bidi.c 56.7 KB
Newer Older
Eli Zaretskii's avatar
Eli Zaretskii committed
1
/* Low-level bidirectional buffer-scanning functions for GNU Emacs.
2 3
   Copyright (C) 2000, 2001, 2004, 2005, 2009, 2010
   Free Software Foundation, Inc.
Eli Zaretskii's avatar
Eli Zaretskii committed
4 5 6

This file is part of GNU Emacs.

Eli Zaretskii's avatar
Eli Zaretskii committed
7
GNU Emacs is free software: you can redistribute it and/or modify
Eli Zaretskii's avatar
Eli Zaretskii committed
8
it under the terms of the GNU General Public License as published by
Eli Zaretskii's avatar
Eli Zaretskii committed
9 10
the Free Software Foundation, either version 3 of the License, or
(at your option) any later version.
Eli Zaretskii's avatar
Eli Zaretskii committed
11 12 13 14 15 16 17

GNU Emacs is distributed in the hope that it will be useful,
but WITHOUT ANY WARRANTY; without even the implied warranty of
MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
GNU General Public License for more details.

You should have received a copy of the GNU General Public License
Eli Zaretskii's avatar
Eli Zaretskii committed
18
along with GNU Emacs.  If not, see <http://www.gnu.org/licenses/>.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
19

20 21 22
/* Written by Eli Zaretskii <eliz@gnu.org>.

   A sequential implementation of the Unicode Bidirectional algorithm,
Eli Zaretskii's avatar
Eli Zaretskii committed
23 24 25
   as per UAX#9, a part of the Unicode Standard.

   Unlike the reference and most other implementations, this one is
26 27
   designed to be called once for every character in the buffer or
   string.
Eli Zaretskii's avatar
Eli Zaretskii committed
28

29
   The main entry point is bidi_move_to_visually_next.  Each time it
Eli Zaretskii's avatar
Eli Zaretskii committed
30 31 32
   is called, it finds the next character in the visual order, and
   returns its information in a special structure.  The caller is then
   expected to process this character for display or any other
33 34 35
   purposes, and call bidi_move_to_visually_next for the next
   character.  See the comments in bidi_move_to_visually_next for more
   details about its algorithm that finds the next visual-order
Eli Zaretskii's avatar
Eli Zaretskii committed
36 37
   character by resolving their levels on the fly.

38 39 40 41 42
   The two other entry points are bidi_paragraph_init and
   bidi_mirror_char.  The first determines the base direction of a
   paragraph, while the second returns the mirrored version of its
   argument character.

43 44 45 46
   If you want to understand the code, you will have to read it
   together with the relevant portions of UAX#9.  The comments include
   references to UAX#9 rules, for that very reason.

Eli Zaretskii's avatar
Eli Zaretskii committed
47 48 49 50 51 52 53 54
   A note about references to UAX#9 rules: if the reference says
   something like "X9/Retaining", it means that you need to refer to
   rule X9 and to its modifications decribed in the "Implementation
   Notes" section of UAX#9, under "Retaining Format Codes".  */

#include <config.h>
#include <stdio.h>
#include <string.h>
55 56
#include <setjmp.h>

Eli Zaretskii's avatar
Eli Zaretskii committed
57 58 59 60 61 62 63
#include "lisp.h"
#include "buffer.h"
#include "character.h"
#include "dispextern.h"

static int bidi_initialized = 0;

64
static Lisp_Object bidi_type_table, bidi_mirror_table;
Eli Zaretskii's avatar
Eli Zaretskii committed
65

66
/* FIXME: Remove these when bidi_explicit_dir_char uses a lookup table.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
67 68 69 70 71 72 73 74 75
#define LRM_CHAR   0x200E
#define RLM_CHAR   0x200F
#define LRE_CHAR   0x202A
#define RLE_CHAR   0x202B
#define PDF_CHAR   0x202C
#define LRO_CHAR   0x202D
#define RLO_CHAR   0x202E

#define BIDI_EOB   -1
76
#define BIDI_BOB   -2		/* FIXME: Is this needed? */
Eli Zaretskii's avatar
Eli Zaretskii committed
77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97

/* Local data structures.  (Look in dispextern.h for the rest.)  */

/* What we need to know about the current paragraph.  */
struct bidi_paragraph_info {
  int start_bytepos;	/* byte position where it begins */
  int end_bytepos;	/* byte position where it ends */
  int embedding_level;	/* its basic embedding level */
  bidi_dir_t base_dir;	/* its base direction */
};

/* Data type for describing the bidirectional character categories.  */
typedef enum {
  UNKNOWN_BC,
  NEUTRAL,
  WEAK,
  STRONG
} bidi_category_t;

int bidi_ignore_explicit_marks_for_paragraph_level = 1;

Eli Zaretskii's avatar
Eli Zaretskii committed
98
static Lisp_Object paragraph_start_re, paragraph_separate_re;
99
static Lisp_Object Qparagraph_start, Qparagraph_separate;
Eli Zaretskii's avatar
Eli Zaretskii committed
100 101

static void
102
bidi_initialize (void)
Eli Zaretskii's avatar
Eli Zaretskii committed
103
{
104 105

#include "biditype.h"
106
#include "bidimirror.h"
107

Eli Zaretskii's avatar
Eli Zaretskii committed
108 109 110
  int i;

  bidi_type_table = Fmake_char_table (Qnil, make_number (STRONG_L));
111
  staticpro (&bidi_type_table);
Eli Zaretskii's avatar
Eli Zaretskii committed
112 113

  for (i = 0; i < sizeof bidi_type / sizeof bidi_type[0]; i++)
114
    char_table_set_range (bidi_type_table, bidi_type[i].from, bidi_type[i].to,
Eli Zaretskii's avatar
Eli Zaretskii committed
115
			  make_number (bidi_type[i].type));
116

117 118 119 120 121 122 123
  bidi_mirror_table = Fmake_char_table (Qnil, Qnil);
  staticpro (&bidi_mirror_table);

  for (i = 0; i < sizeof bidi_mirror / sizeof bidi_mirror[0]; i++)
    char_table_set (bidi_mirror_table, bidi_mirror[i].from,
		    make_number (bidi_mirror[i].to));

124 125
  Qparagraph_start = intern ("paragraph-start");
  staticpro (&Qparagraph_start);
Eli Zaretskii's avatar
Eli Zaretskii committed
126 127 128 129
  paragraph_start_re = Fsymbol_value (Qparagraph_start);
  if (!STRINGP (paragraph_start_re))
    paragraph_start_re = build_string ("\f\\|[ \t]*$");
  staticpro (&paragraph_start_re);
130 131
  Qparagraph_separate = intern ("paragraph-separate");
  staticpro (&Qparagraph_separate);
Eli Zaretskii's avatar
Eli Zaretskii committed
132 133 134 135
  paragraph_separate_re = Fsymbol_value (Qparagraph_separate);
  if (!STRINGP (paragraph_separate_re))
    paragraph_separate_re = build_string ("[ \t\f]*$");
  staticpro (&paragraph_separate_re);
Eli Zaretskii's avatar
Eli Zaretskii committed
136 137 138
  bidi_initialized = 1;
}

139 140
/* Return the bidi type of a character CH, subject to the current
   directional OVERRIDE.  */
141
static INLINE bidi_type_t
142
bidi_get_type (int ch, bidi_dir_t override)
Eli Zaretskii's avatar
Eli Zaretskii committed
143
{
144 145
  bidi_type_t default_type;

