/* Low-level bidirectional buffer-scanning functions for GNU Emacs.
- Copyright (C) 2000, 2001, 2004, 2005 Free Software Foundation, Inc.
+ Copyright (C) 2000, 2001, 2004, 2005, 2009, 2010
+ Free Software Foundation, Inc.
This file is part of GNU Emacs.
-GNU Emacs is free software; you can redistribute it and/or modify
+GNU Emacs is free software: you can redistribute it and/or modify
it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 2, or (at your option)
-any later version.
+the Free Software Foundation, either version 3 of the License, or
+(at your option) any later version.
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
-along with GNU Emacs; see the file COPYING. If not, write to
-the Free Software Foundation, Inc., 51 Franklin Street, Fifth Floor,
-Boston, MA 02110-1301, USA. */
+along with GNU Emacs. If not, see <http://www.gnu.org/licenses/>. */
/* Written by Eli Zaretskii <eliz@gnu.org>.
as per UAX#9, a part of the Unicode Standard.
Unlike the reference and most other implementations, this one is
- designed to be called once for every character in the buffer.
+ designed to be called once for every character in the buffer or
+ string.
The main entry point is bidi_get_next_char_visually. Each time it
is called, it finds the next character in the visual order, and
more details about its algorithm that finds the next visual-order
character by resolving their levels on the fly.
+ 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.
+
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.
#include <string.h>
#endif
+#include <setjmp.h>
+
#include "lisp.h"
#include "buffer.h"
#include "character.h"
int bidi_ignore_explicit_marks_for_paragraph_level = 1;
-/* FIXME: Should be user-definable. */
-bidi_dir_t bidi_overriding_paragraph_direction = L2R;
-
-/* FIXME: Unused? */
-#define ASCII_BIDI_TYPE_SET(STR, TYPE) \
- do { \
- unsigned char *p; \
- for (p = (STR); *p; p++) \
- CHAR_TABLE_SET (bidi_type_table, *p, (TYPE)); \
- } while (0)
+static Lisp_Object fallback_paragraph_start_re, fallback_paragraph_separate_re;
+static Lisp_Object Qparagraph_start, Qparagraph_separate;
static void
bidi_initialize ()
{ 0x0021, 0x0022, NEUTRAL_ON },
{ 0x0023, 0x0025, WEAK_ET },
{ 0x0026, 0x002A, NEUTRAL_ON },
- { 0x002B, 0x0000, WEAK_ET },
+ { 0x002B, 0x0000, WEAK_ES },
{ 0x002C, 0x0000, WEAK_CS },
- { 0x002D, 0x0000, WEAK_ET },
- { 0x002E, 0x0000, WEAK_CS },
- { 0x002F, 0x0000, WEAK_ES },
+ { 0x002D, 0x0000, WEAK_ES },
+ { 0x002E, 0x002F, WEAK_CS },
{ 0x0030, 0x0039, WEAK_EN },
{ 0x003A, 0x0000, WEAK_CS },
{ 0x003B, 0x0040, NEUTRAL_ON },
{ 0x00A1, 0x0000, NEUTRAL_ON },
{ 0x00A2, 0x00A5, WEAK_ET },
{ 0x00A6, 0x00A9, NEUTRAL_ON },
- { 0x00AB, 0x00AF, NEUTRAL_ON },
+ { 0x00AB, 0x00AC, NEUTRAL_ON },
+ { 0x00AD, 0x0000, WEAK_BN },
+ { 0x00AE, 0x00Af, NEUTRAL_ON },
{ 0x00B0, 0x00B1, WEAK_ET },
{ 0x00B2, 0x00B3, WEAK_EN },
{ 0x00B4, 0x0000, NEUTRAL_ON },
{ 0x05C0, 0x0000, STRONG_R },
{ 0x05C1, 0x05C2, WEAK_NSM },
{ 0x05C3, 0x0000, STRONG_R },
- { 0x05C4, 0x0000, WEAK_NSM },
+ { 0x05C4, 0x05C5, WEAK_NSM },
+ { 0x05C6, 0x0000, STRONG_R },
+ { 0x05C7, 0x0000, WEAK_NSM },
{ 0x05D0, 0x05F4, STRONG_R },
{ 0x060C, 0x0000, WEAK_CS },
{ 0x061B, 0x064A, STRONG_AL },
char_table_set_range (bidi_type_table, bidi_type[i].from,
bidi_type[i].to ? bidi_type[i].to : bidi_type[i].from,
make_number (bidi_type[i].type));
+
+ fallback_paragraph_start_re =
+ Fsymbol_value (Fintern_soft (build_string ("paragraph-start"), Qnil));
+ if (!STRINGP (fallback_paragraph_start_re))
+ fallback_paragraph_start_re = build_string ("\f\\|[ \t]*$");
