-/* Low-level bidirectional buffer-scanning functions for GNU Emacs.
+/* Low-level bidirectional buffer/string-scanning functions for GNU Emacs.
Copyright (C) 2000-2001, 2004-2005, 2009-2011
Free Software Foundation, Inc.
/* Written by Eli Zaretskii <eliz@gnu.org>.
A sequential implementation of the Unicode Bidirectional algorithm,
- as per UAX#9, a part of the Unicode Standard.
+ (UBA) 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 or
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
+ 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.
+ A few auxiliary entry points are used to initialize the bidi
+ iterator for iterating an object (buffer or string), push and pop
+ the bidi iterator state, and save and restore the state of the bidi
+ cache.
+
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.
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;
static Lisp_Object paragraph_start_re, paragraph_separate_re;
static Lisp_Object Qparagraph_start, Qparagraph_separate;
-static void
-bidi_initialize (void)
-{
-
-#include "biditype.h"
-#include "bidimirror.h"
-
- int i;
-
- bidi_type_table = Fmake_char_table (Qnil, make_number (STRONG_L));
- 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);
- 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);
- if (!STRINGP (paragraph_start_re))
- paragraph_start_re = build_string ("\f\\|[ \t]*$");
- staticpro (¶graph_start_re);
- Qparagraph_separate = intern ("paragraph-separate");
- staticpro (&Qparagraph_separate);
- paragraph_separate_re = Fsymbol_value (Qparagraph_separate);
- if (!STRINGP (paragraph_separate_re))
- paragraph_separate_re = build_string ("[ \t\f]*$");
- staticpro (¶graph_separate_re);
- bidi_initialized = 1;
-}
+\f
+/***********************************************************************
+ Utilities
+ ***********************************************************************/
/* Return the bidi type of a character CH, subject to the current
directional OVERRIDE. */
-static INLINE bidi_type_t
+static inline bidi_type_t
bidi_get_type (int ch, bidi_dir_t override)
{
bidi_type_t default_type;
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 bidi_category_t
+static inline bidi_category_t
bidi_get_category (bidi_type_t type)
{
switch (type)
val = CHAR_TABLE_REF (bidi_mirror_table, c);
if (INTEGERP (val))
{
- int v = XINT (val);
+ EMACS_INT v = XINT (val);
if (v < 0 || v > MAX_CHAR)
abort ();
return c;
}
+/* 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
+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. */
+ 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->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->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->ignore_bn_limit = -1; /* meaning it's unknown */
+}
+
+/* Push the current embedding level and override status; reset the
+ current level to LEVEL and the current override status to OVERRIDE. */
+static inline void
+bidi_push_embedding_level (struct bidi_it *bidi_it,
+ int level, bidi_dir_t override)
+{
+ bidi_it->stack_idx++;
+ xassert (bidi_it->stack_idx < BIDI_MAXLEVEL);
+ 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. */
+static inline int
+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. */
+static inline void
+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;
+ bidi_check_type (bidi_it->type);
+ saved_info->type_after_w1 = bidi_it->type_after_w1;
+ bidi_check_type (bidi_it->type_after_w1);
+ saved_info->orig_type = bidi_it->orig_type;
+ bidi_check_type (bidi_it->orig_type);
+}
+
/* 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;
to->level_stack[i] = from->level_stack[i];
}
-/* Caching the bidi iterator states. */