146 147
  if (ch == BIDI_EOB)
    return NEUTRAL_B;
148 149
  if (ch < 0 || ch > MAX_CHAR)
    abort ();
150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173 174 175 176 177 178 179 180 181

  default_type = (bidi_type_t) XINT (CHAR_TABLE_REF (bidi_type_table, ch));

  if (override == NEUTRAL_DIR)
    return default_type;

  switch (default_type)
    {
      /* Although UAX#9 does not tell, it doesn't make sense to
	 override NEUTRAL_B and LRM/RLM characters.  */
      case NEUTRAL_B:
      case LRE:
      case LRO:
      case RLE:
      case RLO:
      case PDF:
	return default_type;
      default:
	switch (ch)
	  {
	    case LRM_CHAR:
	    case RLM_CHAR:
	      return default_type;
	    default:
	      if (override == L2R) /* X6 */
		return STRONG_L;
	      else if (override == R2L)
		return STRONG_R;
	      else
		abort ();	/* can't happen: handled above */
	  }
    }
Eli Zaretskii's avatar
Eli Zaretskii committed
182 183
}

184 185 186 187 188 189 190
void
bidi_check_type (bidi_type_t type)
{
  if (type < UNKNOWN_BT || type > NEUTRAL_ON)
    abort ();
}

Eli Zaretskii's avatar
Eli Zaretskii committed
191
/* Given a bidi TYPE of a character, return its category.  */
192
static INLINE bidi_category_t
Eli Zaretskii's avatar
Eli Zaretskii committed
193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225
bidi_get_category (bidi_type_t type)
{
  switch (type)
    {
      case UNKNOWN_BT:
	return UNKNOWN_BC;
      case STRONG_L:
      case STRONG_R:
      case STRONG_AL:
      case LRE:
      case LRO:
      case RLE:
      case RLO:
	return STRONG;
      case PDF:		/* ??? really?? */
      case WEAK_EN:
      case WEAK_ES:
      case WEAK_ET:
      case WEAK_AN:
      case WEAK_CS:
      case WEAK_NSM:
      case WEAK_BN:
	return WEAK;
      case NEUTRAL_B:
      case NEUTRAL_S:
      case NEUTRAL_WS:
      case NEUTRAL_ON:
	return NEUTRAL;
      default:
	abort ();
    }
}

226 227 228 229
/* Return the mirrored character of C, if it has one.  If C has no
   mirrored counterpart, return C.
   Note: The conditions in UAX#9 clause L4 regarding the surrounding
   context must be tested by the caller.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
230 231 232
int
bidi_mirror_char (int c)
{
233 234 235 236 237 238
  Lisp_Object val;

  if (c == BIDI_EOB)
    return c;
  if (c < 0 || c > MAX_CHAR)
    abort ();
Eli Zaretskii's avatar
Eli Zaretskii committed
239

240 241
  val = CHAR_TABLE_REF (bidi_mirror_table, c);
  if (INTEGERP (val))
Eli Zaretskii's avatar
Eli Zaretskii committed
242
    {
243
      int v = XINT (val);
Eli Zaretskii's avatar
Eli Zaretskii committed
244

245 246 247 248
      if (v < 0 || v > MAX_CHAR)
	abort ();

      return v;
Eli Zaretskii's avatar
Eli Zaretskii committed
249
    }
250

Eli Zaretskii's avatar
Eli Zaretskii committed
251 252 253 254 255
  return c;
}

/* Copy the bidi iterator from FROM to TO.  To save cycles, this only
   copies the part of the level stack that is actually in use.  */
256
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
257 258 259 260
bidi_copy_it (struct bidi_it *to, struct bidi_it *from)
{
  int i;

261 262
  /* Copy everything except the level stack and beyond.  */
  memcpy (to, from, ((size_t)&((struct bidi_it *)0)->level_stack[0]));
Eli Zaretskii's avatar
Eli Zaretskii committed
263 264 265 266 267 268 269 270 271

  /* Copy the active part of the level stack.  */
  to->level_stack[0] = from->level_stack[0]; /* level zero is always in use */
  for (i = 1; i <= from->stack_idx; i++)
    to->level_stack[i] = from->level_stack[i];
}

/* Caching the bidi iterator states.  */

272 273 274
#define BIDI_CACHE_CHUNK 200
static struct bidi_it *bidi_cache;
static size_t bidi_cache_size = 0;
275
static size_t elsz = sizeof (struct bidi_it);
276 277
static int bidi_cache_idx;	/* next unused cache slot */
static int bidi_cache_last_idx;	/* slot of last cache hit */
Eli Zaretskii's avatar
Eli Zaretskii committed
278

279
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
280 281 282 283 284 285
bidi_cache_reset (void)
{
  bidi_cache_idx = 0;
  bidi_cache_last_idx = -1;
}

286 287 288 289 290
static INLINE void
bidi_cache_shrink (void)
{
  if (bidi_cache_size > BIDI_CACHE_CHUNK)
    {
291 292 293
      bidi_cache_size = BIDI_CACHE_CHUNK;
      bidi_cache =
	(struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
294 295 296 297
    }
  bidi_cache_reset ();
}

298
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314
bidi_cache_fetch_state (int idx, struct bidi_it *bidi_it)
{
  int current_scan_dir = bidi_it->scan_dir;

  if (idx < 0 || idx >= bidi_cache_idx)
    abort ();

  bidi_copy_it (bidi_it, &bidi_cache[idx]);
  bidi_it->scan_dir = current_scan_dir;
  bidi_cache_last_idx = idx;
}

/* Find a cached state with a given CHARPOS and resolved embedding
   level less or equal to LEVEL.  if LEVEL is -1, disregard the
   resolved levels in cached states.  DIR, if non-zero, means search
   in that direction from the last cache hit.  */
315
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
bidi_cache_search (int charpos, int level, int dir)
{
  int i, i_start;

  if (bidi_cache_idx)
    {
      if (charpos < bidi_cache[bidi_cache_last_idx].charpos)
	dir = -1;
      else if (charpos > bidi_cache[bidi_cache_last_idx].charpos)
	dir = 1;
      if (dir)
	i_start = bidi_cache_last_idx;
      else
	{
	  dir = -1;
	  i_start = bidi_cache_idx - 1;
	}

      if (dir < 0)
	{
	  /* Linear search for now; FIXME!  */
	  for (i = i_start; i >= 0; i--)
	    if (bidi_cache[i].charpos == charpos
		&& (level == -1 || bidi_cache[i].resolved_level <= level))
	      return i;
	}
      else
	{
	  for (i = i_start; i < bidi_cache_idx; i++)
	    if (bidi_cache[i].charpos == charpos
		&& (level == -1 || bidi_cache[i].resolved_level <= level))
	      return i;
	}
    }

  return -1;
}

/* Find a cached state where the resolved level changes to a value
   that is lower than LEVEL, and return its cache slot index.  DIR is
   the direction to search, starting with the last used cache slot.
   BEFORE, if non-zero, means return the index of the slot that is
   ``before'' the level change in the search direction.  That is,
   given the cached levels like this:

