
c5683ceb | 2276 | int comstart_first, syntax, other_syntax; |
f902a008 | 2277 | |
04882296 KH |
2278 | if (from == stop) |
2279 | { | |
ef316cf0 | 2280 | SET_PT_BOTH (from, from_byte); |
b7e6e612 | 2281 | immediate_quit = 0; |
04882296 KH |
2282 | return Qnil; |
2283 | } | |
b7dbcc19 | 2284 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
c5683ceb | 2285 | syntax = SYNTAX_WITH_FLAGS (c); |
b3cfe0c8 | 2286 | code = SYNTAX (c); |
c5683ceb SM |
2287 | comstart_first = SYNTAX_FLAGS_COMSTART_FIRST (syntax); |
2288 | comnested = SYNTAX_FLAGS_COMMENT_NESTED (syntax); | |
2289 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0); | |
6a140a74 | 2290 | INC_BOTH (from, from_byte); |
f902a008 | 2291 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
f902a008 | 2292 | if (from < stop && comstart_first |
b7dbcc19 | 2293 | && (c1 = FETCH_CHAR_AS_MULTIBYTE (from_byte), |
c5683ceb SM |
2294 | other_syntax = SYNTAX_WITH_FLAGS (c1), |
2295 | SYNTAX_FLAGS_COMSTART_SECOND (other_syntax))) | |
b3cfe0c8 | 2296 | { |
7d0393cf | 2297 | /* We have encountered a comment start sequence and we |
7fc8191e | 2298 | are ignoring all text inside comments. We must record |
b3cfe0c8 RS |
2299 | the comment style this sequence begins so that later, |
2300 | only a comment end of the same style actually ends | |
7fc8191e | 2301 | the comment section. */ |
b3cfe0c8 | 2302 | code = Scomment; |
c5683ceb SM |
2303 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (other_syntax, syntax); |
2304 | comnested | |
2305 | = comnested || SYNTAX_FLAGS_COMMENT_NESTED (other_syntax); | |
6a140a74 | 2306 | INC_BOTH (from, from_byte); |
95ff8dfc | 2307 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
b3cfe0c8 | 2308 | } |
04882296 | 2309 | } |
abf8a9ff | 2310 | while (code == Swhitespace || (code == Sendcomment && c == '\n')); |
6a140a74 | 2311 | |
da71ab91 KH |
2312 | if (code == Scomment_fence) |
2313 | comstyle = ST_COMMENT_STYLE; | |
2314 | else if (code != Scomment) | |
04882296 KH |
2315 | { |
2316 | immediate_quit = 0; | |
6a140a74 | 2317 | DEC_BOTH (from, from_byte); |
ef316cf0 | 2318 | SET_PT_BOTH (from, from_byte); |
04882296 KH |
2319 | return Qnil; |
2320 | } | |
2321 | /* We're at the start of a comment. */ | |
95ff8dfc RS |
2322 | found = forw_comment (from, from_byte, stop, comnested, comstyle, 0, |
2323 | &out_charpos, &out_bytepos, &dummy); | |
2324 | from = out_charpos; from_byte = out_bytepos; | |
2325 | if (!found) | |
04882296 | 2326 | { |
95ff8dfc RS |
2327 | immediate_quit = 0; |
2328 | SET_PT_BOTH (from, from_byte); | |
2329 | return Qnil; | |
b3cfe0c8 | 2330 | } |
95ff8dfc RS |
2331 | INC_BOTH (from, from_byte); |
2332 | UPDATE_SYNTAX_TABLE_FORWARD (from); | |
04882296 | 2333 | /* We have skipped one comment. */ |
840f481c | 2334 | count1--; |
b3cfe0c8 RS |
2335 | } |
2336 | ||
840f481c | 2337 | while (count1 < 0) |
b3cfe0c8 | 2338 | { |
b9145dbb | 2339 | while (1) |
b3cfe0c8 | 2340 | { |
c5683ceb | 2341 | int quoted, syntax; |
f902a008 | 2342 | |
b9145dbb RS |
2343 | if (from <= stop) |
2344 | { | |
6a140a74 | 2345 | SET_PT_BOTH (BEGV, BEGV_BYTE); |
b9145dbb RS |
2346 | immediate_quit = 0; |
2347 | return Qnil; | |
2348 | } | |
b3cfe0c8 | 2349 | |
6a140a74 | 2350 | DEC_BOTH (from, from_byte); |
f902a008 | 2351 | /* char_quoted does UPDATE_SYNTAX_TABLE_BACKWARD (from). */ |
6a140a74 | 2352 | quoted = char_quoted (from, from_byte); |
b7dbcc19 | 2353 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
c5683ceb | 2354 | syntax = SYNTAX_WITH_FLAGS (c); |
b3cfe0c8 RS |
2355 | code = SYNTAX (c); |
2356 | comstyle = 0; | |
c5683ceb | 2357 | comnested = SYNTAX_FLAGS_COMMENT_NESTED (syntax); |
7fc8191e | 2358 | if (code == Sendcomment) |
c5683ceb SM |
2359 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0); |
2360 | if (from > stop && SYNTAX_FLAGS_COMEND_SECOND (syntax) | |
f902a008 | 2361 | && prev_char_comend_first (from, from_byte) |
89c6809a | 2362 | && !char_quoted (from - 1, dec_bytepos (from_byte))) |
b3cfe0c8 | 2363 | { |
c5683ceb | 2364 | int other_syntax; |
7fc8191e | 2365 | /* We must record the comment style encountered so that |
b3cfe0c8 | 2366 | later, we can match only the proper comment begin |
7fc8191e | 2367 | sequence of the same style. */ |
6a140a74 | 2368 | DEC_BOTH (from, from_byte); |
f902a008 RS |
2369 | code = Sendcomment; |
2370 | /* Calling char_quoted, above, set up global syntax position | |
2371 | at the new value of FROM. */ | |
b7dbcc19 | 2372 | c1 = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
c5683ceb SM |
2373 | other_syntax = SYNTAX_WITH_FLAGS (c1); |
2374 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (other_syntax, syntax); | |
2375 | comnested | |
2376 | = comnested || SYNTAX_FLAGS_COMMENT_NESTED (other_syntax); | |
b3cfe0c8 RS |
2377 | } |
2378 | ||
195d1361 RS |
2379 | if (code == Scomment_fence) |
2380 | { | |
2381 | /* Skip until first preceding unquoted comment_fence. */ | |
01f44d5a | 2382 | int fence_found = 0; |
d311d28c | 2383 | ptrdiff_t ini = from, ini_byte = from_byte; |
7d0393cf | 2384 | |
6a140a74 | 2385 | while (1) |
195d1361 | 2386 | { |
6a140a74 | 2387 | DEC_BOTH (from, from_byte); |
195d1361 | 2388 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
b7dbcc19 | 2389 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
6a140a74 | 2390 | if (SYNTAX (c) == Scomment_fence |
7d0393cf | 2391 | && !char_quoted (from, from_byte)) |
195d1361 | 2392 | { |
01f44d5a | 2393 | fence_found = 1; |
195d1361 RS |
2394 | break; |
2395 | } | |
fcdd4585 SM |
2396 | else if (from == stop) |
2397 | break; | |
195d1361 | 2398 | } |
01f44d5a | 2399 | if (fence_found == 0) |
195d1361 RS |
2400 | { |
2401 | from = ini; /* Set point to ini + 1. */ | |
6a140a74 | 2402 | from_byte = ini_byte; |
195d1361 RS |
2403 | goto leave; |
2404 | } | |
39c41ad4 SM |
2405 | else |
2406 | /* We have skipped one comment. */ | |
2407 | break; | |
195d1361 RS |
2408 | } |
2409 | else if (code == Sendcomment) | |
b3cfe0c8 | 2410 | { |
95ff8dfc | 2411 | found = back_comment (from, from_byte, stop, comnested, comstyle, |
6a140a74 | 2412 | &out_charpos, &out_bytepos); |
1aa963c8 SM |
2413 | if (found == -1) |
2414 | { | |
abf8a9ff SM |
2415 | if (c == '\n') |
2416 | /* This end-of-line is not an end-of-comment. | |
2417 | Treat it like a whitespace. | |
2418 | CC-mode (and maybe others) relies on this behavior. */ | |
2419 | ; | |
2420 | else | |
2421 | { | |
2422 | /* Failure: we should go back to the end of this | |
2423 | not-quite-endcomment. */ | |
c5683ceb | 2424 | if (SYNTAX (c) != code) |
abf8a9ff SM |
2425 | /* It was a two-char Sendcomment. */ |
2426 | INC_BOTH (from, from_byte); | |
2427 | goto leave; | |
2428 | } | |
1aa963c8 | 2429 | } |
2b927d02 | 2430 | else |
abf8a9ff SM |
2431 | { |
2432 | /* We have skipped one comment. */ | |
2433 | from = out_charpos, from_byte = out_bytepos; | |
2434 | break; | |
2435 | } | |
b3cfe0c8 | 2436 | } |