+\f
+/***********************************************************************
+ Caching the bidi iterator states
+ ***********************************************************************/
#define BIDI_CACHE_CHUNK 200
static struct bidi_it *bidi_cache;
-static size_t bidi_cache_size = 0;
-static size_t 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_size = 0;
+enum { elsz = sizeof (struct bidi_it) };
+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
intact. This is called when the cached information is no more
useful for the current iteration, e.g. when we were reseated to a
new position on the same object. */
-static INLINE void
+static inline void
bidi_cache_reset (void)
{
bidi_cache_idx = bidi_cache_start;
iterator for reordering a buffer or a string that does not come
from display properties, because that means all the previously
cached info is of no further use. */
-static INLINE void
+static inline void
bidi_cache_shrink (void)
{
if (bidi_cache_size > BIDI_CACHE_CHUNK)
{
- bidi_cache_size = BIDI_CACHE_CHUNK;
bidi_cache =
- (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
+ (struct bidi_it *) xrealloc (bidi_cache, BIDI_CACHE_CHUNK * elsz);
+ bidi_cache_size = BIDI_CACHE_CHUNK;
}
bidi_cache_reset ();
}
-static INLINE void
-bidi_cache_fetch_state (int idx, struct bidi_it *bidi_it)
+static inline void
+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 int
-bidi_cache_search (EMACS_INT charpos, int level, int dir)
+static inline ptrdiff_t
+bidi_cache_search (ptrdiff_t charpos, int level, int dir)
{
- int i, i_start;
+ ptrdiff_t i, i_start;
- if (bidi_cache_idx)
+ if (bidi_cache_idx > bidi_cache_start)
{
+ if (bidi_cache_last_idx == -1)
+ bidi_cache_last_idx = bidi_cache_idx - 1;
if (charpos < bidi_cache[bidi_cache_last_idx].charpos)
{
dir = -1;
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
+static ptrdiff_t
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;
+ ptrdiff_t i = dir ? bidi_cache_last_idx : bidi_cache_idx - 1;
int incr = before ? 1 : 0;
+ xassert (!dir || bidi_cache_last_idx >= 0);
+
if (!dir)
dir = -1;
else if (!incr)
return -1;
}
-static INLINE void
+static inline void
+bidi_cache_ensure_space (ptrdiff_t idx)
+{
+ /* Enlarge the cache as needed. */
+ if (idx >= bidi_cache_size)
+ {
+ /* 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)
{
- int idx;
+ ptrdiff_t idx;
/* We should never cache on backward scans. */
if (bidi_it->scan_dir == -1)
if (idx < 0)
{
idx = bidi_cache_idx;
- /* Enlarge the cache as needed. */
- if (idx >= bidi_cache_size)
- {
- bidi_cache_size += BIDI_CACHE_CHUNK;
- bidi_cache =
- (struct bidi_it *) xrealloc (bidi_cache, bidi_cache_size * elsz);
- }
+ bidi_cache_ensure_space (idx);
/* 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 > bidi_cache_start &&
(bidi_it->charpos > (bidi_cache[idx - 1].charpos
+ bidi_cache[idx - 1].nchars)
- || bidi_it->charpos < bidi_cache[0].charpos))
+ || bidi_it->charpos < bidi_cache[bidi_cache_start].charpos))
{
bidi_cache_reset ();
idx = bidi_cache_start;
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;
bidi_cache_idx = idx + 1;
}
-static INLINE bidi_type_t
-bidi_cache_find (EMACS_INT charpos, int level, struct bidi_it *bidi_it)
+static inline bidi_type_t
+bidi_cache_find (ptrdiff_t charpos, int level, struct bidi_it *bidi_it)
{
- 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)
{
return UNKNOWN_BT;
}
-static INLINE int
+static inline int
bidi_peek_at_next_level (struct bidi_it *bidi_it)
{
if (bidi_cache_idx == bidi_cache_start || bidi_cache_last_idx == -1)
return bidi_cache[bidi_cache_last_idx + bidi_it->scan_dir].resolved_level;
}
-/* 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)