	 1122333442211
	  AB        C

   and assuming we are at the position cached at the slot marked with
   C, searching backwards (DIR = -1) for LEVEL = 2 will return the
   index of slot B or A, depending whether BEFORE is, respectively,
   non-zero or zero.  */
static int
bidi_cache_find_level_change (int level, int dir, int before)
{
  if (bidi_cache_idx)
    {
      int i = dir ? bidi_cache_last_idx : bidi_cache_idx - 1;
      int incr = before ? 1 : 0;

      if (!dir)
	dir = -1;
      else if (!incr)
	i += dir;

      if (dir < 0)
	{
	  while (i >= incr)
	    {
	      if (bidi_cache[i - incr].resolved_level >= 0
		  && bidi_cache[i - incr].resolved_level < level)
		return i;
	      i--;
	    }
	}
      else
	{
	  while (i < bidi_cache_idx - incr)
	    {
	      if (bidi_cache[i + incr].resolved_level >= 0
		  && bidi_cache[i + incr].resolved_level < level)
		return i;
	      i++;
	    }
	}
    }

  return -1;
}

406
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
407 408 409 410 411 412 413 414 415 416 417 418
bidi_cache_iterator_state (struct bidi_it *bidi_it, int resolved)
{
  int idx;

  /* We should never cache on backward scans.  */
  if (bidi_it->scan_dir == -1)
    abort ();
  idx = bidi_cache_search (bidi_it->charpos, -1, 1);

  if (idx < 0)
    {
      idx = bidi_cache_idx;
419 420 421
      /* Enlarge the cache as needed.  */
      if (idx >= bidi_cache_size)
	{
422
	  bidi_cache_size += BIDI_CACHE_CHUNK;
423
	  bidi_cache =
424
	    (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
425
	}
426 427 428 429 430 431 432 433 434 435
      /* Character positions should correspond to cache positions 1:1.
	 If we are outside the range of cached positions, the cache is
	 useless and must be reset.  */
      if (idx > 0 &&
	  (bidi_it->charpos > bidi_cache[idx - 1].charpos + 1
	   || bidi_it->charpos < bidi_cache[0].charpos))
	{
	  bidi_cache_reset ();
	  idx = 0;
	}
Eli Zaretskii's avatar
Eli Zaretskii committed
436 437 438 439 440 441 442 443 444
      bidi_copy_it (&bidi_cache[idx], bidi_it);
      if (!resolved)
	bidi_cache[idx].resolved_level = -1;
    }
  else
    {
      /* Copy only the members which could have changed, to avoid
	 costly copying of the entire struct.  */
      bidi_cache[idx].type = bidi_it->type;
445
      bidi_check_type (bidi_it->type);
446 447
      bidi_cache[idx].type_after_w1 = bidi_it->type_after_w1;
      bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463
      if (resolved)
	bidi_cache[idx].resolved_level = bidi_it->resolved_level;
      else
	bidi_cache[idx].resolved_level = -1;
      bidi_cache[idx].invalid_levels = bidi_it->invalid_levels;
      bidi_cache[idx].invalid_rl_levels = bidi_it->invalid_rl_levels;
      bidi_cache[idx].next_for_neutral = bidi_it->next_for_neutral;
      bidi_cache[idx].next_for_ws = bidi_it->next_for_ws;
      bidi_cache[idx].ignore_bn_limit = bidi_it->ignore_bn_limit;
    }

  bidi_cache_last_idx = idx;
  if (idx >= bidi_cache_idx)
    bidi_cache_idx = idx + 1;
}

464
static INLINE bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
465 466 467 468 469 470 471 472
bidi_cache_find (int charpos, int level, struct bidi_it *bidi_it)
{
  int i = bidi_cache_search (charpos, level, bidi_it->scan_dir);

  if (i >= 0)
    {
      bidi_dir_t current_scan_dir = bidi_it->scan_dir;

473
      bidi_copy_it (bidi_it, &bidi_cache[i]);
Eli Zaretskii's avatar
Eli Zaretskii committed
474 475 476 477 478 479 480 481 482 483
      bidi_cache_last_idx = i;
      /* Don't let scan direction from from the cached state override
	 the current scan direction.  */
      bidi_it->scan_dir = current_scan_dir;
      return bidi_it->type;
    }

  return UNKNOWN_BT;
}

484
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
485 486 487 488 489 490 491
bidi_peek_at_next_level (struct bidi_it *bidi_it)
{
  if (bidi_cache_idx == 0 || bidi_cache_last_idx == -1)
    abort ();
  return bidi_cache[bidi_cache_last_idx + bidi_it->scan_dir].resolved_level;
}

492 493 494 495 496
/* Check if buffer position CHARPOS/BYTEPOS is the end of a paragraph.
   Value is the non-negative length of the paragraph separator
   following the buffer position, -1 if position is at the beginning
   of a new paragraph, or -2 if position is neither at beginning nor
   at end of a paragraph.  */
497
static EMACS_INT
498
bidi_at_paragraph_end (EMACS_INT charpos, EMACS_INT bytepos)
Eli Zaretskii's avatar
Eli Zaretskii committed
499
{
Eli Zaretskii's avatar
Eli Zaretskii committed
500 501
  Lisp_Object sep_re;
  Lisp_Object start_re;
502 503
  EMACS_INT val;

Eli Zaretskii's avatar
Eli Zaretskii committed
504 505
  sep_re = paragraph_separate_re;
  start_re = paragraph_start_re;
506 507 508 509 510 511 512 513 514

  val = fast_looking_at (sep_re, charpos, bytepos, ZV, ZV_BYTE, Qnil);
  if (val < 0)
    {
      if (fast_looking_at (start_re, charpos, bytepos, ZV, ZV_BYTE, Qnil) >= 0)
	val = -1;
      else
	val = -2;
    }
Eli Zaretskii's avatar
Eli Zaretskii committed
515

516
  return val;
Eli Zaretskii's avatar
Eli Zaretskii committed
517 518 519 520 521 522
}

/* Determine the start-of-run (sor) directional type given the two
   embedding levels on either side of the run boundary.  Also, update
   the saved info about previously seen characters, since that info is
   generally valid for a single level run.  */
523
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
524 525 526 527 528 529 530 531 532 533 534
bidi_set_sor_type (struct bidi_it *bidi_it, int level_before, int level_after)
{
  int higher_level = level_before > level_after ? level_before : level_after;

  /* The prev_was_pdf gork is required for when we have several PDFs
     in a row.  In that case, we want to compute the sor type for the
     next level run only once: when we see the first PDF.  That's
     because the sor type depends only on the higher of the two levels
     that we find on the two sides of the level boundary (see UAX#9,
     clause X10), and so we don't need to know the final embedding
     level to which we descend after processing all the PDFs.  */
535
  if (!bidi_it->prev_was_pdf || level_before < level_after)
Eli Zaretskii's avatar
Eli Zaretskii committed
536 537 538 539 540 541
    /* FIXME: should the default sor direction be user selectable?  */
    bidi_it->sor = (higher_level & 1) != 0 ? R2L : L2R;
  if (level_before > level_after)
    bidi_it->prev_was_pdf = 1;

  bidi_it->prev.type = UNKNOWN_BT;
542 543
  bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1 =
    bidi_it->last_strong.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
544 545 546
  bidi_it->prev_for_neutral.type = bidi_it->sor == R2L ? STRONG_R : STRONG_L;
  bidi_it->prev_for_neutral.charpos = bidi_it->charpos;
  bidi_it->prev_for_neutral.bytepos = bidi_it->bytepos;
547 548
  bidi_it->next_for_neutral.type = bidi_it->next_for_neutral.type_after_w1 =
    bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
549 550 551
  bidi_it->ignore_bn_limit = 0; /* meaning it's unknown */
}

