Commit | Line | Data |
---|---|---|
b7b65b15 | 1 | /* Low-level bidirectional buffer-scanning functions for GNU Emacs. |
73b0cd50 | 2 | Copyright (C) 2000-2001, 2004-2005, 2009-2011 |
b118e65f | 3 | Free Software Foundation, Inc. |
b7b65b15 EZ |
4 | |
5 | This file is part of GNU Emacs. | |
6 | ||
a8d11bd3 | 7 | GNU Emacs is free software: you can redistribute it and/or modify |
b7b65b15 | 8 | it under the terms of the GNU General Public License as published by |
a8d11bd3 EZ |
9 | the Free Software Foundation, either version 3 of the License, or |
10 | (at your option) any later version. | |
b7b65b15 EZ |
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 | ||
b7b65b15 | 17 | You should have received a copy of the GNU General Public License |
a8d11bd3 | 18 | along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */ |
b7b65b15 | 19 | |
2d6e4628 EZ |
20 | /* Written by Eli Zaretskii <eliz@gnu.org>. |
21 | ||
22 | A sequential implementation of the Unicode Bidirectional algorithm, | |
b7b65b15 EZ |
23 | as per UAX#9, a part of the Unicode Standard. |
24 | ||
25 | Unlike the reference and most other implementations, this one is | |
940afb59 EZ |
26 | designed to be called once for every character in the buffer or |
27 | string. | |
b7b65b15 | 28 | |
4b292a22 | 29 | The main entry point is bidi_move_to_visually_next. Each time it |
b7b65b15 EZ |
30 | is called, it finds the next character in the visual order, and |
31 | returns its information in a special structure. The caller is then | |
32 | expected to process this character for display or any other | |
4b292a22 EZ |
33 | purposes, and call bidi_move_to_visually_next for the next |
34 | character. See the comments in bidi_move_to_visually_next for more | |
35 | details about its algorithm that finds the next visual-order | |
b7b65b15 EZ |
36 | character by resolving their levels on the fly. |
37 | ||
940afb59 EZ |
38 | The two other entry points are bidi_paragraph_init and |
39 | bidi_mirror_char. The first determines the base direction of a | |
40 | paragraph, while the second returns the mirrored version of its | |
41 | argument character. | |
42 | ||
89d3374a EZ |
43 | If you want to understand the code, you will have to read it |
44 | together with the relevant portions of UAX#9. The comments include | |
45 | references to UAX#9 rules, for that very reason. | |
46 | ||
b7b65b15 EZ |
47 | A note about references to UAX#9 rules: if the reference says |
48 | something like "X9/Retaining", it means that you need to refer to | |
49 | rule X9 and to its modifications decribed in the "Implementation | |
50 | Notes" section of UAX#9, under "Retaining Format Codes". */ | |
51 | ||
b7b65b15 | 52 | #include <config.h> |
b7b65b15 | 53 | #include <stdio.h> |
29e3d8d1 EZ |
54 | #include <setjmp.h> |
55 | ||
b7b65b15 EZ |
56 | #include "lisp.h" |
57 | #include "buffer.h" | |
58 | #include "character.h" | |
59 | #include "dispextern.h" | |
60 | ||
61 | static int bidi_initialized = 0; | |
62 | ||
cbc4fd20 | 63 | static Lisp_Object bidi_type_table, bidi_mirror_table; |
b7b65b15 EZ |
64 | |
65 | #define LRM_CHAR 0x200E | |
66 | #define RLM_CHAR 0x200F | |
b7b65b15 | 67 | #define BIDI_EOB -1 |
b7b65b15 | 68 | |
b7b65b15 EZ |
69 | /* Data type for describing the bidirectional character categories. */ |
70 | typedef enum { | |
71 | UNKNOWN_BC, | |
72 | NEUTRAL, | |
73 | WEAK, | |
74 | STRONG | |
75 | } bidi_category_t; | |
76 | ||
e78aecca | 77 | extern int bidi_ignore_explicit_marks_for_paragraph_level EXTERNALLY_VISIBLE; |
b7b65b15 EZ |
78 | int bidi_ignore_explicit_marks_for_paragraph_level = 1; |
79 | ||
372b7a95 | 80 | static Lisp_Object paragraph_start_re, paragraph_separate_re; |
6bff6497 | 81 | static Lisp_Object Qparagraph_start, Qparagraph_separate; |
b7b65b15 | 82 | |
b7b65b15 | 83 | static void |
971de7fb | 84 | bidi_initialize (void) |
b7b65b15 | 85 | { |
317fbf33 EZ |
86 | |
87 | #include "biditype.h" | |
cbc4fd20 | 88 | #include "bidimirror.h" |
317fbf33 | 89 | |
b7b65b15 EZ |
90 | int i; |
91 | ||
92 | bidi_type_table = Fmake_char_table (Qnil, make_number (STRONG_L)); | |
2d6e4628 | 93 | staticpro (&bidi_type_table); |
b7b65b15 EZ |
94 | |
95 | for (i = 0; i < sizeof bidi_type / sizeof bidi_type[0]; i++) | |
317fbf33 | 96 | char_table_set_range (bidi_type_table, bidi_type[i].from, bidi_type[i].to, |
b7b65b15 | 97 | make_number (bidi_type[i].type)); |
6bff6497 | 98 | |
cbc4fd20 EZ |
99 | bidi_mirror_table = Fmake_char_table (Qnil, Qnil); |
100 | staticpro (&bidi_mirror_table); | |
101 | ||
102 | for (i = 0; i < sizeof bidi_mirror / sizeof bidi_mirror[0]; i++) | |
103 | char_table_set (bidi_mirror_table, bidi_mirror[i].from, | |
104 | make_number (bidi_mirror[i].to)); | |
105 | ||
b4bf28b7 SM |
106 | Qparagraph_start = intern ("paragraph-start"); |
107 | staticpro (&Qparagraph_start); | |
372b7a95 EZ |
108 | paragraph_start_re = Fsymbol_value (Qparagraph_start); |
109 | if (!STRINGP (paragraph_start_re)) | |
110 | paragraph_start_re = build_string ("\f\\|[ \t]*$"); | |
111 | staticpro (¶graph_start_re); | |
b4bf28b7 SM |
112 | Qparagraph_separate = intern ("paragraph-separate"); |
113 | staticpro (&Qparagraph_separate); | |
372b7a95 EZ |
114 | paragraph_separate_re = Fsymbol_value (Qparagraph_separate); |
115 | if (!STRINGP (paragraph_separate_re)) | |
116 | paragraph_separate_re = build_string ("[ \t\f]*$"); | |
117 | staticpro (¶graph_separate_re); | |
b7b65b15 EZ |
118 | bidi_initialized = 1; |
119 | } | |
120 | ||
6bff6497 EZ |
121 | /* Return the bidi type of a character CH, subject to the current |
122 | directional OVERRIDE. */ | |
fd3998ff | 123 | static INLINE bidi_type_t |
6bff6497 | 124 | bidi_get_type (int ch, bidi_dir_t override) |
b7b65b15 | 125 | { |
6bff6497 EZ |
126 | bidi_type_t default_type; |
127 | ||
e7402cb2 EZ |
128 | if (ch == BIDI_EOB) |
129 | return NEUTRAL_B; | |
e342a24d EZ |
130 | if (ch < 0 || ch > MAX_CHAR) |
131 | abort (); | |
6bff6497 EZ |
132 | |
133 | default_type = (bidi_type_t) XINT (CHAR_TABLE_REF (bidi_type_table, ch)); | |
134 | ||
135 | if (override == NEUTRAL_DIR) | |
136 | return default_type; | |
137 | ||
138 | switch (default_type) | |
139 | { | |
140 | /* Although UAX#9 does not tell, it doesn't make sense to | |
141 | override NEUTRAL_B and LRM/RLM characters. */ | |
142 | case NEUTRAL_B: | |
143 | case LRE: | |
144 | case LRO: | |
145 | case RLE: | |
146 | case RLO: | |
147 | case PDF: | |
148 | return default_type; | |
149 | default: | |
150 | switch (ch) | |
151 | { | |
152 | case LRM_CHAR: | |
153 | case RLM_CHAR: | |
154 | return default_type; | |
155 | default: | |
156 | if (override == L2R) /* X6 */ | |
157 | return STRONG_L; | |
158 | else if (override == R2L) | |
159 | return STRONG_R; | |
160 | else | |
161 | abort (); /* can't happen: handled above */ | |
162 | } | |
163 | } | |
b7b65b15 EZ |
164 | } |
165 | ||
7d3b3862 | 166 | static void |
2d6e4628 EZ |
167 | bidi_check_type (bidi_type_t type) |
168 | { | |
169 | if (type < UNKNOWN_BT || type > NEUTRAL_ON) | |
170 | abort (); | |
171 | } | |
172 | ||
b7b65b15 | 173 | /* Given a bidi TYPE of a character, return its category. */ |
fd3998ff | 174 | static INLINE bidi_category_t |
b7b65b15 EZ |
175 | bidi_get_category (bidi_type_t type) |
176 | { | |
177 | switch (type) | |
178 | { | |
179 | case UNKNOWN_BT: | |
180 | return UNKNOWN_BC; | |
181 | case STRONG_L: | |
182 | case STRONG_R: | |
183 | case STRONG_AL: | |
184 | case LRE: | |
185 | case LRO: | |
186 | case RLE: | |
187 | case RLO: | |
188 | return STRONG; | |
189 | case PDF: /* ??? really?? */ | |
190 | case WEAK_EN: | |
191 | case WEAK_ES: | |
192 | case WEAK_ET: | |
193 | case WEAK_AN: | |
194 | case WEAK_CS: | |
195 | case WEAK_NSM: | |
196 | case WEAK_BN: | |
197 | return WEAK; | |
198 | case NEUTRAL_B: | |
199 | case NEUTRAL_S: | |
200 | case NEUTRAL_WS: | |
201 | case NEUTRAL_ON: | |
202 | return NEUTRAL; | |
203 | default: | |
204 | abort (); | |
205 | } | |
206 | } | |
207 | ||
a6e8d97c EZ |
208 | /* Return the mirrored character of C, if it has one. If C has no |
209 | mirrored counterpart, return C. | |
210 | Note: The conditions in UAX#9 clause L4 regarding the surrounding | |
211 | context must be tested by the caller. */ | |
b7b65b15 EZ |
212 | int |
213 | bidi_mirror_char (int c) | |
214 | { | |
cbc4fd20 EZ |
215 | Lisp_Object val; |
216 | ||
217 | if (c == BIDI_EOB) | |
218 | return c; | |
219 | if (c < 0 || c > MAX_CHAR) | |
220 | abort (); | |
b7b65b15 | 221 | |
cbc4fd20 EZ |
222 | val = CHAR_TABLE_REF (bidi_mirror_table, c); |
223 | if (INTEGERP (val)) | |
b7b65b15 | 224 | { |
cbc4fd20 | 225 | int v = XINT (val); |
b7b65b15 | 226 | |
cbc4fd20 EZ |
227 | if (v < 0 || v > MAX_CHAR) |
228 | abort (); | |
229 | ||
230 | return v; | |
b7b65b15 | 231 | } |
cbc4fd20 | 232 | |
b7b65b15 EZ |
233 | return c; |
234 | } | |
235 | ||
236 | /* Copy the bidi iterator from FROM to TO. To save cycles, this only | |
237 | copies the part of the level stack that is actually in use. */ | |
fd3998ff | 238 | static INLINE void |
b7b65b15 EZ |
239 | bidi_copy_it (struct bidi_it *to, struct bidi_it *from) |
240 | { | |
241 | int i; | |
242 | ||
e69a9370 | 243 | /* Copy everything except the level stack and beyond. */ |
182ce2d2 | 244 | memcpy (to, from, offsetof (struct bidi_it, level_stack[0])); |
b7b65b15 EZ |
245 | |
246 | /* Copy the active part of the level stack. */ | |
247 | to->level_stack[0] = from->level_stack[0]; /* level zero is always in use */ | |
248 | for (i = 1; i <= from->stack_idx; i++) | |
249 | to->level_stack[i] = from->level_stack[i]; | |
250 | } | |
251 | ||
252 | /* Caching the bidi iterator states. */ | |
253 | ||
2fe72643 EZ |
254 | #define BIDI_CACHE_CHUNK 200 |
255 | static struct bidi_it *bidi_cache; | |
256 | static size_t bidi_cache_size = 0; | |
0416466c | 257 | static size_t elsz = sizeof (struct bidi_it); |
87e67904 EZ |
258 | static EMACS_INT bidi_cache_idx; /* next unused cache slot */ |
259 | static EMACS_INT bidi_cache_last_idx; /* slot of last cache hit */ | |
260 | static EMACS_INT bidi_cache_start = 0; /* start of cache for this | |
261 | "stack" level */ | |
262 | ||
263 | /* Reset the cache state to the empty state. We only reset the part | |
264 | of the cache relevant to iteration of the current object. Previous | |
265 | objects, which are pushed on the display iterator's stack, are left | |
266 | intact. This is called when the cached information is no more | |
267 | useful for the current iteration, e.g. when we were reseated to a | |
268 | new position on the same object. */ | |
fd3998ff | 269 | static INLINE void |
b7b65b15 EZ |
270 | bidi_cache_reset (void) |
271 | { | |
87e67904 | 272 | bidi_cache_idx = bidi_cache_start; |
b7b65b15 EZ |
273 | bidi_cache_last_idx = -1; |
274 | } | |
275 | ||
87e67904 EZ |
276 | /* Shrink the cache to its minimal size. Called when we init the bidi |
277 | iterator for reordering a buffer or a string that does not come | |
278 | from display properties, because that means all the previously | |
