STRONG
} bidi_category_t;
+/* UAX#9 says to search only for L, AL, or R types of characters, and
+ ignore RLE, RLO, LRE, and LRO, when determining the base paragraph
+ level. Yudit indeed ignores them. This variable is therefore set
+ by default to ignore them, but setting it to zero will take them
+ into account. */
extern int bidi_ignore_explicit_marks_for_paragraph_level EXTERNALLY_VISIBLE;
int bidi_ignore_explicit_marks_for_paragraph_level = 1;
abort ();
default_type = (bidi_type_t) XINT (CHAR_TABLE_REF (bidi_type_table, ch));
+ /* Every valid character code, even those that are unassigned by the
+ UCD, have some bidi-class property, according to
+ DerivedBidiClass.txt file. Therefore, if we ever get UNKNOWN_BT
+ (= zero) code from CHAR_TABLE_REF, that's a bug. */
+ if (default_type == UNKNOWN_BT)
+ abort ();
if (override == NEUTRAL_DIR)
return default_type;
}
}
-static void
+static inline void
bidi_check_type (bidi_type_t type)
{
- if (type < UNKNOWN_BT || type > NEUTRAL_ON)
- abort ();
+ xassert (UNKNOWN_BT <= type && type <= NEUTRAL_ON);
}
/* Given a bidi TYPE of a character, return its category. */
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;
+ 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
level to which we descend after processing all the PDFs. */
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;
+ 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;
- bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1 =
- bidi_it->last_strong.orig_type = UNKNOWN_BT;
- bidi_it->prev_for_neutral.type = bidi_it->sor == R2L ? STRONG_R : STRONG_L;
+ bidi_it->last_strong.type = bidi_it->last_strong.type_after_w1
+ = bidi_it->last_strong.orig_type = UNKNOWN_BT;
+ 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;
- bidi_it->next_for_neutral.type = bidi_it->next_for_neutral.type_after_w1 =
- bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
+ bidi_it->next_for_neutral.type = bidi_it->next_for_neutral.type_after_w1
+ = bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
bidi_it->ignore_bn_limit = -1; /* meaning it's unknown */
}
#define BIDI_CACHE_CHUNK 200
static struct bidi_it *bidi_cache;
-static EMACS_INT bidi_cache_size = 0;
+static ptrdiff_t bidi_cache_size = 0;
enum { elsz = sizeof (struct bidi_it) };
-static EMACS_INT bidi_cache_idx; /* next unused cache slot */
-static EMACS_INT bidi_cache_last_idx; /* slot of last cache hit */
-static EMACS_INT bidi_cache_start = 0; /* start of cache for this
+static ptrdiff_t bidi_cache_idx; /* next unused cache slot */
+static ptrdiff_t bidi_cache_last_idx; /* slot of last cache hit */
+static ptrdiff_t bidi_cache_start = 0; /* start of cache for this
"stack" level */
+/* 5-slot stack for saving the start of the previous level of the
+ cache. xdisp.c maintains a 5-slot stack for its iterator state,
+ and we need the same size of our stack. */
+static ptrdiff_t bidi_cache_start_stack[IT_STACK_SIZE];
+static int bidi_cache_sp;
+
+/* Size of header used by bidi_shelve_cache. */
+enum
+ {
+ bidi_shelve_header_size
+ = (sizeof (bidi_cache_idx) + sizeof (bidi_cache_start_stack)
+ + sizeof (bidi_cache_sp) + sizeof (bidi_cache_start)
+ + sizeof (bidi_cache_last_idx))
+ };
+
/* Reset the cache state to the empty state. We only reset the part
of the cache relevant to iteration of the current object. Previous
objects, which are pushed on the display iterator's stack, are left
{
if (bidi_cache_size > BIDI_CACHE_CHUNK)
{
+ bidi_cache
+ = (struct bidi_it *) xrealloc (bidi_cache, BIDI_CACHE_CHUNK * elsz);
bidi_cache_size = BIDI_CACHE_CHUNK;
- bidi_cache =
- (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
}
bidi_cache_reset ();
}
static inline void
