bidi.c 63.9 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 55 56 57 58 59 60 61
   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".  */

#ifdef HAVE_CONFIG_H
#include <config.h>
#endif

#include <stdio.h>

#ifdef HAVE_STRING_H
#include <string.h>
#endif

62 63
#include <setjmp.h>

Eli Zaretskii's avatar
Eli Zaretskii committed
64 65 66 67 68 69 70 71 72
#include "lisp.h"
#include "buffer.h"
#include "character.h"
#include "dispextern.h"

static int bidi_initialized = 0;

static Lisp_Object bidi_type_table;

73
/* FIXME: Remove these when bidi_explicit_dir_char uses a lookup table.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
74 75 76 77 78 79 80 81 82
#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
83
#define BIDI_BOB   -2		/* FIXME: Is this needed? */
Eli Zaretskii's avatar
Eli Zaretskii committed
84 85 86 87 88 89 90 91 92 93 94 95 96 97 98 99 100 101 102 103 104

/* 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
105
static Lisp_Object paragraph_start_re, paragraph_separate_re;
106
static Lisp_Object Qparagraph_start, Qparagraph_separate;
Eli Zaretskii's avatar
Eli Zaretskii committed
107 108 109 110 111 112 113 114 115 116 117 118 119 120 121 122 123 124 125 126 127 128