+\f
+/***********************************************************************
+ Pushing and popping the bidi iterator state
+ ***********************************************************************/
+
+/* Push the bidi iterator state in preparation for reordering a
+ different object, e.g. display string found at certain buffer
+ position. Pushing the bidi iterator boils down to saving its
+ entire state on the cache and starting a new cache "stacked" on top
+ of the current cache. */
+void
+bidi_push_it (struct bidi_it *bidi_it)
{
- Lisp_Object sep_re;
- Lisp_Object start_re;
- EMACS_INT val;
+ /* Save the current iterator state in its entirety after the last
+ used cache slot. */
+ bidi_cache_ensure_space (bidi_cache_idx);
+ memcpy (&bidi_cache[bidi_cache_idx++], bidi_it, sizeof (struct bidi_it));
- sep_re = paragraph_separate_re;
- start_re = paragraph_start_re;
+ /* Push the current cache start onto the stack. */
+ xassert (bidi_cache_sp < IT_STACK_SIZE);
+ bidi_cache_start_stack[bidi_cache_sp++] = bidi_cache_start;
- val = fast_looking_at (sep_re, charpos, bytepos, ZV, ZV_BYTE, Qnil);
- if (val < 0)
+ /* Start a new level of cache, and make it empty. */
+ bidi_cache_start = bidi_cache_idx;
+ bidi_cache_last_idx = -1;
+}
+
+/* Restore the iterator state saved by bidi_push_it and return the
+ cache to the corresponding state. */
+void
+bidi_pop_it (struct bidi_it *bidi_it)
+{
+ if (bidi_cache_start <= 0)
+ abort ();
+
+ /* Reset the next free cache slot index to what it was before the
+ call to bidi_push_it. */
+ bidi_cache_idx = bidi_cache_start - 1;
+
+ /* Restore the bidi iterator state saved in the cache. */
+ memcpy (bidi_it, &bidi_cache[bidi_cache_idx], sizeof (struct bidi_it));
+
+ /* Pop the previous cache start from the stack. */
+ if (bidi_cache_sp <= 0)
+ abort ();
+ bidi_cache_start = bidi_cache_start_stack[--bidi_cache_sp];
+
+ /* Invalidate the last-used cache slot data. */
+ 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;
+
+ 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));
+ memcpy (databuf + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it),
+ bidi_cache_start_stack, sizeof (bidi_cache_start_stack));
+ memcpy (databuf + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it)
+ + sizeof (bidi_cache_start_stack),
+ &bidi_cache_sp, sizeof (bidi_cache_sp));
+ memcpy (databuf + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it)
+ + sizeof (bidi_cache_start_stack) + sizeof (bidi_cache_sp),
+ &bidi_cache_start, sizeof (bidi_cache_start));
+ memcpy (databuf + sizeof (bidi_cache_idx)
+ + bidi_cache_idx * sizeof (struct bidi_it)
+ + sizeof (bidi_cache_start_stack) + sizeof (bidi_cache_sp)
+ + sizeof (bidi_cache_start),
+ &bidi_cache_last_idx, sizeof (bidi_cache_last_idx));
+
+ return databuf;
+}
+
+/* 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, int just_free)
+{
+ unsigned char *p = databuf;
+
+ if (!p)
{
- if (fast_looking_at (start_re, charpos, bytepos, ZV, ZV_BYTE, Qnil) >= 0)
- val = -1;
- else
- val = -2;
+ if (!just_free)
+ {
+ /* A NULL pointer means an empty cache. */
+ bidi_cache_start = 0;
+ bidi_cache_sp = 0;
+ bidi_cache_reset ();
+ }
}
+ else
+ {
+ if (just_free)
+ {
+ ptrdiff_t idx;
- return val;
+ 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);
+ }
}
-/* 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
-bidi_set_sor_type (struct bidi_it *bidi_it, int level_before, int level_after)
+\f
+/***********************************************************************
+ Initialization
+ ***********************************************************************/
+static void
+bidi_initialize (void)
{