279 | cached info is of no further use. */ | |
2fe72643 EZ |
280 | static INLINE void |
281 | bidi_cache_shrink (void) | |
282 | { | |
283 | if (bidi_cache_size > BIDI_CACHE_CHUNK) | |
284 | { | |
0416466c EZ |
285 | bidi_cache_size = BIDI_CACHE_CHUNK; |
286 | bidi_cache = | |
287 | (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz); | |
2fe72643 EZ |
288 | } |
289 | bidi_cache_reset (); | |
290 | } | |
291 | ||
fd3998ff | 292 | static INLINE void |
b7b65b15 EZ |
293 | bidi_cache_fetch_state (int idx, struct bidi_it *bidi_it) |
294 | { | |
295 | int current_scan_dir = bidi_it->scan_dir; | |
296 | ||
87e67904 | 297 | if (idx < bidi_cache_start || idx >= bidi_cache_idx) |
b7b65b15 EZ |
298 | abort (); |
299 | ||
300 | bidi_copy_it (bidi_it, &bidi_cache[idx]); | |
301 | bidi_it->scan_dir = current_scan_dir; | |
302 | bidi_cache_last_idx = idx; | |
303 | } | |
304 | ||
305 | /* Find a cached state with a given CHARPOS and resolved embedding | |
306 | level less or equal to LEVEL. if LEVEL is -1, disregard the | |
307 | resolved levels in cached states. DIR, if non-zero, means search | |
308 | in that direction from the last cache hit. */ | |
fd3998ff | 309 | static INLINE int |
61bfec98 | 310 | bidi_cache_search (EMACS_INT charpos, int level, int dir) |
b7b65b15 EZ |
311 | { |
312 | int i, i_start; | |
313 | ||
314 | if (bidi_cache_idx) | |
315 | { | |
316 | if (charpos < bidi_cache[bidi_cache_last_idx].charpos) | |
182ce2d2 EZ |
317 | { |
318 | dir = -1; | |
319 | i_start = bidi_cache_last_idx - 1; | |
320 | } | |
321 | else if (charpos > (bidi_cache[bidi_cache_last_idx].charpos | |
322 | + bidi_cache[bidi_cache_last_idx].nchars - 1)) | |
323 | { | |
324 | dir = 1; | |
325 | i_start = bidi_cache_last_idx + 1; | |
326 | } | |
327 | else if (dir) | |
b7b65b15 EZ |
328 | i_start = bidi_cache_last_idx; |
329 | else | |
330 | { | |
331 | dir = -1; | |
332 | i_start = bidi_cache_idx - 1; | |
333 | } | |
334 | ||
335 | if (dir < 0) | |
336 | { | |
337 | /* Linear search for now; FIXME! */ | |
87e67904 | 338 | for (i = i_start; i >= bidi_cache_start; i--) |
182ce2d2 EZ |
339 | if (bidi_cache[i].charpos <= charpos |
340 | && charpos < bidi_cache[i].charpos + bidi_cache[i].nchars | |
b7b65b15 EZ |
341 | && (level == -1 || bidi_cache[i].resolved_level <= level)) |
342 | return i; | |
343 | } | |
344 | else | |
345 | { | |
346 | for (i = i_start; i < bidi_cache_idx; i++) | |
182ce2d2 EZ |
347 | if (bidi_cache[i].charpos <= charpos |
348 | && charpos < bidi_cache[i].charpos + bidi_cache[i].nchars | |
b7b65b15 EZ |
349 | && (level == -1 || bidi_cache[i].resolved_level <= level)) |
350 | return i; | |
351 | } | |
352 | } | |
353 | ||
354 | return -1; | |
355 | } | |
356 | ||
357 | /* Find a cached state where the resolved level changes to a value | |
358 | that is lower than LEVEL, and return its cache slot index. DIR is | |
359 | the direction to search, starting with the last used cache slot. | |
87e67904 EZ |
360 | If DIR is zero, we search backwards from the last occupied cache |
361 | slot. BEFORE, if non-zero, means return the index of the slot that | |
362 | is ``before'' the level change in the search direction. That is, | |
b7b65b15 EZ |
363 | given the cached levels like this: |
364 | ||
365 | 1122333442211 | |
366 | AB C | |
367 | ||
368 | and assuming we are at the position cached at the slot marked with | |
369 | C, searching backwards (DIR = -1) for LEVEL = 2 will return the | |
370 | index of slot B or A, depending whether BEFORE is, respectively, | |
371 | non-zero or zero. */ | |
372 | static int | |
373 | bidi_cache_find_level_change (int level, int dir, int before) | |
374 | { | |
375 | if (bidi_cache_idx) | |
376 | { | |
377 | int i = dir ? bidi_cache_last_idx : bidi_cache_idx - 1; | |
378 | int incr = before ? 1 : 0; | |
379 | ||
380 | if (!dir) | |
381 | dir = -1; | |
382 | else if (!incr) | |
383 | i += dir; | |
384 | ||
385 | if (dir < 0) | |
386 | { | |
87e67904 | 387 | while (i >= bidi_cache_start + incr) |
b7b65b15 EZ |
388 | { |
389 | if (bidi_cache[i - incr].resolved_level >= 0 | |
390 | && bidi_cache[i - incr].resolved_level < level) | |
391 | return i; | |
392 | i--; | |
393 | } | |
394 | } | |
395 | else | |
396 | { | |
397 | while (i < bidi_cache_idx - incr) | |
398 | { | |
399 | if (bidi_cache[i + incr].resolved_level >= 0 | |
400 | && bidi_cache[i + incr].resolved_level < level) | |
401 | return i; | |
402 | i++; | |
403 | } | |
404 | } | |
405 | } | |
406 | ||
407 | return -1; | |
408 | } | |
409 | ||
fd3998ff | 410 | static INLINE void |
b7b65b15 EZ |
411 | bidi_cache_iterator_state (struct bidi_it *bidi_it, int resolved) |
412 | { | |
413 | int idx; | |
414 | ||
415 | /* We should never cache on backward scans. */ | |
416 | if (bidi_it->scan_dir == -1) | |
417 | abort (); | |
418 | idx = bidi_cache_search (bidi_it->charpos, -1, 1); | |
419 | ||
420 | if (idx < 0) | |
421 | { | |
422 | idx = bidi_cache_idx; | |
2fe72643 EZ |
423 | /* Enlarge the cache as needed. */ |
424 | if (idx >= bidi_cache_size) | |
425 | { | |
0416466c | 426 | bidi_cache_size += BIDI_CACHE_CHUNK; |
2fe72643 | 427 | bidi_cache = |
0416466c | 428 | (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz); |
2fe72643 | 429 | } |
bd924a5d EZ |
430 | /* Character positions should correspond to cache positions 1:1. |
431 | If we are outside the range of cached positions, the cache is | |
432 | useless and must be reset. */ | |
87e67904 | 433 | if (idx > bidi_cache_start && |
182ce2d2 EZ |
434 | (bidi_it->charpos > (bidi_cache[idx - 1].charpos |
435 | + bidi_cache[idx - 1].nchars) | |
bd924a5d EZ |
436 | || bidi_it->charpos < bidi_cache[0].charpos)) |
437 | { | |
438 | bidi_cache_reset (); | |
87e67904 | 439 | idx = bidi_cache_start; |
bd924a5d | 440 | } |
102ebb00 EZ |
441 | if (bidi_it->nchars <= 0) |
442 | abort (); | |
b7b65b15 EZ |
443 | bidi_copy_it (&bidi_cache[idx], bidi_it); |
444 | if (!resolved) | |
445 | bidi_cache[idx].resolved_level = -1; | |
446 | } | |
447 | else | |
448 | { | |
449 | /* Copy only the members which could have changed, to avoid | |
450 | costly copying of the entire struct. */ | |
451 | bidi_cache[idx].type = bidi_it->type; | |
2d6e4628 | 452 | bidi_check_type (bidi_it->type); |
89d3374a EZ |
453 | bidi_cache[idx].type_after_w1 = bidi_it->type_after_w1; |
454 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 EZ |
455 | if (resolved) |
456 | bidi_cache[idx].resolved_level = bidi_it->resolved_level; | |
457 | else | |
458 | bidi_cache[idx].resolved_level = -1; | |
459 | bidi_cache[idx].invalid_levels = bidi_it->invalid_levels; | |
460 | bidi_cache[idx].invalid_rl_levels = bidi_it->invalid_rl_levels; | |
461 | bidi_cache[idx].next_for_neutral = bidi_it->next_for_neutral; | |
462 | bidi_cache[idx].next_for_ws = bidi_it->next_for_ws; | |
463 | bidi_cache[idx].ignore_bn_limit = bidi_it->ignore_bn_limit; | |
464 | } | |
465 | ||
466 | bidi_cache_last_idx = idx; | |
467 | if (idx >= bidi_cache_idx) | |
468 | bidi_cache_idx = idx + 1; | |
469 | } | |
470 | ||
fd3998ff | 471 | static INLINE bidi_type_t |
61bfec98 | 472 | bidi_cache_find (EMACS_INT charpos, int level, struct bidi_it *bidi_it) |
b7b65b15 EZ |
473 | { |
474 | int i = bidi_cache_search (charpos, level, bidi_it->scan_dir); | |
475 | ||
87e67904 | 476 | if (i >= bidi_cache_start) |
b7b65b15 EZ |
477 | { |
478 | bidi_dir_t current_scan_dir = bidi_it->scan_dir; | |
479 | ||
5009f85e | 480 | bidi_copy_it (bidi_it, &bidi_cache[i]); |
b7b65b15 EZ |
481 | bidi_cache_last_idx = i; |
482 | /* Don't let scan direction from from the cached state override | |
483 | the current scan direction. */ | |
484 | bidi_it->scan_dir = current_scan_dir; | |
485 | return bidi_it->type; | |
486 | } | |
487 | ||
488 | return UNKNOWN_BT; | |
489 | } | |
490 | ||
fd3998ff | 491 | static INLINE int |
b7b65b15 EZ |
492 | bidi_peek_at_next_level (struct bidi_it *bidi_it) |
493 | { | |
87e67904 | 494 | if (bidi_cache_idx == bidi_cache_start || bidi_cache_last_idx == -1) |
b7b65b15 EZ |
495 | abort (); |
496 | return bidi_cache[bidi_cache_last_idx + bidi_it->scan_dir].resolved_level; | |
497 | } | |
498 | ||
be39f003 EZ |
499 | /* Check if buffer position CHARPOS/BYTEPOS is the end of a paragraph. |
500 | Value is the non-negative length of the paragraph separator | |
501 | following the buffer position, -1 if position is at the beginning | |
502 | of a new paragraph, or -2 if position is neither at beginning nor | |
503 | at end of a paragraph. */ | |
fd3998ff | 504 | static EMACS_INT |
6bff6497 | 505 | bidi_at_paragraph_end (EMACS_INT charpos, EMACS_INT bytepos) |
b7b65b15 | 506 | { |
372b7a95 EZ |
507 | Lisp_Object sep_re; |
508 | Lisp_Object start_re; | |
be39f003 EZ |
509 | EMACS_INT val; |
510 | ||
372b7a95 EZ |
511 | sep_re = paragraph_separate_re; |
512 | start_re = paragraph_start_re; | |
be39f003 EZ |
513 | |
514 | val = fast_looking_at (sep_re, charpos, bytepos, ZV, ZV_BYTE, Qnil); | |
515 | if (val < 0) | |
516 | { | |
517 | if (fast_looking_at (start_re, charpos, bytepos, ZV, ZV_BYTE, Qnil) >= 0) | |
518 | val = -1; | |
519 | else | |
520 | val = -2; | |
521 | } | |
b7b65b15 | 522 | |
be39f003 | 523 | return val; |
b7b65b15 EZ |
524 | } |
525 | ||
526 | /* Determine the start-of-run (sor) directional type given the two | |
527 | embedding levels on either side of the run boundary. Also, update | |
528 | the saved info about previously seen characters, since that info is | |
529 | generally valid for a single level run. */ | |
fd3998ff | 530 | static INLINE void |
b7b65b15 EZ |
531 | bidi_set_sor_type (struct bidi_it *bidi_it, int level_before, int level_after) |
532 | { | |
533 | int higher_level = level_before > level_after ? level_before : level_after; | |
534 | ||
535 | /* The prev_was_pdf gork is required for when we have several PDFs | |
536 | in a row. In that case, we want to compute the sor type for the | |
537 | next level run only once: when we see the first PDF. That's | |
538 | because the sor type depends only on the higher of the two levels | |
539 | that we find on the two sides of the level boundary (see UAX#9, | |
540 | clause X10), and so we don't need to know the final embedding | |
541 | level to which we descend after processing all the PDFs. */ | |
e342a24d | 542 | if (!bidi_it->prev_was_pdf || level_before < level_after) |
b7b65b15 EZ |
543 | /* FIXME: should the default sor direction be user selectable? */ |
544 | bidi_it->sor = (higher_level & 1) != 0 ? R2L : L2R; | |
545 | if (level_before > level_after) | |
546 | bidi_it->prev_was_pdf = 1; | |
547 | ||
548 | bidi_it->prev.type = UNKNOWN_BT; | |
89d3374a EZ |
549 | bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1 = |
550 | bidi_it->last_strong.orig_type = UNKNOWN_BT; | |
b7b65b15 EZ |
551 | bidi_it->prev_for_neutral.type = bidi_it->sor == R2L ? STRONG_R : STRONG_L; |
552 | bidi_it->prev_for_neutral.charpos = bidi_it->charpos; | |
553 | bidi_it->prev_for_neutral.bytepos = bidi_it->bytepos; | |