static void
bidi_initialize ()
{
  /* FIXME: This should come from the Unicode Database.  */
  struct {
    int from, to;
    bidi_type_t type;
  } bidi_type[] =
      { { 0x0000, 0x0008, WEAK_BN },
	{ 0x0009, 0x0000, NEUTRAL_S },
	{ 0x000A, 0x0000, NEUTRAL_B },
	{ 0x000B, 0x0000, NEUTRAL_S },
	{ 0x000C, 0x0000, NEUTRAL_WS },
	{ 0x000D, 0x0000, NEUTRAL_B },
	{ 0x000E, 0x001B, WEAK_BN },
	{ 0x001C, 0x001E, NEUTRAL_B },
	{ 0x001F, 0x0000, NEUTRAL_S },
	{ 0x0020, 0x0000, NEUTRAL_WS },
	{ 0x0021, 0x0022, NEUTRAL_ON },
	{ 0x0023, 0x0025, WEAK_ET },
	{ 0x0026, 0x002A, NEUTRAL_ON },
129
	{ 0x002B, 0x0000, WEAK_ES },
Eli Zaretskii's avatar
Eli Zaretskii committed
130
	{ 0x002C, 0x0000, WEAK_CS },
131 132
	{ 0x002D, 0x0000, WEAK_ES },
	{ 0x002E, 0x002F, WEAK_CS },
Eli Zaretskii's avatar
Eli Zaretskii committed
133 134 135 136 137 138 139 140 141 142 143 144
	{ 0x0030, 0x0039, WEAK_EN },
	{ 0x003A, 0x0000, WEAK_CS },
	{ 0x003B, 0x0040, NEUTRAL_ON },
	{ 0x005B, 0x0060, NEUTRAL_ON },
	{ 0x007B, 0x007E, NEUTRAL_ON },
	{ 0x007F, 0x0084, WEAK_BN },
	{ 0x0085, 0x0000, NEUTRAL_B },
	{ 0x0086, 0x009F, WEAK_BN },
	{ 0x00A0, 0x0000, WEAK_CS },
	{ 0x00A1, 0x0000, NEUTRAL_ON },
	{ 0x00A2, 0x00A5, WEAK_ET },
	{ 0x00A6, 0x00A9, NEUTRAL_ON },
145 146 147
	{ 0x00AB, 0x00AC, NEUTRAL_ON },
	{ 0x00AD, 0x0000, WEAK_BN },
	{ 0x00AE, 0x00Af, NEUTRAL_ON },
Eli Zaretskii's avatar
Eli Zaretskii committed
148 149 150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172
	{ 0x00B0, 0x00B1, WEAK_ET },
	{ 0x00B2, 0x00B3, WEAK_EN },
	{ 0x00B4, 0x0000, NEUTRAL_ON },
	{ 0x00B6, 0x00B8, NEUTRAL_ON },
	{ 0x00B9, 0x0000, WEAK_EN },
	{ 0x00BB, 0x00BF, NEUTRAL_ON },
	{ 0x00D7, 0x0000, NEUTRAL_ON },
	{ 0x00F7, 0x0000, NEUTRAL_ON },
	{ 0x02B9, 0x02BA, NEUTRAL_ON },
	{ 0x02C2, 0x02CF, NEUTRAL_ON },
	{ 0x02D2, 0x02DF, NEUTRAL_ON },
	{ 0x02E5, 0x02ED, NEUTRAL_ON },
	{ 0x0300, 0x036F, WEAK_NSM },
	{ 0x0374, 0x0375, NEUTRAL_ON },
	{ 0x037E, 0x0385, NEUTRAL_ON },
	{ 0x0387, 0x0000, NEUTRAL_ON },
	{ 0x03F6, 0x0000, NEUTRAL_ON },
	{ 0x0483, 0x0489, WEAK_NSM },
	{ 0x058A, 0x0000, NEUTRAL_ON },
	{ 0x0591, 0x05BD, WEAK_NSM },
	{ 0x05BE, 0x0000, STRONG_R },
	{ 0x05BF, 0x0000, WEAK_NSM },
	{ 0x05C0, 0x0000, STRONG_R },
	{ 0x05C1, 0x05C2, WEAK_NSM },
	{ 0x05C3, 0x0000, STRONG_R },
173 174 175
	{ 0x05C4, 0x05C5, WEAK_NSM },
	{ 0x05C6, 0x0000, STRONG_R },
	{ 0x05C7, 0x0000, WEAK_NSM },
Eli Zaretskii's avatar
Eli Zaretskii committed
176 177 178 179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196 197 198 199 200 201 202 203 204 205 206 207 208 209 210 211 212 213 214 215 216 217 218 219 220 221 222 223 224 225 226 227 228 229 230 231 232 233 234 235 236 237 238 239 240 241 242 243 244 245 246 247 248 249 250 251 252 253 254 255 256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277 278 279 280 281 282 283 284 285 286 287 288 289 290 291 292 293 294 295 296 297 298 299 300 301 302 303 304 305 306 307 308 309 310 311 312 313 314 315 316 317 318 319 320 321 322 323 324 325 326 327 328 329 330 331 332 333 334 335 336 337 338 339 340 341 342 343 344 345 346 347 348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364 365 366 367 368 369 370 371 372 373 374 375 376 377 378 379 380 381 382 383 384 385 386 387 388 389 390 391 392 393 394
	{ 0x05D0, 0x05F4, STRONG_R },
	{ 0x060C, 0x0000, WEAK_CS },
	{ 0x061B, 0x064A, STRONG_AL },