bb0de084 | 2437 | else if (code != Swhitespace || quoted) |
b3cfe0c8 | 2438 | { |
195d1361 | 2439 | leave: |
b3cfe0c8 | 2440 | immediate_quit = 0; |
6a140a74 | 2441 | INC_BOTH (from, from_byte); |
ef316cf0 | 2442 | SET_PT_BOTH (from, from_byte); |
b3cfe0c8 RS |
2443 | return Qnil; |
2444 | } | |
2445 | } | |
2446 | ||
840f481c | 2447 | count1++; |
b3cfe0c8 RS |
2448 | } |
2449 | ||
ef316cf0 | 2450 | SET_PT_BOTH (from, from_byte); |
b3cfe0c8 RS |
2451 | immediate_quit = 0; |
2452 | return Qt; | |
2453 | } | |
2454 | \f | |
b7dbcc19 | 2455 | /* Return syntax code of character C if C is an ASCII character |
db9f48c3 | 2456 | or `multibyte_symbol_p' is zero. Otherwise, return Ssymbol. */ |
bd25db08 | 2457 | |
b7dbcc19 KH |
2458 | #define SYNTAX_WITH_MULTIBYTE_CHECK(c) \ |
2459 | ((ASCII_CHAR_P (c) || !multibyte_symbol_p) \ | |
bd25db08 KH |
2460 | ? SYNTAX (c) : Ssymbol) |
2461 | ||
6a140a74 | 2462 | static Lisp_Object |
971de7fb | 2463 | scan_lists (register EMACS_INT from, EMACS_INT count, EMACS_INT depth, int sexpflag) |
8489eb67 RS |
2464 | { |
2465 | Lisp_Object val; | |
d311d28c | 2466 | register ptrdiff_t stop = count > 0 ? ZV : BEGV; |
93da5fff KH |
2467 | register int c, c1; |
2468 | int stringterm; | |
8489eb67 RS |
2469 | int quoted; |
2470 | int mathexit = 0; | |
93da5fff | 2471 | register enum syntaxcode code, temp_code; |
d311d28c | 2472 | EMACS_INT min_depth = depth; /* Err out if depth gets less than this. */ |
e5d4f4dc | 2473 | int comstyle = 0; /* style of comment encountered */ |
95ff8dfc | 2474 | int comnested = 0; /* whether the comment is nestable or not */ |
d311d28c | 2475 | ptrdiff_t temp_pos; |
932e6895 | 2476 | EMACS_INT last_good = from; |
195d1361 | 2477 | int found; |
d311d28c PE |
2478 | ptrdiff_t from_byte; |
2479 | ptrdiff_t out_bytepos, out_charpos; | |
2480 | int temp; | |
7c85f529 | 2481 | EMACS_INT dummy; |
bd25db08 | 2482 | int multibyte_symbol_p = sexpflag && multibyte_syntax_as_symbol; |
8489eb67 RS |
2483 | |
2484 | if (depth > 0) min_depth = 0; | |
2485 | ||
c3907a7e RS |
2486 | if (from > ZV) from = ZV; |
2487 | if (from < BEGV) from = BEGV; | |
2488 | ||
2489 | from_byte = CHAR_TO_BYTE (from); | |
2490 | ||
8489eb67 RS |
2491 | immediate_quit = 1; |
2492 | QUIT; | |
2493 | ||
195d1361 | 2494 | SETUP_SYNTAX_TABLE (from, count); |
8489eb67 RS |
2495 | while (count > 0) |
2496 | { | |
8489eb67 RS |
2497 | while (from < stop) |
2498 | { | |
c5683ceb | 2499 | int comstart_first, prefix, syntax, other_syntax; |
195d1361 | 2500 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
b7dbcc19 | 2501 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
c5683ceb | 2502 | syntax = SYNTAX_WITH_FLAGS (c); |
bd25db08 | 2503 | code = SYNTAX_WITH_MULTIBYTE_CHECK (c); |
c5683ceb SM |
2504 | comstart_first = SYNTAX_FLAGS_COMSTART_FIRST (syntax); |
2505 | comnested = SYNTAX_FLAGS_COMMENT_NESTED (syntax); | |
2506 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0); | |
2507 | prefix = SYNTAX_FLAGS_PREFIX (syntax); | |
7bf5e9e4 RS |
2508 | if (depth == min_depth) |
2509 | last_good = from; | |
6a140a74 | 2510 | INC_BOTH (from, from_byte); |
195d1361 | 2511 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
f902a008 | 2512 | if (from < stop && comstart_first |
327719ee | 2513 | && (c = FETCH_CHAR_AS_MULTIBYTE (from_byte), |
c5683ceb SM |
2514 | other_syntax = SYNTAX_WITH_FLAGS (c), |
2515 | SYNTAX_FLAGS_COMSTART_SECOND (other_syntax)) | |
8489eb67 | 2516 | && parse_sexp_ignore_comments) |
e5d4f4dc | 2517 | { |
7d0393cf | 2518 | /* we have encountered a comment start sequence and we |
195d1361 | 2519 | are ignoring all text inside comments. We must record |
e5d4f4dc RS |
2520 | the comment style this sequence begins so that later, |
2521 | only a comment end of the same style actually ends | |
2522 | the comment section */ | |
2523 | code = Scomment; | |
c5683ceb SM |
2524 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (other_syntax, syntax); |
2525 | comnested | |
2526 | = comnested || SYNTAX_FLAGS_COMMENT_NESTED (other_syntax); | |
6a140a74 | 2527 | INC_BOTH (from, from_byte); |
f902a008 | 2528 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
e5d4f4dc | 2529 | } |
7d0393cf | 2530 | |
f902a008 | 2531 | if (prefix) |
8489eb67 RS |
2532 | continue; |
2533 | ||
7393bcbb | 2534 | switch (code) |
8489eb67 RS |
2535 | { |
2536 | case Sescape: | |
2537 | case Scharquote: | |
8b8bf8e6 SM |
2538 | if (from == stop) |
2539 | goto lose; | |
6a140a74 | 2540 | INC_BOTH (from, from_byte); |
8489eb67 RS |
2541 | /* treat following character as a word constituent */ |
2542 | case Sword: | |
2543 | case Ssymbol: | |
2544 | if (depth || !sexpflag) break; | |
195d1361 | 2545 | /* This word counts as a sexp; return at end of it. */ |
8489eb67 RS |
2546 | while (from < stop) |
2547 | { | |
195d1361 | 2548 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
7c7ca3d2 RS |
2549 | |
2550 | /* Some compilers can't handle this inside the switch. */ | |
b7dbcc19 | 2551 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
bd25db08 | 2552 | temp = SYNTAX_WITH_MULTIBYTE_CHECK (c); |
7c7ca3d2 | 2553 | switch (temp) |
8489eb67 RS |
2554 | { |
2555 | case Scharquote: | |
2556 | case Sescape: | |
6a140a74 | 2557 | INC_BOTH (from, from_byte); |
8b8bf8e6 SM |
2558 | if (from == stop) |
2559 | goto lose; | |
8489eb67 RS |
2560 | break; |
2561 | case Sword: | |
2562 | case Ssymbol: | |
2563 | case Squote: | |
2564 | break; | |
2565 | default: | |
2566 | goto done; | |
2567 | } | |
6a140a74 | 2568 | INC_BOTH (from, from_byte); |
8489eb67 RS |
2569 | } |
2570 | goto done; | |
2571 | ||
195d1361 | 2572 | case Scomment_fence: |
95ff8dfc RS |
2573 | comstyle = ST_COMMENT_STYLE; |
2574 | /* FALLTHROUGH */ | |
2575 | case Scomment: | |
8489eb67 | 2576 | if (!parse_sexp_ignore_comments) break; |
95ff8dfc RS |
2577 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
2578 | found = forw_comment (from, from_byte, stop, | |
2579 | comnested, comstyle, 0, | |
2580 | &out_charpos, &out_bytepos, &dummy); | |
2581 | from = out_charpos, from_byte = out_bytepos; | |
2582 | if (!found) | |
8489eb67 | 2583 | { |
95ff8dfc RS |
2584 | if (depth == 0) |
2585 | goto done; | |
2586 | goto lose; | |
8489eb67 | 2587 | } |
95ff8dfc RS |
2588 | INC_BOTH (from, from_byte); |
2589 | UPDATE_SYNTAX_TABLE_FORWARD (from); | |
8489eb67 RS |
2590 | break; |
2591 | ||
2592 | case Smath: | |
2593 | if (!sexpflag) | |
2594 | break; | |
b7dbcc19 | 2595 | if (from != stop && c == FETCH_CHAR_AS_MULTIBYTE (from_byte)) |
6a140a74 RS |
2596 | { |
2597 | INC_BOTH (from, from_byte); | |
2598 | } | |
8489eb67 RS |
2599 | if (mathexit) |
2600 | { | |
2601 | mathexit = 0; | |
2602 | goto close1; | |
2603 | } | |
2604 | mathexit = 1; | |
2605 | ||
2606 | case Sopen: | |
2607 | if (!++depth) goto done; | |
2608 | break; | |
2609 | ||
2610 | case Sclose: | |
2611 | close1: | |
2612 | if (!--depth) goto done; | |
2613 | if (depth < min_depth) | |
9471945b KS |