89d3374a EZ |
554 | bidi_it->next_for_neutral.type = bidi_it->next_for_neutral.type_after_w1 = |
555 | bidi_it->next_for_neutral.orig_type = UNKNOWN_BT; | |
87e67904 | 556 | bidi_it->ignore_bn_limit = -1; /* meaning it's unknown */ |
b7b65b15 EZ |
557 | } |
558 | ||
182ce2d2 | 559 | /* Perform initializations for reordering a new line of bidi text. */ |
6bff6497 | 560 | static void |
be39f003 EZ |
561 | bidi_line_init (struct bidi_it *bidi_it) |
562 | { | |
563 | bidi_it->scan_dir = 1; /* FIXME: do we need to have control on this? */ | |
564 | bidi_it->resolved_level = bidi_it->level_stack[0].level; | |
565 | bidi_it->level_stack[0].override = NEUTRAL_DIR; /* X1 */ | |
566 | bidi_it->invalid_levels = 0; | |
567 | bidi_it->invalid_rl_levels = -1; | |
568 | bidi_it->next_en_pos = -1; | |
569 | bidi_it->next_for_ws.type = UNKNOWN_BT; | |
b44d9321 EZ |
570 | bidi_set_sor_type (bidi_it, |
571 | bidi_it->paragraph_dir == R2L ? 1 : 0, | |
be39f003 EZ |
572 | bidi_it->level_stack[0].level); /* X10 */ |
573 | ||
574 | bidi_cache_reset (); | |
575 | } | |
576 | ||
87e67904 EZ |
577 | /* Count bytes in multibyte string S between BEG/BEGBYTE and END. BEG |
578 | and END are zero-based character positions in S, BEGBYTE is byte | |
579 | position corresponding to BEG. */ | |
580 | static inline EMACS_INT | |
581 | bidi_count_bytes (const unsigned char *s, const EMACS_INT beg, | |
582 | const EMACS_INT begbyte, const EMACS_INT end) | |
583 | { | |
584 | EMACS_INT pos = beg; | |
585 | const unsigned char *p = s + begbyte, *start = p; | |
586 | ||
587 | if (!CHAR_HEAD_P (*p)) | |
588 | abort (); | |
589 | ||
590 | while (pos < end) | |
591 | { | |
592 | p += BYTES_BY_CHAR_HEAD (*p); | |
593 | pos++; | |
594 | } | |
595 | ||
596 | return p - start; | |
597 | } | |
598 | ||
599 | /* Fetch and returns the character at byte position BYTEPOS. If S is | |
600 | non-NULL, fetch the character from string S; otherwise fetch the | |
601 | character from the current buffer. */ | |
602 | static inline int | |
603 | bidi_char_at_pos (EMACS_INT bytepos, const unsigned char *s) | |
604 | { | |
605 | if (s) | |
606 | return STRING_CHAR (s + bytepos); | |
607 | else | |
608 | return FETCH_MULTIBYTE_CHAR (bytepos); | |
609 | } | |
610 | ||
102ebb00 EZ |
611 | /* Fetch and return the character at BYTEPOS/CHARPOS. If that |
612 | character is covered by a display string, treat the entire run of | |
613 | covered characters as a single character u+FFFC, and return their | |
614 | combined length in CH_LEN and NCHARS. DISP_POS specifies the | |
615 | character position of the next display string, or -1 if not yet | |
616 | computed. When the next character is at or beyond that position, | |
617 | the function updates DISP_POS with the position of the next display | |
87e67904 EZ |
618 | string. STRING->s is the string to iterate, or NULL if iterating over |
619 | a buffer. */ | |
620 | static inline int | |
102ebb00 | 621 | bidi_fetch_char (EMACS_INT bytepos, EMACS_INT charpos, EMACS_INT *disp_pos, |
87e67904 | 622 | struct bidi_string_data *string, |
fc6f18ce | 623 | int frame_window_p, EMACS_INT *ch_len, EMACS_INT *nchars) |
182ce2d2 EZ |
624 | { |
625 | int ch; | |
87e67904 | 626 | EMACS_INT endpos = string->s ? string->schars : ZV; |
6db161be | 627 | struct text_pos pos; |
182ce2d2 | 628 | |
182ce2d2 | 629 | /* If we got past the last known position of display string, compute |
87e67904 EZ |
630 | the position of the next one. That position could be at CHARPOS. */ |
631 | if (charpos < endpos && charpos > *disp_pos) | |
6db161be EZ |
632 | { |
633 | SET_TEXT_POS (pos, charpos, bytepos); | |
634 | *disp_pos = compute_display_string_pos (&pos, string, frame_window_p); | |
635 | } | |
102ebb00 EZ |
636 | |
637 | /* Fetch the character at BYTEPOS. */ | |
87e67904 | 638 | if (charpos >= endpos) |
182ce2d2 EZ |
639 | { |
640 | ch = BIDI_EOB; | |
641 | *ch_len = 1; | |
642 | *nchars = 1; | |
87e67904 | 643 | *disp_pos = endpos; |
182ce2d2 | 644 | } |
102ebb00 | 645 | else if (charpos >= *disp_pos) |
182ce2d2 | 646 | { |
7b600102 EZ |
647 | EMACS_INT disp_end_pos; |
648 | ||
649 | /* We don't expect to find ourselves in the middle of a display | |
650 | property. Hopefully, it will never be needed. */ | |
651 | if (charpos > *disp_pos) | |
652 | abort (); | |
653 | /* Return the Unicode Object Replacement Character to represent | |
87e67904 | 654 | the entire run of characters covered by the display string. */ |
7b600102 | 655 | ch = 0xFFFC; |
87e67904 | 656 | disp_end_pos = compute_display_string_end (*disp_pos, string); |
7b600102 | 657 | *nchars = disp_end_pos - *disp_pos; |
87e67904 EZ |
658 | if (string->s) |
659 | *ch_len = bidi_count_bytes (string->s, *disp_pos, bytepos, | |
660 | disp_end_pos); | |
661 | else | |
662 | *ch_len = CHAR_TO_BYTE (disp_end_pos) - bytepos; | |
182ce2d2 | 663 | } |
182ce2d2 EZ |
664 | else |
665 | { | |
87e67904 EZ |
666 | if (string->s) |
667 | { | |
668 | EMACS_INT len; | |
669 | ||
670 | ch = STRING_CHAR_AND_LENGTH (string->s + bytepos, len); | |
671 | *ch_len = len; | |
672 | } | |
673 | else | |
674 | { | |
675 | ch = FETCH_MULTIBYTE_CHAR (bytepos); | |
676 | *ch_len = CHAR_BYTES (ch); | |
677 | } | |
182ce2d2 EZ |
678 | *nchars = 1; |
679 | } | |
680 | ||
681 | /* If we just entered a run of characters covered by a display | |
682 | string, compute the position of the next display string. */ | |
87e67904 | 683 | if (charpos + *nchars <= endpos && charpos + *nchars > *disp_pos) |
6db161be EZ |
684 | { |
685 | SET_TEXT_POS (pos, charpos + *nchars, bytepos + *ch_len); | |
686 | *disp_pos = compute_display_string_pos (&pos, string, frame_window_p); | |
687 | } | |
182ce2d2 EZ |
688 | |
689 | return ch; | |
690 | } | |
691 | ||
be39f003 EZ |
692 | /* Find the beginning of this paragraph by looking back in the buffer. |
693 | Value is the byte position of the paragraph's beginning. */ | |
694 | static EMACS_INT | |
b44d9321 | 695 | bidi_find_paragraph_start (EMACS_INT pos, EMACS_INT pos_byte) |
6bff6497 | 696 | { |
372b7a95 | 697 | Lisp_Object re = paragraph_start_re; |
6bff6497 EZ |
698 | EMACS_INT limit = ZV, limit_byte = ZV_BYTE; |
699 | ||
6bff6497 EZ |
700 | while (pos_byte > BEGV_BYTE |
701 | && fast_looking_at (re, pos, pos_byte, limit, limit_byte, Qnil) < 0) | |
702 | { | |
182ce2d2 EZ |
703 | /* FIXME: What if the paragraph beginning is covered by a |
704 | display string? And what if a display string covering some | |
705 | of the text over which we scan back includes | |
706 | paragraph_start_re? */ | |
be39f003 EZ |
707 | pos = find_next_newline_no_quit (pos - 1, -1); |
708 | pos_byte = CHAR_TO_BYTE (pos); | |
6bff6497 | 709 | } |
be39f003 | 710 | return pos_byte; |
6bff6497 EZ |
711 | } |
712 | ||
bea4f10c | 713 | /* Determine the base direction, a.k.a. base embedding level, of the |
b44d9321 EZ |
714 | paragraph we are about to iterate through. If DIR is either L2R or |
715 | R2L, just use that. Otherwise, determine the paragraph direction | |
bea4f10c EZ |
716 | from the first strong directional character of the paragraph. |
717 | ||
182ce2d2 | 718 | NO_DEFAULT_P non-zero means don't default to L2R if the paragraph |
bea4f10c EZ |
719 | has no strong directional characters and both DIR and |
720 | bidi_it->paragraph_dir are NEUTRAL_DIR. In that case, search back | |
721 | in the buffer until a paragraph is found with a strong character, | |
722 | or until hitting BEGV. In the latter case, fall back to L2R. This | |
723 | flag is used in current-bidi-paragraph-direction. | |
724 | ||
725 | Note that this function gives the paragraph separator the same | |
726 | direction as the preceding paragraph, even though Emacs generally | |
727 | views the separartor as not belonging to any paragraph. */ | |
b7b65b15 | 728 | void |
bea4f10c | 729 | bidi_paragraph_init (bidi_dir_t dir, struct bidi_it *bidi_it, int no_default_p) |
b7b65b15 | 730 | { |
6bff6497 | 731 | EMACS_INT bytepos = bidi_it->bytepos; |
87e67904 | 732 | int string_p = bidi_it->string.s != NULL; |
bea4f10c | 733 | EMACS_INT pstartbyte; |
87e67904 EZ |
734 | /* Note that begbyte is a byte position, while end is a character |
735 | position. Yes, this is ugly, but we are trying to avoid costly | |
736 | calls to BYTE_TO_CHAR and its ilk. */ | |
737 | EMACS_INT begbyte = string_p ? 0 : BEGV_BYTE; | |
738 | EMACS_INT end = string_p ? bidi_it->string.schars : ZV; | |
e7402cb2 | 739 | |
5e65aec0 | 740 | /* Special case for an empty buffer. */ |
87e67904 | 741 | if (bytepos == begbyte && bidi_it->charpos == end) |
5e65aec0 | 742 | dir = L2R; |
9c82e145 | 743 | /* We should never be called at EOB or before BEGV. */ |
87e67904 | 744 | else if (bidi_it->charpos >= end || bytepos < begbyte) |
9c82e145 EZ |
745 | abort (); |
746 | ||
be39f003 EZ |
747 | if (dir == L2R) |
748 | { | |
749 | bidi_it->paragraph_dir = L2R; | |
750 | bidi_it->new_paragraph = 0; | |
751 | } | |
752 | else if (dir == R2L) | |
753 | { | |
754 | bidi_it->paragraph_dir = R2L; | |
755 | bidi_it->new_paragraph = 0; | |
756 | } | |
b7b65b15 EZ |
757 | else if (dir == NEUTRAL_DIR) /* P2 */ |
758 | { | |
182ce2d2 EZ |
759 | int ch; |
760 | EMACS_INT ch_len, nchars; | |
761 | EMACS_INT pos, disp_pos = -1; | |
6bff6497 | 762 | bidi_type_t type; |
be39f003 | 763 | |
d20e1419 EZ |
764 | if (!bidi_initialized) |
765 | bidi_initialize (); | |
766 | ||
be39f003 EZ |
767 | /* If we are inside a paragraph separator, we are just waiting |
768 | for the separator to be exhausted; use the previous paragraph | |
e5a2fec7 EZ |
769 | direction. But don't do that if we have been just reseated, |
770 | because we need to reinitialize below in that case. */ | |
771 | if (!bidi_it->first_elt | |
772 | && bidi_it->charpos < bidi_it->separator_limit) | |
be39f003 EZ |
773 | return; |
774 | ||
b44d9321 | 775 | /* If we are on a newline, get past it to where the next |
5e65aec0 EZ |
776 | paragraph might start. But don't do that at BEGV since then |
777 | we are potentially in a new paragraph that doesn't yet | |
778 | exist. */ | |
c143c213 | 779 | pos = bidi_it->charpos; |
87e67904 EZ |
780 | if (bytepos > begbyte |
781 | && bidi_char_at_pos (bytepos, bidi_it->string.s) == '\n') | |
be39f003 | 782 | { |
b44d9321 | 783 | bytepos++; |
c143c213 | 784 | pos++; |
be39f003 | 785 | } |
b44d9321 EZ |
786 | |
787 | /* We are either at the beginning of a paragraph or in the | |
788 | middle of it. Find where this paragraph starts. */ | |
87e67904 EZ |
789 | if (string_p) |
790 | { | |
791 | /* We don't support changes of paragraph direction inside a | |
792 | string. It is treated as a single paragraph. */ | |
793 | pstartbyte = 0; | |
794 | } | |
795 | else | |
796 | pstartbyte = bidi_find_paragraph_start (pos, bytepos); | |
be39f003 EZ |
797 | bidi_it->separator_limit = -1; |
798 | bidi_it->new_paragraph = 0; | |
bea4f10c EZ |
799 | |
800 | /* The following loop is run more than once only if NO_DEFAULT_P | |
87e67904 | 801 | is non-zero, and only if we are iterating on a buffer. */ |
