walkero
/
harfbuzz
1
0
Fork 0
Code Issues Pull Requests Projects Releases Wiki Activity

Add South-East Asian shaper 

Handles Tai Tham, Cham, and New Tai Lue for now.
This commit is contained in:
Behdad Esfahbod 2013-02-12 12:14:10 -05:00
parent fb96021206
commit 3a83d33ec0
22 changed files with 792 additions and 263 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

12
src/Makefile.am
View File

@ -87,9 +87,11 @@ HBSOURCES += \
hb-ot-shape-complex-indic.cc \
hb-ot-shape-complex-indic-machine.hh \
hb-ot-shape-complex-indic-private.hh \
hb-ot-shape-complex-indic-table.hh \
hb-ot-shape-complex-indic-table.cc \
hb-ot-shape-complex-myanmar.cc \
hb-ot-shape-complex-myanmar-machine.hh \
hb-ot-shape-complex-sea.cc \
hb-ot-shape-complex-sea-machine.hh \
hb-ot-shape-complex-thai.cc \
hb-ot-shape-complex-private.hh \
hb-ot-shape-normalize-private.hh \
@ -242,9 +244,9 @@ EXTRA_DIST += $(GENERATORS)
unicode-tables: arabic-table indic-table
indic-table: gen-indic-table.py IndicSyllabicCategory.txt IndicMatraCategory.txt Blocks.txt
$(AM_V_GEN) $(builddir)/$^ > hb-ot-shape-complex-indic-table.hh.tmp && \
mv hb-ot-shape-complex-indic-table.hh.tmp $(srcdir)/hb-ot-shape-complex-indic-table.hh || \
($(RM) hb-ot-shape-complex-indic-table.hh.tmp; false)
$(AM_V_GEN) $(builddir)/$^ > hb-ot-shape-complex-indic-table.cc.tmp && \
mv hb-ot-shape-complex-indic-table.cc.tmp $(srcdir)/hb-ot-shape-complex-indic-table.cc || \
($(RM) hb-ot-shape-complex-indic-table.cc.tmp; false)
arabic-table: gen-arabic-table.py ArabicShaping.txt UnicodeData.txt
$(AM_V_GEN) $(builddir)/$^ > hb-ot-shape-complex-arabic-table.hh.tmp && \
@ -257,10 +259,12 @@ arabic-table: gen-arabic-table.py ArabicShaping.txt UnicodeData.txt
BUILT_SOURCES += \
hb-ot-shape-complex-indic-machine.hh \
hb-ot-shape-complex-myanmar-machine.hh \
hb-ot-shape-complex-sea-machine.hh \
$(NULL)
EXTRA_DIST += \
hb-ot-shape-complex-indic-machine.rl \
hb-ot-shape-complex-myanmar-machine.rl \
hb-ot-shape-complex-sea-machine.rl \
$(NULL)
%.hh: %.rl
$(AM_V_GEN)$(top_srcdir)/missing --run ragel -e -F1 -o "$@.tmp" "$<" && \

9
src/gen-indic-table.py
View File

@ -75,8 +75,7 @@ for h in headers:
print " * %s" % (l.strip())
print " */"
print
print "#ifndef HB_OT_SHAPE_COMPLEX_INDIC_TABLE_HH"
print "#define HB_OT_SHAPE_COMPLEX_INDIC_TABLE_HH"
print '#include "hb-ot-shape-complex-indic-private.hh"'
print
# Shorten values
@ -182,8 +181,8 @@ print
occupancy = used * 100. / total
print "}; /* Table occupancy: %d%% */" % occupancy
print
print "static INDIC_TABLE_ELEMENT_TYPE"
print "get_indic_categories (hb_codepoint_t u)"
print "INDIC_TABLE_ELEMENT_TYPE"
print "hb_indic_get_categories (hb_codepoint_t u)"
print "{"
for (start,end) in zip (starts, ends):
offset = "indic_offset_0x%04x" % start
@ -202,8 +201,6 @@ for i in range (2):
print "#undef %s_%s" % \
(what_short[i], short[i][v])
print
print "#endif /* HB_OT_SHAPE_COMPLEX_INDIC_TABLE_HH */"
print
print "/* == End of generated table == */"
# Maintain at least 30% occupancy in the table */