-bidi_cache_fetch_state (EMACS_INT idx, struct bidi_it *bidi_it)
+bidi_cache_fetch_state (ptrdiff_t 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 EMACS_INT
+static inline ptrdiff_t
bidi_cache_search (EMACS_INT charpos, int level, int dir)
{
- EMACS_INT i, i_start;
+ ptrdiff_t i, i_start;
if (bidi_cache_idx > bidi_cache_start)
{
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 EMACS_INT
+static ptrdiff_t
bidi_cache_find_level_change (int level, int dir, int before)
{
if (bidi_cache_idx)
{
- EMACS_INT i = dir ? bidi_cache_last_idx : bidi_cache_idx - 1;
+ ptrdiff_t i = dir ? bidi_cache_last_idx : bidi_cache_idx - 1;
int incr = before ? 1 : 0;
xassert (!dir || bidi_cache_last_idx >= 0);
}
static inline void
-bidi_cache_ensure_space (EMACS_INT idx)
+bidi_cache_ensure_space (ptrdiff_t idx)
{
/* Enlarge the cache as needed. */
if (idx >= bidi_cache_size)
{
- while (idx >= bidi_cache_size)
- bidi_cache_size += BIDI_CACHE_CHUNK;
- bidi_cache =
- (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
+ /* The bidi cache cannot be larger than the largest Lisp string
+ or buffer. */
+ ptrdiff_t string_or_buffer_bound
+ = max (BUF_BYTES_MAX, STRING_BYTES_BOUND);
+
+ /* Also, it cannot be larger than what C can represent. */
+ ptrdiff_t c_bound
+ = (min (PTRDIFF_MAX, SIZE_MAX) - bidi_shelve_header_size) / elsz;
+
+ bidi_cache
+ = xpalloc (bidi_cache, &bidi_cache_size,
+ max (BIDI_CACHE_CHUNK, idx - bidi_cache_size + 1),
+ min (string_or_buffer_bound, c_bound), elsz);
}
}
static inline void
bidi_cache_iterator_state (struct bidi_it *bidi_it, int resolved)
{
- EMACS_INT idx;
+ ptrdiff_t idx;
/* We should never cache on backward scans. */
if (bidi_it->scan_dir == -1)
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[idx].disp_pos = bidi_it->disp_pos;
+ bidi_cache[idx].disp_prop = bidi_it->disp_prop;
}
bidi_cache_last_idx = idx;
static inline bidi_type_t
bidi_cache_find (EMACS_INT charpos, int level, struct bidi_it *bidi_it)
{
- EMACS_INT i = bidi_cache_search (charpos, level, bidi_it->scan_dir);
+ ptrdiff_t i = bidi_cache_search (charpos, level, bidi_it->scan_dir);
if (i >= bidi_cache_start)
{
/***********************************************************************
Pushing and popping the bidi iterator state
***********************************************************************/
-/* 5-slot stack for saving the start of the previous level of the
- cache. xdisp.c maintains a 5-slot stack for its iterator state,
- and we need the same size of our stack. */
-static EMACS_INT bidi_cache_start_stack[IT_STACK_SIZE];
-static int bidi_cache_sp;
/* Push the bidi iterator state in preparation for reordering a
different object, e.g. display string found at certain buffer
bidi_cache_last_idx = -1;
}
+static ptrdiff_t bidi_cache_total_alloc;
+
/* Stash away a copy of the cache and its control variables. */
void *
bidi_shelve_cache (void)
{
unsigned char *databuf;
+ ptrdiff_t alloc;
+ /* Empty cache. */
if (bidi_cache_idx == 0)
return NULL;
- databuf = xmalloc (sizeof (bidi_cache_idx)
- + bidi_cache_idx * sizeof (struct bidi_it)
- + sizeof (bidi_cache_start_stack)
- + sizeof (bidi_cache_sp) + sizeof (bidi_cache_start)
- + sizeof (bidi_cache_last_idx));
+ alloc = (bidi_shelve_header_size
+ + bidi_cache_idx * sizeof (struct bidi_it));
+ databuf = xmalloc (alloc);
+ bidi_cache_total_alloc += alloc;
+
memcpy (databuf, &bidi_cache_idx, sizeof (bidi_cache_idx));
memcpy (databuf + sizeof (bidi_cache_idx),
bidi_cache, bidi_cache_idx * sizeof (struct bidi_it));
return databuf;
}
-/* Restore the cache state from a copy stashed away by bidi_shelve_cache. */
+/* Restore the cache state from a copy stashed away by
+ bidi_shelve_cache, and free the buffer used to stash that copy.