bea4f10c EZ |
802 | do { |
803 | bytepos = pstartbyte; | |
87e67904 EZ |
804 | if (!string_p) |
805 | pos = BYTE_TO_CHAR (bytepos); | |
806 | ch = bidi_fetch_char (bytepos, pos, &disp_pos, &bidi_it->string, | |
807 | bidi_it->frame_window_p, &ch_len, &nchars); | |
bea4f10c EZ |
808 | type = bidi_get_type (ch, NEUTRAL_DIR); |
809 | ||
182ce2d2 | 810 | for (pos += nchars, bytepos += ch_len; |
bea4f10c EZ |
811 | /* NOTE: UAX#9 says to search only for L, AL, or R types |
812 | of characters, and ignore RLE, RLO, LRE, and LRO. | |
813 | However, I'm not sure it makes sense to omit those 4; | |
814 | should try with and without that to see the effect. */ | |
815 | (bidi_get_category (type) != STRONG) | |
816 | || (bidi_ignore_explicit_marks_for_paragraph_level | |
817 | && (type == RLE || type == RLO | |
818 | || type == LRE || type == LRO)); | |
819 | type = bidi_get_type (ch, NEUTRAL_DIR)) | |
820 | { | |
87e67904 EZ |
821 | if (!string_p |
822 | && type == NEUTRAL_B | |
823 | && bidi_at_paragraph_end (pos, bytepos) >= -1) | |
21fce5ab | 824 | break; |
87e67904 | 825 | if (pos >= end) |
bea4f10c EZ |
826 | { |
827 | /* Pretend there's a paragraph separator at end of | |
87e67904 | 828 | buffer/string. */ |
bea4f10c EZ |
829 | type = NEUTRAL_B; |
830 | break; | |
831 | } | |
102ebb00 | 832 | /* Fetch next character and advance to get past it. */ |
87e67904 | 833 | ch = bidi_fetch_char (bytepos, pos, &disp_pos, &bidi_it->string, |
fc6f18ce | 834 | bidi_it->frame_window_p, &ch_len, &nchars); |
182ce2d2 EZ |
835 | pos += nchars; |
836 | bytepos += ch_len; | |
bea4f10c EZ |
837 | } |
838 | if (type == STRONG_R || type == STRONG_AL) /* P3 */ | |
839 | bidi_it->paragraph_dir = R2L; | |
840 | else if (type == STRONG_L) | |
841 | bidi_it->paragraph_dir = L2R; | |
87e67904 EZ |
842 | if (!string_p |
843 | && no_default_p && bidi_it->paragraph_dir == NEUTRAL_DIR) | |
bea4f10c EZ |
844 | { |
845 | /* If this paragraph is at BEGV, default to L2R. */ | |
846 | if (pstartbyte == BEGV_BYTE) | |
847 | bidi_it->paragraph_dir = L2R; /* P3 and HL1 */ | |
848 | else | |
849 | { | |
850 | EMACS_INT prevpbyte = pstartbyte; | |
851 | EMACS_INT p = BYTE_TO_CHAR (pstartbyte), pbyte = pstartbyte; | |
852 | ||
853 | /* Find the beginning of the previous paragraph, if any. */ | |
854 | while (pbyte > BEGV_BYTE && prevpbyte >= pstartbyte) | |
855 | { | |
182ce2d2 EZ |
856 | /* FXIME: What if p is covered by a display |
857 | string? See also a FIXME inside | |
858 | bidi_find_paragraph_start. */ | |
bea4f10c EZ |
859 | p--; |
860 | pbyte = CHAR_TO_BYTE (p); | |
861 | prevpbyte = bidi_find_paragraph_start (p, pbyte); | |
862 | } | |
863 | pstartbyte = prevpbyte; | |
864 | } | |
865 | } | |
87e67904 EZ |
866 | } while (!string_p |
867 | && no_default_p && bidi_it->paragraph_dir == NEUTRAL_DIR); | |
b7b65b15 | 868 | } |
be39f003 EZ |
869 | else |
870 | abort (); | |
b7b65b15 | 871 | |
b44d9321 EZ |
872 | /* Contrary to UAX#9 clause P3, we only default the paragraph |
873 | direction to L2R if we have no previous usable paragraph | |
bea4f10c | 874 | direction. This is allowed by the HL1 clause. */ |
d20e1419 | 875 | if (bidi_it->paragraph_dir != L2R && bidi_it->paragraph_dir != R2L) |
bea4f10c | 876 | bidi_it->paragraph_dir = L2R; /* P3 and HL1 ``higher-level protocols'' */ |
be39f003 | 877 | if (bidi_it->paragraph_dir == R2L) |
b44d9321 | 878 | bidi_it->level_stack[0].level = 1; |
be39f003 | 879 | else |
b44d9321 | 880 | bidi_it->level_stack[0].level = 0; |
be39f003 EZ |
881 | |
882 | bidi_line_init (bidi_it); | |
b7b65b15 EZ |
883 | } |
884 | ||
6bff6497 EZ |
885 | /* Do whatever UAX#9 clause X8 says should be done at paragraph's |
886 | end. */ | |
fd3998ff | 887 | static INLINE void |
b7b65b15 EZ |
888 | bidi_set_paragraph_end (struct bidi_it *bidi_it) |
889 | { | |
890 | bidi_it->invalid_levels = 0; | |
891 | bidi_it->invalid_rl_levels = -1; | |
892 | bidi_it->stack_idx = 0; | |
893 | bidi_it->resolved_level = bidi_it->level_stack[0].level; | |
b7b65b15 EZ |
894 | } |
895 | ||
182ce2d2 | 896 | /* Initialize the bidi iterator from buffer/string position CHARPOS. */ |
b7b65b15 | 897 | void |
fc6f18ce EZ |
898 | bidi_init_it (EMACS_INT charpos, EMACS_INT bytepos, int frame_window_p, |
899 | struct bidi_it *bidi_it) | |
b7b65b15 EZ |
900 | { |
901 | if (! bidi_initialized) | |
902 | bidi_initialize (); | |
87e67904 EZ |
903 | if (charpos >= 0) |
904 | bidi_it->charpos = charpos; | |
905 | if (bytepos >= 0) | |
906 | bidi_it->bytepos = bytepos; | |
fc6f18ce | 907 | bidi_it->frame_window_p = frame_window_p; |
182ce2d2 | 908 | bidi_it->nchars = -1; /* to be computed in bidi_resolve_explicit_1 */ |
6bff6497 EZ |
909 | bidi_it->first_elt = 1; |
910 | bidi_set_paragraph_end (bidi_it); | |
911 | bidi_it->new_paragraph = 1; | |
be39f003 | 912 | bidi_it->separator_limit = -1; |
b7b65b15 | 913 | bidi_it->type = NEUTRAL_B; |
ebb5722e EZ |
914 | bidi_it->type_after_w1 = NEUTRAL_B; |
915 | bidi_it->orig_type = NEUTRAL_B; | |
b7b65b15 | 916 | bidi_it->prev_was_pdf = 0; |
ebb5722e EZ |
917 | bidi_it->prev.type = bidi_it->prev.type_after_w1 = |
918 | bidi_it->prev.orig_type = UNKNOWN_BT; | |
89d3374a EZ |
919 | bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1 = |
920 | bidi_it->last_strong.orig_type = UNKNOWN_BT; | |
b7b65b15 EZ |
921 | bidi_it->next_for_neutral.charpos = -1; |
922 | bidi_it->next_for_neutral.type = | |
89d3374a EZ |
923 | bidi_it->next_for_neutral.type_after_w1 = |
924 | bidi_it->next_for_neutral.orig_type = UNKNOWN_BT; | |
b7b65b15 EZ |
925 | bidi_it->prev_for_neutral.charpos = -1; |
926 | bidi_it->prev_for_neutral.type = | |
89d3374a EZ |
927 | bidi_it->prev_for_neutral.type_after_w1 = |
928 | bidi_it->prev_for_neutral.orig_type = UNKNOWN_BT; | |
b7b65b15 | 929 | bidi_it->sor = L2R; /* FIXME: should it be user-selectable? */ |
182ce2d2 | 930 | bidi_it->disp_pos = -1; /* invalid/unknown */ |
87e67904 EZ |
931 | /* We can only shrink the cache if we are at the bottom level of its |
932 | "stack". */ | |
933 | if (bidi_cache_start == 0) | |
934 | bidi_cache_shrink (); | |
b7b65b15 EZ |
935 | } |
936 | ||
937 | /* Push the current embedding level and override status; reset the | |
938 | current level to LEVEL and the current override status to OVERRIDE. */ | |
fd3998ff | 939 | static INLINE void |
b7b65b15 EZ |
940 | bidi_push_embedding_level (struct bidi_it *bidi_it, |
941 | int level, bidi_dir_t override) | |
942 | { | |
943 | bidi_it->stack_idx++; | |
944 | if (bidi_it->stack_idx >= BIDI_MAXLEVEL) | |
945 | abort (); | |
946 | bidi_it->level_stack[bidi_it->stack_idx].level = level; | |
947 | bidi_it->level_stack[bidi_it->stack_idx].override = override; | |
948 | } | |
949 | ||
950 | /* Pop the embedding level and directional override status from the | |
951 | stack, and return the new level. */ | |
fd3998ff | 952 | static INLINE int |
b7b65b15 EZ |
953 | bidi_pop_embedding_level (struct bidi_it *bidi_it) |
954 | { | |
955 | /* UAX#9 says to ignore invalid PDFs. */ | |
956 | if (bidi_it->stack_idx > 0) | |
957 | bidi_it->stack_idx--; | |
958 | return bidi_it->level_stack[bidi_it->stack_idx].level; | |
959 | } | |
960 | ||
961 | /* Record in SAVED_INFO the information about the current character. */ | |
fd3998ff | 962 | static INLINE void |
b7b65b15 EZ |
963 | bidi_remember_char (struct bidi_saved_info *saved_info, |
964 | struct bidi_it *bidi_it) | |
965 | { | |
966 | saved_info->charpos = bidi_it->charpos; | |
967 | saved_info->bytepos = bidi_it->bytepos; | |
968 | saved_info->type = bidi_it->type; | |
2d6e4628 | 969 | bidi_check_type (bidi_it->type); |
89d3374a EZ |
970 | saved_info->type_after_w1 = bidi_it->type_after_w1; |
971 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 | 972 | saved_info->orig_type = bidi_it->orig_type; |
2d6e4628 | 973 | bidi_check_type (bidi_it->orig_type); |
b7b65b15 EZ |
974 | } |
975 | ||
976 | /* Resolve the type of a neutral character according to the type of | |
977 | surrounding strong text and the current embedding level. */ | |
fd3998ff | 978 | static INLINE bidi_type_t |
b7b65b15 EZ |
979 | bidi_resolve_neutral_1 (bidi_type_t prev_type, bidi_type_t next_type, int lev) |
980 | { | |
981 | /* N1: European and Arabic numbers are treated as though they were R. */ | |
982 | if (next_type == WEAK_EN || next_type == WEAK_AN) | |
983 | next_type = STRONG_R; | |
984 | if (prev_type == WEAK_EN || prev_type == WEAK_AN) | |
985 | prev_type = STRONG_R; | |
986 | ||
987 | if (next_type == prev_type) /* N1 */ | |
988 | return next_type; | |
989 | else if ((lev & 1) == 0) /* N2 */ | |
990 | return STRONG_L; | |
991 | else | |
992 | return STRONG_R; | |
993 | } | |
994 | ||
fd3998ff | 995 | static INLINE int |
182ce2d2 | 996 | bidi_explicit_dir_char (int ch) |
b7b65b15 | 997 | { |
182ce2d2 EZ |
998 | bidi_type_t ch_type; |
999 | ||
1000 | if (!bidi_initialized) | |
1001 | abort (); | |
1002 | ch_type = (bidi_type_t) XINT (CHAR_TABLE_REF (bidi_type_table, ch)); | |
1003 | return (ch_type == LRE || ch_type == LRO | |
1004 | || ch_type == RLE || ch_type == RLO | |
1005 | || ch_type == PDF); | |
b7b65b15 EZ |
1006 | } |
1007 | ||
1008 | /* A helper function for bidi_resolve_explicit. It advances to the | |
1009 | next character in logical order and determines the new embedding | |
1010 | level and directional override, but does not take into account | |
1011 | empty embeddings. */ | |
1012 | static int | |
1013 | bidi_resolve_explicit_1 (struct bidi_it *bidi_it) | |
1014 | { | |
1015 | int curchar; | |
1016 | bidi_type_t type; | |
1017 | int current_level; | |
1018 | int new_level; | |
1019 | bidi_dir_t override; | |
87e67904 | 1020 | int string_p = bidi_it->string.s != NULL; |
b7b65b15 | 1021 | |
182ce2d2 EZ |
1022 | /* If reseat()'ed, don't advance, so as to start iteration from the |
1023 | position where we were reseated. bidi_it->bytepos can be less | |
1024 | than BEGV_BYTE after reseat to BEGV. */ | |
87e67904 | 1025 | if (bidi_it->bytepos < (string_p ? 0 : BEGV_BYTE) |
9c82e145 | 1026 | || bidi_it->first_elt) |
e7402cb2 | 1027 | { |
9c82e145 | 1028 | bidi_it->first_elt = 0; |
87e67904 EZ |
1029 | if (string_p) |
1030 | { | |
1031 | if (bidi_it->charpos < 0) | |
1032 | bidi_it->charpos = 0; | |
1033 | bidi_it->bytepos = bidi_count_bytes (bidi_it->string.s, 0, 0, | |
1034 | bidi_it->charpos); | |
1035 | } | |
1036 | else | |
1037 | { | |
1038 | if (bidi_it->charpos < BEGV) | |
1039 | bidi_it->charpos = BEGV; | |
1040 | bidi_it->bytepos = CHAR_TO_BYTE (bidi_it->charpos); | |
1041 | } | |
e7402cb2 | 1042 | } |
87e67904 EZ |
1043 | /* Don't move at end of buffer/string. */ |
1044 | else if (bidi_it->charpos < (string_p ? bidi_it->string.schars : ZV)) | |
b7b65b15 | 1045 | { |
182ce2d2 EZ |
1046 | /* Advance to the next character, skipping characters covered by |
1047 | display strings (nchars > 1). */ | |
102ebb00 EZ |
1048 | if (bidi_it->nchars <= 0) |
1049 | abort (); | |
182ce2d2 | 1050 | bidi_it->charpos += bidi_it->nchars; |
e7402cb2 EZ |