	{ 0x064B, 0x0655, WEAK_NSM },
	{ 0x0660, 0x0669, WEAK_AN },
	{ 0x066A, 0x0000, WEAK_ET },
	{ 0x066B, 0x066C, WEAK_AN },
	{ 0x066D, 0x066F, STRONG_AL },
	{ 0x0670, 0x0000, WEAK_NSM },
	{ 0x0671, 0x06D5, STRONG_AL },
	{ 0x06D6, 0x06DC, WEAK_NSM },
	{ 0x06DD, 0x0000, STRONG_AL },
	{ 0x06DE, 0x06E4, WEAK_NSM },
	{ 0x06E5, 0x06E6, STRONG_AL },
	{ 0x06E7, 0x06E8, WEAK_NSM },
	{ 0x06E9, 0x0000, NEUTRAL_ON },
	{ 0x06EA, 0x06ED, WEAK_NSM },
	{ 0x06F0, 0x06F9, WEAK_EN },
	{ 0x06FA, 0x070D, STRONG_AL },
	{ 0x070F, 0x0000, WEAK_BN },
	{ 0x0710, 0x0000, STRONG_AL },
	{ 0x0711, 0x0000, WEAK_NSM },
	{ 0x0712, 0x072C, STRONG_AL },
	{ 0x0730, 0x074A, WEAK_NSM },
	{ 0x0780, 0x07A5, STRONG_AL },
	{ 0x07A6, 0x07B0, WEAK_NSM },
	{ 0x07B1, 0x0000, STRONG_AL },
	{ 0x0901, 0x0902, WEAK_NSM },
	{ 0x093C, 0x0000, WEAK_NSM },
	{ 0x0941, 0x0948, WEAK_NSM },
	{ 0x094D, 0x0000, WEAK_NSM },
	{ 0x0951, 0x0954, WEAK_NSM },
	{ 0x0962, 0x0963, WEAK_NSM },
	{ 0x0981, 0x0000, WEAK_NSM },
	{ 0x09BC, 0x0000, WEAK_NSM },
	{ 0x09C1, 0x09C4, WEAK_NSM },
	{ 0x09CD, 0x0000, WEAK_NSM },
	{ 0x09E2, 0x09E3, WEAK_NSM },
	{ 0x09F2, 0x09F3, WEAK_ET },
	{ 0x0A02, 0x0000, WEAK_NSM },
	{ 0x0A3C, 0x0000, WEAK_NSM },
	{ 0x0A41, 0x0A4D, WEAK_NSM },
	{ 0x0A70, 0x0A71, WEAK_NSM },
	{ 0x0A81, 0x0A82, WEAK_NSM },
	{ 0x0ABC, 0x0000, WEAK_NSM },
	{ 0x0AC1, 0x0AC8, WEAK_NSM },
	{ 0x0ACD, 0x0000, WEAK_NSM },
	{ 0x0B01, 0x0000, WEAK_NSM },
	{ 0x0B3C, 0x0000, WEAK_NSM },
	{ 0x0B3F, 0x0000, WEAK_NSM },
	{ 0x0B41, 0x0B43, WEAK_NSM },
	{ 0x0B4D, 0x0B56, WEAK_NSM },
	{ 0x0B82, 0x0000, WEAK_NSM },
	{ 0x0BC0, 0x0000, WEAK_NSM },
	{ 0x0BCD, 0x0000, WEAK_NSM },
	{ 0x0C3E, 0x0C40, WEAK_NSM },
	{ 0x0C46, 0x0C56, WEAK_NSM },
	{ 0x0CBF, 0x0000, WEAK_NSM },
	{ 0x0CC6, 0x0000, WEAK_NSM },
	{ 0x0CCC, 0x0CCD, WEAK_NSM },
	{ 0x0D41, 0x0D43, WEAK_NSM },
	{ 0x0D4D, 0x0000, WEAK_NSM },
	{ 0x0DCA, 0x0000, WEAK_NSM },
	{ 0x0DD2, 0x0DD6, WEAK_NSM },
	{ 0x0E31, 0x0000, WEAK_NSM },
	{ 0x0E34, 0x0E3A, WEAK_NSM },
	{ 0x0E3F, 0x0000, WEAK_ET },
	{ 0x0E47, 0x0E4E, WEAK_NSM },
	{ 0x0EB1, 0x0000, WEAK_NSM },
	{ 0x0EB4, 0x0EBC, WEAK_NSM },
	{ 0x0EC8, 0x0ECD, WEAK_NSM },
	{ 0x0F18, 0x0F19, WEAK_NSM },
	{ 0x0F35, 0x0000, WEAK_NSM },
	{ 0x0F37, 0x0000, WEAK_NSM },
	{ 0x0F39, 0x0000, WEAK_NSM },
	{ 0x0F3A, 0x0F3D, NEUTRAL_ON },
	{ 0x0F71, 0x0F7E, WEAK_NSM },
	{ 0x0F80, 0x0F84, WEAK_NSM },
	{ 0x0F86, 0x0F87, WEAK_NSM },
	{ 0x0F90, 0x0FBC, WEAK_NSM },
	{ 0x0FC6, 0x0000, WEAK_NSM },
	{ 0x102D, 0x1030, WEAK_NSM },
	{ 0x1032, 0x1037, WEAK_NSM },
	{ 0x1039, 0x0000, WEAK_NSM },
	{ 0x1058, 0x1059, WEAK_NSM },
	{ 0x1680, 0x0000, NEUTRAL_WS },
	{ 0x169B, 0x169C, NEUTRAL_ON },
	{ 0x1712, 0x1714, WEAK_NSM },
	{ 0x1732, 0x1734, WEAK_NSM },
	{ 0x1752, 0x1753, WEAK_NSM },
	{ 0x1772, 0x1773, WEAK_NSM },
	{ 0x17B7, 0x17BD, WEAK_NSM },
	{ 0x17C6, 0x0000, WEAK_NSM },
	{ 0x17C9, 0x17D3, WEAK_NSM },
	{ 0x17DB, 0x0000, WEAK_ET },
	{ 0x1800, 0x180A, NEUTRAL_ON },
	{ 0x180B, 0x180D, WEAK_NSM },
	{ 0x180E, 0x0000, WEAK_BN },
	{ 0x18A9, 0x0000, WEAK_NSM },
	{ 0x1FBD, 0x0000, NEUTRAL_ON },
	{ 0x1FBF, 0x1FC1, NEUTRAL_ON },
	{ 0x1FCD, 0x1FCF, NEUTRAL_ON },
	{ 0x1FDD, 0x1FDF, NEUTRAL_ON },
	{ 0x1FED, 0x1FEF, NEUTRAL_ON },
	{ 0x1FFD, 0x1FFE, NEUTRAL_ON },
	{ 0x2000, 0x200A, NEUTRAL_WS },
	{ 0x200B, 0x200D, WEAK_BN },
	{ 0x200F, 0x0000, STRONG_R },
	{ 0x2010, 0x2027, NEUTRAL_ON },
	{ 0x2028, 0x0000, NEUTRAL_WS },