241
src/hb-ot-shape-complex-indic-private.hh
View File

@ -34,11 +34,6 @@
#include "hb-ot-shape-private.hh" /* XXX Remove */
/* buffer var allocations */
#define indic_category() complex_var_u8_0() /* indic_category_t */
#define indic_position() complex_var_u8_1() /* indic_matra_category_t */
#define INDIC_TABLE_ELEMENT_TYPE uint16_t
/* Cateories used in the OpenType spec:
@ -150,239 +145,7 @@ enum indic_matra_category_t {
ASSERT_STATIC_EXPR_ZERO (S < 255 && M < 255) + \
((M << 8) | S))
#include "hb-ot-shape-complex-indic-table.hh"
#define IN_HALF_BLOCK(u, Base) (((u) & ~0x7F) == (Base))
#define IS_DEVA(u) (IN_HALF_BLOCK (u, 0x0900))
#define IS_BENG(u) (IN_HALF_BLOCK (u, 0x0980))
#define IS_GURU(u) (IN_HALF_BLOCK (u, 0x0A00))
#define IS_GUJR(u) (IN_HALF_BLOCK (u, 0x0A80))
#define IS_ORYA(u) (IN_HALF_BLOCK (u, 0x0B00))
#define IS_TAML(u) (IN_HALF_BLOCK (u, 0x0B80))
#define IS_TELU(u) (IN_HALF_BLOCK (u, 0x0C00))
#define IS_KNDA(u) (IN_HALF_BLOCK (u, 0x0C80))
#define IS_MLYM(u) (IN_HALF_BLOCK (u, 0x0D00))
#define IS_SINH(u) (IN_HALF_BLOCK (u, 0x0D80))
#define IS_KHMR(u) (IN_HALF_BLOCK (u, 0x1780))
#define MATRA_POS_LEFT(u) POS_PRE_M
#define MATRA_POS_RIGHT(u) ( \
IS_DEVA(u) ? POS_AFTER_SUB : \
IS_BENG(u) ? POS_AFTER_POST : \
IS_GURU(u) ? POS_AFTER_POST : \
IS_GUJR(u) ? POS_AFTER_POST : \
IS_ORYA(u) ? POS_AFTER_POST : \
IS_TAML(u) ? POS_AFTER_POST : \
IS_TELU(u) ? (u <= 0x0C42 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
IS_KNDA(u) ? (u < 0x0CC3 || u > 0xCD6 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
IS_MLYM(u) ? POS_AFTER_POST : \
IS_SINH(u) ? POS_AFTER_SUB : \
IS_KHMR(u) ? POS_AFTER_POST : \
/*default*/ POS_AFTER_SUB \
)
#define MATRA_POS_TOP(u) ( /* BENG and MLYM don't have top matras. */ \
IS_DEVA(u) ? POS_AFTER_SUB : \
IS_GURU(u) ? POS_AFTER_POST : /* Deviate from spec */ \
IS_GUJR(u) ? POS_AFTER_SUB : \
IS_ORYA(u) ? POS_AFTER_MAIN : \
IS_TAML(u) ? POS_AFTER_SUB : \
IS_TELU(u) ? POS_BEFORE_SUB : \
IS_KNDA(u) ? POS_BEFORE_SUB : \
IS_SINH(u) ? POS_AFTER_SUB : \
IS_KHMR(u) ? POS_AFTER_POST : \
/*default*/ POS_AFTER_SUB \
)
#define MATRA_POS_BOTTOM(u) ( \
IS_DEVA(u) ? POS_AFTER_SUB : \
IS_BENG(u) ? POS_AFTER_SUB : \
IS_GURU(u) ? POS_AFTER_POST : \
IS_GUJR(u) ? POS_AFTER_POST : \
IS_ORYA(u) ? POS_AFTER_SUB : \
IS_TAML(u) ? POS_AFTER_POST : \
IS_TELU(u) ? POS_BEFORE_SUB : \
IS_KNDA(u) ? POS_BEFORE_SUB : \
IS_MLYM(u) ? POS_AFTER_POST : \
IS_SINH(u) ? POS_AFTER_SUB : \
IS_KHMR(u) ? POS_AFTER_POST : \
/*default*/ POS_AFTER_SUB \
)
static inline indic_position_t
matra_position (hb_codepoint_t u, indic_position_t side)
{
switch ((int) side)
{
case POS_PRE_C: return MATRA_POS_LEFT (u);
case POS_POST_C: return MATRA_POS_RIGHT (u);
case POS_ABOVE_C: return MATRA_POS_TOP (u);
case POS_BELOW_C: return MATRA_POS_BOTTOM (u);
};
return side;
}
/* XXX
* This is a hack for now. We should move this data into the main Indic table.
* Or completely remove it and just check in the tables.
*/
static const hb_codepoint_t ra_chars[] = {
0x0930, /* Devanagari */
0x09B0, /* Bengali */
0x09F0, /* Bengali */
0x0A30, /* Gurmukhi */ /* No Reph */
0x0AB0, /* Gujarati */
0x0B30, /* Oriya */
0x0BB0, /* Tamil */ /* No Reph */
0x0C30, /* Telugu */ /* Reph formed only with ZWJ */
0x0CB0, /* Kannada */
0x0D30, /* Malayalam */ /* No Reph, Logical Repha */
0x0DBB, /* Sinhala */ /* Reph formed only with ZWJ */
0x179A, /* Khmer */ /* No Reph, Visual Repha */
};
static inline indic_position_t
consonant_position (hb_codepoint_t u)
{
if ((u & ~0x007F) == 0x1780)