1051 | if (bidi_it->ch_len == 0) |
1052 | abort (); | |
b7b65b15 EZ |
1053 | bidi_it->bytepos += bidi_it->ch_len; |
1054 | } | |
1055 | ||
1056 | current_level = bidi_it->level_stack[bidi_it->stack_idx].level; /* X1 */ | |
1057 | override = bidi_it->level_stack[bidi_it->stack_idx].override; | |
1058 | new_level = current_level; | |
1059 | ||
87e67904 | 1060 | if (bidi_it->charpos >= (string_p ? bidi_it->string.schars : ZV)) |
e7402cb2 EZ |
1061 | { |
1062 | curchar = BIDI_EOB; | |
1063 | bidi_it->ch_len = 1; | |
182ce2d2 | 1064 | bidi_it->nchars = 1; |
87e67904 | 1065 | bidi_it->disp_pos = (string_p ? bidi_it->string.schars : ZV); |
e7402cb2 EZ |
1066 | } |
1067 | else | |
1068 | { | |
182ce2d2 EZ |
1069 | /* Fetch the character at BYTEPOS. If it is covered by a |
1070 | display string, treat the entire run of covered characters as | |
1071 | a single character u+FFFC. */ | |
102ebb00 | 1072 | curchar = bidi_fetch_char (bidi_it->bytepos, bidi_it->charpos, |
87e67904 EZ |
1073 | &bidi_it->disp_pos, &bidi_it->string, |
1074 | bidi_it->frame_window_p, | |
102ebb00 | 1075 | &bidi_it->ch_len, &bidi_it->nchars); |
e7402cb2 | 1076 | } |
b7b65b15 | 1077 | bidi_it->ch = curchar; |
b7b65b15 | 1078 | |
6bff6497 EZ |
1079 | /* Don't apply directional override here, as all the types we handle |
1080 | below will not be affected by the override anyway, and we need | |
1081 | the original type unaltered. The override will be applied in | |
1082 | bidi_resolve_weak. */ | |
1083 | type = bidi_get_type (curchar, NEUTRAL_DIR); | |
89d3374a EZ |
1084 | bidi_it->orig_type = type; |
1085 | bidi_check_type (bidi_it->orig_type); | |
b7b65b15 EZ |
1086 | |
1087 | if (type != PDF) | |
1088 | bidi_it->prev_was_pdf = 0; | |
1089 | ||
89d3374a | 1090 | bidi_it->type_after_w1 = UNKNOWN_BT; |
b7b65b15 EZ |
1091 | |
1092 | switch (type) | |
1093 | { | |
1094 | case RLE: /* X2 */ | |
1095 | case RLO: /* X4 */ | |
89d3374a EZ |
1096 | bidi_it->type_after_w1 = type; |
1097 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 | 1098 | type = WEAK_BN; /* X9/Retaining */ |
87e67904 | 1099 | if (bidi_it->ignore_bn_limit <= -1) |
b7b65b15 EZ |
1100 | { |
1101 | if (current_level <= BIDI_MAXLEVEL - 4) | |
1102 | { | |
1103 | /* Compute the least odd embedding level greater than | |
1104 | the current level. */ | |
1105 | new_level = ((current_level + 1) & ~1) + 1; | |
89d3374a | 1106 | if (bidi_it->type_after_w1 == RLE) |
b7b65b15 EZ |
1107 | override = NEUTRAL_DIR; |
1108 | else | |
1109 | override = R2L; | |
1110 | if (current_level == BIDI_MAXLEVEL - 4) | |
1111 | bidi_it->invalid_rl_levels = 0; | |
1112 | bidi_push_embedding_level (bidi_it, new_level, override); | |
1113 | } | |
1114 | else | |
1115 | { | |
1116 | bidi_it->invalid_levels++; | |
1117 | /* See the commentary about invalid_rl_levels below. */ | |
1118 | if (bidi_it->invalid_rl_levels < 0) | |
1119 | bidi_it->invalid_rl_levels = 0; | |
1120 | bidi_it->invalid_rl_levels++; | |
1121 | } | |
1122 | } | |
89d3374a | 1123 | else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */ |
b7b65b15 EZ |
1124 | || bidi_it->next_en_pos > bidi_it->charpos) |
1125 | type = WEAK_EN; | |
1126 | break; | |
1127 | case LRE: /* X3 */ | |
1128 | case LRO: /* X5 */ | |
89d3374a EZ |
1129 | bidi_it->type_after_w1 = type; |
1130 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 | 1131 | type = WEAK_BN; /* X9/Retaining */ |
87e67904 | 1132 | if (bidi_it->ignore_bn_limit <= -1) |
b7b65b15 EZ |
1133 | { |
1134 | if (current_level <= BIDI_MAXLEVEL - 5) | |
1135 | { | |
1136 | /* Compute the least even embedding level greater than | |
1137 | the current level. */ | |
1138 | new_level = ((current_level + 2) & ~1); | |
89d3374a | 1139 | if (bidi_it->type_after_w1 == LRE) |
b7b65b15 EZ |
1140 | override = NEUTRAL_DIR; |
1141 | else | |
1142 | override = L2R; | |
1143 | bidi_push_embedding_level (bidi_it, new_level, override); | |
1144 | } | |
1145 | else | |
1146 | { | |
1147 | bidi_it->invalid_levels++; | |
1148 | /* invalid_rl_levels counts invalid levels encountered | |
1149 | while the embedding level was already too high for | |
1150 | LRE/LRO, but not for RLE/RLO. That is because | |
1151 | there may be exactly one PDF which we should not | |
1152 | ignore even though invalid_levels is non-zero. | |
1153 | invalid_rl_levels helps to know what PDF is | |
1154 | that. */ | |
1155 | if (bidi_it->invalid_rl_levels >= 0) | |
1156 | bidi_it->invalid_rl_levels++; | |
1157 | } | |
1158 | } | |
89d3374a | 1159 | else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */ |
b7b65b15 EZ |
1160 | || bidi_it->next_en_pos > bidi_it->charpos) |
1161 | type = WEAK_EN; | |
1162 | break; | |
1163 | case PDF: /* X7 */ | |
89d3374a EZ |
1164 | bidi_it->type_after_w1 = type; |
1165 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 | 1166 | type = WEAK_BN; /* X9/Retaining */ |
87e67904 | 1167 | if (bidi_it->ignore_bn_limit <= -1) |
b7b65b15 EZ |
1168 | { |
1169 | if (!bidi_it->invalid_rl_levels) | |
1170 | { | |
1171 | new_level = bidi_pop_embedding_level (bidi_it); | |
1172 | bidi_it->invalid_rl_levels = -1; | |
1173 | if (bidi_it->invalid_levels) | |
1174 | bidi_it->invalid_levels--; | |
1175 | /* else nothing: UAX#9 says to ignore invalid PDFs */ | |
1176 | } | |
1177 | if (!bidi_it->invalid_levels) | |
1178 | new_level = bidi_pop_embedding_level (bidi_it); | |
1179 | else | |
1180 | { | |
1181 | bidi_it->invalid_levels--; | |
1182 | bidi_it->invalid_rl_levels--; | |
1183 | } | |
1184 | } | |
89d3374a | 1185 | else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */ |
b7b65b15 EZ |
1186 | || bidi_it->next_en_pos > bidi_it->charpos) |
1187 | type = WEAK_EN; | |
1188 | break; | |
1189 | default: | |
1190 | /* Nothing. */ | |
1191 | break; | |
1192 | } | |
1193 | ||
1194 | bidi_it->type = type; | |
2d6e4628 | 1195 | bidi_check_type (bidi_it->type); |
b7b65b15 EZ |
1196 | |
1197 | return new_level; | |
1198 | } | |
1199 | ||
1200 | /* Given an iterator state in BIDI_IT, advance one character position | |
182ce2d2 EZ |
1201 | in the buffer/string to the next character (in the logical order), |
1202 | resolve any explicit embeddings and directional overrides, and | |
1203 | return the embedding level of the character after resolving | |
1204 | explicit directives and ignoring empty embeddings. */ | |
b7b65b15 EZ |
1205 | static int |
1206 | bidi_resolve_explicit (struct bidi_it *bidi_it) | |
1207 | { | |
1208 | int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level; | |
1209 | int new_level = bidi_resolve_explicit_1 (bidi_it); | |
87e67904 | 1210 | EMACS_INT eob = bidi_it->string.s ? bidi_it->string.schars : ZV; |
b7b65b15 EZ |
1211 | |
1212 | if (prev_level < new_level | |
1213 | && bidi_it->type == WEAK_BN | |
87e67904 EZ |
1214 | && bidi_it->ignore_bn_limit == -1 /* only if not already known */ |
1215 | && bidi_it->charpos < eob /* not already at EOB */ | |
1216 | && bidi_explicit_dir_char (bidi_char_at_pos (bidi_it->bytepos | |
1217 | + bidi_it->ch_len, | |
1218 | bidi_it->string.s))) | |
b7b65b15 EZ |
1219 | { |
1220 | /* Avoid pushing and popping embedding levels if the level run | |
1221 | is empty, as this breaks level runs where it shouldn't. | |
1222 | UAX#9 removes all the explicit embedding and override codes, | |
1223 | so empty embeddings disappear without a trace. We need to | |
1224 | behave as if we did the same. */ | |
1225 | struct bidi_it saved_it; | |
1226 | int level = prev_level; | |
1227 | ||
1228 | bidi_copy_it (&saved_it, bidi_it); | |
1229 | ||
87e67904 EZ |
1230 | while (bidi_explicit_dir_char (bidi_char_at_pos (bidi_it->bytepos |
1231 | + bidi_it->ch_len, | |
1232 | bidi_it->string.s))) | |
b7b65b15 | 1233 | { |
182ce2d2 EZ |
1234 | /* This advances to the next character, skipping any |
1235 | characters covered by display strings. */ | |
b7b65b15 EZ |
1236 | level = bidi_resolve_explicit_1 (bidi_it); |
1237 | } | |
1238 | ||
102ebb00 EZ |
1239 | if (bidi_it->nchars <= 0) |
1240 | abort (); | |
b7b65b15 | 1241 | if (level == prev_level) /* empty embedding */ |
182ce2d2 | 1242 | saved_it.ignore_bn_limit = bidi_it->charpos + bidi_it->nchars; |
b7b65b15 | 1243 | else /* this embedding is non-empty */ |
87e67904 | 1244 | saved_it.ignore_bn_limit = -2; |
b7b65b15 EZ |
1245 | |
1246 | bidi_copy_it (bidi_it, &saved_it); | |
87e67904 | 1247 | if (bidi_it->ignore_bn_limit > -1) |
b7b65b15 EZ |
1248 | { |
1249 | /* We pushed a level, but we shouldn't have. Undo that. */ | |
1250 | if (!bidi_it->invalid_rl_levels) | |
1251 | { | |
1252 | new_level = bidi_pop_embedding_level (bidi_it); | |
1253 | bidi_it->invalid_rl_levels = -1; | |
1254 | if (bidi_it->invalid_levels) | |
1255 | bidi_it->invalid_levels--; | |
1256 | } | |
1257 | if (!bidi_it->invalid_levels) | |
1258 | new_level = bidi_pop_embedding_level (bidi_it); | |
1259 | else | |
1260 | { | |
1261 | bidi_it->invalid_levels--; | |
1262 | bidi_it->invalid_rl_levels--; | |
1263 | } | |
1264 | } | |
1265 | } | |
1266 | ||
b7b65b15 EZ |
1267 | if (bidi_it->type == NEUTRAL_B) /* X8 */ |
1268 | { | |
21fce5ab | 1269 | bidi_set_paragraph_end (bidi_it); |
6bff6497 EZ |
1270 | /* This is needed by bidi_resolve_weak below, and in L1. */ |
1271 | bidi_it->type_after_w1 = bidi_it->type; | |
89d3374a | 1272 | bidi_check_type (bidi_it->type_after_w1); |
b7b65b15 EZ |
1273 | } |
1274 | ||
1275 | return new_level; | |
1276 | } | |
1277 | ||
182ce2d2 EZ |
1278 | /* Advance in the buffer/string, resolve weak types and return the |
1279 | type of the next character after weak type resolution. */ | |
fd3998ff | 1280 | static bidi_type_t |
b7b65b15 EZ |
1281 | bidi_resolve_weak (struct bidi_it *bidi_it) |
1282 | { | |
1283 | bidi_type_t type; | |
1284 | bidi_dir_t override; | |
1285 | int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level; | |
1286 | int new_level = bidi_resolve_explicit (bidi_it); | |
1287 | int next_char; | |
1288 | bidi_type_t type_of_next; | |
1289 | struct bidi_it saved_it; | |
87e67904 | 1290 | EMACS_INT eob = bidi_it->string.s ? bidi_it->string.schars : ZV; |
b7b65b15 EZ |
1291 | |
1292 | type = bidi_it->type; | |
1293 | override = bidi_it->level_stack[bidi_it->stack_idx].override; | |
1294 | ||
1295 | if (type == UNKNOWN_BT | |
1296 | || type == LRE | |
1297 | || type == LRO | |
1298 | || type == RLE | |
1299 | || type == RLO | |
1300 | || type == PDF) | |
1301 | abort (); | |
1302 | ||
1303 | if (new_level != prev_level | |
1304 | || bidi_it->type == NEUTRAL_B) | |
1305 | { | |
1306 | /* We've got a new embedding level run, compute the directional | |
1307 | type of sor and initialize per-run variables (UAX#9, clause | |
1308 | X10). */ | |
1309 | bidi_set_sor_type (bidi_it, prev_level, new_level); | |
1310 | } | |
1311 | else if (type == NEUTRAL_S || type == NEUTRAL_WS | |
1312 | || type == WEAK_BN || type == STRONG_AL) | |
89d3374a EZ |
1313 | bidi_it->type_after_w1 = type; /* needed in L1 */ |
1314 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 EZ |
1315 | |