	{ 0x2029, 0x0000, NEUTRAL_B },
	{ 0x202A, 0x0000, LRE },
	{ 0x202B, 0x0000, RLE },
	{ 0x202C, 0x0000, PDF },
	{ 0x202D, 0x0000, LRO },
	{ 0x202E, 0x0000, RLO },
	{ 0x202F, 0x0000, NEUTRAL_WS },
	{ 0x2030, 0x2034, WEAK_ET },
	{ 0x2035, 0x2057, NEUTRAL_ON },
	{ 0x205F, 0x0000, NEUTRAL_WS },
	{ 0x2060, 0x206F, WEAK_BN },
	{ 0x2070, 0x0000, WEAK_EN },
	{ 0x2074, 0x2079, WEAK_EN },
	{ 0x207A, 0x207B, WEAK_ET },
	{ 0x207C, 0x207E, NEUTRAL_ON },
	{ 0x2080, 0x2089, WEAK_EN },
	{ 0x208A, 0x208B, WEAK_ET },
	{ 0x208C, 0x208E, NEUTRAL_ON },
	{ 0x20A0, 0x20B1, WEAK_ET },
	{ 0x20D0, 0x20EA, WEAK_NSM },
	{ 0x2100, 0x2101, NEUTRAL_ON },
	{ 0x2103, 0x2106, NEUTRAL_ON },
	{ 0x2108, 0x2109, NEUTRAL_ON },
	{ 0x2114, 0x0000, NEUTRAL_ON },
	{ 0x2116, 0x2118, NEUTRAL_ON },
	{ 0x211E, 0x2123, NEUTRAL_ON },
	{ 0x2125, 0x0000, NEUTRAL_ON },
	{ 0x2127, 0x0000, NEUTRAL_ON },
	{ 0x2129, 0x0000, NEUTRAL_ON },
	{ 0x212E, 0x0000, WEAK_ET },
	{ 0x2132, 0x0000, NEUTRAL_ON },
	{ 0x213A, 0x0000, NEUTRAL_ON },
	{ 0x2140, 0x2144, NEUTRAL_ON },
	{ 0x214A, 0x215F, NEUTRAL_ON },
	{ 0x2190, 0x2211, NEUTRAL_ON },
	{ 0x2212, 0x2213, WEAK_ET },
	{ 0x2214, 0x2335, NEUTRAL_ON },
	{ 0x237B, 0x2394, NEUTRAL_ON },
	{ 0x2396, 0x244A, NEUTRAL_ON },
	{ 0x2460, 0x249B, WEAK_EN },
	{ 0x24EA, 0x0000, WEAK_EN },
	{ 0x24EB, 0x2FFB, NEUTRAL_ON },
	{ 0x3000, 0x0000, NEUTRAL_WS },
	{ 0x3001, 0x3004, NEUTRAL_ON },
	{ 0x3008, 0x3020, NEUTRAL_ON },
	{ 0x302A, 0x302F, WEAK_NSM },
	{ 0x3030, 0x0000, NEUTRAL_ON },
	{ 0x3036, 0x3037, NEUTRAL_ON },
	{ 0x303D, 0x303F, NEUTRAL_ON },
	{ 0x3099, 0x309A, WEAK_NSM },
	{ 0x309B, 0x309C, NEUTRAL_ON },
	{ 0x30A0, 0x0000, NEUTRAL_ON },
	{ 0x30FB, 0x0000, NEUTRAL_ON },
	{ 0x3251, 0x325F, NEUTRAL_ON },
	{ 0x32B1, 0x32BF, NEUTRAL_ON },
	{ 0xA490, 0xA4C6, NEUTRAL_ON },
	{ 0xFB1D, 0x0000, STRONG_R },
	{ 0xFB1E, 0x0000, WEAK_NSM },
	{ 0xFB1F, 0xFB28, STRONG_R },
	{ 0xFB29, 0x0000, WEAK_ET },
	{ 0xFB2A, 0xFB4F, STRONG_R },
	{ 0xFB50, 0xFD3D, STRONG_AL },
	{ 0xFD3E, 0xFD3F, NEUTRAL_ON },
	{ 0xFD50, 0xFDFC, STRONG_AL },
	{ 0xFE00, 0xFE23, WEAK_NSM },
	{ 0xFE30, 0xFE4F, NEUTRAL_ON },
	{ 0xFE50, 0x0000, WEAK_CS },
	{ 0xFE51, 0x0000, NEUTRAL_ON },
	{ 0xFE52, 0x0000, WEAK_CS },
	{ 0xFE54, 0x0000, NEUTRAL_ON },
	{ 0xFE55, 0x0000, WEAK_CS },
	{ 0xFE56, 0xFE5E, NEUTRAL_ON },
	{ 0xFE5F, 0x0000, WEAK_ET },
	{ 0xFE60, 0xFE61, NEUTRAL_ON },
	{ 0xFE62, 0xFE63, WEAK_ET },
	{ 0xFE64, 0xFE68, NEUTRAL_ON },
	{ 0xFE69, 0xFE6A, WEAK_ET },
	{ 0xFE6B, 0x0000, NEUTRAL_ON },
	{ 0xFE70, 0xFEFC, STRONG_AL },
	{ 0xFEFF, 0x0000, WEAK_BN },
	{ 0xFF01, 0xFF02, NEUTRAL_ON },
	{ 0xFF03, 0xFF05, WEAK_ET },
	{ 0xFF06, 0xFF0A, NEUTRAL_ON },
	{ 0xFF0B, 0x0000, WEAK_ET },
	{ 0xFF0C, 0x0000, WEAK_CS },
	{ 0xFF0D, 0x0000, WEAK_ET },
	{ 0xFF0E, 0x0000, WEAK_CS },
	{ 0xFF0F, 0x0000, WEAK_ES },
	{ 0xFF10, 0xFF19, WEAK_EN },
	{ 0xFF1A, 0x0000, WEAK_CS },
	{ 0xFF1B, 0xFF20, NEUTRAL_ON },
	{ 0xFF3B, 0xFF40, NEUTRAL_ON },
	{ 0xFF5B, 0xFF65, NEUTRAL_ON },
	{ 0xFFE0, 0xFFE1, WEAK_ET },
	{ 0xFFE2, 0xFFE4, NEUTRAL_ON },
	{ 0xFFE5, 0xFFE6, WEAK_ET },
	{ 0xFFE8, 0xFFEE, NEUTRAL_ON },
	{ 0xFFF9, 0xFFFB, WEAK_BN },
	{ 0xFFFC, 0xFFFD, NEUTRAL_ON },
	{ 0x1D167, 0x1D169, WEAK_NSM },
	{ 0x1D173, 0x1D17A, WEAK_BN },
	{ 0x1D17B, 0x1D182, WEAK_NSM },
	{ 0x1D185, 0x1D18B, WEAK_NSM },
	{ 0x1D1AA, 0x1D1AD, WEAK_NSM },
	{ 0x1D7CE, 0x1D7FF, WEAK_EN },
	{ 0xE0001, 0xE007F, WEAK_BN } };
  int i;