- int higher_level = level_before > level_after ? level_before : level_after;
+ bidi_type_table = uniprop_table (intern ("bidi-class"));
+ if (NILP (bidi_type_table))
+ abort ();
+ staticpro (&bidi_type_table);
- /* 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. */
- 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->prev_was_pdf = 1;
+ bidi_mirror_table = uniprop_table (intern ("mirroring"));
+ if (NILP (bidi_mirror_table))
+ abort ();
+ staticpro (&bidi_mirror_table);
- bidi_it->prev.type = UNKNOWN_BT;
+ Qparagraph_start = intern ("paragraph-start");
+ staticpro (&Qparagraph_start);
+ paragraph_start_re = Fsymbol_value (Qparagraph_start);
+ if (!STRINGP (paragraph_start_re))
+ paragraph_start_re = build_string ("\f\\|[ \t]*$");
+ staticpro (¶graph_start_re);
+ Qparagraph_separate = intern ("paragraph-separate");
+ staticpro (&Qparagraph_separate);
+ paragraph_separate_re = Fsymbol_value (Qparagraph_separate);
+ if (!STRINGP (paragraph_separate_re))
+ paragraph_separate_re = build_string ("[ \t\f]*$");
+ staticpro (¶graph_separate_re);
+
+ bidi_cache_sp = 0;
+ bidi_cache_total_alloc = 0;
+
+ bidi_initialized = 1;
+}
+
+/* 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;
+}
+
+/* Initialize the bidi iterator from buffer/string position CHARPOS. */
+void
+bidi_init_it (ptrdiff_t charpos, ptrdiff_t bytepos, int frame_window_p,
+ struct bidi_it *bidi_it)
+{
+ if (! bidi_initialized)
+ bidi_initialize ();
+ if (charpos >= 0)
+ bidi_it->charpos = charpos;
+ if (bytepos >= 0)
+ bidi_it->bytepos = bytepos;
+ bidi_it->frame_window_p = frame_window_p;
+ bidi_it->nchars = -1; /* to be computed in bidi_resolve_explicit_1 */
+ 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 = 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_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.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->ignore_bn_limit = -1; /* meaning it's unknown */
+ 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->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_cache_shrink ();
+ else
+ bidi_cache_reset ();
}
/* Perform initializations for reordering a new line of bidi text. */
bidi_cache_reset ();
}
-/* Count bytes in multibyte string S between BEG/BEGBYTE and END. BEG
- and END are zero-based character positions in S, BEGBYTE is byte
- position corresponding to BEG. */
-static inline EMACS_INT
-bidi_count_bytes (const unsigned char *s, const EMACS_INT beg,
- const EMACS_INT begbyte, const EMACS_INT end)
+\f
+/***********************************************************************
+ Fetching characters
+ ***********************************************************************/
+
+/* Count bytes in string S between BEG/BEGBYTE and END. BEG and END
+ are zero-based character positions in S, BEGBYTE is byte position
+ corresponding to BEG. UNIBYTE, if non-zero, means S is a unibyte
+ string. */
+static inline ptrdiff_t
+bidi_count_bytes (const unsigned char *s, const ptrdiff_t beg,
+ const ptrdiff_t begbyte, const ptrdiff_t end, int unibyte)
{
- EMACS_INT pos = beg;
+ ptrdiff_t pos = beg;
const unsigned char *p = s + begbyte, *start = p;
- if (!CHAR_HEAD_P (*p))
- abort ();
-
- while (pos < end)
+ if (unibyte)
+ p = s + end;
+ else
{
- p += BYTES_BY_CHAR_HEAD (*p);
- pos++;
+ if (!CHAR_HEAD_P (*p))
+ abort ();
+
+ while (pos < end)
+ {
+ p += BYTES_BY_CHAR_HEAD (*p);
+ pos++;
+ }
}
return p - start;
/* Fetch and returns the character at byte position BYTEPOS. If S is
non-NULL, fetch the character from string S; otherwise fetch the
- character from the current buffer. */
+ character from the current buffer. UNIBYTE non-zero means S is a
+ unibyte string. */
static inline int