+ JUST_FREE non-zero means free the buffer, but don't restore the
+ cache; used when the corresponding iterator is discarded instead of
+ being restored. */
void
-bidi_unshelve_cache (void *databuf)
+bidi_unshelve_cache (void *databuf, int just_free)
{
unsigned char *p = databuf;
if (!p)
{
- /* A NULL pointer means an empty cache. */
- bidi_cache_start = 0;
- bidi_cache_sp = 0;
- bidi_cache_reset ();
+ if (!just_free)
+ {
+ /* A NULL pointer means an empty cache. */
+ bidi_cache_start = 0;
+ bidi_cache_sp = 0;
+ bidi_cache_reset ();
+ }
}
else
{
- memcpy (&bidi_cache_idx, p, sizeof (bidi_cache_idx));
- bidi_cache_ensure_space (bidi_cache_idx);
- memcpy (bidi_cache, p + sizeof (bidi_cache_idx),
- bidi_cache_idx * sizeof (struct bidi_it));
- memcpy (bidi_cache_start_stack,
- p + sizeof (bidi_cache_idx)
- + bidi_cache_idx * sizeof (struct bidi_it),
- sizeof (bidi_cache_start_stack));
- memcpy (&bidi_cache_sp,
- p + sizeof (bidi_cache_idx)
- + bidi_cache_idx * sizeof (struct bidi_it)
- + sizeof (bidi_cache_start_stack),
- sizeof (bidi_cache_sp));
- memcpy (&bidi_cache_start,
- p + sizeof (bidi_cache_idx)
- + bidi_cache_idx * sizeof (struct bidi_it)
- + sizeof (bidi_cache_start_stack) + sizeof (bidi_cache_sp),
- sizeof (bidi_cache_start));
- memcpy (&bidi_cache_last_idx,
- p + sizeof (bidi_cache_idx)
- + bidi_cache_idx * sizeof (struct bidi_it)
- + sizeof (bidi_cache_start_stack) + sizeof (bidi_cache_sp)
- + sizeof (bidi_cache_start),
- sizeof (bidi_cache_last_idx));
+ if (just_free)
+ {
+ ptrdiff_t idx;
+
+ memcpy (&idx, p, sizeof (bidi_cache_idx));
+ bidi_cache_total_alloc
+ -= bidi_shelve_header_size + idx * sizeof (struct bidi_it);
+ }
+ else
+ {
+ memcpy (&bidi_cache_idx, p, sizeof (bidi_cache_idx));
+ bidi_cache_ensure_space (bidi_cache_idx);
+ memcpy (bidi_cache, p + sizeof (bidi_cache_idx),
+ bidi_cache_idx * sizeof (struct bidi_it));
+ memcpy (bidi_cache_start_stack,
+ p + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it),
+ sizeof (bidi_cache_start_stack));
+ memcpy (&bidi_cache_sp,
+ p + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it)
+ + sizeof (bidi_cache_start_stack),
+ sizeof (bidi_cache_sp));
+ memcpy (&bidi_cache_start,
+ p + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it)
+ + sizeof (bidi_cache_start_stack) + sizeof (bidi_cache_sp),
+ sizeof (bidi_cache_start));
+ memcpy (&bidi_cache_last_idx,
+ p + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it)
+ + sizeof (bidi_cache_start_stack) + sizeof (bidi_cache_sp)
+ + sizeof (bidi_cache_start),
+ sizeof (bidi_cache_last_idx));
+ bidi_cache_total_alloc
+ -= (bidi_shelve_header_size
+ + bidi_cache_idx * sizeof (struct bidi_it));
+ }
xfree (p);
}
static void
bidi_initialize (void)
{
-
-#include "biditype.h"
-#include "bidimirror.h"
-
- int i;
-
- bidi_type_table = Fmake_char_table (Qnil, make_number (STRONG_L));
+ bidi_type_table = uniprop_table (intern ("bidi-class"));
+ if (NILP (bidi_type_table))
+ abort ();
staticpro (&bidi_type_table);
- for (i = 0; i < sizeof bidi_type / sizeof bidi_type[0]; i++)
- char_table_set_range (bidi_type_table, bidi_type[i].from, bidi_type[i].to,
- make_number (bidi_type[i].type));
-
- bidi_mirror_table = Fmake_char_table (Qnil, Qnil);
+ bidi_mirror_table = uniprop_table (intern ("mirroring"));
+ if (NILP (bidi_mirror_table))
+ abort ();
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));
-
Qparagraph_start = intern ("paragraph-start");
staticpro (&Qparagraph_start);
paragraph_start_re = Fsymbol_value (Qparagraph_start);