  bidi_type_table = Fmake_char_table (Qnil, make_number (STRONG_L));
395
  staticpro (&bidi_type_table);
Eli Zaretskii's avatar
Eli Zaretskii committed
396 397

  for (i = 0; i < sizeof bidi_type / sizeof bidi_type[0]; i++)
398 399
    char_table_set_range (bidi_type_table, bidi_type[i].from,
			  bidi_type[i].to ? bidi_type[i].to : bidi_type[i].from,
Eli Zaretskii's avatar
Eli Zaretskii committed
400
			  make_number (bidi_type[i].type));
401

402 403
  Qparagraph_start = intern ("paragraph-start");
  staticpro (&Qparagraph_start);
Eli Zaretskii's avatar
Eli Zaretskii committed
404 405 406 407
  paragraph_start_re = Fsymbol_value (Qparagraph_start);
  if (!STRINGP (paragraph_start_re))
    paragraph_start_re = build_string ("\f\\|[ \t]*$");
  staticpro (&paragraph_start_re);
408 409
  Qparagraph_separate = intern ("paragraph-separate");
  staticpro (&Qparagraph_separate);
Eli Zaretskii's avatar
Eli Zaretskii committed
410 411 412 413
  paragraph_separate_re = Fsymbol_value (Qparagraph_separate);
  if (!STRINGP (paragraph_separate_re))
    paragraph_separate_re = build_string ("[ \t\f]*$");
  staticpro (&paragraph_separate_re);
Eli Zaretskii's avatar
Eli Zaretskii committed
414 415 416
  bidi_initialized = 1;
}

417 418
/* Return the bidi type of a character CH, subject to the current
   directional OVERRIDE.  */
419
static INLINE bidi_type_t
420
bidi_get_type (int ch, bidi_dir_t override)
Eli Zaretskii's avatar
Eli Zaretskii committed
421
{
422 423
  bidi_type_t default_type;

424 425
  if (ch == BIDI_EOB)
    return NEUTRAL_B;
426 427
  if (ch < 0 || ch > MAX_CHAR)
    abort ();
428 429 430 431 432 433 434 435 436 437 438 439 440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459

  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
460 461
}

462 463 464 465 466 467 468
void
bidi_check_type (bidi_type_t type)
{
  if (type < UNKNOWN_BT || type > NEUTRAL_ON)
    abort ();
}

Eli Zaretskii's avatar
Eli Zaretskii committed
469
/* Given a bidi TYPE of a character, return its category.  */
470
static INLINE bidi_category_t
Eli Zaretskii's avatar
Eli Zaretskii committed
471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502 503
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 ();
    }
}

504 505 506 507
/* Return the mirrored character of C, if any.

   Note: The conditions in UAX#9 clause L4 must be tested by the
   caller.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
508 509 510 511 512 513
/* FIXME: exceedingly temporary!  Should consult the Unicode database
   of character properties.  */
int
bidi_mirror_char (int c)
{
  static const char mirrored_pairs[] = "()<>[]{}";
514
  const char *p = c > 0 && c < 128 ? strchr (mirrored_pairs, c) : NULL;
Eli Zaretskii's avatar
Eli Zaretskii committed
515 516 517 518 519

  if (p)
    {
      size_t i = p - mirrored_pairs;

520
      return mirrored_pairs [(i ^ 1)];
Eli Zaretskii's avatar
Eli Zaretskii committed
521 522 523 524 525 526
    }
  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.  */
527
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
528 529 530 531
bidi_copy_it (struct bidi_it *to, struct bidi_it *from)
{
  int i;

532 533
  /* 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
534 535 536 537 538 539 540 541 542

  /* 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.  */

543 544 545
#define BIDI_CACHE_CHUNK 200
static struct bidi_it *bidi_cache;
static size_t bidi_cache_size = 0;
546
static size_t elsz = sizeof (struct bidi_it);
547 548
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
549

550
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
551 552 553 554 555 556
bidi_cache_reset (void)
{
  bidi_cache_idx = 0;
  bidi_cache_last_idx = -1;
}

557 558 559 560 561
static INLINE void
bidi_cache_shrink (void)
{
  if (bidi_cache_size > BIDI_CACHE_CHUNK)
    {
562 563 564
      bidi_cache_size = BIDI_CACHE_CHUNK;
      bidi_cache =
	(struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
565 566 567 568
    }
  bidi_cache_reset ();
}

569
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
570 571 572 573 574 575 576 577 578 579 580 581 582 583 584 585
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.  */
586
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641 642 643 644 645 646 647 648 649 650 651 652 653 654 655 656 657 658 659 660 661 662 663 664 665 666 667 668 669 670 671 672 673 674 675 676
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;
}