+ staticpro (&fallback_paragraph_start_re);
+ Qparagraph_start = intern ("paragraph-start");
+ staticpro (&Qparagraph_start);
+ fallback_paragraph_separate_re =
+ Fsymbol_value (Fintern_soft (build_string ("paragraph-separate"), Qnil));
+ if (!STRINGP (fallback_paragraph_separate_re))
+ fallback_paragraph_separate_re = build_string ("[ \t\f]*$");
+ staticpro (&fallback_paragraph_separate_re);
+ Qparagraph_separate = intern ("paragraph-separate");
+ staticpro (&Qparagraph_separate);
bidi_initialized = 1;
}
-static int
-bidi_is_arabic_number (int ch)
+/* Return the bidi type of a character CH, subject to the current
+ directional OVERRIDE. */
+static INLINE bidi_type_t
+bidi_get_type (int ch, bidi_dir_t override)
{
- return 0; /* FIXME! */
-}
+ bidi_type_t default_type;
-/* Return the bidi type of a character CH. */
-bidi_type_t
-bidi_get_type (int ch)
-{
if (ch == BIDI_EOB)
return NEUTRAL_B;
- return (bidi_type_t) XINT (CHAR_TABLE_REF (bidi_type_table, ch));
+ if (ch < 0 || ch > MAX_CHAR)
+ abort ();
+
+ 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 */
+ }
+ }
}
void
}
/* Given a bidi TYPE of a character, return its category. */
-bidi_category_t
+static INLINE bidi_category_t
bidi_get_category (bidi_type_t type)
{
switch (type)
}
}
+/* Return the mirrored character of C, if any.
+
+ Note: The conditions in UAX#9 clause L4 must be tested by the
+ caller. */
/* FIXME: exceedingly temporary! Should consult the Unicode database
of character properties. */
int
bidi_mirror_char (int c)
{
static const char mirrored_pairs[] = "()<>[]{}";
- const char *p = strchr (mirrored_pairs, c);
+ const char *p = c > 0 && c < 128 ? strchr (mirrored_pairs, c) : NULL;
if (p)
{
size_t i = p - mirrored_pairs;
- if ((i & 1) == 0)
- return mirrored_pairs[i + 1];
- else
- return mirrored_pairs[i - 1];
+ return mirrored_pairs [(i ^ 1)];
}
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. */
-static inline void
+static INLINE void
bidi_copy_it (struct bidi_it *to, struct bidi_it *from)
{
int i;
- /* Copy everything except the level stack. */
- memcpy (to, from, ((int)&((struct bidi_it *)0)->level_stack[0]));
+ /* Copy everything except the level stack and beyond. */
+ memcpy (to, from, ((size_t)&((struct bidi_it *)0)->level_stack[0]));
/* Copy the active part of the level stack. */
to->level_stack[0] = from->level_stack[0]; /* level zero is always in use */
static int bidi_cache_idx;
static int bidi_cache_last_idx;
-static inline void
+static INLINE void
bidi_cache_reset (void)
{
bidi_cache_idx = 0;
bidi_cache_last_idx = -1;
}
-static inline void
+static INLINE void
bidi_cache_fetch_state (int idx, struct bidi_it *bidi_it)
{
int current_scan_dir = bidi_it->scan_dir;
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. */
-static inline int
+static INLINE int
bidi_cache_search (int charpos, int level, int dir)
{
int i, i_start;
return -1;
}
-static inline void
+static INLINE void
bidi_cache_iterator_state (struct bidi_it *bidi_it, int resolved)
{
int idx;
if (idx < 0)
{
idx = bidi_cache_idx;
+ /* Don't overrun the cache limit. */
if (idx > sizeof (bidi_cache) / sizeof (bidi_cache[0]) - 1)
abort ();
+ /* 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;
+ }
bidi_copy_it (&bidi_cache[idx], bidi_it);
if (!resolved)
bidi_cache[idx].resolved_level = -1;
+ bidi_cache[idx].new_paragraph = 0;
}
else
{
bidi_cache_idx = idx + 1;
}
-static inline bidi_type_t
+static INLINE bidi_type_t
bidi_cache_find (int charpos, int level, struct bidi_it *bidi_it)
{
int i = bidi_cache_search (charpos, level, bidi_it->scan_dir);
{