2614 | xsignal3 (Qscan_error, |
2615 | build_string ("Containing expression ends prematurely"), | |
2616 | make_number (last_good), make_number (from)); | |
8489eb67 RS |
2617 | break; |
2618 | ||
2619 | case Sstring: | |
195d1361 | 2620 | case Sstring_fence: |
ef316cf0 | 2621 | temp_pos = dec_bytepos (from_byte); |
b7dbcc19 | 2622 | stringterm = FETCH_CHAR_AS_MULTIBYTE (temp_pos); |
8489eb67 RS |
2623 | while (1) |
2624 | { | |
8b8bf8e6 SM |
2625 | if (from >= stop) |
2626 | goto lose; | |
195d1361 | 2627 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
b7dbcc19 | 2628 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
42ebfa31 SM |
2629 | if (code == Sstring |
2630 | ? (c == stringterm | |
2631 | && SYNTAX_WITH_MULTIBYTE_CHECK (c) == Sstring) | |
bd25db08 | 2632 | : SYNTAX_WITH_MULTIBYTE_CHECK (c) == Sstring_fence) |
195d1361 | 2633 | break; |
7c7ca3d2 RS |
2634 | |
2635 | /* Some compilers can't handle this inside the switch. */ | |
bd25db08 | 2636 | temp = SYNTAX_WITH_MULTIBYTE_CHECK (c); |
7c7ca3d2 | 2637 | switch (temp) |
8489eb67 RS |
2638 | { |
2639 | case Scharquote: | |
2640 | case Sescape: | |
6a140a74 | 2641 | INC_BOTH (from, from_byte); |
8489eb67 | 2642 | } |
6a140a74 | 2643 | INC_BOTH (from, from_byte); |
8489eb67 | 2644 | } |
6a140a74 | 2645 | INC_BOTH (from, from_byte); |
8489eb67 RS |
2646 | if (!depth && sexpflag) goto done; |
2647 | break; | |
240c43e8 SM |
2648 | default: |
2649 | /* Ignore whitespace, punctuation, quote, endcomment. */ | |
2650 | break; | |
8489eb67 RS |
2651 | } |
2652 | } | |
2653 | ||
2654 | /* Reached end of buffer. Error if within object, return nil if between */ | |
8b8bf8e6 SM |
2655 | if (depth) |
2656 | goto lose; | |
8489eb67 RS |
2657 | |
2658 | immediate_quit = 0; | |
2659 | return Qnil; | |
2660 | ||
2661 | /* End of object reached */ | |
2662 | done: | |
2663 | count--; | |
2664 | } | |
2665 | ||
2666 | while (count < 0) | |
2667 | { | |
8489eb67 RS |
2668 | while (from > stop) |
2669 | { | |
c5683ceb | 2670 | int syntax; |
6a140a74 | 2671 | DEC_BOTH (from, from_byte); |
195d1361 | 2672 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
b7dbcc19 | 2673 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
c5683ceb | 2674 | syntax= SYNTAX_WITH_FLAGS (c); |
bd25db08 | 2675 | code = SYNTAX_WITH_MULTIBYTE_CHECK (c); |
7bf5e9e4 RS |
2676 | if (depth == min_depth) |
2677 | last_good = from; | |
7fc8191e | 2678 | comstyle = 0; |
c5683ceb | 2679 | comnested = SYNTAX_FLAGS_COMMENT_NESTED (syntax); |
7fc8191e | 2680 | if (code == Sendcomment) |
c5683ceb SM |
2681 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (syntax, 0); |
2682 | if (from > stop && SYNTAX_FLAGS_COMEND_SECOND (syntax) | |
1674d9a2 | 2683 | && prev_char_comend_first (from, from_byte) |
8489eb67 | 2684 | && parse_sexp_ignore_comments) |
e5d4f4dc | 2685 | { |
f902a008 | 2686 | /* We must record the comment style encountered so that |
e5d4f4dc | 2687 | later, we can match only the proper comment begin |
f902a008 | 2688 | sequence of the same style. */ |
01f44d5a | 2689 | int c2, other_syntax; |
6a140a74 | 2690 | DEC_BOTH (from, from_byte); |
f902a008 RS |
2691 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
2692 | code = Sendcomment; | |
01f44d5a PE |
2693 | c2 = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
2694 | other_syntax = SYNTAX_WITH_FLAGS (c2); | |
c5683ceb SM |
2695 | comstyle = SYNTAX_FLAGS_COMMENT_STYLE (other_syntax, syntax); |
2696 | comnested | |
2697 | = comnested || SYNTAX_FLAGS_COMMENT_NESTED (other_syntax); | |
e5d4f4dc | 2698 | } |
7d0393cf | 2699 | |
9828a477 | 2700 | /* Quoting turns anything except a comment-ender |
e6365855 | 2701 | into a word character. Note that this cannot be true |
f902a008 | 2702 | if we decremented FROM in the if-statement above. */ |
9828a477 | 2703 | if (code != Sendcomment && char_quoted (from, from_byte)) |
240c43e8 SM |
2704 | { |
2705 | DEC_BOTH (from, from_byte); | |
2706 | code = Sword; | |
2707 | } | |
c5683ceb | 2708 | else if (SYNTAX_FLAGS_PREFIX (syntax)) |
8489eb67 RS |
2709 | continue; |
2710 | ||
7393bcbb | 2711 | switch (code) |
8489eb67 RS |
2712 | { |
2713 | case Sword: | |
2714 | case Ssymbol: | |
9828a477 KH |
2715 | case Sescape: |
2716 | case Scharquote: | |
8489eb67 | 2717 | if (depth || !sexpflag) break; |
195d1361 RS |
2718 | /* This word counts as a sexp; count object finished |
2719 | after passing it. */ | |
8489eb67 RS |
2720 | while (from > stop) |
2721 | { | |
6a140a74 | 2722 | temp_pos = from_byte; |
4b4deea2 | 2723 | if (! NILP (BVAR (current_buffer, enable_multibyte_characters))) |
ef316cf0 RS |
2724 | DEC_POS (temp_pos); |
2725 | else | |
2726 | temp_pos--; | |
6a140a74 | 2727 | UPDATE_SYNTAX_TABLE_BACKWARD (from - 1); |
b7dbcc19 | 2728 | c1 = FETCH_CHAR_AS_MULTIBYTE (temp_pos); |
bd25db08 | 2729 | temp_code = SYNTAX_WITH_MULTIBYTE_CHECK (c1); |
9828a477 KH |
2730 | /* Don't allow comment-end to be quoted. */ |
2731 | if (temp_code == Sendcomment) | |
2732 | goto done2; | |
6a140a74 | 2733 | quoted = char_quoted (from - 1, temp_pos); |
8489eb67 | 2734 | if (quoted) |
93da5fff | 2735 | { |
6a140a74 | 2736 | DEC_BOTH (from, from_byte); |
ef316cf0 | 2737 | temp_pos = dec_bytepos (temp_pos); |
6a140a74 | 2738 | UPDATE_SYNTAX_TABLE_BACKWARD (from - 1); |
93da5fff | 2739 | } |
b7dbcc19 | 2740 | c1 = FETCH_CHAR_AS_MULTIBYTE (temp_pos); |
bd25db08 | 2741 | temp_code = SYNTAX_WITH_MULTIBYTE_CHECK (c1); |
93da5fff KH |
2742 | if (! (quoted || temp_code == Sword |
2743 | || temp_code == Ssymbol | |
2744 | || temp_code == Squote)) | |
8489eb67 | 2745 | goto done2; |
6a140a74 | 2746 | DEC_BOTH (from, from_byte); |
8489eb67 RS |
2747 | } |
2748 | goto done2; | |
2749 | ||
2750 | case Smath: | |
2751 | if (!sexpflag) | |
2752 | break; | |
ef316cf0 | 2753 | temp_pos = dec_bytepos (from_byte); |
6a140a74 | 2754 | UPDATE_SYNTAX_TABLE_BACKWARD (from - 1); |
b7dbcc19 | 2755 | if (from != stop && c == FETCH_CHAR_AS_MULTIBYTE (temp_pos)) |
6a140a74 | 2756 | DEC_BOTH (from, from_byte); |
8489eb67 RS |
2757 | if (mathexit) |
2758 | { | |
2759 | mathexit = 0; | |
2760 | goto open2; | |
2761 | } | |
2762 | mathexit = 1; | |
2763 | ||
2764 | case Sclose: | |
2765 | if (!++depth) goto done2; | |
2766 | break; | |
2767 | ||
2768 | case Sopen: | |
2769 | open2: | |
2770 | if (!--depth) goto done2; | |
2771 | if (depth < min_depth) | |
9471945b KS |
2772 | xsignal3 (Qscan_error, |
2773 | build_string ("Containing expression ends prematurely"), | |
2774 | make_number (last_good), make_number (from)); | |
8489eb67 RS |
2775 | break; |
2776 | ||
2777 | case Sendcomment: | |
2778 | if (!parse_sexp_ignore_comments) | |
2779 | break; | |
95ff8dfc | 2780 | found = back_comment (from, from_byte, stop, comnested, comstyle, |
6a140a74 | 2781 | &out_charpos, &out_bytepos); |
1aa963c8 SM |
2782 | /* FIXME: if found == -1, then it really wasn't a comment-end. |
2783 | For single-char Sendcomment, we can't do much about it apart | |