staticpro (¶graph_separate_re);
bidi_cache_sp = 0;
+ bidi_cache_total_alloc = 0;
bidi_initialized = 1;
}
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 =
- 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->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;
- bidi_it->next_for_neutral.type =
- bidi_it->next_for_neutral.type_after_w1 =
- bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
+ bidi_it->next_for_neutral.type
+ = bidi_it->next_for_neutral.type_after_w1
+ = bidi_it->next_for_neutral.orig_type = UNKNOWN_BT;
bidi_it->prev_for_neutral.charpos = -1;
- bidi_it->prev_for_neutral.type =
- bidi_it->prev_for_neutral.type_after_w1 =
- bidi_it->prev_for_neutral.orig_type = UNKNOWN_BT;
+ bidi_it->prev_for_neutral.type
+ = bidi_it->prev_for_neutral.type_after_w1
+ = bidi_it->prev_for_neutral.orig_type = UNKNOWN_BT;
bidi_it->sor = L2R; /* FIXME: should it be user-selectable? */
bidi_it->disp_pos = -1; /* invalid/unknown */
+ bidi_it->disp_prop = 0;
/* We can only shrink the cache if we are at the bottom level of its
"stack". */
if (bidi_cache_start == 0)
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;
+ /* Setting this to zero will force its recomputation the first time
+ we need it for W5. */
+ bidi_it->next_en_pos = 0;
bidi_it->next_for_ws.type = UNKNOWN_BT;
bidi_set_sor_type (bidi_it,
- bidi_it->paragraph_dir == R2L ? 1 : 0,
+ (bidi_it->paragraph_dir == R2L ? 1 : 0),
bidi_it->level_stack[0].level); /* X10 */
bidi_cache_reset ();
/* Fetch and return the character at BYTEPOS/CHARPOS. If that
character is covered by a display string, treat the entire run of
- covered characters as a single character u+FFFC, and return their
- combined length in CH_LEN and NCHARS. DISP_POS specifies the
- character position of the next display string, or -1 if not yet
- computed. When the next character is at or beyond that position,
- the function updates DISP_POS with the position of the next display
- string. STRING->s is the C string to iterate, or NULL if iterating
- over a buffer or a Lisp string; in the latter case, STRING->lstring
- is the Lisp string. */
+ covered characters as a single character, either u+2029 or u+FFFC,
+ and return their combined length in CH_LEN and NCHARS. DISP_POS
+ specifies the character position of the next display string, or -1
+ if not yet computed. When the next character is at or beyond that
+ position, the function updates DISP_POS with the position of the
+ next display string. DISP_PROP non-zero means that there's really
+ a display string at DISP_POS, as opposed to when we searched till
+ DISP_POS without finding one. If DISP_PROP is 2, it means the
+ display spec is of the form `(space ...)', which is replaced with
+ u+2029 to handle it as a paragraph separator. STRING->s is the C
+ string to iterate, or NULL if iterating over a buffer or a Lisp
+ string; in the latter case, STRING->lstring is the Lisp string. */
static inline int
bidi_fetch_char (EMACS_INT bytepos, EMACS_INT charpos, EMACS_INT *disp_pos,
- struct bidi_string_data *string,
+ int *disp_prop, struct bidi_string_data *string,
int frame_window_p, EMACS_INT *ch_len, EMACS_INT *nchars)
{
int ch;
- EMACS_INT endpos =
- (string->s || STRINGP (string->lstring)) ? string->schars : ZV;
+ EMACS_INT endpos
+ = (string->s || STRINGP (string->lstring)) ? string->schars : ZV;
struct text_pos pos;
/* If we got past the last known position of display string, compute
if (charpos < endpos && charpos > *disp_pos)
{
SET_TEXT_POS (pos, charpos, bytepos);