bidi_dir_t current_scan_dir = bidi_it->scan_dir;
- *bidi_it = bidi_cache[i];
+ bidi_copy_it (bidi_it, &bidi_cache[i]);
bidi_cache_last_idx = i;
/* Don't let scan direction from from the cached state override
the current scan direction. */
return UNKNOWN_BT;
}
-static inline int
+static INLINE int
bidi_peek_at_next_level (struct bidi_it *bidi_it)
{
if (bidi_cache_idx == 0 || bidi_cache_last_idx == -1)
return bidi_cache[bidi_cache_last_idx + bidi_it->scan_dir].resolved_level;
}
-/* Return non-zero if buffer's byte position POS is the last character
- of a paragraph. THIS_CH is the character preceding the one at POS in
- the buffer. */
-int
-bidi_at_paragraph_end (int this_ch, int pos)
+/* 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. */
+static EMACS_INT
+bidi_at_paragraph_end (EMACS_INT charpos, EMACS_INT bytepos)
{
- int next_ch;
-
- if (pos >= ZV_BYTE)
- return 1;
+ Lisp_Object sep_re = Fbuffer_local_value (Qparagraph_separate,
+ Fcurrent_buffer ());
+ Lisp_Object start_re = Fbuffer_local_value (Qparagraph_start,
+ Fcurrent_buffer ());
+ EMACS_INT val;
+
+ if (!STRINGP (sep_re))
+ sep_re = fallback_paragraph_separate_re;
+ if (!STRINGP (start_re))
+ start_re = fallback_paragraph_start_re;
+
+ 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;
+ }
- next_ch = FETCH_CHAR (pos);
- /* FIXME: This should support all Unicode characters that can end a
- paragraph. */
- return (this_ch == '\n' && next_ch == '\n');
+ return val;
}
/* 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. */
-static inline void
+static INLINE void
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;
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. */
- if (level_before < level_after || !bidi_it->prev_was_pdf)
+ if (!bidi_it->prev_was_pdf || level_before < level_after)
/* 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->ignore_bn_limit = 0; /* meaning it's unknown */
}
+static void
+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;
+ bidi_set_sor_type (bidi_it,
+ bidi_it->paragraph_dir == R2L ? 1 : 0,
+ 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
+bidi_find_paragraph_start (EMACS_INT pos, EMACS_INT pos_byte)
+{
+ Lisp_Object re = Fbuffer_local_value (Qparagraph_start, Fcurrent_buffer ());
+ EMACS_INT limit = ZV, limit_byte = ZV_BYTE;
+
+ if (!STRINGP (re))
+ re = fallback_paragraph_start_re;
+ while (pos_byte > BEGV_BYTE
+ && fast_looking_at (re, pos, pos_byte, limit, limit_byte, Qnil) < 0)
+ {
+ pos = find_next_newline_no_quit (pos - 1, -1);
+ pos_byte = CHAR_TO_BYTE (pos);
+ }
+ return pos_byte;
+}
+
+/* Determine the direction, a.k.a. base embedding level, of the
+ 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. */
void
bidi_paragraph_init (bidi_dir_t dir, struct bidi_it *bidi_it)
{
- int pos = bidi_it->charpos, bytepos = bidi_it->bytepos;
- int ch;
+ EMACS_INT bytepos = bidi_it->bytepos;
- /* We should never be called at EOB. */
- if (bytepos >= ZV_BYTE)
+ /* Special case for an empty buffer. */
+ if (bytepos == BEGV_BYTE && bytepos == ZV_BYTE)
+ dir = L2R;
+ /* We should never be called at EOB or before BEGV. */
+ else if (bytepos >= ZV_BYTE || bytepos < BEGV_BYTE)
abort ();
- ch = FETCH_CHAR (bytepos);
- bidi_it->ch_len = CHAR_BYTES (ch);
- bidi_it->level_stack[0].level = 0; /* default for L2R */
- if (dir == R2L)
- bidi_it->level_stack[0].level = 1;
+ 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;
+ }
else if (dir == NEUTRAL_DIR) /* P2 */
{
+ int ch, ch_len;
+ EMACS_INT pos;
bidi_type_t type;
- /* FIXME: should actually go to where the paragraph begins and