return POS_BELOW_C; /* In Khmer coeng model, post and below forms should not be reordered. */
return POS_BASE_C; /* Will recategorize later based on font lookups. */
}
static inline bool
is_ra (hb_codepoint_t u)
{
for (unsigned int i = 0; i < ARRAY_LENGTH (ra_chars); i++)
if (u == ra_chars[i])
return true;
return false;
}
static inline bool
is_one_of (const hb_glyph_info_t &info, unsigned int flags)
{
/* If it ligated, all bets are off. */
if (is_a_ligature (info)) return false;
return !!(FLAG (info.indic_category()) & flags);
}
#define JOINER_FLAGS (FLAG (OT_ZWJ) | FLAG (OT_ZWNJ))
static inline bool
is_joiner (const hb_glyph_info_t &info)
{
return is_one_of (info, JOINER_FLAGS);
}
/* Note:
*
* We treat Vowels and placeholders as if they were consonants. This is safe because Vowels
* cannot happen in a consonant syllable. The plus side however is, we can call the
* consonant syllable logic from the vowel syllable function and get it all right! */
#define CONSONANT_FLAGS (FLAG (OT_C) | FLAG (OT_CM) | FLAG (OT_Ra) | FLAG (OT_V) | FLAG (OT_NBSP) | FLAG (OT_DOTTEDCIRCLE))
static inline bool
is_consonant (const hb_glyph_info_t &info)
{
return is_one_of (info, CONSONANT_FLAGS);
}
#define HALANT_OR_COENG_FLAGS (FLAG (OT_H) | FLAG (OT_Coeng))
static inline bool
is_halant_or_coeng (const hb_glyph_info_t &info)
{
return is_one_of (info, HALANT_OR_COENG_FLAGS);
}
static inline void
set_indic_properties (hb_glyph_info_t &info)
{
hb_codepoint_t u = info.codepoint;
unsigned int type = get_indic_categories (u);
indic_category_t cat = (indic_category_t) (type & 0x7F);
indic_position_t pos = (indic_position_t) (type >> 8);
/*
* Re-assign category
*/
/* The spec says U+0952 is OT_A. However, testing shows that Uniscribe
* treats U+0951..U+0952 all as OT_VD.
* TESTS:
* U+092E,U+0947,U+0952
* U+092E,U+0952,U+0947
* U+092E,U+0947,U+0951
* U+092E,U+0951,U+0947
* */
if (unlikely (hb_in_range<hb_codepoint_t> (u, 0x0951, 0x0954)))
cat = OT_VD;
if (unlikely (u == 0x17D1))
cat = OT_X;
if (cat == OT_X &&
unlikely (hb_in_range<hb_codepoint_t> (u, 0x17CB, 0x17D3))) /* Khmer Various signs */
{
/* These are like Top Matras. */
cat = OT_M;
pos = POS_ABOVE_C;
}
if (u == 0x17C6) /* Khmer Bindu doesn't like to be repositioned. */
cat = OT_N;
if (unlikely (u == 0x17D2)) cat = OT_Coeng; /* Khmer coeng */
else if (unlikely (u == 0x200C)) cat = OT_ZWNJ;
else if (unlikely (u == 0x200D)) cat = OT_ZWJ;
else if (unlikely (u == 0x25CC)) cat = OT_DOTTEDCIRCLE;
else if (unlikely (u == 0x0A71)) cat = OT_SM; /* GURMUKHI ADDAK. More like consonant medial. like 0A75. */
if (cat == OT_Repha) {
/* There are two kinds of characters marked as Repha:
* - The ones that are GenCat=Mn are already positioned visually, ie. after base. (eg. Khmer)
* - The ones that are GenCat=Lo is encoded logically, ie. beginning of syllable. (eg. Malayalam)
*
* We recategorize the first kind to look like a Nukta and attached to the base directly.
*/
if (_hb_glyph_info_get_general_category (&info) == HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK)
cat = OT_N;
}
/*
* Re-assign position.
*/
if ((FLAG (cat) & CONSONANT_FLAGS))
{
pos = consonant_position (u);
if (is_ra (u))
cat = OT_Ra;
}
else if (cat == OT_M)
{
pos = matra_position (u, pos);
}
else if (cat == OT_SM || cat == OT_VD)
{
pos = POS_SMVD;
}
if (unlikely (u == 0x0B01)) pos = POS_BEFORE_SUB; /* Oriya Bindu is BeforeSub in the spec. */
info.indic_category() = cat;
info.indic_position() = pos;
}
HB_INTERNAL INDIC_TABLE_ELEMENT_TYPE
hb_indic_get_categories (hb_codepoint_t u);
#endif /* HB_OT_SHAPE_COMPLEX_INDIC_PRIVATE_HH */