1316 | /* Level and directional override status are already recorded in | |
1317 | bidi_it, and do not need any change; see X6. */ | |
1318 | if (override == R2L) /* X6 */ | |
1319 | type = STRONG_R; | |
1320 | else if (override == L2R) | |
1321 | type = STRONG_L; | |
bc5a45f3 | 1322 | else |
b7b65b15 | 1323 | { |
bc5a45f3 | 1324 | if (type == WEAK_NSM) /* W1 */ |
b7b65b15 | 1325 | { |
bc5a45f3 | 1326 | /* Note that we don't need to consider the case where the |
5930fe97 EZ |
1327 | prev character has its type overridden by an RLO or LRO, |
1328 | because then either the type of this NSM would have been | |
1329 | also overridden, or the previous character is outside the | |
1330 | current level run, and thus not relevant to this NSM. | |
1331 | This is why NSM gets the type_after_w1 of the previous | |
1332 | character. */ | |
ebb5722e EZ |
1333 | if (bidi_it->prev.type_after_w1 != UNKNOWN_BT |
1334 | /* if type_after_w1 is NEUTRAL_B, this NSM is at sor */ | |
1335 | && bidi_it->prev.type_after_w1 != NEUTRAL_B) | |
5930fe97 | 1336 | type = bidi_it->prev.type_after_w1; |
bc5a45f3 EZ |
1337 | else if (bidi_it->sor == R2L) |
1338 | type = STRONG_R; | |
1339 | else if (bidi_it->sor == L2R) | |
1340 | type = STRONG_L; | |
1341 | else /* shouldn't happen! */ | |
1342 | abort (); | |
b7b65b15 | 1343 | } |
bc5a45f3 EZ |
1344 | if (type == WEAK_EN /* W2 */ |
1345 | && bidi_it->last_strong.type_after_w1 == STRONG_AL) | |
1346 | type = WEAK_AN; | |
1347 | else if (type == STRONG_AL) /* W3 */ | |
1348 | type = STRONG_R; | |
1349 | else if ((type == WEAK_ES /* W4 */ | |
1350 | && bidi_it->prev.type_after_w1 == WEAK_EN | |
1351 | && bidi_it->prev.orig_type == WEAK_EN) | |
1352 | || (type == WEAK_CS | |
1353 | && ((bidi_it->prev.type_after_w1 == WEAK_EN | |
1354 | && bidi_it->prev.orig_type == WEAK_EN) | |
1355 | || bidi_it->prev.type_after_w1 == WEAK_AN))) | |
b7b65b15 | 1356 | { |
e7402cb2 | 1357 | next_char = |
87e67904 EZ |
1358 | bidi_it->charpos + bidi_it->nchars >= eob |
1359 | ? BIDI_EOB | |
1360 | : bidi_char_at_pos (bidi_it->bytepos + bidi_it->ch_len, | |
1361 | bidi_it->string.s); | |
6bff6497 | 1362 | type_of_next = bidi_get_type (next_char, override); |
b7b65b15 | 1363 | |
bc5a45f3 | 1364 | if (type_of_next == WEAK_BN |
b7b65b15 EZ |
1365 | || bidi_explicit_dir_char (next_char)) |
1366 | { | |
1367 | bidi_copy_it (&saved_it, bidi_it); | |
1368 | while (bidi_resolve_explicit (bidi_it) == new_level | |
bc5a45f3 | 1369 | && bidi_it->type == WEAK_BN) |
b7b65b15 EZ |
1370 | ; |
1371 | type_of_next = bidi_it->type; | |
b7b65b15 EZ |
1372 | bidi_copy_it (bidi_it, &saved_it); |
1373 | } | |
bc5a45f3 EZ |
1374 | |
1375 | /* If the next character is EN, but the last strong-type | |
1376 | character is AL, that next EN will be changed to AN when | |
1377 | we process it in W2 above. So in that case, this ES | |
1378 | should not be changed into EN. */ | |
1379 | if (type == WEAK_ES | |
1380 | && type_of_next == WEAK_EN | |
1381 | && bidi_it->last_strong.type_after_w1 != STRONG_AL) | |
1382 | type = WEAK_EN; | |
1383 | else if (type == WEAK_CS) | |
b7b65b15 | 1384 | { |
bc5a45f3 EZ |
1385 | if (bidi_it->prev.type_after_w1 == WEAK_AN |
1386 | && (type_of_next == WEAK_AN | |
1387 | /* If the next character is EN, but the last | |
1388 | strong-type character is AL, EN will be later | |
1389 | changed to AN when we process it in W2 above. | |
1390 | So in that case, this ES should not be | |
1391 | changed into EN. */ | |
1392 | || (type_of_next == WEAK_EN | |
1393 | && bidi_it->last_strong.type_after_w1 == STRONG_AL))) | |
1394 | type = WEAK_AN; | |
1395 | else if (bidi_it->prev.type_after_w1 == WEAK_EN | |
1396 | && type_of_next == WEAK_EN | |
1397 | && bidi_it->last_strong.type_after_w1 != STRONG_AL) | |
1398 | type = WEAK_EN; | |
1399 | } | |
1400 | } | |
1401 | else if (type == WEAK_ET /* W5: ET with EN before or after it */ | |
1402 | || type == WEAK_BN) /* W5/Retaining */ | |
1403 | { | |
1404 | if (bidi_it->prev.type_after_w1 == WEAK_EN /* ET/BN w/EN before it */ | |
1405 | || bidi_it->next_en_pos > bidi_it->charpos) | |
1406 | type = WEAK_EN; | |
1407 | else /* W5: ET/BN with EN after it. */ | |
1408 | { | |
182ce2d2 | 1409 | EMACS_INT en_pos = bidi_it->charpos + bidi_it->nchars; |
bc5a45f3 | 1410 | |
102ebb00 EZ |
1411 | if (bidi_it->nchars <= 0) |
1412 | abort (); | |
bc5a45f3 | 1413 | next_char = |
87e67904 EZ |
1414 | bidi_it->charpos + bidi_it->nchars >= eob |
1415 | ? BIDI_EOB | |
1416 | : bidi_char_at_pos (bidi_it->bytepos + bidi_it->ch_len, | |
1417 | bidi_it->string.s); | |
bc5a45f3 EZ |
1418 | type_of_next = bidi_get_type (next_char, override); |
1419 | ||
1420 | if (type_of_next == WEAK_ET | |
1421 | || type_of_next == WEAK_BN | |
1422 | || bidi_explicit_dir_char (next_char)) | |
1423 | { | |
1424 | bidi_copy_it (&saved_it, bidi_it); | |
1425 | while (bidi_resolve_explicit (bidi_it) == new_level | |
1426 | && (bidi_it->type == WEAK_BN | |
1427 | || bidi_it->type == WEAK_ET)) | |
1428 | ; | |
1429 | type_of_next = bidi_it->type; | |
1430 | en_pos = bidi_it->charpos; | |
1431 | bidi_copy_it (bidi_it, &saved_it); | |
1432 | } | |
1433 | if (type_of_next == WEAK_EN) | |
b7b65b15 | 1434 | { |
bc5a45f3 EZ |
1435 | /* If the last strong character is AL, the EN we've |
1436 | found will become AN when we get to it (W2). */ | |
1437 | if (bidi_it->last_strong.type_after_w1 != STRONG_AL) | |
1438 | { | |
1439 | type = WEAK_EN; | |
1440 | /* Remember this EN position, to speed up processing | |
1441 | of the next ETs. */ | |
1442 | bidi_it->next_en_pos = en_pos; | |
1443 | } | |
1444 | else if (type == WEAK_BN) | |
1445 | type = NEUTRAL_ON; /* W6/Retaining */ | |
b7b65b15 | 1446 | } |
b7b65b15 EZ |
1447 | } |
1448 | } | |
1449 | } | |
1450 | ||
1451 | if (type == WEAK_ES || type == WEAK_ET || type == WEAK_CS /* W6 */ | |
89d3374a EZ |
1452 | || (type == WEAK_BN |
1453 | && (bidi_it->prev.type_after_w1 == WEAK_CS /* W6/Retaining */ | |
1454 | || bidi_it->prev.type_after_w1 == WEAK_ES | |
1455 | || bidi_it->prev.type_after_w1 == WEAK_ET))) | |
b7b65b15 EZ |
1456 | type = NEUTRAL_ON; |
1457 | ||
1458 | /* Store the type we've got so far, before we clobber it with strong | |
1459 | types in W7 and while resolving neutral types. But leave alone | |
1460 | the original types that were recorded above, because we will need | |
1461 | them for the L1 clause. */ | |
89d3374a EZ |
1462 | if (bidi_it->type_after_w1 == UNKNOWN_BT) |
1463 | bidi_it->type_after_w1 = type; | |
1464 | bidi_check_type (bidi_it->type_after_w1); | |
b7b65b15 EZ |
1465 | |
1466 | if (type == WEAK_EN) /* W7 */ | |
1467 | { | |
89d3374a | 1468 | if ((bidi_it->last_strong.type_after_w1 == STRONG_L) |
b7b65b15 EZ |
1469 | || (bidi_it->last_strong.type == UNKNOWN_BT && bidi_it->sor == L2R)) |
1470 | type = STRONG_L; | |
1471 | } | |
1472 | ||
1473 | bidi_it->type = type; | |
2d6e4628 | 1474 | bidi_check_type (bidi_it->type); |
b7b65b15 EZ |
1475 | return type; |
1476 | } | |
1477 | ||
fd3998ff | 1478 | static bidi_type_t |
b7b65b15 EZ |
1479 | bidi_resolve_neutral (struct bidi_it *bidi_it) |
1480 | { | |
1481 | int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level; | |
1482 | bidi_type_t type = bidi_resolve_weak (bidi_it); | |
1483 | int current_level = bidi_it->level_stack[bidi_it->stack_idx].level; | |
1484 | ||
1485 | if (!(type == STRONG_R | |
1486 | || type == STRONG_L | |
1487 | || type == WEAK_BN | |
1488 | || type == WEAK_EN | |
1489 | || type == WEAK_AN | |
1490 | || type == NEUTRAL_B | |
1491 | || type == NEUTRAL_S | |
1492 | || type == NEUTRAL_WS | |
1493 | || type == NEUTRAL_ON)) | |
1494 | abort (); | |
1495 | ||
1496 | if (bidi_get_category (type) == NEUTRAL | |
1497 | || (type == WEAK_BN && prev_level == current_level)) | |
1498 | { | |
1499 | if (bidi_it->next_for_neutral.type != UNKNOWN_BT) | |
1500 | type = bidi_resolve_neutral_1 (bidi_it->prev_for_neutral.type, | |
1501 | bidi_it->next_for_neutral.type, | |
1502 | current_level); | |
1503 | else | |
1504 | { | |
1505 | /* Arrrgh!! The UAX#9 algorithm is too deeply entrenched in | |
1506 | the assumption of batch-style processing; see clauses W4, | |
1507 | W5, and especially N1, which require to look far forward | |
182ce2d2 EZ |
1508 | (as well as back) in the buffer/string. May the fleas of |
1509 | a thousand camels infest the armpits of those who design | |
b7b65b15 EZ |
1510 | supposedly general-purpose algorithms by looking at their |
1511 | own implementations, and fail to consider other possible | |
1512 | implementations! */ | |
1513 | struct bidi_it saved_it; | |
1514 | bidi_type_t next_type; | |
1515 | ||
1516 | if (bidi_it->scan_dir == -1) | |
1517 | abort (); | |
1518 | ||
1519 | bidi_copy_it (&saved_it, bidi_it); | |
1520 | /* Scan the text forward until we find the first non-neutral | |
1521 | character, and then use that to resolve the neutral we | |
1522 | are dealing with now. We also cache the scanned iterator | |
1523 | states, to salvage some of the effort later. */ | |
1524 | bidi_cache_iterator_state (bidi_it, 0); | |
1525 | do { | |
1526 | /* Record the info about the previous character, so that | |
1527 | it will be cached below with this state. */ | |
89d3374a | 1528 | if (bidi_it->type_after_w1 != WEAK_BN /* W1/Retaining */ |
b7b65b15 EZ |
1529 | && bidi_it->type != WEAK_BN) |
1530 | bidi_remember_char (&bidi_it->prev, bidi_it); | |
1531 | type = bidi_resolve_weak (bidi_it); | |
1532 | /* Paragraph separators have their levels fully resolved | |
1533 | at this point, so cache them as resolved. */ | |
1534 | bidi_cache_iterator_state (bidi_it, type == NEUTRAL_B); | |
1535 | /* FIXME: implement L1 here, by testing for a newline and | |
1536 | resetting the level for any sequence of whitespace | |
1537 | characters adjacent to it. */ | |
1538 | } while (!(type == NEUTRAL_B | |
1539 | || (type != WEAK_BN | |
1540 | && bidi_get_category (type) != NEUTRAL) | |
1541 | /* This is all per level run, so stop when we | |
1542 | reach the end of this level run. */ | |
1543 | || bidi_it->level_stack[bidi_it->stack_idx].level != | |
1544 | current_level)); | |
1545 | ||
1546 | bidi_remember_char (&saved_it.next_for_neutral, bidi_it); | |
1547 | ||
1548 | switch (type) | |
1549 | { | |
1550 | case STRONG_L: | |
1551 | case STRONG_R: | |
1552 | case STRONG_AL: | |
1553 | next_type = type; | |
1554 | break; | |
1555 | case WEAK_EN: | |
1556 | case WEAK_AN: | |
1557 | /* N1: ``European and Arabic numbers are treated as | |
1558 | though they were R.'' */ | |
1559 | next_type = STRONG_R; | |
1560 | saved_it.next_for_neutral.type = STRONG_R; | |
1561 | break; | |
1562 | case WEAK_BN: | |
1563 | if (!bidi_explicit_dir_char (bidi_it->ch)) | |
1564 | abort (); /* can't happen: BNs are skipped */ | |
1565 | /* FALLTHROUGH */ | |
1566 | case NEUTRAL_B: | |
1567 | /* Marched all the way to the end of this level run. | |
1568 | We need to use the eor type, whose information is | |
1569 | stored by bidi_set_sor_type in the prev_for_neutral | |
1570 | member. */ | |