- start the loop below from there, since UAX#9 says to find the
- first strong directional character in the paragraph. */
+ /* If we are inside a paragraph separator, we are just waiting
+ for the separator to be exhausted; use the previous paragraph
+ 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)
+ return;
+
+ /* If we are on a newline, get past it to where the next
+ paragraph might start. But don't do that at BEGV since then
+ we are potentially in a new paragraph that doesn't yet
+ exist. */
+ pos = bidi_it->charpos;
+ if (bytepos > BEGV_BYTE && FETCH_CHAR (bytepos) == '\n')
+ {
+ bytepos++;
+ pos++;
+ }
+
+ /* 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);
+
+ /* We should always be at the beginning of a new line at this
+ point. */
+ if (!(bytepos == BEGV_BYTE || FETCH_CHAR (bytepos - 1) == '\n'))
+ abort ();
+
+ bidi_it->separator_limit = -1;
+ bidi_it->new_paragraph = 0;
+ ch = FETCH_CHAR (bytepos);
+ ch_len = CHAR_BYTES (ch);
+ pos = BYTE_TO_CHAR (bytepos);
+ type = bidi_get_type (ch, NEUTRAL_DIR);
- for (type = bidi_get_type (ch), pos++, bytepos += bidi_it->ch_len;
+ 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
|| (bidi_ignore_explicit_marks_for_paragraph_level
&& (type == RLE || type == RLO
|| type == LRE || type == LRO));
- type = bidi_get_type (ch))
+ type = bidi_get_type (ch, NEUTRAL_DIR))
{
- if (type == NEUTRAL_B || bidi_at_paragraph_end (ch, bytepos))
+ if (type == NEUTRAL_B && bidi_at_paragraph_end (pos, bytepos) >= -1)
break;
+ 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->level_stack[0].level = 1;
+ bidi_it->paragraph_dir = R2L;
+ else if (type == STRONG_L)
+ bidi_it->paragraph_dir = L2R;
}
- 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->new_paragraph = 0;
- bidi_it->next_en_pos = -1;
- bidi_it->next_for_ws.type = UNKNOWN_BT;
- bidi_set_sor_type (bidi_it, bidi_it->level_stack[0].level, 0); /* X10 */
+ else
+ abort ();
- bidi_cache_reset ();
+ /* Contrary to UAX#9 clause P3, we only default the paragraph
+ direction to L2R if we have no previous usable paragraph
+ direction. */
+ if (bidi_it->paragraph_dir == NEUTRAL_DIR)
+ bidi_it->paragraph_dir = L2R; /* P3 and ``higher protocols'' */
+ if (bidi_it->paragraph_dir == R2L)
+ bidi_it->level_stack[0].level = 1;
+ else
+ bidi_it->level_stack[0].level = 0;
+
+ bidi_line_init (bidi_it);
}
-/* Do whatever UAX#9 clause X8 says should be done at paragraph's end,
- and set the new paragraph flag in the iterator. */
-static inline void
+/* Do whatever UAX#9 clause X8 says should be done at paragraph's
+ end. */
+static INLINE void
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;
- bidi_it->new_paragraph = 1;
}
/* Initialize the bidi iterator from buffer position CHARPOS. */
void
-bidi_init_it (int charpos, int bytepos, struct bidi_it *bidi_it)
+bidi_init_it (EMACS_INT charpos, EMACS_INT bytepos, struct bidi_it *bidi_it)
{
if (! bidi_initialized)
bidi_initialize ();
- bidi_set_paragraph_end (bidi_it);
bidi_it->charpos = charpos;
bidi_it->bytepos = bytepos;
+ bidi_it->first_elt = 1;
+ bidi_set_paragraph_end (bidi_it);
+ bidi_it->new_paragraph = 1;
+ bidi_it->separator_limit = -1;
bidi_it->type = NEUTRAL_B;
- bidi_it->type_after_w1 = UNKNOWN_BT;
- bidi_it->orig_type = UNKNOWN_BT;
+ bidi_it->type_after_w1 = NEUTRAL_B;
+ bidi_it->orig_type = NEUTRAL_B;
bidi_it->prev_was_pdf = 0;
- bidi_it->prev.type = bidi_it->prev.type_after_w1 = UNKNOWN_BT;
+ bidi_it->prev.type = bidi_it->prev.type_after_w1 =
+ bidi_it->prev.orig_type = UNKNOWN_BT;
bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1 =