-bidi_char_at_pos (EMACS_INT bytepos, const unsigned char *s)
+bidi_char_at_pos (ptrdiff_t bytepos, const unsigned char *s, int unibyte)
{
if (s)
- return STRING_CHAR (s + bytepos);
+ {
+ if (unibyte)
+ return s[bytepos];
+ else
+ return STRING_CHAR (s + bytepos);
+ }
else
return FETCH_MULTIBYTE_CHAR (bytepos);
}
/* 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 string to iterate, or NULL if iterating over
- a buffer. */
+ 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 frame_window_p, EMACS_INT *ch_len, EMACS_INT *nchars)
+bidi_fetch_char (ptrdiff_t bytepos, ptrdiff_t charpos, ptrdiff_t *disp_pos,
+ int *disp_prop, struct bidi_string_data *string,
+ int frame_window_p, ptrdiff_t *ch_len, ptrdiff_t *nchars)
{
int ch;
- EMACS_INT endpos = string->s ? string->schars : ZV;
+ ptrdiff_t 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
the position of the next one. That position could be at CHARPOS. */
if (charpos < endpos && charpos > *disp_pos)
- *disp_pos = compute_display_string_pos (charpos, string, frame_window_p);
+ {
+ SET_TEXT_POS (pos, charpos, bytepos);
+ *disp_pos = compute_display_string_pos (&pos, string, frame_window_p,
+ disp_prop);
+ }
/* Fetch the character at BYTEPOS. */
if (charpos >= endpos)
*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;
+ ptrdiff_t disp_end_pos;
/* We don't expect to find ourselves in the middle of a display
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);
*nchars = disp_end_pos - *disp_pos;
+ if (*nchars <= 0)
+ abort ();
if (string->s)
*ch_len = bidi_count_bytes (string->s, *disp_pos, bytepos,
- disp_end_pos);
+ disp_end_pos, string->unibyte);
+ else if (STRINGP (string->lstring))
+ *ch_len = bidi_count_bytes (SDATA (string->lstring), *disp_pos,
+ bytepos, disp_end_pos, string->unibyte);
else
*ch_len = CHAR_TO_BYTE (disp_end_pos) - bytepos;
}
{
if (string->s)
{
- EMACS_INT len;
+ int len;
+
+ if (!string->unibyte)
+ {
+ ch = STRING_CHAR_AND_LENGTH (string->s + bytepos, len);
+ *ch_len = len;
+ }
+ else
+ {
+ ch = UNIBYTE_TO_CHAR (string->s[bytepos]);
+ *ch_len = 1;
+ }
+ }
+ else if (STRINGP (string->lstring))
+ {
+ int len;
- ch = STRING_CHAR_AND_LENGTH (string->s + bytepos, len);
- *ch_len = len;
+ if (!string->unibyte)
+ {
+ ch = STRING_CHAR_AND_LENGTH (SDATA (string->lstring) + bytepos,
+ len);
+ *ch_len = len;
+ }
+ else
+ {
+ ch = UNIBYTE_TO_CHAR (SREF (string->lstring, bytepos));
+ *ch_len = 1;
+ }
}
else
{
/* 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)
- *disp_pos = compute_display_string_pos (charpos + *nchars, string,
- frame_window_p);
+ 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_prop);
+ }
return ch;
}
+\f
+/***********************************************************************
+ Determining paragraph direction
+ ***********************************************************************/
+
+/* 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 ptrdiff_t
+bidi_at_paragraph_end (ptrdiff_t charpos, ptrdiff_t bytepos)
+{
+ Lisp_Object sep_re;
+ Lisp_Object start_re;
+ ptrdiff_t val;
+
+ sep_re = paragraph_separate_re;
+ start_re = 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;
+ }
+
+ return val;
+}
+
/* 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)
+static ptrdiff_t
+bidi_find_paragraph_start (ptrdiff_t pos, ptrdiff_t pos_byte)
{
Lisp_Object re = paragraph_start_re;
- EMACS_INT limit = ZV, limit_byte = ZV_BYTE;
+ ptrdiff_t limit = ZV, limit_byte = ZV_BYTE;
while (pos_byte > BEGV_BYTE
&& fast_looking_at (re, pos, pos_byte, limit, limit_byte, Qnil) < 0)
void
bidi_paragraph_init (bidi_dir_t dir, struct bidi_it *bidi_it, int no_default_p)
{
- EMACS_INT bytepos = bidi_it->bytepos;