- *disp_pos = compute_display_string_pos (&pos, string, frame_window_p);
+ *disp_pos = compute_display_string_pos (&pos, string, frame_window_p,
+ disp_prop);
}
/* Fetch the character at BYTEPOS. */
*ch_len = 1;
*nchars = 1;
*disp_pos = endpos;
+ *disp_prop = 0;
}
- else if (charpos >= *disp_pos)
+ else if (charpos >= *disp_pos && *disp_prop)
{
EMACS_INT disp_end_pos;
property. Hopefully, it will never be needed. */
if (charpos > *disp_pos)
abort ();
- /* Return the Unicode Object Replacement Character to represent
- the entire run of characters covered by the display string. */
- ch = 0xFFFC;
+ /* Text covered by `display' properties and overlays with
+ display properties or display strings is handled as a single
+ character that represents the entire run of characters
+ covered by the display property. */
+ if (*disp_prop == 2)
+ {
+ /* `(space ...)' display specs are handled as paragraph
+ separators for the purposes of the reordering; see UAX#9
+ section 3 and clause HL1 in section 4.3 there. */
+ ch = 0x2029;
+ }
+ else
+ {
+ /* All other display specs are handled as the Unicode Object
+ Replacement Character. */
+ ch = 0xFFFC;
+ }
disp_end_pos = compute_display_string_end (*disp_pos, string);
+ if (disp_end_pos < 0)
+ {
+ /* Somebody removed the display string from the buffer
+ behind our back. Recover by processing this buffer
+ position as if no display property were present there to
+ begin with. */
+ *disp_prop = 0;
+ goto normal_char;
+ }
*nchars = disp_end_pos - *disp_pos;
if (*nchars <= 0)
abort ();
}
else
{
+ normal_char:
if (string->s)
{
int len;
/* If we just entered a run of characters covered by a display
string, compute the position of the next display string. */
- if (charpos + *nchars <= endpos && charpos + *nchars > *disp_pos)
+ if (charpos + *nchars <= endpos && charpos + *nchars > *disp_pos
+ && *disp_prop)
{
SET_TEXT_POS (pos, charpos + *nchars, bytepos + *ch_len);
- *disp_pos = compute_display_string_pos (&pos, string, frame_window_p);
+ *disp_pos = compute_display_string_pos (&pos, string, frame_window_p,
+ disp_prop);
}
return ch;
return val;
}
+/* On my 2005-vintage machine, searching back for paragraph start
+ takes ~1 ms per line. And bidi_paragraph_init is called 4 times
+ when user types C-p. The number below limits each call to
+ bidi_paragraph_init to about 10 ms. */
+#define MAX_PARAGRAPH_SEARCH 7500
+
/* Find the beginning of this paragraph by looking back in the buffer.
- Value is the byte position of the paragraph's beginning. */
+ Value is the byte position of the paragraph's beginning, or
+ BEGV_BYTE if paragraph_start_re is still not found after looking
+ back MAX_PARAGRAPH_SEARCH lines in the buffer. */
static EMACS_INT
bidi_find_paragraph_start (EMACS_INT pos, EMACS_INT pos_byte)
{
Lisp_Object re = paragraph_start_re;
EMACS_INT limit = ZV, limit_byte = ZV_BYTE;
+ EMACS_INT n = 0;
while (pos_byte > BEGV_BYTE
+ && n++ < MAX_PARAGRAPH_SEARCH
&& fast_looking_at (re, pos, pos_byte, limit, limit_byte, Qnil) < 0)
{
/* FIXME: What if the paragraph beginning is covered by a
pos = find_next_newline_no_quit (pos - 1, -1);
pos_byte = CHAR_TO_BYTE (pos);
}
+ if (n >= MAX_PARAGRAPH_SEARCH)
+ pos_byte = BEGV_BYTE;
return pos_byte;
}
int ch;
EMACS_INT ch_len, nchars;
EMACS_INT pos, disp_pos = -1;
+ int disp_prop = 0;
bidi_type_t type;
const unsigned char *s;
we are potentially in a new paragraph that doesn't yet
exist. */
pos = bidi_it->charpos;
- s = STRINGP (bidi_it->string.lstring) ?