552
static void
553 554 555 556 557 558 559 560 561
bidi_line_init (struct bidi_it *bidi_it)
{
  bidi_it->scan_dir = 1; /* FIXME: do we need to have control on this? */
  bidi_it->resolved_level = bidi_it->level_stack[0].level;
  bidi_it->level_stack[0].override = NEUTRAL_DIR; /* X1 */
  bidi_it->invalid_levels = 0;
  bidi_it->invalid_rl_levels = -1;
  bidi_it->next_en_pos = -1;
  bidi_it->next_for_ws.type = UNKNOWN_BT;
562 563
  bidi_set_sor_type (bidi_it,
		     bidi_it->paragraph_dir == R2L ? 1 : 0,
564 565 566 567 568 569 570 571
		     bidi_it->level_stack[0].level); /* X10 */

  bidi_cache_reset ();
}

/* Find the beginning of this paragraph by looking back in the buffer.
   Value is the byte position of the paragraph's beginning.  */
static EMACS_INT
572
bidi_find_paragraph_start (EMACS_INT pos, EMACS_INT pos_byte)
573
{
Eli Zaretskii's avatar
Eli Zaretskii committed
574
  Lisp_Object re = paragraph_start_re;
575 576 577 578 579
  EMACS_INT limit = ZV, limit_byte = ZV_BYTE;

  while (pos_byte > BEGV_BYTE
	 && fast_looking_at (re, pos, pos_byte, limit, limit_byte, Qnil) < 0)
    {
580 581
      pos = find_next_newline_no_quit (pos - 1, -1);
      pos_byte = CHAR_TO_BYTE (pos);
582
    }
583
  return pos_byte;
584 585
}

586
/* Determine the base direction, a.k.a. base embedding level, of the
587 588
   paragraph we are about to iterate through.  If DIR is either L2R or
   R2L, just use that.  Otherwise, determine the paragraph direction
589 590 591 592 593 594 595 596 597 598 599 600
   from the first strong directional character of the paragraph.

   NO_DEFAULT_P non-nil means don't default to L2R if the paragraph
   has no strong directional characters and both DIR and
   bidi_it->paragraph_dir are NEUTRAL_DIR.  In that case, search back
   in the buffer until a paragraph is found with a strong character,
   or until hitting BEGV.  In the latter case, fall back to L2R.  This
   flag is used in current-bidi-paragraph-direction.

   Note that this function gives the paragraph separator the same
   direction as the preceding paragraph, even though Emacs generally
   views the separartor as not belonging to any paragraph.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
601
void
602
bidi_paragraph_init (bidi_dir_t dir, struct bidi_it *bidi_it, int no_default_p)
Eli Zaretskii's avatar
Eli Zaretskii committed
603
{
604
  EMACS_INT bytepos = bidi_it->bytepos;
605
  EMACS_INT pstartbyte;
606

607 608 609
  /* Special case for an empty buffer. */
  if (bytepos == BEGV_BYTE && bytepos == ZV_BYTE)
    dir = L2R;
610
  /* We should never be called at EOB or before BEGV.  */
611
  else if (bytepos >= ZV_BYTE || bytepos < BEGV_BYTE)
612 613
    abort ();

614 615 616 617 618 619 620 621 622 623
  if (dir == L2R)
    {
      bidi_it->paragraph_dir = L2R;
      bidi_it->new_paragraph = 0;
    }
  else if (dir == R2L)
    {
      bidi_it->paragraph_dir = R2L;
      bidi_it->new_paragraph = 0;
    }
Eli Zaretskii's avatar
Eli Zaretskii committed
624 625
  else if (dir == NEUTRAL_DIR)	/* P2 */
    {
626 627 628
      int ch, ch_len;
      EMACS_INT pos;
      bidi_type_t type;
629

630 631 632
      if (!bidi_initialized)
	bidi_initialize ();

633 634
      /* If we are inside a paragraph separator, we are just waiting
	 for the separator to be exhausted; use the previous paragraph
635 636 637 638
	 direction.  But don't do that if we have been just reseated,
	 because we need to reinitialize below in that case.  */
      if (!bidi_it->first_elt
	  && bidi_it->charpos < bidi_it->separator_limit)
639 640
	return;

641
      /* If we are on a newline, get past it to where the next
642 643 644
	 paragraph might start.  But don't do that at BEGV since then
	 we are potentially in a new paragraph that doesn't yet
	 exist.  */
645
      pos = bidi_it->charpos;
646
      if (bytepos > BEGV_BYTE && FETCH_CHAR (bytepos) == '\n')
647
	{
648
	  bytepos++;
649
	  pos++;
650
	}
651 652 653

      /* We are either at the beginning of a paragraph or in the
	 middle of it.  Find where this paragraph starts.  */
654
      pstartbyte = bidi_find_paragraph_start (pos, bytepos);
655 656
      bidi_it->separator_limit = -1;
      bidi_it->new_paragraph = 0;
657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676 677 678

      /* The following loop is run more than once only if NO_DEFAULT_P
	 is non-zero.  */
      do {
	bytepos = pstartbyte;
	ch = FETCH_CHAR (bytepos);
	ch_len = CHAR_BYTES (ch);
	pos = BYTE_TO_CHAR (bytepos);
	type = bidi_get_type (ch, NEUTRAL_DIR);

	for (pos++, bytepos += ch_len;
	     /* NOTE: UAX#9 says to search only for L, AL, or R types
		of characters, and ignore RLE, RLO, LRE, and LRO.
		However, I'm not sure it makes sense to omit those 4;
		should try with and without that to see the effect.  */
	     (bidi_get_category (type) != STRONG)
	       || (bidi_ignore_explicit_marks_for_paragraph_level
		   && (type == RLE || type == RLO
		       || type == LRE || type == LRO));
	     type = bidi_get_type (ch, NEUTRAL_DIR))
	  {
	    if (type == NEUTRAL_B && bidi_at_paragraph_end (pos, bytepos) >= -1)
679
	      break;
680 681 682 683 684 685 686 687 688 689 690 691 692 693 694 695 696 697 698 699 700 701 702 703 704 705 706 707 708 709 710 711 712 713
	    if (bytepos >= ZV_BYTE)
	      {
		/* Pretend there's a paragraph separator at end of
		   buffer.  */
		type = NEUTRAL_B;
		break;
	      }
	    FETCH_CHAR_ADVANCE (ch, pos, bytepos);
	  }
	if (type == STRONG_R || type == STRONG_AL) /* P3 */
	  bidi_it->paragraph_dir = R2L;
	else if (type == STRONG_L)
	  bidi_it->paragraph_dir = L2R;
	if (no_default_p && bidi_it->paragraph_dir == NEUTRAL_DIR)
	  {
	    /* If this paragraph is at BEGV, default to L2R.  */
	    if (pstartbyte == BEGV_BYTE)
	      bidi_it->paragraph_dir = L2R; /* P3 and HL1 */
	    else
	      {
		EMACS_INT prevpbyte = pstartbyte;
		EMACS_INT p = BYTE_TO_CHAR (pstartbyte), pbyte = pstartbyte;

		/* Find the beginning of the previous paragraph, if any.  */
		while (pbyte > BEGV_BYTE && prevpbyte >= pstartbyte)
		  {
		    p--;
		    pbyte = CHAR_TO_BYTE (p);
		    prevpbyte = bidi_find_paragraph_start (p, pbyte);
		  }
		pstartbyte = prevpbyte;
	      }
	  }
      } while (no_default_p && bidi_it->paragraph_dir == NEUTRAL_DIR);
Eli Zaretskii's avatar
Eli Zaretskii committed
714
    }
715 716
  else
    abort ();
Eli Zaretskii's avatar
Eli Zaretskii committed
717