- int string_p = bidi_it->string.s != NULL;
- EMACS_INT pstartbyte;
+ ptrdiff_t bytepos = bidi_it->bytepos;
+ int string_p = bidi_it->string.s != NULL || STRINGP (bidi_it->string.lstring);
+ ptrdiff_t pstartbyte;
/* Note that begbyte is a byte position, while end is a character
position. Yes, this is ugly, but we are trying to avoid costly
calls to BYTE_TO_CHAR and its ilk. */
- EMACS_INT begbyte = string_p ? 0 : BEGV_BYTE;
- EMACS_INT end = string_p ? bidi_it->string.schars : ZV;
+ ptrdiff_t begbyte = string_p ? 0 : BEGV_BYTE;
+ ptrdiff_t end = string_p ? bidi_it->string.schars : ZV;
/* Special case for an empty buffer. */
if (bytepos == begbyte && bidi_it->charpos == end)
else if (dir == NEUTRAL_DIR) /* P2 */
{
int ch;
- EMACS_INT ch_len, nchars;
- EMACS_INT pos, disp_pos = -1;
+ ptrdiff_t ch_len, nchars;
+ ptrdiff_t pos, disp_pos = -1;
+ int disp_prop = 0;
bidi_type_t type;
+ const unsigned char *s;
if (!bidi_initialized)
bidi_initialize ();
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;
if (bytepos > begbyte
- && bidi_char_at_pos (bytepos, bidi_it->string.s) == '\n')
+ && bidi_char_at_pos (bytepos, s, bidi_it->string.unibyte) == '\n')
{
bytepos++;
pos++;
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
|| type == LRE || type == LRO));
type = bidi_get_type (ch, NEUTRAL_DIR))
{
- if (!string_p
- && type == NEUTRAL_B
- && bidi_at_paragraph_end (pos, bytepos) >= -1)
- break;
if (pos >= end)
{
/* Pretend there's a paragraph separator at end of
type = NEUTRAL_B;
break;
}
+ if (!string_p
+ && type == NEUTRAL_B
+ && 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->paragraph_dir = L2R; /* P3 and HL1 */
else
{
- EMACS_INT prevpbyte = pstartbyte;
- EMACS_INT p = BYTE_TO_CHAR (pstartbyte), pbyte = pstartbyte;
+ ptrdiff_t prevpbyte = pstartbyte;
+ ptrdiff_t p = BYTE_TO_CHAR (pstartbyte), pbyte = pstartbyte;
/* Find the beginning of the previous paragraph, if any. */
while (pbyte > BEGV_BYTE && prevpbyte >= pstartbyte)
bidi_line_init (bidi_it);
}
-/* 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;
-}
-
-/* Initialize the bidi iterator from buffer/string position CHARPOS. */
-void
-bidi_init_it (EMACS_INT charpos, EMACS_INT bytepos, int frame_window_p,
- struct bidi_it *bidi_it)
-{
- if (! bidi_initialized)
- bidi_initialize ();
- if (charpos >= 0)
- bidi_it->charpos = charpos;
- if (bytepos >= 0)
- bidi_it->bytepos = bytepos;
- bidi_it->frame_window_p = frame_window_p;
- bidi_it->nchars = -1; /* to be computed in bidi_resolve_explicit_1 */
- 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 = 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->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->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->sor = L2R; /* FIXME: should it be user-selectable? */
- bidi_it->disp_pos = -1; /* invalid/unknown */
- /* We can only shrink the cache if we are at the bottom level of its
- "stack". */
- if (bidi_cache_start == 0)
- bidi_cache_shrink ();
-}
-
-/* Push the current embedding level and override status; reset the
- current level to LEVEL and the current override status to OVERRIDE. */
-static INLINE void
-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. */
-static INLINE int
-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. */
-static INLINE void
-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;
- bidi_check_type (bidi_it->type);
- saved_info->type_after_w1 = bidi_it->type_after_w1;
- bidi_check_type (bidi_it->type_after_w1);
- saved_info->orig_type = bidi_it->orig_type;
- bidi_check_type (bidi_it->orig_type);
-}
+\f
+/***********************************************************************
+ Resolving explicit and implicit levels.