- SDATA (bidi_it->string.lstring) : bidi_it->string.s;
+ s = (STRINGP (bidi_it->string.lstring)
+ ? SDATA (bidi_it->string.lstring)
+ : bidi_it->string.s);
if (bytepos > begbyte
&& bidi_char_at_pos (bytepos, s, bidi_it->string.unibyte) == '\n')
{
bytepos = pstartbyte;
if (!string_p)
pos = BYTE_TO_CHAR (bytepos);
- ch = bidi_fetch_char (bytepos, pos, &disp_pos, &bidi_it->string,
+ ch = bidi_fetch_char (bytepos, pos, &disp_pos, &disp_prop,
+ &bidi_it->string,
bidi_it->frame_window_p, &ch_len, &nchars);
type = bidi_get_type (ch, NEUTRAL_DIR);
for (pos += nchars, 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
&& bidi_at_paragraph_end (pos, bytepos) >= -1)
break;
/* Fetch next character and advance to get past it. */
- ch = bidi_fetch_char (bytepos, pos, &disp_pos, &bidi_it->string,
+ ch = bidi_fetch_char (bytepos, pos, &disp_pos,
+ &disp_prop, &bidi_it->string,
bidi_it->frame_window_p, &ch_len, &nchars);
pos += nchars;
bytepos += ch_len;
}
- if (type == STRONG_R || type == STRONG_AL) /* P3 */
+ if ((type == STRONG_R || type == STRONG_AL) /* P3 */
+ || (!bidi_ignore_explicit_marks_for_paragraph_level
+ && (type == RLO || type == RLE)))
bidi_it->paragraph_dir = R2L;
- else if (type == STRONG_L)
+ else if (type == STRONG_L
+ || (!bidi_ignore_explicit_marks_for_paragraph_level
+ && (type == LRO || type == LRE)))
bidi_it->paragraph_dir = L2R;
if (!string_p
&& no_default_p && bidi_it->paragraph_dir == NEUTRAL_DIR)
bidi_it->first_elt = 0;
if (string_p)
{
- const unsigned char *p =
- STRINGP (bidi_it->string.lstring)
- ? SDATA (bidi_it->string.lstring) : bidi_it->string.s;
+ const unsigned char *p
+ = (STRINGP (bidi_it->string.lstring)
+ ? SDATA (bidi_it->string.lstring)
+ : bidi_it->string.s);
if (bidi_it->charpos < 0)
bidi_it->charpos = 0;
bidi_it->ch_len = 1;
bidi_it->nchars = 1;
bidi_it->disp_pos = (string_p ? bidi_it->string.schars : ZV);
+ bidi_it->disp_prop = 0;
}
else
{
display string, treat the entire run of covered characters as
a single character u+FFFC. */
curchar = bidi_fetch_char (bidi_it->bytepos, bidi_it->charpos,
- &bidi_it->disp_pos, &bidi_it->string,
- bidi_it->frame_window_p,
+ &bidi_it->disp_pos, &bidi_it->disp_prop,
+ &bidi_it->string, bidi_it->frame_window_p,
&bidi_it->ch_len, &bidi_it->nchars);
}
bidi_it->ch = curchar;
int prev_level = bidi_it->level_stack[bidi_it->stack_idx].level;
int new_level = bidi_resolve_explicit_1 (bidi_it);
EMACS_INT eob = bidi_it->string.s ? bidi_it->string.schars : ZV;
- const unsigned char *s = STRINGP (bidi_it->string.lstring)
- ? SDATA (bidi_it->string.lstring) : bidi_it->string.s;
+ const unsigned char *s
+ = (STRINGP (bidi_it->string.lstring)
+ ? SDATA (bidi_it->string.lstring)
+ : bidi_it->string.s);
if (prev_level < new_level
&& bidi_it->type == WEAK_BN
int next_char;
bidi_type_t type_of_next;
struct bidi_it saved_it;
- EMACS_INT eob =
- (STRINGP (bidi_it->string.lstring) || bidi_it->string.s)
- ? bidi_it->string.schars : ZV;
+ EMACS_INT eob
+ = ((STRINGP (bidi_it->string.lstring) || bidi_it->string.s)
+ ? bidi_it->string.schars : ZV);
type = bidi_it->type;
override = bidi_it->level_stack[bidi_it->stack_idx].override;
&& bidi_it->prev.orig_type == WEAK_EN)
|| bidi_it->prev.type_after_w1 == WEAK_AN)))
{
- const unsigned char *s =
- STRINGP (bidi_it->string.lstring)
- ? SDATA (bidi_it->string.lstring) : bidi_it->string.s;
-
- next_char =
- bidi_it->charpos + bidi_it->nchars >= eob
- ? BIDI_EOB
- : bidi_char_at_pos (bidi_it->bytepos + bidi_it->ch_len, s,
- bidi_it->string.unibyte);
+ const unsigned char *s
+ = (STRINGP (bidi_it->string.lstring)
+ ? SDATA (bidi_it->string.lstring)
+ : bidi_it->string.s);
+
+ next_char = (bidi_it->charpos + bidi_it->nchars >= eob
+ ? BIDI_EOB
+ : bidi_char_at_pos (bidi_it->bytepos + bidi_it->ch_len,
+ s, bidi_it->string.unibyte));
type_of_next = bidi_get_type (next_char, override);