9
src/hb-ot-shape-complex-indic-table.hh → src/hb-ot-shape-complex-indic-table.cc
View File

@ -14,8 +14,7 @@
* # Date: 2012-05-14, 22:42:00 GMT [KW, LI]
*/
#ifndef HB_OT_SHAPE_COMPLEX_INDIC_TABLE_HH
#define HB_OT_SHAPE_COMPLEX_INDIC_TABLE_HH
#include "hb-ot-shape-complex-indic-private.hh"
#define ISC_A INDIC_SYLLABIC_CATEGORY_AVAGRAHA /* 11 chars; Avagraha */
@ -807,8 +806,8 @@ static const INDIC_TABLE_ELEMENT_TYPE indic_table[] = {
}; /* Table occupancy: 60% */
static INDIC_TABLE_ELEMENT_TYPE
get_indic_categories (hb_codepoint_t u)
INDIC_TABLE_ELEMENT_TYPE
hb_indic_get_categories (hb_codepoint_t u)
{
if (0x0900 <= u && u <= 0x10A0) return indic_table[u - 0x0900 + indic_offset_0x0900];
if (0x1700 <= u && u <= 0x1800) return indic_table[u - 0x1700 + indic_offset_0x1700];
@ -867,6 +866,4 @@ get_indic_categories (hb_codepoint_t u)
#undef IMC_TR
#undef IMC_VOL
#endif /* HB_OT_SHAPE_COMPLEX_INDIC_TABLE_HH */
/* == End of generated table == */