+ The rest of this file constitutes the core of the UBA implementation.
+ ***********************************************************************/
-/* 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
-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;
-}
-
-static INLINE int
+static inline int
bidi_explicit_dir_char (int ch)
{
bidi_type_t ch_type;
int current_level;
int new_level;
bidi_dir_t override;
- int string_p = bidi_it->string.s != NULL;
+ int string_p = bidi_it->string.s != NULL || STRINGP (bidi_it->string.lstring);
/* If reseat()'ed, don't advance, so as to start iteration from the
position where we were reseated. bidi_it->bytepos can be less
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;
+
if (bidi_it->charpos < 0)
bidi_it->charpos = 0;
- bidi_it->bytepos = bidi_count_bytes (bidi_it->string.s, 0, 0,
- bidi_it->charpos);
+ bidi_it->bytepos = bidi_count_bytes (p, 0, 0, bidi_it->charpos,
+ bidi_it->string.unibyte);
}
else
{
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;
+ ptrdiff_t 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;
if (prev_level < new_level
&& bidi_it->type == WEAK_BN
&& bidi_it->ignore_bn_limit == -1 /* only if not already known */
&& bidi_it->charpos < eob /* not already at EOB */
&& bidi_explicit_dir_char (bidi_char_at_pos (bidi_it->bytepos
- + bidi_it->ch_len,
- bidi_it->string.s)))
+ + bidi_it->ch_len, s,
+ bidi_it->string.unibyte)))
{
/* Avoid pushing and popping embedding levels if the level run
is empty, as this breaks level runs where it shouldn't.
bidi_copy_it (&saved_it, bidi_it);
while (bidi_explicit_dir_char (bidi_char_at_pos (bidi_it->bytepos
- + bidi_it->ch_len,
- bidi_it->string.s)))
+ + bidi_it->ch_len, s,
+ bidi_it->string.unibyte)))
{
/* This advances to the next character, skipping any
characters covered by display strings. */
level = bidi_resolve_explicit_1 (bidi_it);
+ /* If string.lstring was relocated inside bidi_resolve_explicit_1,
+ a pointer to its data is no longer valid. */
+ if (STRINGP (bidi_it->string.lstring))
+ s = SDATA (bidi_it->string.lstring);
}
if (bidi_it->nchars <= 0)
int next_char;
bidi_type_t type_of_next;
struct bidi_it saved_it;
- EMACS_INT eob = bidi_it->string.s ? bidi_it->string.schars : ZV;
+ ptrdiff_t 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,
- bidi_it->string.s);
+ : 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
type = WEAK_EN;
else /* W5: ET/BN with EN after it. */
{
- EMACS_INT en_pos = bidi_it->charpos + bidi_it->nchars;
+ ptrdiff_t 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;
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,
- bidi_it->string.s);
+ : 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
return type;
}
+/* 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
+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;
+}
+
static bidi_type_t
bidi_resolve_neutral (struct bidi_it *bidi_it)
{
bidi_type_t type;
int level, prev_level = -1;
struct bidi_saved_info next_for_neutral;
- EMACS_INT next_char_pos = -1;
+ ptrdiff_t next_char_pos = -2;
if (bidi_it->scan_dir == 1)
{
+ ptrdiff_t 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 >= (bidi_it->string.s ? bidi_it->string.schars : ZV))