if (type_of_next == WEAK_BN
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. */
+ else if (bidi_it->next_en_pos >=0) /* W5: ET/BN with EN after it. */
{
EMACS_INT en_pos = bidi_it->charpos + bidi_it->nchars;
- const unsigned char *s =
- STRINGP (bidi_it->string.lstring)
- ? SDATA (bidi_it->string.lstring) : bidi_it->string.s;
+ const unsigned char *s = (STRINGP (bidi_it->string.lstring)
+ ? SDATA (bidi_it->string.lstring)
+ : bidi_it->string.s);
if (bidi_it->nchars <= 0)
abort ();
- next_char =
- bidi_it->charpos + bidi_it->nchars >= eob
- ? BIDI_EOB
- : bidi_char_at_pos (bidi_it->bytepos + bidi_it->ch_len, s,
- bidi_it->string.unibyte);
+ next_char
+ = (bidi_it->charpos + bidi_it->nchars >= eob
+ ? BIDI_EOB
+ : bidi_char_at_pos (bidi_it->bytepos + bidi_it->ch_len, s,
+ bidi_it->string.unibyte));
type_of_next = bidi_get_type (next_char, override);
if (type_of_next == WEAK_ET
else if (type == WEAK_BN)
type = NEUTRAL_ON; /* W6/Retaining */
}
+ else if (type_of_next == NEUTRAL_B)
+ /* Record the fact that there are no more ENs from
+ here to the end of paragraph, to avoid entering the
+ loop above ever again in this paragraph. */
+ bidi_it->next_en_pos = -1;
}
}
}
|| type == NEUTRAL_ON))
abort ();
- if (bidi_get_category (type) == NEUTRAL
+ if ((type != NEUTRAL_B /* Don't risk entering the long loop below if
+ we are already at paragraph end. */
+ && 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);
+ /* The next two "else if" clauses are shortcuts for the
+ important special case when we have a long sequence of
+ neutral or WEAK_BN characters, such as whitespace or nulls or
+ other control characters, on the base embedding level of the
+ paragraph, and that sequence goes all the way to the end of
+ the paragraph and follows a character whose resolved
+ directionality is identical to the base embedding level.
+ (This is what happens in a buffer with plain L2R text that
+ happens to include long sequences of control characters.) By
+ virtue of N1, the result of examining this long sequence will
+ always be either STRONG_L or STRONG_R, depending on the base
+ embedding level. So we use this fact directly instead of
+ entering the expensive loop in the "else" clause. */
+ else if (current_level == 0
+ && bidi_it->prev_for_neutral.type == STRONG_L
+ && !bidi_explicit_dir_char (bidi_it->ch))
+ type = bidi_resolve_neutral_1 (bidi_it->prev_for_neutral.type,
+ STRONG_L, current_level);
+ else if (/* current level is 1 */
+ current_level == 1
+ /* base embedding level is also 1 */
+ && bidi_it->level_stack[0].level == 1
+ /* previous character is one of those considered R for
+ the purposes of W5 */
+ && (bidi_it->prev_for_neutral.type == STRONG_R
+ || bidi_it->prev_for_neutral.type == WEAK_EN
+ || bidi_it->prev_for_neutral.type == WEAK_AN)
+ && !bidi_explicit_dir_char (bidi_it->ch))
+ type = bidi_resolve_neutral_1 (bidi_it->prev_for_neutral.type,
+ STRONG_R, current_level);
else
{
/* Arrrgh!! The UAX#9 algorithm is too deeply entrenched in
&& 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_it->level_stack[bidi_it->stack_idx].level
+ != current_level)));
bidi_remember_char (&saved_it.next_for_neutral, bidi_it);
case STRONG_L:
case STRONG_R:
case STRONG_AL:
+ /* Actually, STRONG_AL cannot happen here, because
+ bidi_resolve_weak converts it to STRONG_R, per W3. */
+ xassert (type != STRONG_AL);
next_type = type;
break;
case WEAK_EN:
/* 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))
member. */
if (saved_it.type != WEAK_BN
|| bidi_get_category (bidi_it->prev.type_after_w1) == NEUTRAL)
- {
- next_type = bidi_it->prev_for_neutral.type;
- saved_it.next_for_neutral.type = next_type;
- bidi_check_type (next_type);
- }
+ next_type = bidi_it->prev_for_neutral.type;
else
{
/* This is a BN which does not adjoin neutrals.