2784 | from skipping the char. | |
2785 | For 2-char endcomments, we could try again, taking both | |
2786 | chars as separate entities, but it's a lot of trouble | |
2787 | for very little gain, so we don't bother either. -sm */ | |
6a140a74 RS |
2788 | if (found != -1) |
2789 | from = out_charpos, from_byte = out_bytepos; | |
8489eb67 RS |
2790 | break; |
2791 | ||
195d1361 RS |
2792 | case Scomment_fence: |
2793 | case Sstring_fence: | |
2794 | while (1) | |
2795 | { | |
8b8bf8e6 SM |
2796 | if (from == stop) |
2797 | goto lose; | |
6b61353c | 2798 | DEC_BOTH (from, from_byte); |
195d1361 | 2799 | UPDATE_SYNTAX_TABLE_BACKWARD (from); |
7d0393cf | 2800 | if (!char_quoted (from, from_byte) |
b7dbcc19 | 2801 | && (c = FETCH_CHAR_AS_MULTIBYTE (from_byte), |
bd25db08 | 2802 | SYNTAX_WITH_MULTIBYTE_CHECK (c) == code)) |
195d1361 RS |
2803 | break; |
2804 | } | |
2805 | if (code == Sstring_fence && !depth && sexpflag) goto done2; | |
2806 | break; | |
7d0393cf | 2807 | |
8489eb67 | 2808 | case Sstring: |
b7dbcc19 | 2809 | stringterm = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
8489eb67 RS |
2810 | while (1) |
2811 | { | |
8b8bf8e6 SM |
2812 | if (from == stop) |
2813 | goto lose; | |
6b61353c KH |
2814 | DEC_BOTH (from, from_byte); |
2815 | UPDATE_SYNTAX_TABLE_BACKWARD (from); | |
2816 | if (!char_quoted (from, from_byte) | |
2817 | && (stringterm | |
2818 | == (c = FETCH_CHAR_AS_MULTIBYTE (from_byte))) | |
42ebfa31 | 2819 | && SYNTAX_WITH_MULTIBYTE_CHECK (c) == Sstring) |
8489eb67 | 2820 | break; |
8489eb67 | 2821 | } |
8489eb67 RS |
2822 | if (!depth && sexpflag) goto done2; |
2823 | break; | |
240c43e8 SM |
2824 | default: |
2825 | /* Ignore whitespace, punctuation, quote, endcomment. */ | |
2826 | break; | |
8489eb67 RS |
2827 | } |
2828 | } | |
2829 | ||
2830 | /* Reached start of buffer. Error if within object, return nil if between */ | |
8b8bf8e6 SM |
2831 | if (depth) |
2832 | goto lose; | |
8489eb67 RS |
2833 | |
2834 | immediate_quit = 0; | |
2835 | return Qnil; | |
2836 | ||
2837 | done2: | |
2838 | count++; | |
2839 | } | |
2840 | ||
2841 | ||
2842 | immediate_quit = 0; | |
1e142fb7 | 2843 | XSETFASTINT (val, from); |
8489eb67 RS |
2844 | return val; |
2845 | ||
2846 | lose: | |
9471945b KS |
2847 | xsignal3 (Qscan_error, |
2848 | build_string ("Unbalanced parentheses"), | |
2849 | make_number (last_good), make_number (from)); | |
8489eb67 RS |
2850 | } |
2851 | ||
8489eb67 | 2852 | DEFUN ("scan-lists", Fscan_lists, Sscan_lists, 3, 3, 0, |
fdb82f93 | 2853 | doc: /* Scan from character number FROM by COUNT lists. |
7b926f3f CY |
2854 | Scan forward if COUNT is positive, backward if COUNT is negative. |
2855 | Return the character number of the position thus found. | |
2856 | ||
2857 | A \"list", in this context, refers to a balanced parenthetical | |
2858 | grouping, as determined by the syntax table. | |
fdb82f93 | 2859 | |
7b926f3f CY |
2860 | If DEPTH is nonzero, treat that as the nesting depth of the starting |
2861 | point (i.e. the starting point is DEPTH parentheses deep). This | |
2862 | function scans over parentheses until the depth goes to zero COUNT | |
2863 | times. Hence, positive DEPTH moves out that number of levels of | |
221cd413 | 2864 | parentheses, while negative DEPTH moves to a deeper level. |
fdb82f93 PJ |
2865 | |
2866 | Comments are ignored if `parse-sexp-ignore-comments' is non-nil. | |
2867 | ||
7b926f3f CY |
2868 | If we reach the beginning or end of the accessible part of the buffer |
2869 | before we have scanned over COUNT lists, return nil if the depth at | |
2870 | that point is zero, and signal a error if the depth is nonzero. */) | |
5842a27b | 2871 | (Lisp_Object from, Lisp_Object count, Lisp_Object depth) |
8489eb67 | 2872 | { |
b7826503 PJ |
2873 | CHECK_NUMBER (from); |
2874 | CHECK_NUMBER (count); | |
2875 | CHECK_NUMBER (depth); | |
8489eb67 RS |
2876 | |
2877 | return scan_lists (XINT (from), XINT (count), XINT (depth), 0); | |
2878 | } | |
2879 | ||
2880 | DEFUN ("scan-sexps", Fscan_sexps, Sscan_sexps, 2, 2, 0, | |
fdb82f93 PJ |
2881 | doc: /* Scan from character number FROM by COUNT balanced expressions. |
2882 | If COUNT is negative, scan backwards. | |
2883 | Returns the character number of the position thus found. | |
2884 | ||
2885 | Comments are ignored if `parse-sexp-ignore-comments' is non-nil. | |
2886 | ||
2887 | If the beginning or end of (the accessible part of) the buffer is reached | |
2888 | in the middle of a parenthetical grouping, an error is signaled. | |
2889 | If the beginning or end is reached between groupings | |
2890 | but before count is used up, nil is returned. */) | |
5842a27b | 2891 | (Lisp_Object from, Lisp_Object count) |
8489eb67 | 2892 | { |
b7826503 PJ |
2893 | CHECK_NUMBER (from); |
2894 | CHECK_NUMBER (count); | |
8489eb67 RS |
2895 | |
2896 | return scan_lists (XINT (from), XINT (count), 0, 1); | |
2897 | } | |
2898 | ||
2899 | DEFUN ("backward-prefix-chars", Fbackward_prefix_chars, Sbackward_prefix_chars, | |
fdb82f93 PJ |
2900 | 0, 0, 0, |
2901 | doc: /* Move point backward over any number of chars with prefix syntax. | |
2902 | This includes chars with "quote" or "prefix" syntax (' or p). */) | |
5842a27b | 2903 | (void) |
8489eb67 | 2904 | { |
d311d28c PE |
2905 | ptrdiff_t beg = BEGV; |
2906 | ptrdiff_t opoint = PT; | |
2907 | ptrdiff_t opoint_byte = PT_BYTE; | |
2908 | ptrdiff_t pos = PT; | |
2909 | ptrdiff_t pos_byte = PT_BYTE; | |
93da5fff | 2910 | int c; |
93da5fff | 2911 | |
7d0393cf | 2912 | if (pos <= beg) |
195d1361 | 2913 | { |
f902a008 RS |
2914 | SET_PT_BOTH (opoint, opoint_byte); |
2915 | ||
2916 | return Qnil; | |
195d1361 | 2917 | } |
8489eb67 | 2918 | |
f902a008 RS |
2919 | SETUP_SYNTAX_TABLE (pos, -1); |
2920 | ||
6a140a74 RS |
2921 | DEC_BOTH (pos, pos_byte); |
2922 | ||
1fd3172d | 2923 | while (!char_quoted (pos, pos_byte) |
195d1361 | 2924 | /* Previous statement updates syntax table. */ |
b7dbcc19 | 2925 | && ((c = FETCH_CHAR_AS_MULTIBYTE (pos_byte), SYNTAX (c) == Squote) |
93da5fff KH |
2926 | || SYNTAX_PREFIX (c))) |
2927 | { | |
1fd3172d RS |
2928 | opoint = pos; |
2929 | opoint_byte = pos_byte; | |
2930 | ||
2931 | if (pos + 1 > beg) | |
2932 | DEC_BOTH (pos, pos_byte); | |
93da5fff | 2933 | } |
8489eb67 | 2934 | |
6a140a74 | 2935 | SET_PT_BOTH (opoint, opoint_byte); |
8489eb67 RS |
2936 | |
2937 | return Qnil; | |
2938 | } | |
2939 | \f | |
6a140a74 | 2940 | /* Parse forward from FROM / FROM_BYTE to END, |
e5d4f4dc RS |
2941 | assuming that FROM has state OLDSTATE (nil means FROM is start of function), |
2942 | and return a description of the state of the parse at END. | |
c81a3712 | 2943 | If STOPBEFORE is nonzero, stop at the start of an atom. |
644ea4df RS |
2944 | If COMMENTSTOP is 1, stop at the start of a comment. |
2945 | If COMMENTSTOP is -1, stop at the start or end of a comment, | |
2946 | after the beginning of a string, or after the end of a string. */ | |