718 719
  /* Contrary to UAX#9 clause P3, we only default the paragraph
     direction to L2R if we have no previous usable paragraph
720
     direction.  This is allowed by the HL1 clause.  */
721
  if (bidi_it->paragraph_dir != L2R && bidi_it->paragraph_dir != R2L)
722
    bidi_it->paragraph_dir = L2R; /* P3 and HL1 ``higher-level protocols'' */
723
  if (bidi_it->paragraph_dir == R2L)
724
    bidi_it->level_stack[0].level = 1;
725
  else
726
    bidi_it->level_stack[0].level = 0;
727 728

  bidi_line_init (bidi_it);
Eli Zaretskii's avatar
Eli Zaretskii committed
729 730
}

731 732
/* Do whatever UAX#9 clause X8 says should be done at paragraph's
   end.  */
733
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
734 735 736 737 738 739 740 741
bidi_set_paragraph_end (struct bidi_it *bidi_it)
{
  bidi_it->invalid_levels = 0;
  bidi_it->invalid_rl_levels = -1;
  bidi_it->stack_idx = 0;
  bidi_it->resolved_level = bidi_it->level_stack[0].level;
}

742
/* Initialize the bidi iterator from buffer position CHARPOS.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
743
void
744
bidi_init_it (EMACS_INT charpos, EMACS_INT bytepos, struct bidi_it *bidi_it)
Eli Zaretskii's avatar
Eli Zaretskii committed
745 746 747
{
  if (! bidi_initialized)
    bidi_initialize ();
748 749
  bidi_it->charpos = charpos;
  bidi_it->bytepos = bytepos;
750 751 752
  bidi_it->first_elt = 1;
  bidi_set_paragraph_end (bidi_it);
  bidi_it->new_paragraph = 1;
753
  bidi_it->separator_limit = -1;
Eli Zaretskii's avatar
Eli Zaretskii committed
754
  bidi_it->type = NEUTRAL_B;
755 756
  bidi_it->type_after_w1 = NEUTRAL_B;
  bidi_it->orig_type = NEUTRAL_B;
Eli Zaretskii's avatar
Eli Zaretskii committed
757
  bidi_it->prev_was_pdf = 0;
758 759
  bidi_it->prev.type = bidi_it->prev.type_after_w1 =
    bidi_it->prev.orig_type = UNKNOWN_BT;
760 761
  bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1 =
    bidi_it->last_strong.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
762 763
  bidi_it->next_for_neutral.charpos = -1;
  bidi_it->next_for_neutral.type =
764 765
    bidi_it->next_for_neutral.type_after_w1 =
    bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
766 767
  bidi_it->prev_for_neutral.charpos = -1;
  bidi_it->prev_for_neutral.type =
768 769
    bidi_it->prev_for_neutral.type_after_w1 =
    bidi_it->prev_for_neutral.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
770
  bidi_it->sor = L2R;	 /* FIXME: should it be user-selectable? */
771
  bidi_cache_shrink ();
Eli Zaretskii's avatar
Eli Zaretskii committed
772 773 774 775
}

/* Push the current embedding level and override status; reset the
   current level to LEVEL and the current override status to OVERRIDE.  */
776
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
777 778 779 780 781 782 783 784 785 786 787 788
bidi_push_embedding_level (struct bidi_it *bidi_it,
			   int level, bidi_dir_t override)
{
  bidi_it->stack_idx++;
  if (bidi_it->stack_idx >= BIDI_MAXLEVEL)
    abort ();
  bidi_it->level_stack[bidi_it->stack_idx].level = level;
  bidi_it->level_stack[bidi_it->stack_idx].override = override;
}

/* Pop the embedding level and directional override status from the
   stack, and return the new level.  */
789
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
790 791 792 793 794 795 796 797 798
bidi_pop_embedding_level (struct bidi_it *bidi_it)
{
  /* UAX#9 says to ignore invalid PDFs.  */
  if (bidi_it->stack_idx > 0)
    bidi_it->stack_idx--;
  return bidi_it->level_stack[bidi_it->stack_idx].level;
}

/* Record in SAVED_INFO the information about the current character.  */
799
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
800 801 802 803 804 805
bidi_remember_char (struct bidi_saved_info *saved_info,
		    struct bidi_it *bidi_it)
{
  saved_info->charpos = bidi_it->charpos;
  saved_info->bytepos = bidi_it->bytepos;
  saved_info->type = bidi_it->type;
806
  bidi_check_type (bidi_it->type);
807 808
  saved_info->type_after_w1 = bidi_it->type_after_w1;
  bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
809
  saved_info->orig_type = bidi_it->orig_type;
810
  bidi_check_type (bidi_it->orig_type);
Eli Zaretskii's avatar
Eli Zaretskii committed
811 812 813 814
}

/* Resolve the type of a neutral character according to the type of
   surrounding strong text and the current embedding level.  */
815
static INLINE bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831
bidi_resolve_neutral_1 (bidi_type_t prev_type, bidi_type_t next_type, int lev)
{
  /* N1: European and Arabic numbers are treated as though they were R.  */
  if (next_type == WEAK_EN || next_type == WEAK_AN)
    next_type = STRONG_R;
  if (prev_type == WEAK_EN || prev_type == WEAK_AN)
    prev_type = STRONG_R;

  if (next_type == prev_type)	/* N1 */
    return next_type;
  else if ((lev & 1) == 0)	/* N2 */
    return STRONG_L;
  else
    return STRONG_R;
}

832
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853
bidi_explicit_dir_char (int c)
{
  /* FIXME: this should be replaced with a lookup table with suitable
     bits set, like standard C ctype macros do.  */
  return (c == LRE_CHAR || c == LRO_CHAR
	  || c == RLE_CHAR || c == RLO_CHAR || c == PDF_CHAR);
}

/* A helper function for bidi_resolve_explicit.  It advances to the
   next character in logical order and determines the new embedding
   level and directional override, but does not take into account
   empty embeddings.  */
static int
bidi_resolve_explicit_1 (struct bidi_it *bidi_it)
{
  int curchar;
  bidi_type_t type;
  int current_level;
  int new_level;
  bidi_dir_t override;

854 855
  if (bidi_it->bytepos < BEGV_BYTE	/* after reseat to BEGV? */
      || bidi_it->first_elt)
856
    {
857 858 859 860
      bidi_it->first_elt = 0;
      if (bidi_it->charpos < BEGV)
	bidi_it->charpos = BEGV;
      bidi_it->bytepos = CHAR_TO_BYTE (bidi_it->charpos);
861
    }
862
  else if (bidi_it->bytepos < ZV_BYTE)	/* don't move at ZV */
Eli Zaretskii's avatar
Eli Zaretskii committed
863 864
    {
      bidi_it->charpos++;
865 866
      if (bidi_it->ch_len == 0)
	abort ();
Eli Zaretskii's avatar
Eli Zaretskii committed
867 868 869 870 871 872 873 874 875
      bidi_it->bytepos += bidi_it->ch_len;
    }

  current_level = bidi_it->level_stack[bidi_it->stack_idx].level; /* X1 */
  override = bidi_it->level_stack[bidi_it->stack_idx].override;
  new_level = current_level;