bidi_it->last_strong.orig_type = UNKNOWN_BT;
bidi_it->next_for_neutral.charpos = -1;
/* Push the current embedding level and override status; reset the
current level to LEVEL and the current override status to OVERRIDE. */
-static inline void
+static INLINE void
bidi_push_embedding_level (struct bidi_it *bidi_it,
int level, bidi_dir_t override)
{
/* Pop the embedding level and directional override status from the
stack, and return the new level. */
-static inline int
+static INLINE int
bidi_pop_embedding_level (struct bidi_it *bidi_it)
{
/* UAX#9 says to ignore invalid PDFs. */
}
/* Record in SAVED_INFO the information about the current character. */
-static inline void
+static INLINE void
bidi_remember_char (struct bidi_saved_info *saved_info,
struct bidi_it *bidi_it)
{
/* Resolve the type of a neutral character according to the type of
surrounding strong text and the current embedding level. */
-static inline bidi_type_t
+static INLINE bidi_type_t
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. */
return STRONG_R;
}
-static inline int
+static INLINE int
bidi_explicit_dir_char (int c)
{
/* FIXME: this should be replaced with a lookup table with suitable
int new_level;
bidi_dir_t override;
- if (bidi_it->bytepos < BEGV_BYTE) /* after reseat to BEGV */
+ if (bidi_it->bytepos < BEGV_BYTE /* after reseat to BEGV? */
+ || bidi_it->first_elt)
{
- bidi_it->charpos = BEGV;
- bidi_it->bytepos = BEGV_BYTE;
+ bidi_it->first_elt = 0;
+ if (bidi_it->charpos < BEGV)
+ bidi_it->charpos = BEGV;
+ bidi_it->bytepos = CHAR_TO_BYTE (bidi_it->charpos);
}
- else if (bidi_it->bytepos < ZV_BYTE) /* don't move at ZV */
+ else if (bidi_it->bytepos < ZV_BYTE) /* don't move at ZV */
{
bidi_it->charpos++;
if (bidi_it->ch_len == 0)
}
bidi_it->ch = curchar;
- type = bidi_get_type (curchar);
+ /* 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);
bidi_it->orig_type = type;
bidi_check_type (bidi_it->orig_type);
if (prev_level < new_level
&& bidi_it->type == WEAK_BN
&& bidi_it->ignore_bn_limit == 0 /* only if not already known */
- && bidi_it->ch != BIDI_EOB /* not already at EOB */
+ && bidi_it->bytepos < ZV_BYTE /* not already at EOB */
&& bidi_explicit_dir_char (FETCH_CHAR (bidi_it->bytepos
+ bidi_it->ch_len)))
{
}
}
- /* For when the paragraph end is defined by anything other than a
- special Unicode character (a.k.a. ``higher protocols''). */
- if (bidi_it->type != NEUTRAL_B)
- if (bidi_at_paragraph_end (bidi_it->ch,
- bidi_it->bytepos + bidi_it->ch_len))
- bidi_it->type = NEUTRAL_B;
-
if (bidi_it->type == NEUTRAL_B) /* X8 */
{
bidi_set_paragraph_end (bidi_it);
- bidi_it->type_after_w1 = bidi_it->type; /* needed below and in L1 */
+ /* This is needed by bidi_resolve_weak below, and in L1. */
+ bidi_it->type_after_w1 = bidi_it->type;
bidi_check_type (bidi_it->type_after_w1);
}
/* Advance in the buffer, resolve weak types and return the type of
the next character after weak type resolution. */
-bidi_type_t
+static bidi_type_t
bidi_resolve_weak (struct bidi_it *bidi_it)
{
bidi_type_t type;
type = STRONG_R;
else if (override == L2R)
type = STRONG_L;
- else if (type == STRONG_AL)
- type = STRONG_R; /* W3 */
- else if (type == WEAK_NSM) /* W1 */
- {
- /* Note that we don't need to consider the case where the prev
- character has its type overridden by an RLO or LRO: such
- characters are outside the current level run, and thus not
- relevant to this NSM. Thus, NSM gets the orig_type of the
- previous character. */
- if (bidi_it->prev.type != UNKNOWN_BT)
- type = bidi_it->prev.orig_type;
- else if (bidi_it->sor == R2L)
- type = STRONG_R;