8489eb67 | 2947 | |
340f92b5 | 2948 | static void |
d5a3eaaf | 2949 | scan_sexps_forward (struct lisp_parse_state *stateptr, |
d311d28c PE |
2950 | ptrdiff_t from, ptrdiff_t from_byte, ptrdiff_t end, |
2951 | EMACS_INT targetdepth, int stopbefore, | |
d5a3eaaf | 2952 | Lisp_Object oldstate, int commentstop) |
8489eb67 RS |
2953 | { |
2954 | struct lisp_parse_state state; | |
2955 | ||
2956 | register enum syntaxcode code; | |
95ff8dfc RS |
2957 | int c1; |
2958 | int comnested; | |
d311d28c | 2959 | struct level { ptrdiff_t last, prev; }; |
8489eb67 RS |
2960 | struct level levelstart[100]; |
2961 | register struct level *curlevel = levelstart; | |
2962 | struct level *endlevel = levelstart + 100; | |
d311d28c | 2963 | register EMACS_INT depth; /* Paren depth of current scanning location. |
8489eb67 RS |
2964 | level - levelstart equals this except |
2965 | when the depth becomes negative. */ | |
d311d28c | 2966 | EMACS_INT mindepth; /* Lowest DEPTH value seen. */ |
8489eb67 RS |
2967 | int start_quoted = 0; /* Nonzero means starting after a char quote */ |
2968 | Lisp_Object tem; | |
d311d28c PE |
2969 | ptrdiff_t prev_from; /* Keep one character before FROM. */ |
2970 | ptrdiff_t prev_from_byte; | |
1fd3172d | 2971 | int prev_from_syntax; |
195d1361 RS |
2972 | int boundary_stop = commentstop == -1; |
2973 | int nofence; | |
95ff8dfc | 2974 | int found; |
d311d28c | 2975 | ptrdiff_t out_bytepos, out_charpos; |
7c7ca3d2 | 2976 | int temp; |
93da5fff KH |
2977 | |
2978 | prev_from = from; | |
6a140a74 RS |
2979 | prev_from_byte = from_byte; |
2980 | if (from != BEGV) | |
2981 | DEC_BOTH (prev_from, prev_from_byte); | |
93da5fff KH |
2982 | |
2983 | /* Use this macro instead of `from++'. */ | |
6a140a74 RS |
2984 | #define INC_FROM \ |
2985 | do { prev_from = from; \ | |
2986 | prev_from_byte = from_byte; \ | |
327719ee | 2987 | temp = FETCH_CHAR_AS_MULTIBYTE (prev_from_byte); \ |
ab229fdd | 2988 | prev_from_syntax = SYNTAX_WITH_FLAGS (temp); \ |
ef316cf0 | 2989 | INC_BOTH (from, from_byte); \ |
3ceb4629 SM |
2990 | if (from < end) \ |
2991 | UPDATE_SYNTAX_TABLE_FORWARD (from); \ | |
6a140a74 | 2992 | } while (0) |
8489eb67 RS |
2993 | |
2994 | immediate_quit = 1; | |
2995 | QUIT; | |
2996 | ||
265a9e55 | 2997 | if (NILP (oldstate)) |
8489eb67 RS |
2998 | { |
2999 | depth = 0; | |
3000 | state.instring = -1; | |
3001 | state.incomment = 0; | |
195d1361 RS |
3002 | state.comstyle = 0; /* comment style a by default. */ |
3003 | state.comstr_start = -1; /* no comment/string seen. */ | |
8489eb67 RS |
3004 | } |
3005 | else | |
3006 | { | |
3007 | tem = Fcar (oldstate); | |
265a9e55 | 3008 | if (!NILP (tem)) |
8489eb67 RS |
3009 | depth = XINT (tem); |
3010 | else | |
3011 | depth = 0; | |
3012 | ||
3013 | oldstate = Fcdr (oldstate); | |
3014 | oldstate = Fcdr (oldstate); | |
3015 | oldstate = Fcdr (oldstate); | |
3016 | tem = Fcar (oldstate); | |
195d1361 | 3017 | /* Check whether we are inside string_fence-style string: */ |
7d0393cf | 3018 | state.instring = (!NILP (tem) |
d311d28c | 3019 | ? (CHARACTERP (tem) ? XFASTINT (tem) : ST_STRING_STYLE) |
e76f1c44 | 3020 | : -1); |
8489eb67 RS |
3021 | |
3022 | oldstate = Fcdr (oldstate); | |
3023 | tem = Fcar (oldstate); | |
e76f1c44 GM |
3024 | state.incomment = (!NILP (tem) |
3025 | ? (INTEGERP (tem) ? XINT (tem) : -1) | |
95ff8dfc | 3026 | : 0); |
8489eb67 RS |
3027 | |
3028 | oldstate = Fcdr (oldstate); | |
3029 | tem = Fcar (oldstate); | |
265a9e55 | 3030 | start_quoted = !NILP (tem); |
e5d4f4dc | 3031 | |
95ff8dfc | 3032 | /* if the eighth element of the list is nil, we are in comment |
195d1361 RS |
3033 | style a. If it is non-nil, we are in comment style b */ |
3034 | oldstate = Fcdr (oldstate); | |
e5d4f4dc | 3035 | oldstate = Fcdr (oldstate); |
195d1361 | 3036 | tem = Fcar (oldstate); |
c5683ceb SM |
3037 | state.comstyle = (NILP (tem) |
3038 | ? 0 | |
d311d28c PE |
3039 | : (RANGED_INTEGERP (0, tem, ST_COMMENT_STYLE) |
3040 | ? XINT (tem) | |
3041 | : ST_COMMENT_STYLE)); | |
195d1361 | 3042 | |
e5d4f4dc | 3043 | oldstate = Fcdr (oldstate); |
e5d4f4dc | 3044 | tem = Fcar (oldstate); |
d311d28c PE |
3045 | state.comstr_start = |
3046 | RANGED_INTEGERP (PTRDIFF_MIN, tem, PTRDIFF_MAX) ? XINT (tem) : -1; | |
1a102a58 RS |
3047 | oldstate = Fcdr (oldstate); |
3048 | tem = Fcar (oldstate); | |
3049 | while (!NILP (tem)) /* >= second enclosing sexps. */ | |
3050 | { | |
d311d28c PE |
3051 | Lisp_Object temhd = Fcar (tem); |
3052 | if (RANGED_INTEGERP (PTRDIFF_MIN, temhd, PTRDIFF_MAX)) | |
3053 | curlevel->last = XINT (temhd); | |
1a102a58 | 3054 | if (++curlevel == endlevel) |
02010917 | 3055 | curlevel--; /* error ("Nesting too deep for parser"); */ |
1a102a58 RS |
3056 | curlevel->prev = -1; |
3057 | curlevel->last = -1; | |
3058 | tem = Fcdr (tem); | |
3059 | } | |
8489eb67 RS |
3060 | } |
3061 | state.quoted = 0; | |
3062 | mindepth = depth; | |
3063 | ||
3064 | curlevel->prev = -1; | |
3065 | curlevel->last = -1; | |
3066 | ||
326b283b | 3067 | SETUP_SYNTAX_TABLE (prev_from, 1); |
ab229fdd AS |
3068 | temp = FETCH_CHAR (prev_from_byte); |
3069 | prev_from_syntax = SYNTAX_WITH_FLAGS (temp); | |
3ceb4629 | 3070 | UPDATE_SYNTAX_TABLE_FORWARD (from); |
326b283b | 3071 | |
195d1361 | 3072 | /* Enter the loop at a place appropriate for initial state. */ |
8489eb67 | 3073 | |
326b283b RS |
3074 | if (state.incomment) |
3075 | goto startincomment; | |
8489eb67 RS |
3076 | if (state.instring >= 0) |
3077 | { | |
195d1361 | 3078 | nofence = state.instring != ST_STRING_STYLE; |
326b283b RS |
3079 | if (start_quoted) |
3080 | goto startquotedinstring; | |
8489eb67 RS |
3081 | goto startinstring; |
3082 | } | |
326b283b RS |
3083 | else if (start_quoted) |
3084 | goto startquoted; | |
1fd3172d | 3085 | |
8489eb67 RS |
3086 | while (from < end) |
3087 | { | |
c5683ceb | 3088 | int syntax; |
93da5fff | 3089 | INC_FROM; |
1fd3172d | 3090 | code = prev_from_syntax & 0xff; |
4c920633 | 3091 | |
c5d4b96f SM |
3092 | if (from < end |
3093 | && SYNTAX_FLAGS_COMSTART_FIRST (prev_from_syntax) | |
3094 | && (c1 = FETCH_CHAR (from_byte), | |
c5683ceb SM |
3095 | syntax = SYNTAX_WITH_FLAGS (c1), |
3096 | SYNTAX_FLAGS_COMSTART_SECOND (syntax))) | |
c5d4b96f SM |
3097 | /* Duplicate code to avoid a complex if-expression |
3098 | which causes trouble for the SGI compiler. */ | |
95ff8dfc | 3099 | { |
c5d4b96f SM |
3100 | /* Record the comment style we have entered so that only |
3101 | the comment-end sequence of the same style actually | |
3102 | terminates the comment section. */ | |
c5683ceb SM |
3103 | state.comstyle |
3104 | = SYNTAX_FLAGS_COMMENT_STYLE (syntax, prev_from_syntax); | |
c5d4b96f | 3105 | comnested = SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax); |
c5683ceb | 3106 | comnested = comnested || SYNTAX_FLAGS_COMMENT_NESTED (syntax); |
c5d4b96f | 3107 | state.incomment = comnested ? 1 : -1; |
95ff8dfc | 3108 | state.comstr_start = prev_from; |