677
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
678 679 680 681 682 683 684 685 686 687 688 689
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;
690 691 692
      /* Enlarge the cache as needed.  */
      if (idx >= bidi_cache_size)
	{
693
	  bidi_cache_size += BIDI_CACHE_CHUNK;
694
	  bidi_cache =
695
	    (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
696
	}
697 698 699 700 701 702 703 704 705 706
      /* 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
707 708 709 710 711 712 713 714 715
      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;
716
      bidi_check_type (bidi_it->type);
717 718
      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
719 720 721 722 723 724 725 726 727 728 729 730 731 732 733 734
      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;
}

735
static INLINE bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
736 737 738 739 740 741 742 743
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;

744
      bidi_copy_it (bidi_it, &bidi_cache[i]);
Eli Zaretskii's avatar
Eli Zaretskii committed
745 746 747 748 749 750 751 752 753 754
      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;
}

755
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
756 757 758 759 760 761 762
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;
}

763 764 765 766 767
/* 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.  */
768
static EMACS_INT
769
bidi_at_paragraph_end (EMACS_INT charpos, EMACS_INT bytepos)
Eli Zaretskii's avatar
Eli Zaretskii committed
770
{
771
  /* FIXME: Why Fbuffer_local_value rather than just Fsymbol_value?  */
Eli Zaretskii's avatar
Eli Zaretskii committed
772 773
  Lisp_Object sep_re;
  Lisp_Object start_re;
774 775
  EMACS_INT val;

Eli Zaretskii's avatar
Eli Zaretskii committed
776 777
  sep_re = paragraph_separate_re;
  start_re = paragraph_start_re;
778 779 780 781 782 783 784 785 786

  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
787

788
  return val;
Eli Zaretskii's avatar
Eli Zaretskii committed
789 790 791 792 793 794
}

/* 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.  */
795
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
796 797 798 799 800 801 802 803 804 805 806
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.  */
807
  if (!bidi_it->prev_was_pdf || level_before < level_after)
Eli Zaretskii's avatar
Eli Zaretskii committed
808 809 810 811 812 813
    /* 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;
814 815
  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
816 817 818
  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;
819 820
  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
821 822 823
  bidi_it->ignore_bn_limit = 0; /* meaning it's unknown */
}

824
static void
825 826 827 828 829 830 831 832 833
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;
834 835
  bidi_set_sor_type (bidi_it,
		     bidi_it->paragraph_dir == R2L ? 1 : 0,
836 837 838 839 840 841 842 843
		     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
844
bidi_find_paragraph_start (EMACS_INT pos, EMACS_INT pos_byte)
845
{
Eli Zaretskii's avatar
Eli Zaretskii committed
846
  Lisp_Object re = paragraph_start_re;
847 848 849 850 851
  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)
    {
852 853
      pos = find_next_newline_no_quit (pos - 1, -1);
      pos_byte = CHAR_TO_BYTE (pos);
854
    }
855
  return pos_byte;
856 857
}

858
/* Determine the direction, a.k.a. base embedding level, of the
859 860 861 862 863 864 865
   paragraph we are about to iterate through.  If DIR is either L2R or
   R2L, just use that.  Otherwise, determine the paragraph direction
   from the first strong character of the paragraph.

   Note that this 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
866 867 868
void
bidi_paragraph_init (bidi_dir_t dir, struct bidi_it *bidi_it)
{
869
  EMACS_INT bytepos = bidi_it->bytepos;
870

871 872 873
  /* Special case for an empty buffer. */
  if (bytepos == BEGV_BYTE && bytepos == ZV_BYTE)
    dir = L2R;
874
  /* We should never be called at EOB or before BEGV.  */
875
  else if (bytepos >= ZV_BYTE || bytepos < BEGV_BYTE)
876 877
    abort ();

878 879 880 881 882 883 884 885 886 887
  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
888 889
  else if (dir == NEUTRAL_DIR)	/* P2 */
    {
890 891 892
      int ch, ch_len;
      EMACS_INT pos;
      bidi_type_t type;
893

894 895 896
      if (!bidi_initialized)
	bidi_initialize ();

897 898
      /* If we are inside a paragraph separator, we are just waiting
	 for the separator to be exhausted; use the previous paragraph
899 900 901 902
	 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)
903 904
	return;

905
      /* If we are on a newline, get past it to where the next
906 907 908
	 paragraph might start.  But don't do that at BEGV since then
	 we are potentially in a new paragraph that doesn't yet
	 exist.  */
909
      pos = bidi_it->charpos;
910
      if (bytepos > BEGV_BYTE && FETCH_CHAR (bytepos) == '\n')
911
	{
912
	  bytepos++;
913
	  pos++;
914
	}
915 916 917 918

      /* We are either at the beginning of a paragraph or in the
	 middle of it.  Find where this paragraph starts.  */
      bytepos = bidi_find_paragraph_start (pos, bytepos);
919

920 921
      bidi_it->separator_limit = -1;
      bidi_it->new_paragraph = 0;
922 923
      ch = FETCH_CHAR (bytepos);
      ch_len = CHAR_BYTES (ch);
924
      pos = BYTE_TO_CHAR (bytepos);
925
      type = bidi_get_type (ch, NEUTRAL_DIR);
Eli Zaretskii's avatar
Eli Zaretskii committed
926