  /* in case it is a unibyte character (not yet implemented) */
  /* _fetch_multibyte_char_len = 1; */
876 877 878 879 880 881 882 883 884 885
  if (bidi_it->bytepos >= ZV_BYTE)
    {
      curchar = BIDI_EOB;
      bidi_it->ch_len = 1;
    }
  else
    {
      curchar = FETCH_CHAR (bidi_it->bytepos);
      bidi_it->ch_len = CHAR_BYTES (curchar);
    }
Eli Zaretskii's avatar
Eli Zaretskii committed
886 887
  bidi_it->ch = curchar;

888 889 890 891 892
  /* Don't apply directional override here, as all the types we handle
     below will not be affected by the override anyway, and we need
     the original type unaltered.  The override will be applied in
     bidi_resolve_weak.  */
  type = bidi_get_type (curchar, NEUTRAL_DIR);
893 894
  bidi_it->orig_type = type;
  bidi_check_type (bidi_it->orig_type);
Eli Zaretskii's avatar
Eli Zaretskii committed
895 896 897 898

  if (type != PDF)
    bidi_it->prev_was_pdf = 0;

899
  bidi_it->type_after_w1 = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
900 901 902 903 904

  switch (type)
    {
      case RLE:	/* X2 */
      case RLO:	/* X4 */
905 906
	bidi_it->type_after_w1 = type;
	bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
907 908 909 910 911 912 913 914
	type = WEAK_BN; /* X9/Retaining */
	if (bidi_it->ignore_bn_limit <= 0)
	  {
	    if (current_level <= BIDI_MAXLEVEL - 4)
	      {
		/* Compute the least odd embedding level greater than
		   the current level.  */
		new_level = ((current_level + 1) & ~1) + 1;
915
		if (bidi_it->type_after_w1 == RLE)
Eli Zaretskii's avatar
Eli Zaretskii committed
916 917 918 919 920 921 922 923 924 925 926 927 928 929 930 931
		  override = NEUTRAL_DIR;
		else
		  override = R2L;
		if (current_level == BIDI_MAXLEVEL - 4)
		  bidi_it->invalid_rl_levels = 0;
		bidi_push_embedding_level (bidi_it, new_level, override);
	      }
	    else
	      {
		bidi_it->invalid_levels++;
		/* See the commentary about invalid_rl_levels below.  */
		if (bidi_it->invalid_rl_levels < 0)
		  bidi_it->invalid_rl_levels = 0;
		bidi_it->invalid_rl_levels++;
	      }
	  }
932
	else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
933 934 935 936 937
		 || bidi_it->next_en_pos > bidi_it->charpos)
	  type = WEAK_EN;
	break;
      case LRE:	/* X3 */
      case LRO:	/* X5 */
938 939
	bidi_it->type_after_w1 = type;
	bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
940 941 942 943 944 945 946 947
	type = WEAK_BN; /* X9/Retaining */
	if (bidi_it->ignore_bn_limit <= 0)
	  {
	    if (current_level <= BIDI_MAXLEVEL - 5)
	      {
		/* Compute the least even embedding level greater than
		   the current level.  */
		new_level = ((current_level + 2) & ~1);
948
		if (bidi_it->type_after_w1 == LRE)
Eli Zaretskii's avatar
Eli Zaretskii committed
949 950 951 952 953 954 955 956 957 958 959 960 961 962 963 964 965 966 967
		  override = NEUTRAL_DIR;
		else
		  override = L2R;
		bidi_push_embedding_level (bidi_it, new_level, override);
	      }
	    else
	      {
		bidi_it->invalid_levels++;
		/* invalid_rl_levels counts invalid levels encountered
		   while the embedding level was already too high for
		   LRE/LRO, but not for RLE/RLO.  That is because
		   there may be exactly one PDF which we should not
		   ignore even though invalid_levels is non-zero.
		   invalid_rl_levels helps to know what PDF is
		   that.  */
		if (bidi_it->invalid_rl_levels >= 0)
		  bidi_it->invalid_rl_levels++;
	      }
	  }
968
	else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
969 970 971 972
		 || bidi_it->next_en_pos > bidi_it->charpos)
	  type = WEAK_EN;
	break;
      case PDF:	/* X7 */
973 974
	bidi_it->type_after_w1 = type;
	bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
975 976 977 978 979 980 981 982 983 984 985 986 987 988 989 990 991 992 993
	type = WEAK_BN; /* X9/Retaining */
	if (bidi_it->ignore_bn_limit <= 0)
	  {
	    if (!bidi_it->invalid_rl_levels)
	      {
		new_level = bidi_pop_embedding_level (bidi_it);
		bidi_it->invalid_rl_levels = -1;
		if (bidi_it->invalid_levels)
		  bidi_it->invalid_levels--;
		/* else nothing: UAX#9 says to ignore invalid PDFs */
	      }
	    if (!bidi_it->invalid_levels)
	      new_level = bidi_pop_embedding_level (bidi_it);
	    else
	      {
		bidi_it->invalid_levels--;
		bidi_it->invalid_rl_levels--;
	      }
	  }
994
	else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
995 996 997 998 999 1000 1001 1002 1003
		 || bidi_it->next_en_pos > bidi_it->charpos)
	  type = WEAK_EN;
	break;
      default:
	/* Nothing.  */
	break;
    }

  bidi_it->type = type;
1004
  bidi_check_type (bidi_it->type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1005 1006 1007 1008 1009 1010 1011 1012 1013 1014 1015 1016 1017 1018 1019 1020 1021 1022

  return new_level;
}

/* Given an iterator state in BIDI_IT, advance one character position
   in the buffer to the next character (in the logical order), resolve
   any explicit embeddings and directional overrides, and return the
   embedding level of the character after resolving explicit
   directives and ignoring empty embeddings.  */
static int
bidi_resolve_explicit (struct bidi_it *bidi_it)
{
  int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level;
  int new_level  = bidi_resolve_explicit_1 (bidi_it);

  if (prev_level < new_level
      && bidi_it->type == WEAK_BN
      && bidi_it->ignore_bn_limit == 0 /* only if not already known */
1023
      && bidi_it->bytepos < ZV_BYTE    /* not already at EOB */
Eli Zaretskii's avatar
Eli Zaretskii committed
1024 1025 1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047 1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069 1070
      && bidi_explicit_dir_char (FETCH_CHAR (bidi_it->bytepos
					     + bidi_it->ch_len)))
    {
      /* Avoid pushing and popping embedding levels if the level run
	 is empty, as this breaks level runs where it shouldn't.
	 UAX#9 removes all the explicit embedding and override codes,
	 so empty embeddings disappear without a trace.  We need to
	 behave as if we did the same.  */
      struct bidi_it saved_it;
      int level = prev_level;

      bidi_copy_it (&saved_it, bidi_it);

      while (bidi_explicit_dir_char (FETCH_CHAR (bidi_it->bytepos
						 + bidi_it->ch_len)))
	{
	  level = bidi_resolve_explicit_1 (bidi_it);
	}

      if (level == prev_level)	/* empty embedding */
	saved_it.ignore_bn_limit = bidi_it->charpos + 1;
      else			/* this embedding is non-empty */
	saved_it.ignore_bn_limit = -1;

      bidi_copy_it (bidi_it, &saved_it);
      if (bidi_it->ignore_bn_limit > 0)
	{
	  /* We pushed a level, but we shouldn't have.  Undo that. */
	  if (!bidi_it->invalid_rl_levels)
	    {
	      new_level = bidi_pop_embedding_level (bidi_it);
	      bidi_it->invalid_rl_levels = -1;
	      if (bidi_it->invalid_levels)
		bidi_it->invalid_levels--;
	    }
	  if (!bidi_it->invalid_levels)
	    new_level = bidi_pop_embedding_level (bidi_it);
	  else
	    {
	      bidi_it->invalid_levels--;
	      bidi_it->invalid_rl_levels--;
	    }
	}
    }