- else if (bidi_it->sor == L2R)
- type = STRONG_L;
- else /* shouldn't happen! */
- abort ();
- if (type == WEAK_EN /* W2 after W1 */
- && bidi_it->last_strong.type_after_w1 == STRONG_AL)
- type = WEAK_AN;
- }
- else if (type == WEAK_EN /* W2 */
- && bidi_it->last_strong.type_after_w1 == STRONG_AL)
- type = WEAK_AN;
- else if ((type == WEAK_ES
- && (bidi_it->prev.type_after_w1 == WEAK_EN /* W4 */
- && (bidi_it->prev.orig_type == WEAK_EN
- || bidi_it->prev.orig_type == WEAK_NSM))) /* aft W1 */
- || (type == WEAK_CS
- && ((bidi_it->prev.type_after_w1 == WEAK_EN
- && (bidi_it->prev.orig_type == WEAK_EN /* W4 */
- || bidi_it->prev.orig_type == WEAK_NSM)) /* a/W1 */
- || bidi_it->prev.type_after_w1 == WEAK_AN))) /* W4 */
+ else
{
- 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);
-
- if (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)
- ;
- type_of_next = bidi_it->type;
- bidi_copy_it (bidi_it, &saved_it);
- }
-
- /* 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)
+ if (type == WEAK_NSM) /* W1 */
{
- 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;
+ /* Note that we don't need to consider the case where the
+ 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. */
+ 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)
+ type = bidi_it->prev.type_after_w1;
+ else if (bidi_it->sor == R2L)
+ type = STRONG_R;
+ else if (bidi_it->sor == L2R)
+ type = STRONG_L;
+ else /* shouldn't happen! */
+ abort ();
}
- }
- 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 with EN before it */
- || bidi_it->next_en_pos > bidi_it->charpos)
- type = WEAK_EN;
- /* W5: ET with EN after it. */
- else
+ 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)))
{
- 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);
+ type_of_next = bidi_get_type (next_char, override);
- if (type_of_next == WEAK_ET
- || type_of_next == WEAK_BN
+ if (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))
+ && bidi_it->type == WEAK_BN)
;
type_of_next = bidi_it->type;
- en_pos = bidi_it->charpos;
bidi_copy_it (bidi_it, &saved_it);
}
- if (type_of_next == WEAK_EN)
+
+ /* 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)
+ {
+ 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. */
{
- /* 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)
+ 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)
{
- type = WEAK_EN;
- /* Remember this EN position, to speed up processing
- of the next ETs. */
- bidi_it->next_en_pos = en_pos;
+ /* 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 */
}
- else if (type == WEAK_BN)
- type = NEUTRAL_ON; /* W6/Retaining */
}
}
}
return type;
}
-bidi_type_t
+static bidi_type_t
bidi_resolve_neutral (struct bidi_it *bidi_it)
{
int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level;
/* Given an iterator state in BIDI_IT, advance one character position
in the buffer to the next character (in the logical order), resolve
the bidi type of that next character, and return that type. */
-bidi_type_t
+static bidi_type_t
bidi_type_of_next_char (struct bidi_it *bidi_it)
{
bidi_type_t type;
the buffer to the next character (in the logical order), resolve
the embedding and implicit levels of that next character, and
return the resulting level. */
-int
+static int
bidi_level_of_next_char (struct bidi_it *bidi_it)
{
bidi_type_t type;
if (bidi_it->scan_dir == 1)
{
/* There's no sense in trying to advance if we hit end of text. */
- if (bidi_it->ch == BIDI_EOB)
+ if (bidi_it->bytepos >= ZV_BYTE)
return bidi_it->resolved_level;
/* Record the info about the previous character. */
{