927
      for (pos++, bytepos += ch_len;
Eli Zaretskii's avatar
Eli Zaretskii committed
928 929 930 931 932 933 934 935
	   /* 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));
936
	   type = bidi_get_type (ch, NEUTRAL_DIR))
Eli Zaretskii's avatar
Eli Zaretskii committed
937
	{
938
	  if (type == NEUTRAL_B && bidi_at_paragraph_end (pos, bytepos) >= -1)
Eli Zaretskii's avatar
Eli Zaretskii committed
939
	    break;
940 941 942 943 944 945
	  if (bytepos >= ZV_BYTE)
	    {
	      /* Pretend there's a paragraph separator at end of buffer.  */
	      type = NEUTRAL_B;
	      break;
	    }
Eli Zaretskii's avatar
Eli Zaretskii committed
946 947 948
	  FETCH_CHAR_ADVANCE (ch, pos, bytepos);
	}
      if (type == STRONG_R || type == STRONG_AL) /* P3 */
949 950 951
	bidi_it->paragraph_dir = R2L;
      else if (type == STRONG_L)
	bidi_it->paragraph_dir = L2R;
Eli Zaretskii's avatar
Eli Zaretskii committed
952
    }
953 954
  else
    abort ();
Eli Zaretskii's avatar
Eli Zaretskii committed
955

956 957 958
  /* Contrary to UAX#9 clause P3, we only default the paragraph
     direction to L2R if we have no previous usable paragraph
     direction.  */
959
  if (bidi_it->paragraph_dir != L2R && bidi_it->paragraph_dir != R2L)
960
    bidi_it->paragraph_dir = L2R; /* P3 and ``higher protocols'' */
961
  if (bidi_it->paragraph_dir == R2L)
962
    bidi_it->level_stack[0].level = 1;
963
  else
964
    bidi_it->level_stack[0].level = 0;
965 966

  bidi_line_init (bidi_it);
Eli Zaretskii's avatar
Eli Zaretskii committed
967 968
}

969 970
/* Do whatever UAX#9 clause X8 says should be done at paragraph's
   end.  */
971
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
972 973 974 975 976 977 978 979
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;
}

980
/* Initialize the bidi iterator from buffer position CHARPOS.  */
Eli Zaretskii's avatar
Eli Zaretskii committed
981
void
982
bidi_init_it (EMACS_INT charpos, EMACS_INT bytepos, struct bidi_it *bidi_it)
Eli Zaretskii's avatar
Eli Zaretskii committed
983 984 985
{
  if (! bidi_initialized)
    bidi_initialize ();
986 987
  bidi_it->charpos = charpos;
  bidi_it->bytepos = bytepos;
988 989 990
  bidi_it->first_elt = 1;
  bidi_set_paragraph_end (bidi_it);
  bidi_it->new_paragraph = 1;
991
  bidi_it->separator_limit = -1;
Eli Zaretskii's avatar
Eli Zaretskii committed
992
  bidi_it->type = NEUTRAL_B;
993 994
  bidi_it->type_after_w1 = NEUTRAL_B;
  bidi_it->orig_type = NEUTRAL_B;
Eli Zaretskii's avatar
Eli Zaretskii committed
995
  bidi_it->prev_was_pdf = 0;
996 997
  bidi_it->prev.type = bidi_it->prev.type_after_w1 =
    bidi_it->prev.orig_type = UNKNOWN_BT;
998 999
  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
1000 1001
  bidi_it->next_for_neutral.charpos = -1;
  bidi_it->next_for_neutral.type =
1002 1003
    bidi_it->next_for_neutral.type_after_w1 =
    bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
1004 1005
  bidi_it->prev_for_neutral.charpos = -1;
  bidi_it->prev_for_neutral.type =
1006 1007
    bidi_it->prev_for_neutral.type_after_w1 =
    bidi_it->prev_for_neutral.orig_type = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
1008
  bidi_it->sor = L2R;	 /* FIXME: should it be user-selectable? */
1009
  bidi_cache_shrink ();
Eli Zaretskii's avatar
Eli Zaretskii committed
1010 1011 1012 1013
}

/* Push the current embedding level and override status; reset the
   current level to LEVEL and the current override status to OVERRIDE.  */
1014
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
1015 1016 1017 1018 1019 1020 1021 1022 1023 1024 1025 1026
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.  */
1027
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
1028 1029 1030 1031 1032 1033 1034 1035 1036
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.  */
1037
static INLINE void
Eli Zaretskii's avatar
Eli Zaretskii committed
1038 1039 1040 1041 1042 1043
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;
1044
  bidi_check_type (bidi_it->type);
1045 1046
  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
1047
  saved_info->orig_type = bidi_it->orig_type;
1048
  bidi_check_type (bidi_it->orig_type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1049 1050 1051 1052
}

/* Resolve the type of a neutral character according to the type of
   surrounding strong text and the current embedding level.  */
1053
static INLINE bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
1054 1055 1056 1057 1058 1059 1060 1061 1062 1063 1064 1065 1066 1067 1068 1069
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;
}