1571 | if (saved_it.type != WEAK_BN | |
89d3374a | 1572 | || bidi_get_category (bidi_it->prev.type_after_w1) == NEUTRAL) |
b7b65b15 EZ |
1573 | { |
1574 | next_type = bidi_it->prev_for_neutral.type; | |
1575 | saved_it.next_for_neutral.type = next_type; | |
2d6e4628 | 1576 | bidi_check_type (next_type); |
b7b65b15 EZ |
1577 | } |
1578 | else | |
1579 | { | |
1580 | /* This is a BN which does not adjoin neutrals. | |
1581 | Leave its type alone. */ | |
1582 | bidi_copy_it (bidi_it, &saved_it); | |
1583 | return bidi_it->type; | |
1584 | } | |
1585 | break; | |
1586 | default: | |
1587 | abort (); | |
1588 | } | |
1589 | type = bidi_resolve_neutral_1 (saved_it.prev_for_neutral.type, | |
1590 | next_type, current_level); | |
1591 | saved_it.type = type; | |
2d6e4628 | 1592 | bidi_check_type (type); |
b7b65b15 EZ |
1593 | bidi_copy_it (bidi_it, &saved_it); |
1594 | } | |
1595 | } | |
1596 | return type; | |
1597 | } | |
1598 | ||
1599 | /* Given an iterator state in BIDI_IT, advance one character position | |
182ce2d2 EZ |
1600 | in the buffer/string to the next character (in the logical order), |
1601 | resolve the bidi type of that next character, and return that | |
1602 | type. */ | |
fd3998ff | 1603 | static bidi_type_t |
b7b65b15 EZ |
1604 | bidi_type_of_next_char (struct bidi_it *bidi_it) |
1605 | { | |
1606 | bidi_type_t type; | |
1607 | ||
1608 | /* This should always be called during a forward scan. */ | |
1609 | if (bidi_it->scan_dir != 1) | |
1610 | abort (); | |
1611 | ||
1612 | /* Reset the limit until which to ignore BNs if we step out of the | |
1613 | area where we found only empty levels. */ | |
87e67904 | 1614 | if ((bidi_it->ignore_bn_limit > -1 |
b7b65b15 | 1615 | && bidi_it->ignore_bn_limit <= bidi_it->charpos) |
87e67904 | 1616 | || (bidi_it->ignore_bn_limit == -2 |
b7b65b15 | 1617 | && !bidi_explicit_dir_char (bidi_it->ch))) |
87e67904 | 1618 | bidi_it->ignore_bn_limit = -1; |
b7b65b15 EZ |
1619 | |
1620 | type = bidi_resolve_neutral (bidi_it); | |
1621 | ||
1622 | return type; | |
1623 | } | |
1624 | ||
1625 | /* Given an iterator state BIDI_IT, advance one character position in | |
182ce2d2 EZ |
1626 | the buffer/string to the next character (in the current scan |
1627 | direction), resolve the embedding and implicit levels of that next | |
1628 | character, and return the resulting level. */ | |
fd3998ff | 1629 | static int |
b7b65b15 EZ |
1630 | bidi_level_of_next_char (struct bidi_it *bidi_it) |
1631 | { | |
1632 | bidi_type_t type; | |
1633 | int level, prev_level = -1; | |
1634 | struct bidi_saved_info next_for_neutral; | |
87e67904 | 1635 | EMACS_INT next_char_pos = -1; |
b7b65b15 EZ |
1636 | |
1637 | if (bidi_it->scan_dir == 1) | |
1638 | { | |
1639 | /* There's no sense in trying to advance if we hit end of text. */ | |
87e67904 | 1640 | if (bidi_it->charpos >= (bidi_it->string.s ? bidi_it->string.schars : ZV)) |
b7b65b15 EZ |
1641 | return bidi_it->resolved_level; |
1642 | ||
1643 | /* Record the info about the previous character. */ | |
89d3374a | 1644 | if (bidi_it->type_after_w1 != WEAK_BN /* W1/Retaining */ |
b7b65b15 EZ |
1645 | && bidi_it->type != WEAK_BN) |
1646 | bidi_remember_char (&bidi_it->prev, bidi_it); | |
89d3374a EZ |
1647 | if (bidi_it->type_after_w1 == STRONG_R |
1648 | || bidi_it->type_after_w1 == STRONG_L | |
1649 | || bidi_it->type_after_w1 == STRONG_AL) | |
b7b65b15 EZ |
1650 | bidi_remember_char (&bidi_it->last_strong, bidi_it); |
1651 | /* FIXME: it sounds like we don't need both prev and | |
1652 | prev_for_neutral members, but I'm leaving them both for now. */ | |
1653 | if (bidi_it->type == STRONG_R || bidi_it->type == STRONG_L | |
1654 | || bidi_it->type == WEAK_EN || bidi_it->type == WEAK_AN) | |
1655 | bidi_remember_char (&bidi_it->prev_for_neutral, bidi_it); | |
1656 | ||
1657 | /* If we overstepped the characters used for resolving neutrals | |
1658 | and whitespace, invalidate their info in the iterator. */ | |
1659 | if (bidi_it->charpos >= bidi_it->next_for_neutral.charpos) | |
1660 | bidi_it->next_for_neutral.type = UNKNOWN_BT; | |
1661 | if (bidi_it->next_en_pos >= 0 | |
1662 | && bidi_it->charpos >= bidi_it->next_en_pos) | |
1663 | bidi_it->next_en_pos = -1; | |
1664 | if (bidi_it->next_for_ws.type != UNKNOWN_BT | |
1665 | && bidi_it->charpos >= bidi_it->next_for_ws.charpos) | |
1666 | bidi_it->next_for_ws.type = UNKNOWN_BT; | |
1667 | ||
1668 | /* This must be taken before we fill the iterator with the info | |
1669 | about the next char. If we scan backwards, the iterator | |
1670 | state must be already cached, so there's no need to know the | |
1671 | embedding level of the previous character, since we will be | |
1672 | returning to our caller shortly. */ | |
1673 | prev_level = bidi_it->level_stack[bidi_it->stack_idx].level; | |
1674 | } | |
1675 | next_for_neutral = bidi_it->next_for_neutral; | |
1676 | ||
182ce2d2 EZ |
1677 | /* Perhaps the character we want is already cached. If it is, the |
1678 | call to bidi_cache_find below will return a type other than | |
1679 | UNKNOWN_BT. */ | |
87e67904 | 1680 | if (bidi_cache_idx > bidi_cache_start && !bidi_it->first_elt) |
102ebb00 EZ |
1681 | { |
1682 | if (bidi_it->scan_dir > 0) | |
1683 | { | |
1684 | if (bidi_it->nchars <= 0) | |
1685 | abort (); | |
1686 | next_char_pos = bidi_it->charpos + bidi_it->nchars; | |
1687 | } | |
bb269206 EZ |
1688 | else if (bidi_it->charpos >= (bidi_it->string.s ? 0 : 1)) |
1689 | /* Implementation note: we allow next_char_pos to be as low as | |
1690 | 0 for buffers or -1 for strings, and that is okay because | |
1691 | that's the "position" of the sentinel iterator state we | |
1692 | cached at the beginning of the iteration. */ | |
102ebb00 | 1693 | next_char_pos = bidi_it->charpos - 1; |
87e67904 EZ |
1694 | if (next_char_pos >= 0) |
1695 | type = bidi_cache_find (next_char_pos, -1, bidi_it); | |
1696 | else | |
1697 | type = UNKNOWN_BT; | |
102ebb00 | 1698 | } |
182ce2d2 | 1699 | else |
102ebb00 | 1700 | type = UNKNOWN_BT; |
b7b65b15 EZ |
1701 | if (type != UNKNOWN_BT) |
1702 | { | |
1703 | /* Don't lose the information for resolving neutrals! The | |
1704 | cached states could have been cached before their | |
1705 | next_for_neutral member was computed. If we are on our way | |
1706 | forward, we can simply take the info from the previous | |
1707 | state. */ | |
1708 | if (bidi_it->scan_dir == 1 | |
1709 | && bidi_it->next_for_neutral.type == UNKNOWN_BT) | |
1710 | bidi_it->next_for_neutral = next_for_neutral; | |
1711 | ||
1712 | /* If resolved_level is -1, it means this state was cached | |
1713 | before it was completely resolved, so we cannot return | |
1714 | it. */ | |
1715 | if (bidi_it->resolved_level != -1) | |
1716 | return bidi_it->resolved_level; | |
1717 | } | |
1718 | if (bidi_it->scan_dir == -1) | |
1719 | /* If we are going backwards, the iterator state is already cached | |
1720 | from previous scans, and should be fully resolved. */ | |
1721 | abort (); | |
1722 | ||
1723 | if (type == UNKNOWN_BT) | |
1724 | type = bidi_type_of_next_char (bidi_it); | |
1725 | ||
1726 | if (type == NEUTRAL_B) | |
1727 | return bidi_it->resolved_level; | |
1728 | ||
1729 | level = bidi_it->level_stack[bidi_it->stack_idx].level; | |
1730 | if ((bidi_get_category (type) == NEUTRAL /* && type != NEUTRAL_B */) | |
1731 | || (type == WEAK_BN && prev_level == level)) | |
1732 | { | |
1733 | if (bidi_it->next_for_neutral.type == UNKNOWN_BT) | |
1734 | abort (); | |
1735 | ||
1736 | /* If the cached state shows a neutral character, it was not | |
1737 | resolved by bidi_resolve_neutral, so do it now. */ | |
1738 | type = bidi_resolve_neutral_1 (bidi_it->prev_for_neutral.type, | |
1739 | bidi_it->next_for_neutral.type, | |
1740 | level); | |
1741 | } | |
1742 | ||
1743 | if (!(type == STRONG_R | |
1744 | || type == STRONG_L | |
1745 | || type == WEAK_BN | |
1746 | || type == WEAK_EN | |
1747 | || type == WEAK_AN)) | |
1748 | abort (); | |
1749 | bidi_it->type = type; | |
2d6e4628 | 1750 | bidi_check_type (bidi_it->type); |
b7b65b15 EZ |
1751 | |
1752 | /* For L1 below, we need to know, for each WS character, whether | |
0d327994 | 1753 | it belongs to a sequence of WS characters preceding a newline |
b7b65b15 | 1754 | or a TAB or a paragraph separator. */ |
89d3374a | 1755 | if (bidi_it->orig_type == NEUTRAL_WS |
b7b65b15 EZ |
1756 | && bidi_it->next_for_ws.type == UNKNOWN_BT) |
1757 | { | |
1758 | int ch; | |
9d68c2a9 | 1759 | EMACS_INT clen = bidi_it->ch_len; |
6bff6497 EZ |
1760 | EMACS_INT bpos = bidi_it->bytepos; |
1761 | EMACS_INT cpos = bidi_it->charpos; | |
182ce2d2 | 1762 | EMACS_INT disp_pos = bidi_it->disp_pos; |
102ebb00 | 1763 | EMACS_INT nc = bidi_it->nchars; |
87e67904 | 1764 | struct bidi_string_data bs = bidi_it->string; |
b7b65b15 | 1765 | bidi_type_t chtype; |
fc6f18ce | 1766 | int fwp = bidi_it->frame_window_p; |
b7b65b15 | 1767 | |
102ebb00 EZ |
1768 | if (bidi_it->nchars <= 0) |
1769 | abort (); | |
b7b65b15 | 1770 | do { |
87e67904 | 1771 | ch = bidi_fetch_char (bpos += clen, cpos += nc, &disp_pos, &bs, fwp, |
fc6f18ce | 1772 | &clen, &nc); |
e7402cb2 | 1773 | if (ch == '\n' || ch == BIDI_EOB /* || ch == LINESEP_CHAR */) |
b7b65b15 EZ |
1774 | chtype = NEUTRAL_B; |
1775 | else | |
6bff6497 | 1776 | chtype = bidi_get_type (ch, NEUTRAL_DIR); |
b7b65b15 EZ |
1777 | } while (chtype == NEUTRAL_WS || chtype == WEAK_BN |
1778 | || bidi_explicit_dir_char (ch)); /* L1/Retaining */ | |
1779 | bidi_it->next_for_ws.type = chtype; | |
2d6e4628 | 1780 | bidi_check_type (bidi_it->next_for_ws.type); |
b7b65b15 EZ |
1781 | bidi_it->next_for_ws.charpos = cpos; |
1782 | bidi_it->next_for_ws.bytepos = bpos; | |
1783 | } | |
1784 | ||
1785 | /* Resolve implicit levels, with a twist: PDFs get the embedding | |
1786 | level of the enbedding they terminate. See below for the | |
1787 | reason. */ | |
89d3374a | 1788 | if (bidi_it->orig_type == PDF |
b7b65b15 EZ |
1789 | /* Don't do this if this formatting code didn't change the |
1790 | embedding level due to invalid or empty embeddings. */ | |
1791 | && prev_level != level) | |
1792 | { | |
1793 | /* Don't look in UAX#9 for the reason for this: it's our own | |
1794 | private quirk. The reason is that we want the formatting | |
1795 | codes to be delivered so that they bracket the text of their | |
1796 | embedding. For example, given the text | |
1797 | ||
1798 | {RLO}teST{PDF} | |
1799 | ||
1800 | we want it to be displayed as | |
1801 | ||
0d68907d | 1802 | {PDF}STet{RLO} |
b7b65b15 EZ |
1803 | |
1804 | not as | |
1805 | ||
1806 | STet{RLO}{PDF} | |
1807 | ||
1808 | which will result because we bump up the embedding level as | |
1809 | soon as we see the RLO and pop it as soon as we see the PDF, | |
1810 | so RLO itself has the same embedding level as "teST", and | |
1811 | thus would be normally delivered last, just before the PDF. | |
1812 | The switch below fiddles with the level of PDF so that this | |
1813 | ugly side effect does not happen. | |
1814 | ||