int ch;
int clen = bidi_it->ch_len;
- int bpos = bidi_it->bytepos;
- int cpos = bidi_it->charpos;
+ EMACS_INT bpos = bidi_it->bytepos;
+ EMACS_INT cpos = bidi_it->charpos;
bidi_type_t chtype;
do {
if (ch == '\n' || ch == BIDI_EOB /* || ch == LINESEP_CHAR */)
chtype = NEUTRAL_B;
else
- chtype = bidi_get_type (ch);
+ chtype = bidi_get_type (ch, NEUTRAL_DIR);
} while (chtype == NEUTRAL_WS || chtype == WEAK_BN
|| bidi_explicit_dir_char (ch)); /* L1/Retaining */
bidi_it->next_for_ws.type = chtype;
bidi_get_next_char_visually (struct bidi_it *bidi_it)
{
int old_level, new_level, next_level;
- struct bidi_it prev_bidi_it;
+ struct bidi_it sentinel;
if (bidi_it->scan_dir == 0)
{
bidi_it->scan_dir = 1; /* default to logical order */
}
- if (bidi_it->new_paragraph)
- bidi_paragraph_init (bidi_overriding_paragraph_direction, bidi_it);
+ /* If we just passed a newline, initialize for the next line. */
+ if (!bidi_it->first_elt && bidi_it->orig_type == NEUTRAL_B)
+ bidi_line_init (bidi_it);
+
+ /* Prepare the sentinel iterator state. */
if (bidi_cache_idx == 0)
- bidi_copy_it (&prev_bidi_it, bidi_it);
+ {
+ bidi_copy_it (&sentinel, bidi_it);
+ if (bidi_it->first_elt)
+ {
+ sentinel.charpos--; /* cached charpos needs to be monotonic */
+ sentinel.bytepos--;
+ sentinel.ch = '\n'; /* doesn't matter, but why not? */
+ sentinel.ch_len = 1;
+ }
+ }
old_level = bidi_it->resolved_level;
new_level = bidi_level_of_next_char (bidi_it);
- if (bidi_it->ch == BIDI_EOB)
- return;
/* Reordering of resolved levels (clause L2) is implemented by
jumping to the other edge of the level and flipping direction of
- scanning the buffer whenever we find a level change. */
+ scanning the text whenever we find a level change. */
if (new_level != old_level)
{
int ascending = new_level > old_level;
int expected_next_level = old_level + incr;
/* If we don't have anything cached yet, we need to cache the
- previous character we've seen, since we'll need it to record
- where to jump when the last non-base level is exhausted. */
+ sentinel state, since we'll need it to record where to jump
+ when the last non-base level is exhausted. */
if (bidi_cache_idx == 0)
- bidi_cache_iterator_state (&prev_bidi_it, 1);
+ bidi_cache_iterator_state (&sentinel, 1);
/* Jump (or walk) to the other edge of this level. */
bidi_find_other_level_edge (bidi_it, level_to_search, !ascending);
/* Switch scan direction and peek at the next character in the
next_level = bidi_level_of_next_char (bidi_it);
}
+ /* Take note when we have just processed the newline that precedes
+ the end of the paragraph. The next time we are about to be
+ called, set_iterator_to_next will automatically reinit the
+ paragraph direction, if needed. We do this at the newline before
+ the paragraph separator, because the next character might not be
+ the first character of the next paragraph, due to the bidi
+ reordering, whereas we _must_ know the paragraph base direction
+ _before_ we process the paragraph's text, since the base
+ direction affects the reordering. */
+ if (bidi_it->scan_dir == 1
+ && bidi_it->orig_type == NEUTRAL_B
+ && bidi_it->bytepos < ZV_BYTE)
+ {
+ EMACS_INT sep_len =
+ bidi_at_paragraph_end (bidi_it->charpos + 1,
+ bidi_it->bytepos + bidi_it->ch_len);
+ if (sep_len >= 0)
+ {
+ bidi_it->new_paragraph = 1;
+ /* Record the buffer position of the last character of the
+ paragraph separator. */
+ bidi_it->separator_limit = bidi_it->charpos + 1 + sep_len;
+ }
+ }
+
if (bidi_it->scan_dir == 1 && bidi_cache_idx)
{
/* If we are at paragraph's base embedding level and beyond the