239
src/hb-ot-shape-complex-indic.cc
View File

@ -27,6 +27,245 @@
#include "hb-ot-shape-complex-indic-private.hh"
#include "hb-ot-layout-private.hh"
/* buffer var allocations */
#define indic_category() complex_var_u8_0() /* indic_category_t */
#define indic_position() complex_var_u8_1() /* indic_position_t */
/*
* Indic shaper.
*/
#define IN_HALF_BLOCK(u, Base) (((u) & ~0x7F) == (Base))
#define IS_DEVA(u) (IN_HALF_BLOCK (u, 0x0900))
#define IS_BENG(u) (IN_HALF_BLOCK (u, 0x0980))
#define IS_GURU(u) (IN_HALF_BLOCK (u, 0x0A00))
#define IS_GUJR(u) (IN_HALF_BLOCK (u, 0x0A80))
#define IS_ORYA(u) (IN_HALF_BLOCK (u, 0x0B00))
#define IS_TAML(u) (IN_HALF_BLOCK (u, 0x0B80))
#define IS_TELU(u) (IN_HALF_BLOCK (u, 0x0C00))
#define IS_KNDA(u) (IN_HALF_BLOCK (u, 0x0C80))
#define IS_MLYM(u) (IN_HALF_BLOCK (u, 0x0D00))
#define IS_SINH(u) (IN_HALF_BLOCK (u, 0x0D80))
#define IS_KHMR(u) (IN_HALF_BLOCK (u, 0x1780))
#define MATRA_POS_LEFT(u) POS_PRE_M
#define MATRA_POS_RIGHT(u) ( \
IS_DEVA(u) ? POS_AFTER_SUB : \
IS_BENG(u) ? POS_AFTER_POST : \
IS_GURU(u) ? POS_AFTER_POST : \
IS_GUJR(u) ? POS_AFTER_POST : \
IS_ORYA(u) ? POS_AFTER_POST : \
IS_TAML(u) ? POS_AFTER_POST : \
IS_TELU(u) ? (u <= 0x0C42 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
IS_KNDA(u) ? (u < 0x0CC3 || u > 0xCD6 ? POS_BEFORE_SUB : POS_AFTER_SUB) : \
IS_MLYM(u) ? POS_AFTER_POST : \
IS_SINH(u) ? POS_AFTER_SUB : \
IS_KHMR(u) ? POS_AFTER_POST : \
/*default*/ POS_AFTER_SUB \
)
#define MATRA_POS_TOP(u) ( /* BENG and MLYM don't have top matras. */ \
IS_DEVA(u) ? POS_AFTER_SUB : \
IS_GURU(u) ? POS_AFTER_POST : /* Deviate from spec */ \
IS_GUJR(u) ? POS_AFTER_SUB : \
IS_ORYA(u) ? POS_AFTER_MAIN : \
IS_TAML(u) ? POS_AFTER_SUB : \
IS_TELU(u) ? POS_BEFORE_SUB : \
IS_KNDA(u) ? POS_BEFORE_SUB : \
IS_SINH(u) ? POS_AFTER_SUB : \
IS_KHMR(u) ? POS_AFTER_POST : \
/*default*/ POS_AFTER_SUB \
)
#define MATRA_POS_BOTTOM(u) ( \
IS_DEVA(u) ? POS_AFTER_SUB : \
IS_BENG(u) ? POS_AFTER_SUB : \
IS_GURU(u) ? POS_AFTER_POST : \
IS_GUJR(u) ? POS_AFTER_POST : \
IS_ORYA(u) ? POS_AFTER_SUB : \
IS_TAML(u) ? POS_AFTER_POST : \
IS_TELU(u) ? POS_BEFORE_SUB : \
IS_KNDA(u) ? POS_BEFORE_SUB : \
IS_MLYM(u) ? POS_AFTER_POST : \
IS_SINH(u) ? POS_AFTER_SUB : \
IS_KHMR(u) ? POS_AFTER_POST : \
/*default*/ POS_AFTER_SUB \
)
static inline indic_position_t
matra_position (hb_codepoint_t u, indic_position_t side)
{
switch ((int) side)
{
case POS_PRE_C: return MATRA_POS_LEFT (u);
case POS_POST_C: return MATRA_POS_RIGHT (u);
case POS_ABOVE_C: return MATRA_POS_TOP (u);