  if (bidi_it->type == NEUTRAL_B)	/* X8 */
    {
1071
      bidi_set_paragraph_end (bidi_it);
1072 1073
      /* This is needed by bidi_resolve_weak below, and in L1.  */
      bidi_it->type_after_w1 = bidi_it->type;
1074
      bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1075 1076 1077 1078 1079 1080 1081
    }

  return new_level;
}

/* Advance in the buffer, resolve weak types and return the type of
   the next character after weak type resolution.  */
1082
static bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
1083 1084 1085 1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113
bidi_resolve_weak (struct bidi_it *bidi_it)
{
  bidi_type_t type;
  bidi_dir_t override;
  int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level;
  int new_level  = bidi_resolve_explicit (bidi_it);
  int next_char;
  bidi_type_t type_of_next;
  struct bidi_it saved_it;

  type = bidi_it->type;
  override = bidi_it->level_stack[bidi_it->stack_idx].override;

  if (type == UNKNOWN_BT
      || type == LRE
      || type == LRO
      || type == RLE
      || type == RLO
      || type == PDF)
    abort ();

  if (new_level != prev_level
      || bidi_it->type == NEUTRAL_B)
    {
      /* We've got a new embedding level run, compute the directional
         type of sor and initialize per-run variables (UAX#9, clause
         X10).  */
      bidi_set_sor_type (bidi_it, prev_level, new_level);
    }
  else if (type == NEUTRAL_S || type == NEUTRAL_WS
	   || type == WEAK_BN || type == STRONG_AL)
1114 1115
    bidi_it->type_after_w1 = type;	/* needed in L1 */
  bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1116 1117 1118 1119 1120 1121 1122

  /* Level and directional override status are already recorded in
     bidi_it, and do not need any change; see X6.  */
  if (override == R2L)		/* X6 */
    type = STRONG_R;
  else if (override == L2R)
    type = STRONG_L;
1123
  else
Eli Zaretskii's avatar
Eli Zaretskii committed
1124
    {
1125
      if (type == WEAK_NSM)	/* W1 */
Eli Zaretskii's avatar
Eli Zaretskii committed
1126
	{
1127
	  /* Note that we don't need to consider the case where the
1128 1129 1130 1131 1132 1133
	     prev character has its type overridden by an RLO or LRO,
	     because then either the type of this NSM would have been
	     also overridden, or the previous character is outside the
	     current level run, and thus not relevant to this NSM.
	     This is why NSM gets the type_after_w1 of the previous
	     character.  */
1134 1135 1136
	  if (bidi_it->prev.type_after_w1 != UNKNOWN_BT
	      /* if type_after_w1 is NEUTRAL_B, this NSM is at sor */
	      && bidi_it->prev.type_after_w1 != NEUTRAL_B)
1137
	    type = bidi_it->prev.type_after_w1;
1138 1139 1140 1141 1142 1143
	  else if (bidi_it->sor == R2L)
	    type = STRONG_R;
	  else if (bidi_it->sor == L2R)
	    type = STRONG_L;
	  else /* shouldn't happen! */
	    abort ();
Eli Zaretskii's avatar
Eli Zaretskii committed
1144
	}
1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156
      if (type == WEAK_EN	/* W2 */
	  && bidi_it->last_strong.type_after_w1 == STRONG_AL)
	type = WEAK_AN;
      else if (type == STRONG_AL) /* W3 */
	type = STRONG_R;
      else if ((type == WEAK_ES	/* W4 */
		&& bidi_it->prev.type_after_w1 == WEAK_EN
		&& bidi_it->prev.orig_type == WEAK_EN)
	       || (type == WEAK_CS
		   && ((bidi_it->prev.type_after_w1 == WEAK_EN
			&& bidi_it->prev.orig_type == WEAK_EN)
		       || bidi_it->prev.type_after_w1 == WEAK_AN)))
Eli Zaretskii's avatar
Eli Zaretskii committed
1157
	{
1158 1159 1160
	  next_char =
	    bidi_it->bytepos + bidi_it->ch_len >= ZV_BYTE
	    ? BIDI_EOB : FETCH_CHAR (bidi_it->bytepos + bidi_it->ch_len);
1161
	  type_of_next = bidi_get_type (next_char, override);
Eli Zaretskii's avatar
Eli Zaretskii committed
1162

1163
	  if (type_of_next == WEAK_BN
Eli Zaretskii's avatar
Eli Zaretskii committed
1164 1165 1166 1167
	      || bidi_explicit_dir_char (next_char))
	    {
	      bidi_copy_it (&saved_it, bidi_it);
	      while (bidi_resolve_explicit (bidi_it) == new_level
1168
		     && bidi_it->type == WEAK_BN)
Eli Zaretskii's avatar
Eli Zaretskii committed
1169 1170 1171 1172
		;
	      type_of_next = bidi_it->type;
	      bidi_copy_it (bidi_it, &saved_it);
	    }
1173 1174 1175 1176 1177 1178 1179 1180 1181 1182

	  /* If the next character is EN, but the last strong-type
	     character is AL, that next EN will be changed to AN when
	     we process it in W2 above.  So in that case, this ES
	     should not be changed into EN.  */
	  if (type == WEAK_ES
	      && type_of_next == WEAK_EN
	      && bidi_it->last_strong.type_after_w1 != STRONG_AL)
	    type = WEAK_EN;
	  else if (type == WEAK_CS)
Eli Zaretskii's avatar
Eli Zaretskii committed
1183
	    {
1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
	      if (bidi_it->prev.type_after_w1 == WEAK_AN
		  && (type_of_next == WEAK_AN
		      /* If the next character is EN, but the last
			 strong-type character is AL, EN will be later
			 changed to AN when we process it in W2 above.
			 So in that case, this ES should not be
			 changed into EN.  */
		      || (type_of_next == WEAK_EN
			  && bidi_it->last_strong.type_after_w1 == STRONG_AL)))
		type = WEAK_AN;
	      else if (bidi_it->prev.type_after_w1 == WEAK_EN
		       && type_of_next == WEAK_EN
		       && bidi_it->last_strong.type_after_w1 != STRONG_AL)
		type = WEAK_EN;
	    }
	}
      else if (type == WEAK_ET	/* W5: ET with EN before or after it */
	       || type == WEAK_BN)	/* W5/Retaining */
	{
	  if (bidi_it->prev.type_after_w1 == WEAK_EN /* ET/BN w/EN before it */
	      || bidi_it->next_en_pos > bidi_it->charpos)
	    type = WEAK_EN;
	  else			/* W5: ET/BN with EN after it.  */
	    {
	      EMACS_INT en_pos = bidi_it->charpos + 1;

	      next_char =
		bidi_it->bytepos + bidi_it->ch_len >= ZV_BYTE
		? BIDI_EOB : FETCH_CHAR (bidi_it->bytepos + bidi_it->ch_len);
	      type_of_next = bidi_get_type (next_char, override);