+ if (bidi_it->charpos >= eob)
return bidi_it->resolved_level;
/* Record the info about the previous character. */
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;
+
if (bidi_it->scan_dir > 0)
{
if (bidi_it->nchars <= 0)
abort ();
next_char_pos = bidi_it->charpos + bidi_it->nchars;
}
- else if (bidi_it->charpos >= (bidi_it->string.s ? 0 : 1))
+ else if (bidi_it->charpos >= bob)
/* Implementation note: we allow next_char_pos to be as low as
0 for buffers or -1 for strings, and that is okay because
that's the "position" of the sentinel iterator state we
cached at the beginning of the iteration. */
next_char_pos = bidi_it->charpos - 1;
- if (next_char_pos >= 0)
+ if (next_char_pos >= bob - 1)
type = bidi_cache_find (next_char_pos, -1, bidi_it);
else
type = UNKNOWN_BT;
&& bidi_it->next_for_ws.type == UNKNOWN_BT)
{
int ch;
- EMACS_INT clen = bidi_it->ch_len;
- EMACS_INT bpos = bidi_it->bytepos;
- EMACS_INT cpos = bidi_it->charpos;
- EMACS_INT disp_pos = bidi_it->disp_pos;
- EMACS_INT nc = bidi_it->nchars;
+ ptrdiff_t clen = bidi_it->ch_len;
+ ptrdiff_t bpos = bidi_it->bytepos;
+ ptrdiff_t cpos = bidi_it->charpos;
+ ptrdiff_t disp_pos = bidi_it->disp_pos;
+ ptrdiff_t nc = bidi_it->nchars;
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);
+ ch = bidi_fetch_char (bpos += clen, cpos += nc, &disp_pos, &dpp, &bs,
+ fwp, &clen, &nc);
if (ch == '\n' || ch == BIDI_EOB /* || ch == LINESEP_CHAR */)
chtype = NEUTRAL_B;
else
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;
- int idx;
+ ptrdiff_t idx;
/* Try the cache first. */
if ((idx = bidi_cache_find_level_change (level, dir, end_flag))
{
int old_level, new_level, next_level;
struct bidi_it sentinel;
+ struct gcpro gcpro1;
if (bidi_it->charpos < 0 || bidi_it->bytepos < 0)
abort ();
bidi_it->scan_dir = 1; /* default to logical order */
}
+ /* The code below can call eval, and thus cause GC. If we are
+ iterating a Lisp string, make sure it won't be GCed. */
+ if (STRINGP (bidi_it->string.lstring))
+ 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)
bidi_line_init (bidi_it);
separator limit to the end of the string prevents
bidi_paragraph_init from being called automatically on this
string. */
- if (bidi_it->string.s)
+ if (bidi_it->string.s || STRINGP (bidi_it->string.lstring))
bidi_it->separator_limit = bidi_it->string.schars;
else if (bidi_it->bytepos < ZV_BYTE)
{
- EMACS_INT sep_len =
+ ptrdiff_t sep_len =
bidi_at_paragraph_end (bidi_it->charpos + bidi_it->nchars,
bidi_it->bytepos + bidi_it->ch_len);
if (bidi_it->nchars <= 0)
else
bidi_cache_iterator_state (bidi_it, 1);
}
+
+ if (STRINGP (bidi_it->string.lstring))
+ UNGCPRO;
}
/* This is meant to be called from within the debugger, whenever you
void
bidi_dump_cached_states (void)
{
- int i;
+ ptrdiff_t i;
int ndigits = 1;
if (bidi_cache_idx == 0)
fprintf (stderr, "The cache is empty.\n");
return;
}
- fprintf (stderr, "Total of %d state%s in cache:\n",
+ fprintf (stderr, "Total of %"pD"d state%s in cache:\n",
bidi_cache_idx, bidi_cache_idx == 1 ? "" : "s");
for (i = bidi_cache[bidi_cache_idx - 1].charpos; i > 0; i /= 10)