c5d4b96f SM |
3109 | INC_FROM; |
3110 | code = Scomment; | |
95ff8dfc | 3111 | } |
4c920633 | 3112 | else if (code == Scomment_fence) |
e5d4f4dc RS |
3113 | { |
3114 | /* Record the comment style we have entered so that only | |
3115 | the comment-end sequence of the same style actually | |
3116 | terminates the comment section. */ | |
95ff8dfc RS |
3117 | state.comstyle = ST_COMMENT_STYLE; |
3118 | state.incomment = -1; | |
195d1361 | 3119 | state.comstr_start = prev_from; |
e5d4f4dc | 3120 | code = Scomment; |
e5d4f4dc | 3121 | } |
c5d4b96f SM |
3122 | else if (code == Scomment) |
3123 | { | |
c5683ceb | 3124 | state.comstyle = SYNTAX_FLAGS_COMMENT_STYLE (prev_from_syntax, 0); |
c5d4b96f SM |
3125 | state.incomment = (SYNTAX_FLAGS_COMMENT_NESTED (prev_from_syntax) ? |
3126 | 1 : -1); | |
3127 | state.comstr_start = prev_from; | |
3128 | } | |
e5d4f4dc | 3129 | |
1fd3172d | 3130 | if (SYNTAX_FLAGS_PREFIX (prev_from_syntax)) |
8489eb67 | 3131 | continue; |
7393bcbb | 3132 | switch (code) |
8489eb67 RS |
3133 | { |
3134 | case Sescape: | |
3135 | case Scharquote: | |
3136 | if (stopbefore) goto stop; /* this arg means stop at sexp start */ | |
93da5fff | 3137 | curlevel->last = prev_from; |
8489eb67 RS |
3138 | startquoted: |
3139 | if (from == end) goto endquoted; | |
93da5fff | 3140 | INC_FROM; |
8489eb67 RS |
3141 | goto symstarted; |
3142 | /* treat following character as a word constituent */ | |
3143 | case Sword: | |
3144 | case Ssymbol: | |
3145 | if (stopbefore) goto stop; /* this arg means stop at sexp start */ | |
93da5fff | 3146 | curlevel->last = prev_from; |
8489eb67 RS |
3147 | symstarted: |
3148 | while (from < end) | |
3149 | { | |
7c7ca3d2 | 3150 | /* Some compilers can't handle this inside the switch. */ |
327719ee | 3151 | temp = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
ab229fdd | 3152 | temp = SYNTAX (temp); |
7c7ca3d2 | 3153 | switch (temp) |
8489eb67 RS |
3154 | { |
3155 | case Scharquote: | |
3156 | case Sescape: | |
93da5fff | 3157 | INC_FROM; |
8489eb67 RS |
3158 | if (from == end) goto endquoted; |
3159 | break; | |
3160 | case Sword: | |
3161 | case Ssymbol: | |
3162 | case Squote: | |
3163 | break; | |
3164 | default: | |
3165 | goto symdone; | |
3166 | } | |
93da5fff | 3167 | INC_FROM; |
8489eb67 RS |
3168 | } |
3169 | symdone: | |
3170 | curlevel->prev = curlevel->last; | |
3171 | break; | |
3172 | ||
240c43e8 | 3173 | case Scomment_fence: /* Can't happen because it's handled above. */ |
8489eb67 | 3174 | case Scomment: |
195d1361 | 3175 | if (commentstop || boundary_stop) goto done; |
d355bd8a SM |
3176 | startincomment: |
3177 | /* The (from == BEGV) test was to enter the loop in the middle so | |
95ff8dfc | 3178 | that we find a 2-char comment ender even if we start in the |
e6365855 | 3179 | middle of it. We don't want to do that if we're just at the |
d355bd8a | 3180 | beginning of the comment (think of (*) ... (*)). */ |
95ff8dfc RS |
3181 | found = forw_comment (from, from_byte, end, |
3182 | state.incomment, state.comstyle, | |
e6365855 SM |
3183 | (from == BEGV || from < state.comstr_start + 3) |
3184 | ? 0 : prev_from_syntax, | |
95ff8dfc RS |
3185 | &out_charpos, &out_bytepos, &state.incomment); |
3186 | from = out_charpos; from_byte = out_bytepos; | |
3187 | /* Beware! prev_from and friends are invalid now. | |
3188 | Luckily, the `done' doesn't use them and the INC_FROM | |
3189 | sets them to a sane value without looking at them. */ | |
3190 | if (!found) goto done; | |
d355bd8a | 3191 | INC_FROM; |
8489eb67 | 3192 | state.incomment = 0; |
e5d4f4dc | 3193 | state.comstyle = 0; /* reset the comment style */ |
195d1361 | 3194 | if (boundary_stop) goto done; |
8489eb67 RS |
3195 | break; |
3196 | ||
3197 | case Sopen: | |
3198 | if (stopbefore) goto stop; /* this arg means stop at sexp start */ | |
3199 | depth++; | |
3200 | /* curlevel++->last ran into compiler bug on Apollo */ | |
93da5fff | 3201 | curlevel->last = prev_from; |
8489eb67 | 3202 | if (++curlevel == endlevel) |
02010917 | 3203 | curlevel--; /* error ("Nesting too deep for parser"); */ |
8489eb67 RS |
3204 | curlevel->prev = -1; |
3205 | curlevel->last = -1; | |
30844415 | 3206 | if (targetdepth == depth) goto done; |
8489eb67 RS |
3207 | break; |
3208 | ||
3209 | case Sclose: | |
3210 | depth--; | |
3211 | if (depth < mindepth) | |
3212 | mindepth = depth; | |
3213 | if (curlevel != levelstart) | |
3214 | curlevel--; | |
3215 | curlevel->prev = curlevel->last; | |
30844415 | 3216 | if (targetdepth == depth) goto done; |
8489eb67 RS |
3217 | break; |
3218 | ||
3219 | case Sstring: | |
195d1361 RS |
3220 | case Sstring_fence: |
3221 | state.comstr_start = from - 1; | |
8489eb67 | 3222 | if (stopbefore) goto stop; /* this arg means stop at sexp start */ |
93da5fff | 3223 | curlevel->last = prev_from; |
7d0393cf | 3224 | state.instring = (code == Sstring |
b7dbcc19 | 3225 | ? (FETCH_CHAR_AS_MULTIBYTE (prev_from_byte)) |
195d1361 RS |
3226 | : ST_STRING_STYLE); |
3227 | if (boundary_stop) goto done; | |
8489eb67 | 3228 | startinstring: |
195d1361 | 3229 | { |
644ea4df | 3230 | nofence = state.instring != ST_STRING_STYLE; |
7d0393cf | 3231 | |
644ea4df RS |
3232 | while (1) |
3233 | { | |
3234 | int c; | |
195d1361 | 3235 | |
644ea4df | 3236 | if (from >= end) goto done; |
b7dbcc19 | 3237 | c = FETCH_CHAR_AS_MULTIBYTE (from_byte); |
7c7ca3d2 RS |
3238 | /* Some compilers can't handle this inside the switch. */ |
3239 | temp = SYNTAX (c); | |
ccf89641 KH |
3240 | |
3241 | /* Check TEMP here so that if the char has | |
3242 | a syntax-table property which says it is NOT | |
3243 | a string character, it does not end the string. */ | |
3244 | if (nofence && c == state.instring && temp == Sstring) | |
3245 | break; | |
3246 | ||
7c7ca3d2 | 3247 | switch (temp) |
644ea4df RS |
3248 | { |
3249 | case Sstring_fence: | |
3250 | if (!nofence) goto string_end; | |
3251 | break; | |
3252 | case Scharquote: | |
3253 | case Sescape: | |
3254 | INC_FROM; | |
3255 | startquotedinstring: | |
3256 | if (from >= end) goto endquoted; | |
195d1361 | 3257 | } |
644ea4df RS |
3258 | INC_FROM; |
3259 | } | |
195d1361 RS |
3260 | } |
3261 | string_end: | |
8489eb67 RS |
3262 | state.instring = -1; |
3263 | curlevel->prev = curlevel->last; | |
93da5fff | 3264 | INC_FROM; |
195d1361 | 3265 | if (boundary_stop) goto done; |
8489eb67 RS |
3266 | break; |
3267 | ||
3268 | case Smath: | |
240c43e8 SM |
3269 | /* FIXME: We should do something with it. */ |
3270 | break; | |
3271 | default: | |
3272 | /* Ignore whitespace, punctuation, quote, endcomment. */ | |
8489eb67 RS |
3273 | break; |
3274 | } | |
3275 | } | |
3276 | goto done; | |
3277 | ||
3278 | stop: /* Here if stopping before start of sexp. */ | |
93da5fff | 3279 | from = prev_from; /* We have just fetched the char that starts it; */ |
8489eb67 RS |
3280 | goto done; /* but return the position before it. */ |
3281 | ||
3282 | endquoted: | |