	      if (type_of_next == WEAK_ET
		  || type_of_next == WEAK_BN
		  || bidi_explicit_dir_char (next_char))
		{
		  bidi_copy_it (&saved_it, bidi_it);
		  while (bidi_resolve_explicit (bidi_it) == new_level
			 && (bidi_it->type == WEAK_BN
			     || bidi_it->type == WEAK_ET))
		    ;
		  type_of_next = bidi_it->type;
		  en_pos = bidi_it->charpos;
		  bidi_copy_it (bidi_it, &saved_it);
		}
	      if (type_of_next == WEAK_EN)
Eli Zaretskii's avatar
Eli Zaretskii committed
1229
		{
1230 1231 1232 1233 1234 1235 1236 1237 1238 1239 1240
		  /* If the last strong character is AL, the EN we've
		     found will become AN when we get to it (W2). */
		  if (bidi_it->last_strong.type_after_w1 != STRONG_AL)
		    {
		      type = WEAK_EN;
		      /* Remember this EN position, to speed up processing
			 of the next ETs.  */
		      bidi_it->next_en_pos = en_pos;
		    }
		  else if (type == WEAK_BN)
		    type = NEUTRAL_ON; /* W6/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
1241 1242 1243 1244 1245 1246
		}
	    }
	}
    }

  if (type == WEAK_ES || type == WEAK_ET || type == WEAK_CS /* W6 */
1247 1248 1249 1250
      || (type == WEAK_BN
	  && (bidi_it->prev.type_after_w1 == WEAK_CS	    /* W6/Retaining */
	      || bidi_it->prev.type_after_w1 == WEAK_ES
	      || bidi_it->prev.type_after_w1 == WEAK_ET)))
Eli Zaretskii's avatar
Eli Zaretskii committed
1251 1252 1253 1254 1255 1256
    type = NEUTRAL_ON;

  /* Store the type we've got so far, before we clobber it with strong
     types in W7 and while resolving neutral types.  But leave alone
     the original types that were recorded above, because we will need
     them for the L1 clause.  */
1257 1258 1259
  if (bidi_it->type_after_w1 == UNKNOWN_BT)
    bidi_it->type_after_w1 = type;
  bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1260 1261 1262

  if (type == WEAK_EN)	/* W7 */
    {
1263
      if ((bidi_it->last_strong.type_after_w1 == STRONG_L)
Eli Zaretskii's avatar
Eli Zaretskii committed
1264 1265 1266 1267 1268
	  || (bidi_it->last_strong.type == UNKNOWN_BT && bidi_it->sor == L2R))
	type = STRONG_L;
    }

  bidi_it->type = type;
1269
  bidi_check_type (bidi_it->type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1270 1271 1272
  return type;
}

1273
static bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
1274 1275 1276 1277 1278 1279 1280 1281 1282 1283 1284 1285 1286 1287 1288 1289 1290 1291 1292 1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313 1314 1315 1316 1317 1318 1319 1320 1321 1322
bidi_resolve_neutral (struct bidi_it *bidi_it)
{
  int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level;
  bidi_type_t type = bidi_resolve_weak (bidi_it);
  int current_level = bidi_it->level_stack[bidi_it->stack_idx].level;

  if (!(type == STRONG_R
	|| type == STRONG_L
	|| type == WEAK_BN
	|| type == WEAK_EN
	|| type == WEAK_AN
	|| type == NEUTRAL_B
	|| type == NEUTRAL_S
	|| type == NEUTRAL_WS
	|| type == NEUTRAL_ON))
    abort ();

  if (bidi_get_category (type) == NEUTRAL
      || (type == WEAK_BN && prev_level == current_level))
    {
      if (bidi_it->next_for_neutral.type != UNKNOWN_BT)
	type = bidi_resolve_neutral_1 (bidi_it->prev_for_neutral.type,
				       bidi_it->next_for_neutral.type,
				       current_level);
      else
	{
	  /* Arrrgh!!  The UAX#9 algorithm is too deeply entrenched in
	     the assumption of batch-style processing; see clauses W4,
	     W5, and especially N1, which require to look far forward
	     (as well as back) in the buffer.  May the fleas of a
	     thousand camels infest the armpits of those who design
	     supposedly general-purpose algorithms by looking at their
	     own implementations, and fail to consider other possible
	     implementations!  */
	  struct bidi_it saved_it;
	  bidi_type_t next_type;

	  if (bidi_it->scan_dir == -1)
	    abort ();

	  bidi_copy_it (&saved_it, bidi_it);
	  /* Scan the text forward until we find the first non-neutral
	     character, and then use that to resolve the neutral we
	     are dealing with now.  We also cache the scanned iterator
	     states, to salvage some of the effort later.  */
	  bidi_cache_iterator_state (bidi_it, 0);
	  do {
	    /* Record the info about the previous character, so that
	       it will be cached below with this state.  */
1323
	    if (bidi_it->type_after_w1 != WEAK_BN /* W1/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334 1335 1336 1337 1338 1339 1340 1341 1342 1343 1344 1345 1346 1347 1348 1349 1350 1351 1352 1353 1354 1355 1356 1357 1358 1359 1360 1361 1362 1363 1364 1365 1366
		&& bidi_it->type != WEAK_BN)
	      bidi_remember_char (&bidi_it->prev, bidi_it);
	    type = bidi_resolve_weak (bidi_it);
	    /* Paragraph separators have their levels fully resolved
	       at this point, so cache them as resolved.  */
	    bidi_cache_iterator_state (bidi_it, type == NEUTRAL_B);
	    /* FIXME: implement L1 here, by testing for a newline and
	       resetting the level for any sequence of whitespace
	       characters adjacent to it.  */
	  } while (!(type == NEUTRAL_B
		     || (type != WEAK_BN
			 && bidi_get_category (type) != NEUTRAL)
		     /* This is all per level run, so stop when we
			reach the end of this level run.  */
		     || bidi_it->level_stack[bidi_it->stack_idx].level !=
		     current_level));

	  bidi_remember_char (&saved_it.next_for_neutral, bidi_it);

	  switch (type)
	    {
	      case STRONG_L:
	      case STRONG_R:
	      case STRONG_AL:
		next_type = type;
		break;
	      case WEAK_EN:
	      case WEAK_AN:
		/* N1: ``European and Arabic numbers are treated as
		   though they were R.''  */
		next_type = STRONG_R;
		saved_it.next_for_neutral.type = STRONG_R;
		break;
	      case WEAK_BN:
		if (!bidi_explicit_dir_char (bidi_it->ch))
		  abort ();		/* can't happen: BNs are skipped */
		/* FALLTHROUGH */
	      case NEUTRAL_B:
		/* Marched all the way to the end of this level run.
		   We need to use the eor type, whose information is
		   stored by bidi_set_sor_type in the prev_for_neutral
		   member.  */
		if (saved_it.type != WEAK_BN
1367
		    || bidi_get_category (bidi_it->prev.type_after_w1) == NEUTRAL)
Eli Zaretskii's avatar
Eli Zaretskii committed
1368 1369 1370
		  {
		    next_type = bidi_it->prev_for_neutral.type;
		    saved_it.next_for_neutral.type = next_type;
1371
		    bidi_check_type (next_type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1372 1373 1374 1375 1376 1377 1378 1379 1380 1381 1382 1383 1384 1385 1386
		  }
		else
		  {
		    /* This is a BN which does not adjoin neutrals.
		       Leave its type alone.  */
		    bidi_copy_it (bidi_it, &saved_it);
		    return bidi_it->type;
		  }
		break;
	      default:
		abort ();
	    }
	  type = bidi_resolve_neutral_1 (saved_it.prev_for_neutral.type,
					 next_type, current_level);
	  saved_it.type = type;
1387
	  bidi_check_type (type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1388 1389 1390