1070
static INLINE int
Eli Zaretskii's avatar
Eli Zaretskii committed
1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084 1085 1086 1087 1088 1089 1090 1091
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;

1092 1093
  if (bidi_it->bytepos < BEGV_BYTE	/* after reseat to BEGV? */
      || bidi_it->first_elt)
1094
    {
1095 1096 1097 1098
      bidi_it->first_elt = 0;
      if (bidi_it->charpos < BEGV)
	bidi_it->charpos = BEGV;
      bidi_it->bytepos = CHAR_TO_BYTE (bidi_it->charpos);
1099
    }
1100
  else if (bidi_it->bytepos < ZV_BYTE)	/* don't move at ZV */
Eli Zaretskii's avatar
Eli Zaretskii committed
1101 1102
    {
      bidi_it->charpos++;
1103 1104
      if (bidi_it->ch_len == 0)
	abort ();
Eli Zaretskii's avatar
Eli Zaretskii committed
1105 1106 1107 1108 1109 1110 1111 1112 1113
      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; */
1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
  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
1124 1125
  bidi_it->ch = curchar;

1126 1127 1128 1129 1130
  /* 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);
1131 1132
  bidi_it->orig_type = type;
  bidi_check_type (bidi_it->orig_type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1133 1134 1135 1136

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

1137
  bidi_it->type_after_w1 = UNKNOWN_BT;
Eli Zaretskii's avatar
Eli Zaretskii committed
1138 1139 1140 1141 1142

  switch (type)
    {
      case RLE:	/* X2 */
      case RLO:	/* X4 */
1143 1144
	bidi_it->type_after_w1 = type;
	bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1145 1146 1147 1148 1149 1150 1151 1152
	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;
1153
		if (bidi_it->type_after_w1 == RLE)
Eli Zaretskii's avatar
Eli Zaretskii committed
1154 1155 1156 1157 1158 1159 1160 1161 1162 1163 1164 1165 1166 1167 1168 1169
		  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++;
	      }
	  }
1170
	else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
1171 1172 1173 1174 1175
		 || bidi_it->next_en_pos > bidi_it->charpos)
	  type = WEAK_EN;
	break;
      case LRE:	/* X3 */
      case LRO:	/* X5 */
1176 1177
	bidi_it->type_after_w1 = type;
	bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1178 1179 1180 1181 1182 1183 1184 1185
	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);
1186
		if (bidi_it->type_after_w1 == LRE)
Eli Zaretskii's avatar
Eli Zaretskii committed
1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205
		  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++;
	      }
	  }
1206
	else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
1207 1208 1209 1210
		 || bidi_it->next_en_pos > bidi_it->charpos)
	  type = WEAK_EN;
	break;
      case PDF:	/* X7 */
1211 1212
	bidi_it->type_after_w1 = type;
	bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1213 1214 1215 1216 1217 1218 1219 1220 1221 1222 1223 1224 1225 1226 1227 1228 1229 1230 1231
	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--;
	      }
	  }
1232
	else if (bidi_it->prev.type_after_w1 == WEAK_EN /* W5/Retaining */
Eli Zaretskii's avatar
Eli Zaretskii committed
1233 1234 1235 1236 1237 1238 1239 1240 1241
		 || bidi_it->next_en_pos > bidi_it->charpos)
	  type = WEAK_EN;
	break;
      default:
	/* Nothing.  */
	break;
    }

  bidi_it->type = type;
1242
  bidi_check_type (bidi_it->type);
Eli Zaretskii's avatar
Eli Zaretskii committed
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260

  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 */
1261
      && bidi_it->bytepos < ZV_BYTE    /* not already at EOB */
Eli Zaretskii's avatar
Eli Zaretskii committed
1262 1263 1264 1265 1266 1267 1268 1269 1270 1271 1272 1273 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
      && 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 */
    {
1309
      bidi_set_paragraph_end (bidi_it);
1310 1311
      /* This is needed by bidi_resolve_weak below, and in L1.  */
      bidi_it->type_after_w1 = bidi_it->type;
1312
      bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1313 1314 1315 1316 1317 1318 1319
    }

  return new_level;
}

/* Advance in the buffer, resolve weak types and return the type of
   the next character after weak type resolution.  */
1320
static bidi_type_t
Eli Zaretskii's avatar
Eli Zaretskii committed
1321 1322 1323 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
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)
1352 1353
    bidi_it->type_after_w1 = type;	/* needed in L1 */
  bidi_check_type (bidi_it->type_after_w1);
Eli Zaretskii's avatar
Eli Zaretskii committed
1354 1355 1356 1357 1358 1359 1360

  /* 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;
1361
  else
Eli Zaretskii's avatar
Eli Zaretskii committed
1362
    {
1363
      if (type == WEAK_NSM)	/* W1 */
Eli Zaretskii's avatar
Eli Zaretskii committed
1364
	{
1365
	  /* Note that we don't need to consider the case where the
1366 1367 1368 1369 1370 1371
	     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.  */
1372 1373 1374
	  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)
1375
	    type = bidi_it->prev.type_after_w1;
1376 1377 1378 1379 1380 1381
	  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
1382
	}
Eli Zaretskii's avatar