3283 | state.quoted = 1; | |
3284 | done: | |
3285 | state.depth = depth; | |
3286 | state.mindepth = mindepth; | |
3287 | state.thislevelstart = curlevel->prev; | |
3288 | state.prevlevelstart | |
3289 | = (curlevel == levelstart) ? -1 : (curlevel - 1)->last; | |
3290 | state.location = from; | |
1a102a58 | 3291 | state.levelstarts = Qnil; |
eb4d412d PE |
3292 | while (curlevel > levelstart) |
3293 | state.levelstarts = Fcons (make_number ((--curlevel)->last), | |
3294 | state.levelstarts); | |
8489eb67 RS |
3295 | immediate_quit = 0; |
3296 | ||
e5d4f4dc | 3297 | *stateptr = state; |
8489eb67 RS |
3298 | } |
3299 | ||
c81a3712 | 3300 | DEFUN ("parse-partial-sexp", Fparse_partial_sexp, Sparse_partial_sexp, 2, 6, 0, |
fdb82f93 PJ |
3301 | doc: /* Parse Lisp syntax starting at FROM until TO; return status of parse at TO. |
3302 | Parsing stops at TO or when certain criteria are met; | |
3303 | point is set to where parsing stops. | |
3304 | If fifth arg OLDSTATE is omitted or nil, | |
3305 | parsing assumes that FROM is the beginning of a function. | |
0e6228bc | 3306 | Value is a list of elements describing final state of parsing: |
fdb82f93 PJ |
3307 | 0. depth in parens. |
3308 | 1. character address of start of innermost containing list; nil if none. | |
3309 | 2. character address of start of last complete sexp terminated. | |
3310 | 3. non-nil if inside a string. | |
3311 | (it is the character that will terminate the string, | |
3312 | or t if the string should be terminated by a generic string delimiter.) | |
7d0393cf | 3313 | 4. nil if outside a comment, t if inside a non-nestable comment, |
fdb82f93 PJ |
3314 | else an integer (the current comment nesting). |
3315 | 5. t if following a quote character. | |
3316 | 6. the minimum paren-depth encountered during this scan. | |
c5683ceb | 3317 | 7. style of comment, if any. |
fdb82f93 PJ |
3318 | 8. character address of start of comment or string; nil if not in one. |
3319 | 9. Intermediate data for continuation of parsing (subject to change). | |
3320 | If third arg TARGETDEPTH is non-nil, parsing stops if the depth | |
3321 | in parentheses becomes equal to TARGETDEPTH. | |
3322 | Fourth arg STOPBEFORE non-nil means stop when come to | |
3323 | any character that starts a sexp. | |
0e6228bc | 3324 | Fifth arg OLDSTATE is a list like what this function returns. |
fdb82f93 | 3325 | It is used to initialize the state of the parse. Elements number 1, 2, 6 |
d311d28c | 3326 | are ignored. |
fdb82f93 PJ |
3327 | Sixth arg COMMENTSTOP non-nil means stop at the start of a comment. |
3328 | If it is symbol `syntax-table', stop after the start of a comment or a | |
3329 | string, or after end of a comment or a string. */) | |
5842a27b | 3330 | (Lisp_Object from, Lisp_Object to, Lisp_Object targetdepth, Lisp_Object stopbefore, Lisp_Object oldstate, Lisp_Object commentstop) |
8489eb67 RS |
3331 | { |
3332 | struct lisp_parse_state state; | |
d311d28c | 3333 | EMACS_INT target; |
8489eb67 | 3334 | |
265a9e55 | 3335 | if (!NILP (targetdepth)) |
8489eb67 | 3336 | { |
b7826503 | 3337 | CHECK_NUMBER (targetdepth); |
8489eb67 RS |
3338 | target = XINT (targetdepth); |
3339 | } | |
3340 | else | |
d311d28c | 3341 | target = TYPE_MINIMUM (EMACS_INT); /* We won't reach this depth */ |
8489eb67 RS |
3342 | |
3343 | validate_region (&from, &to); | |
6a140a74 RS |
3344 | scan_sexps_forward (&state, XINT (from), CHAR_TO_BYTE (XINT (from)), |
3345 | XINT (to), | |
c81a3712 | 3346 | target, !NILP (stopbefore), oldstate, |
7d0393cf | 3347 | (NILP (commentstop) |
195d1361 | 3348 | ? 0 : (EQ (commentstop, Qsyntax_table) ? -1 : 1))); |
8489eb67 RS |
3349 | |
3350 | SET_PT (state.location); | |
7d0393cf | 3351 | |
8489eb67 | 3352 | return Fcons (make_number (state.depth), |
c5683ceb SM |
3353 | Fcons (state.prevlevelstart < 0 |
3354 | ? Qnil : make_number (state.prevlevelstart), | |
3355 | Fcons (state.thislevelstart < 0 | |
3356 | ? Qnil : make_number (state.thislevelstart), | |
7d0393cf JB |
3357 | Fcons (state.instring >= 0 |
3358 | ? (state.instring == ST_STRING_STYLE | |
195d1361 | 3359 | ? Qt : make_number (state.instring)) : Qnil, |
95ff8dfc RS |
3360 | Fcons (state.incomment < 0 ? Qt : |
3361 | (state.incomment == 0 ? Qnil : | |
3362 | make_number (state.incomment)), | |
8489eb67 | 3363 | Fcons (state.quoted ? Qt : Qnil, |
bff37d2a | 3364 | Fcons (make_number (state.mindepth), |
7d0393cf | 3365 | Fcons ((state.comstyle |
bff37d2a | 3366 | ? (state.comstyle == ST_COMMENT_STYLE |
c5683ceb SM |
3367 | ? Qsyntax_table |
3368 | : make_number (state.comstyle)) | |
3369 | : Qnil), | |
0beaf54f DL |
3370 | Fcons (((state.incomment |
3371 | || (state.instring >= 0)) | |
bff37d2a RS |
3372 | ? make_number (state.comstr_start) |
3373 | : Qnil), | |
1a102a58 | 3374 | Fcons (state.levelstarts, Qnil)))))))))); |
8489eb67 RS |
3375 | } |
3376 | \f | |
dfcf069d | 3377 | void |
971de7fb | 3378 | init_syntax_once (void) |
8489eb67 | 3379 | { |
78f9a1f7 | 3380 | register int i, c; |
8ea151b2 | 3381 | Lisp_Object temp; |
8489eb67 | 3382 | |
5ebaddf5 | 3383 | /* This has to be done here, before we call Fmake_char_table. */ |
cd3520a4 | 3384 | DEFSYM (Qsyntax_table, "syntax-table"); |
5ebaddf5 | 3385 | |
d67b4f80 | 3386 | /* Intern_C_String this now in case it isn't already done. |
5ebaddf5 RS |
3387 | Setting this variable twice is harmless. |
3388 | But don't staticpro it here--that is done in alloc.c. */ | |
d67b4f80 | 3389 | Qchar_table_extra_slots = intern_c_string ("char-table-extra-slots"); |
5ebaddf5 | 3390 | |
93da5fff | 3391 | /* Create objects which can be shared among syntax tables. */ |
42ebfa31 | 3392 | Vsyntax_code_object = Fmake_vector (make_number (Smax), Qnil); |
77b37c05 | 3393 | for (i = 0; i < ASIZE (Vsyntax_code_object); i++) |
28be1ada | 3394 | ASET (Vsyntax_code_object, i, Fcons (make_number (i), Qnil)); |
93da5fff | 3395 | |
5ebaddf5 RS |
3396 | /* Now we are ready to set up this property, so we can |
3397 | create syntax tables. */ | |
3398 | Fput (Qsyntax_table, Qchar_table_extra_slots, make_number (0)); | |
3399 | ||
28be1ada | 3400 | temp = AREF (Vsyntax_code_object, (int) Swhitespace); |
8489eb67 | 3401 | |
5ebaddf5 | 3402 | Vstandard_syntax_table = Fmake_char_table (Qsyntax_table, temp); |
8489eb67 | 3403 | |
aa7b08b4 | 3404 | /* Control characters should not be whitespace. */ |
28be1ada | 3405 | temp = AREF (Vsyntax_code_object, (int) Spunct); |
aa7b08b4 RS |
3406 | for (i = 0; i <= ' ' - 1; i++) |
3407 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp); | |
3408 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, 0177, temp); | |
3409 | ||
3410 | /* Except that a few really are whitespace. */ | |
28be1ada | 3411 | temp = AREF (Vsyntax_code_object, (int) Swhitespace); |
aa7b08b4 RS |
3412 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, ' ', temp); |
3413 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '\t', temp); | |