1815 | (This is, of course, only important if the formatting codes | |
e7402cb2 EZ |
1816 | are actually displayed, but Emacs does need to display them |
1817 | if the user wants to.) */ | |
b7b65b15 EZ |
1818 | level = prev_level; |
1819 | } | |
89d3374a EZ |
1820 | else if (bidi_it->orig_type == NEUTRAL_B /* L1 */ |
1821 | || bidi_it->orig_type == NEUTRAL_S | |
e7402cb2 EZ |
1822 | || bidi_it->ch == '\n' || bidi_it->ch == BIDI_EOB |
1823 | /* || bidi_it->ch == LINESEP_CHAR */ | |
89d3374a | 1824 | || (bidi_it->orig_type == NEUTRAL_WS |
b7b65b15 EZ |
1825 | && (bidi_it->next_for_ws.type == NEUTRAL_B |
1826 | || bidi_it->next_for_ws.type == NEUTRAL_S))) | |
1827 | level = bidi_it->level_stack[0].level; | |
1828 | else if ((level & 1) == 0) /* I1 */ | |
1829 | { | |
1830 | if (type == STRONG_R) | |
1831 | level++; | |
1832 | else if (type == WEAK_EN || type == WEAK_AN) | |
1833 | level += 2; | |
1834 | } | |
1835 | else /* I2 */ | |
1836 | { | |
1837 | if (type == STRONG_L || type == WEAK_EN || type == WEAK_AN) | |
1838 | level++; | |
1839 | } | |
1840 | ||
1841 | bidi_it->resolved_level = level; | |
1842 | return level; | |
1843 | } | |
1844 | ||
1845 | /* Move to the other edge of a level given by LEVEL. If END_FLAG is | |
1846 | non-zero, we are at the end of a level, and we need to prepare to | |
1847 | resume the scan of the lower level. | |
1848 | ||
1849 | If this level's other edge is cached, we simply jump to it, filling | |
1850 | the iterator structure with the iterator state on the other edge. | |
182ce2d2 EZ |
1851 | Otherwise, we walk the buffer or string until we come back to the |
1852 | same level as LEVEL. | |
b7b65b15 EZ |
1853 | |
1854 | Note: we are not talking here about a ``level run'' in the UAX#9 | |
1855 | sense of the term, but rather about a ``level'' which includes | |
1856 | all the levels higher than it. In other words, given the levels | |
1857 | like this: | |
1858 | ||
1859 | 11111112222222333333334443343222222111111112223322111 | |
1860 | A B C | |
1861 | ||
1862 | and assuming we are at point A scanning left to right, this | |
1863 | function moves to point C, whereas the UAX#9 ``level 2 run'' ends | |
1864 | at point B. */ | |
1865 | static void | |
1866 | bidi_find_other_level_edge (struct bidi_it *bidi_it, int level, int end_flag) | |
1867 | { | |
1868 | int dir = end_flag ? -bidi_it->scan_dir : bidi_it->scan_dir; | |
1869 | int idx; | |
1870 | ||
1871 | /* Try the cache first. */ | |
87e67904 EZ |
1872 | if ((idx = bidi_cache_find_level_change (level, dir, end_flag)) |
1873 | >= bidi_cache_start) | |
b7b65b15 EZ |
1874 | bidi_cache_fetch_state (idx, bidi_it); |
1875 | else | |
1876 | { | |
1877 | int new_level; | |
1878 | ||
1879 | if (end_flag) | |
1880 | abort (); /* if we are at end of level, its edges must be cached */ | |
1881 | ||
1882 | bidi_cache_iterator_state (bidi_it, 1); | |
1883 | do { | |
1884 | new_level = bidi_level_of_next_char (bidi_it); | |
1885 | bidi_cache_iterator_state (bidi_it, 1); | |
1886 | } while (new_level >= level); | |
1887 | } | |
1888 | } | |
1889 | ||
1890 | void | |
4b292a22 | 1891 | bidi_move_to_visually_next (struct bidi_it *bidi_it) |
b7b65b15 EZ |
1892 | { |
1893 | int old_level, new_level, next_level; | |
9c82e145 | 1894 | struct bidi_it sentinel; |
b7b65b15 | 1895 | |
87e67904 EZ |
1896 | if (bidi_it->charpos < 0 || bidi_it->bytepos < 0) |
1897 | abort (); | |
1898 | ||
b7b65b15 EZ |
1899 | if (bidi_it->scan_dir == 0) |
1900 | { | |
1901 | bidi_it->scan_dir = 1; /* default to logical order */ | |
1902 | } | |
1903 | ||
be39f003 EZ |
1904 | /* If we just passed a newline, initialize for the next line. */ |
1905 | if (!bidi_it->first_elt && bidi_it->orig_type == NEUTRAL_B) | |
1906 | bidi_line_init (bidi_it); | |
1907 | ||
6dcfd253 EZ |
1908 | /* Prepare the sentinel iterator state, and cache it. When we bump |
1909 | into it, scanning backwards, we'll know that the last non-base | |
1910 | level is exhausted. */ | |
87e67904 | 1911 | if (bidi_cache_idx == bidi_cache_start) |
9c82e145 EZ |
1912 | { |
1913 | bidi_copy_it (&sentinel, bidi_it); | |
1914 | if (bidi_it->first_elt) | |
1915 | { | |
1916 | sentinel.charpos--; /* cached charpos needs to be monotonic */ | |
1917 | sentinel.bytepos--; | |
1918 | sentinel.ch = '\n'; /* doesn't matter, but why not? */ | |
1919 | sentinel.ch_len = 1; | |
182ce2d2 | 1920 | sentinel.nchars = 1; |
9c82e145 | 1921 | } |
6dcfd253 | 1922 | bidi_cache_iterator_state (&sentinel, 1); |
9c82e145 | 1923 | } |
b7b65b15 EZ |
1924 | |
1925 | old_level = bidi_it->resolved_level; | |
1926 | new_level = bidi_level_of_next_char (bidi_it); | |
b7b65b15 EZ |
1927 | |
1928 | /* Reordering of resolved levels (clause L2) is implemented by | |
1929 | jumping to the other edge of the level and flipping direction of | |
c0546589 | 1930 | scanning the text whenever we find a level change. */ |
b7b65b15 EZ |
1931 | if (new_level != old_level) |
1932 | { | |
1933 | int ascending = new_level > old_level; | |
1934 | int level_to_search = ascending ? old_level + 1 : old_level; | |
1935 | int incr = ascending ? 1 : -1; | |
1936 | int expected_next_level = old_level + incr; | |
1937 | ||
b7b65b15 EZ |
1938 | /* Jump (or walk) to the other edge of this level. */ |
1939 | bidi_find_other_level_edge (bidi_it, level_to_search, !ascending); | |
1940 | /* Switch scan direction and peek at the next character in the | |
1941 | new direction. */ | |
1942 | bidi_it->scan_dir = -bidi_it->scan_dir; | |
1943 | ||
1944 | /* The following loop handles the case where the resolved level | |
1945 | jumps by more than one. This is typical for numbers inside a | |
1946 | run of text with left-to-right embedding direction, but can | |
1947 | also happen in other situations. In those cases the decision | |
1948 | where to continue after a level change, and in what direction, | |
1949 | is tricky. For example, given a text like below: | |
1950 | ||
1951 | abcdefgh | |
1952 | 11336622 | |
1953 | ||
1954 | (where the numbers below the text show the resolved levels), | |
1955 | the result of reordering according to UAX#9 should be this: | |
1956 | ||
1957 | efdcghba | |
1958 | ||
1959 | This is implemented by the loop below which flips direction | |
1960 | and jumps to the other edge of the level each time it finds | |
1961 | the new level not to be the expected one. The expected level | |
1962 | is always one more or one less than the previous one. */ | |
1963 | next_level = bidi_peek_at_next_level (bidi_it); | |
1964 | while (next_level != expected_next_level) | |
1965 | { | |
1966 | expected_next_level += incr; | |
1967 | level_to_search += incr; | |
1968 | bidi_find_other_level_edge (bidi_it, level_to_search, !ascending); | |
1969 | bidi_it->scan_dir = -bidi_it->scan_dir; | |
1970 | next_level = bidi_peek_at_next_level (bidi_it); | |
1971 | } | |
1972 | ||
1973 | /* Finally, deliver the next character in the new direction. */ | |
1974 | next_level = bidi_level_of_next_char (bidi_it); | |
1975 | } | |
1976 | ||
b44d9321 EZ |
1977 | /* Take note when we have just processed the newline that precedes |
1978 | the end of the paragraph. The next time we are about to be | |
1979 | called, set_iterator_to_next will automatically reinit the | |
1980 | paragraph direction, if needed. We do this at the newline before | |
1981 | the paragraph separator, because the next character might not be | |
1982 | the first character of the next paragraph, due to the bidi | |
c0546589 EZ |
1983 | reordering, whereas we _must_ know the paragraph base direction |
1984 | _before_ we process the paragraph's text, since the base | |
1985 | direction affects the reordering. */ | |
87e67904 | 1986 | if (bidi_it->scan_dir == 1 && bidi_it->orig_type == NEUTRAL_B) |
be39f003 | 1987 | { |
87e67904 EZ |
1988 | /* The paragraph direction of the entire string, once |
1989 | determined, is in effect for the entire string. Setting the | |
1990 | separator limit to the end of the string prevents | |
1991 | bidi_paragraph_init from being called automatically on this | |
1992 | string. */ | |
1993 | if (bidi_it->string.s) | |
1994 | bidi_it->separator_limit = bidi_it->string.schars; | |
1995 | else if (bidi_it->bytepos < ZV_BYTE) | |
be39f003 | 1996 | { |
87e67904 EZ |
1997 | EMACS_INT sep_len = |
1998 | bidi_at_paragraph_end (bidi_it->charpos + bidi_it->nchars, | |
1999 | bidi_it->bytepos + bidi_it->ch_len); | |
2000 | if (bidi_it->nchars <= 0) | |
2001 | abort (); | |
2002 | if (sep_len >= 0) | |
2003 | { | |
2004 | bidi_it->new_paragraph = 1; | |
2005 | /* Record the buffer position of the last character of the | |
2006 | paragraph separator. */ | |
2007 | bidi_it->separator_limit = | |
2008 | bidi_it->charpos + bidi_it->nchars + sep_len; | |
2009 | } | |
be39f003 EZ |
2010 | } |
2011 | } | |
6bff6497 | 2012 | |
87e67904 | 2013 | if (bidi_it->scan_dir == 1 && bidi_cache_idx > bidi_cache_start) |
b7b65b15 EZ |
2014 | { |
2015 | /* If we are at paragraph's base embedding level and beyond the | |
2016 | last cached position, the cache's job is done and we can | |
2017 | discard it. */ | |
2018 | if (bidi_it->resolved_level == bidi_it->level_stack[0].level | |
182ce2d2 EZ |
2019 | && bidi_it->charpos > (bidi_cache[bidi_cache_idx - 1].charpos |
2020 | + bidi_cache[bidi_cache_idx - 1].nchars - 1)) | |
b7b65b15 EZ |
2021 | bidi_cache_reset (); |
2022 | /* But as long as we are caching during forward scan, we must | |
2023 | cache each state, or else the cache integrity will be | |
2024 | compromised: it assumes cached states correspond to buffer | |
2025 | positions 1:1. */ | |
2026 | else | |
2027 | bidi_cache_iterator_state (bidi_it, 1); | |
2028 | } | |
2029 | } | |
2030 | ||
2031 | /* This is meant to be called from within the debugger, whenever you | |
2032 | wish to examine the cache contents. */ | |
e78aecca | 2033 | void bidi_dump_cached_states (void) EXTERNALLY_VISIBLE; |
b7b65b15 EZ |
2034 | void |
2035 | bidi_dump_cached_states (void) | |
2036 | { | |
2037 | int i; | |
2038 | int ndigits = 1; | |
2039 | ||
2040 | if (bidi_cache_idx == 0) | |
2041 | { | |
2042 | fprintf (stderr, "The cache is empty.\n"); | |
2043 | return; | |
2044 | } | |
2045 | fprintf (stderr, "Total of %d state%s in cache:\n", | |
2046 | bidi_cache_idx, bidi_cache_idx == 1 ? "" : "s"); | |
2047 | ||
2048 | for (i = bidi_cache[bidi_cache_idx - 1].charpos; i > 0; i /= 10) | |
2049 | ndigits++; | |
2050 | fputs ("ch ", stderr); | |
2051 | for (i = 0; i < bidi_cache_idx; i++) | |
2052 | fprintf (stderr, "%*c", ndigits, bidi_cache[i].ch); | |
2053 | fputs ("\n", stderr); | |
2054 | fputs ("lvl ", stderr); | |
2055 | for (i = 0; i < bidi_cache_idx; i++) | |
2056 | fprintf (stderr, "%*d", ndigits, bidi_cache[i].resolved_level); | |
2057 | fputs ("\n", stderr); | |
2058 | fputs ("pos ", stderr); | |
2059 | for (i = 0; i < bidi_cache_idx; i++) | |
c2982e87 | 2060 | fprintf (stderr, "%*"pI"d", ndigits, bidi_cache[i].charpos); |
b7b65b15 EZ |
2061 | fputs ("\n", stderr); |
2062 | } |