}
type = bidi_resolve_neutral_1 (saved_it.prev_for_neutral.type,
next_type, current_level);
+ saved_it.next_for_neutral.type = next_type;
saved_it.type = type;
+ bidi_check_type (next_type);
bidi_check_type (type);
bidi_copy_it (bidi_it, &saved_it);
}
if (bidi_it->scan_dir == 1)
{
- EMACS_INT eob =
- (bidi_it->string.s || STRINGP (bidi_it->string.lstring))
- ? bidi_it->string.schars : ZV;
+ EMACS_INT eob
+ = ((bidi_it->string.s || STRINGP (bidi_it->string.lstring))
+ ? bidi_it->string.schars : ZV);
/* There's no sense in trying to advance if we hit end of text. */
if (bidi_it->charpos >= eob)
bidi_it->next_for_neutral.type = UNKNOWN_BT;
if (bidi_it->next_en_pos >= 0
&& bidi_it->charpos >= bidi_it->next_en_pos)
- bidi_it->next_en_pos = -1;
+ bidi_it->next_en_pos = 0;
if (bidi_it->next_for_ws.type != UNKNOWN_BT
&& bidi_it->charpos >= bidi_it->next_for_ws.charpos)
bidi_it->next_for_ws.type = UNKNOWN_BT;
UNKNOWN_BT. */
if (bidi_cache_idx > bidi_cache_start && !bidi_it->first_elt)
{
- int bob =
- (bidi_it->string.s || STRINGP (bidi_it->string.lstring)) ? 0 : 1;
-
+ int bob = ((bidi_it->string.s || STRINGP (bidi_it->string.lstring))
+ ? 0 : 1);
if (bidi_it->scan_dir > 0)
{
if (bidi_it->nchars <= 0)
struct bidi_string_data bs = bidi_it->string;
bidi_type_t chtype;
int fwp = bidi_it->frame_window_p;
+ int dpp = bidi_it->disp_prop;
if (bidi_it->nchars <= 0)
abort ();
do {
- ch = bidi_fetch_char (bpos += clen, cpos += nc, &disp_pos, &bs, fwp,
- &clen, &nc);
- if (ch == '\n' || ch == BIDI_EOB /* || ch == LINESEP_CHAR */)
+ ch = bidi_fetch_char (bpos += clen, cpos += nc, &disp_pos, &dpp, &bs,
+ fwp, &clen, &nc);
+ if (ch == '\n' || ch == BIDI_EOB)
chtype = NEUTRAL_B;
else
chtype = bidi_get_type (ch, NEUTRAL_DIR);
else if (bidi_it->orig_type == NEUTRAL_B /* L1 */
|| bidi_it->orig_type == NEUTRAL_S
|| bidi_it->ch == '\n' || bidi_it->ch == BIDI_EOB
- /* || bidi_it->ch == LINESEP_CHAR */
|| (bidi_it->orig_type == NEUTRAL_WS
&& (bidi_it->next_for_ws.type == NEUTRAL_B
|| bidi_it->next_for_ws.type == NEUTRAL_S)))
bidi_find_other_level_edge (struct bidi_it *bidi_it, int level, int end_flag)
{
int dir = end_flag ? -bidi_it->scan_dir : bidi_it->scan_dir;
- EMACS_INT idx;
+ ptrdiff_t idx;
/* Try the cache first. */
if ((idx = bidi_cache_find_level_change (level, dir, end_flag))
GCPRO1 (bidi_it->string.lstring);
/* If we just passed a newline, initialize for the next line. */
- if (!bidi_it->first_elt && bidi_it->orig_type == NEUTRAL_B)
+ if (!bidi_it->first_elt
+ && (bidi_it->ch == '\n' || bidi_it->ch == BIDI_EOB))
bidi_line_init (bidi_it);
/* Prepare the sentinel iterator state, and cache it. When we bump
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)
+ if (bidi_it->scan_dir == 1
+ && (bidi_it->ch == '\n' || bidi_it->ch == BIDI_EOB))
{
/* The paragraph direction of the entire string, once
determined, is in effect for the entire string. Setting the
bidi_it->separator_limit = bidi_it->string.schars;
else if (bidi_it->bytepos < ZV_BYTE)
{
- EMACS_INT sep_len =
- bidi_at_paragraph_end (bidi_it->charpos + bidi_it->nchars,
- bidi_it->bytepos + bidi_it->ch_len);
+ EMACS_INT sep_len
+ = bidi_at_paragraph_end (bidi_it->charpos + bidi_it->nchars,
+ bidi_it->bytepos + bidi_it->ch_len);
if (bidi_it->nchars <= 0)
abort ();
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 + bidi_it->nchars + sep_len;
+ bidi_it->separator_limit
+ = bidi_it->charpos + bidi_it->nchars + sep_len;
}
}
}
void
bidi_dump_cached_states (void)
{
- int i;
+ ptrdiff_t i;
int ndigits = 1;
if (bidi_cache_idx == 0)
HCoop / bpt / emacs.git / blobdiff