3414 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '\n', temp); | |
3415 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, 015, temp); | |
3416 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, 014, temp); | |
3417 | ||
28be1ada | 3418 | temp = AREF (Vsyntax_code_object, (int) Sword); |
8489eb67 | 3419 | for (i = 'a'; i <= 'z'; i++) |
8ea151b2 | 3420 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp); |
8489eb67 | 3421 | for (i = 'A'; i <= 'Z'; i++) |
8ea151b2 | 3422 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp); |
8489eb67 | 3423 | for (i = '0'; i <= '9'; i++) |
8ea151b2 RS |
3424 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, i, temp); |
3425 | ||
3426 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '$', temp); | |
3427 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '%', temp); | |
3428 | ||
3429 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '(', | |
3430 | Fcons (make_number (Sopen), make_number (')'))); | |
3431 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, ')', | |
3432 | Fcons (make_number (Sclose), make_number ('('))); | |
3433 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '[', | |
3434 | Fcons (make_number (Sopen), make_number (']'))); | |
3435 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, ']', | |
3436 | Fcons (make_number (Sclose), make_number ('['))); | |
3437 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '{', | |
3438 | Fcons (make_number (Sopen), make_number ('}'))); | |
3439 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '}', | |
3440 | Fcons (make_number (Sclose), make_number ('{'))); | |
3441 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '"', | |
3442 | Fcons (make_number ((int) Sstring), Qnil)); | |
3443 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, '\\', | |
3444 | Fcons (make_number ((int) Sescape), Qnil)); | |
3445 | ||
28be1ada | 3446 | temp = AREF (Vsyntax_code_object, (int) Ssymbol); |
8489eb67 | 3447 | for (i = 0; i < 10; i++) |
78f9a1f7 KH |
3448 | { |
3449 | c = "_-+*/&|<>="[i]; | |
3450 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, c, temp); | |
3451 | } | |
8489eb67 | 3452 | |
28be1ada | 3453 | temp = AREF (Vsyntax_code_object, (int) Spunct); |
8489eb67 | 3454 | for (i = 0; i < 12; i++) |
78f9a1f7 KH |
3455 | { |
3456 | c = ".,;:?!#@~^'`"[i]; | |
3457 | SET_RAW_SYNTAX_ENTRY (Vstandard_syntax_table, c, temp); | |
3458 | } | |
bd25db08 KH |
3459 | |
3460 | /* All multibyte characters have syntax `word' by default. */ | |
28be1ada | 3461 | temp = AREF (Vsyntax_code_object, (int) Sword); |
5c7b02ab | 3462 | char_table_set_range (Vstandard_syntax_table, 0x80, MAX_CHAR, temp); |
8489eb67 RS |
3463 | } |
3464 | ||
dfcf069d | 3465 | void |
971de7fb | 3466 | syms_of_syntax (void) |
8489eb67 | 3467 | { |
cd3520a4 | 3468 | DEFSYM (Qsyntax_table_p, "syntax-table-p"); |
8489eb67 | 3469 | |
93da5fff KH |
3470 | staticpro (&Vsyntax_code_object); |
3471 | ||
a2994c46 KS |
3472 | staticpro (&gl_state.object); |
3473 | staticpro (&gl_state.global_code); | |
3474 | staticpro (&gl_state.current_syntax_table); | |
3475 | staticpro (&gl_state.old_prop); | |
3476 | ||
3477 | /* Defined in regex.c */ | |
3478 | staticpro (&re_match_object); | |
3479 | ||
cd3520a4 | 3480 | DEFSYM (Qscan_error, "scan-error"); |
7bf5e9e4 | 3481 | Fput (Qscan_error, Qerror_conditions, |
3438fe21 | 3482 | listn (CONSTYPE_PURE, 2, Qscan_error, Qerror)); |
7bf5e9e4 | 3483 | Fput (Qscan_error, Qerror_message, |
2a0213a6 | 3484 | build_pure_c_string ("Scan error")); |
7bf5e9e4 | 3485 | |
29208e82 | 3486 | DEFVAR_BOOL ("parse-sexp-ignore-comments", parse_sexp_ignore_comments, |
fdb82f93 | 3487 | doc: /* Non-nil means `forward-sexp', etc., should treat comments as whitespace. */); |
8489eb67 | 3488 | |
29208e82 | 3489 | DEFVAR_BOOL ("parse-sexp-lookup-properties", parse_sexp_lookup_properties, |
fdb82f93 PJ |
3490 | doc: /* Non-nil means `forward-sexp', etc., obey `syntax-table' property. |
3491 | Otherwise, that text property is simply ignored. | |
3492 | See the info node `(elisp)Syntax Properties' for a description of the | |
3493 | `syntax-table' property. */); | |
195d1361 | 3494 | |
8489eb67 | 3495 | words_include_escapes = 0; |
29208e82 | 3496 | DEFVAR_BOOL ("words-include-escapes", words_include_escapes, |
fdb82f93 | 3497 | doc: /* Non-nil means `forward-word', etc., should treat escape chars part of words. */); |
8489eb67 | 3498 | |
29208e82 | 3499 | DEFVAR_BOOL ("multibyte-syntax-as-symbol", multibyte_syntax_as_symbol, |
fdb82f93 | 3500 | doc: /* Non-nil means `scan-sexps' treats all multibyte characters as symbol. */); |
bd25db08 KH |
3501 | multibyte_syntax_as_symbol = 0; |
3502 | ||
f4ed767f | 3503 | DEFVAR_BOOL ("open-paren-in-column-0-is-defun-start", |
29208e82 | 3504 | open_paren_in_column_0_is_defun_start, |
fb7ada5f | 3505 | doc: /* Non-nil means an open paren in column 0 denotes the start of a defun. */); |
f4ed767f GM |
3506 | open_paren_in_column_0_is_defun_start = 1; |
3507 | ||
8f924df7 KH |
3508 | |
3509 | DEFVAR_LISP ("find-word-boundary-function-table", | |
29208e82 | 3510 | Vfind_word_boundary_function_table, |
869bb237 | 3511 | doc: /* |
8f924df7 | 3512 | Char table of functions to search for the word boundary. |
869bb237 KH |
3513 | Each function is called with two arguments; POS and LIMIT. |
3514 | POS and LIMIT are character positions in the current buffer. | |
3515 | ||
3516 | If POS is less than LIMIT, POS is at the first character of a word, | |
3517 | and the return value of a function is a position after the last | |
3518 | character of that word. | |
3519 | ||
3520 | If POS is not less than LIMIT, POS is at the last character of a word, | |
3521 | and the return value of a function is a position at the first | |
3522 | character of that word. | |
3523 | ||
3524 | In both cases, LIMIT bounds the search. */); | |
8f924df7 | 3525 | Vfind_word_boundary_function_table = Fmake_char_table (Qnil, Qnil); |
869bb237 | 3526 | |
8489eb67 RS |
3527 | defsubr (&Ssyntax_table_p); |
3528 | defsubr (&Ssyntax_table); | |
3529 | defsubr (&Sstandard_syntax_table); | |
3530 | defsubr (&Scopy_syntax_table); | |
3531 | defsubr (&Sset_syntax_table); | |
3532 | defsubr (&Schar_syntax); | |
beefa22e | 3533 | defsubr (&Smatching_paren); |
c65adb44 | 3534 | defsubr (&Sstring_to_syntax); |
8489eb67 | 3535 | defsubr (&Smodify_syntax_entry); |
62abe9cb | 3536 | defsubr (&Sinternal_describe_syntax_value); |
8489eb67 RS |
3537 | |
3538 | defsubr (&Sforward_word); | |
3539 | ||
195d1361 RS |
3540 | defsubr (&Sskip_chars_forward); |
3541 | defsubr (&Sskip_chars_backward); | |
3542 | defsubr (&Sskip_syntax_forward); | |
3543 | defsubr (&Sskip_syntax_backward); | |
3544 | ||
b3cfe0c8 | 3545 | defsubr (&Sforward_comment); |
8489eb67 RS |
3546 | defsubr (&Sscan_lists); |
3547 | defsubr (&Sscan_sexps); | |
3548 | defsubr (&Sbackward_prefix_chars); | |
3549 | defsubr (&Sparse_partial_sexp); | |
3550 | } |