case POS_BELOW_C: return MATRA_POS_BOTTOM (u);
};
return side;
}
/* XXX
* This is a hack for now. We should move this data into the main Indic table.
* Or completely remove it and just check in the tables.
*/
static const hb_codepoint_t ra_chars[] = {
0x0930, /* Devanagari */
0x09B0, /* Bengali */
0x09F0, /* Bengali */
0x0A30, /* Gurmukhi */ /* No Reph */
0x0AB0, /* Gujarati */
0x0B30, /* Oriya */
0x0BB0, /* Tamil */ /* No Reph */
0x0C30, /* Telugu */ /* Reph formed only with ZWJ */
0x0CB0, /* Kannada */
0x0D30, /* Malayalam */ /* No Reph, Logical Repha */
0x0DBB, /* Sinhala */ /* Reph formed only with ZWJ */
0x179A, /* Khmer */ /* No Reph, Visual Repha */
};
static inline indic_position_t
consonant_position (hb_codepoint_t u)
{
if ((u & ~0x007F) == 0x1780)
return POS_BELOW_C; /* In Khmer coeng model, post and below forms should not be reordered. */
return POS_BASE_C; /* Will recategorize later based on font lookups. */
}
static inline bool
is_ra (hb_codepoint_t u)
{
for (unsigned int i = 0; i < ARRAY_LENGTH (ra_chars); i++)
if (u == ra_chars[i])
return true;
return false;
}
static inline bool
is_one_of (const hb_glyph_info_t &info, unsigned int flags)
{
/* If it ligated, all bets are off. */
if (is_a_ligature (info)) return false;
return !!(FLAG (info.indic_category()) & flags);
}
#define JOINER_FLAGS (FLAG (OT_ZWJ) | FLAG (OT_ZWNJ))
static inline bool
is_joiner (const hb_glyph_info_t &info)
{
return is_one_of (info, JOINER_FLAGS);
}
/* Note:
*
* We treat Vowels and placeholders as if they were consonants. This is safe because Vowels
* cannot happen in a consonant syllable. The plus side however is, we can call the
* consonant syllable logic from the vowel syllable function and get it all right! */
#define CONSONANT_FLAGS (FLAG (OT_C) | FLAG (OT_CM) | FLAG (OT_Ra) | FLAG (OT_V) | FLAG (OT_NBSP) | FLAG (OT_DOTTEDCIRCLE))
static inline bool
is_consonant (const hb_glyph_info_t &info)
{
return is_one_of (info, CONSONANT_FLAGS);
}
#define HALANT_OR_COENG_FLAGS (FLAG (OT_H) | FLAG (OT_Coeng))
static inline bool
is_halant_or_coeng (const hb_glyph_info_t &info)
{
return is_one_of (info, HALANT_OR_COENG_FLAGS);
}
static inline void
set_indic_properties (hb_glyph_info_t &info)
{
hb_codepoint_t u = info.codepoint;
unsigned int type = hb_indic_get_categories (u);
indic_category_t cat = (indic_category_t) (type & 0x7F);
indic_position_t pos = (indic_position_t) (type >> 8);
/*
* Re-assign category
*/
/* The spec says U+0952 is OT_A. However, testing shows that Uniscribe
* treats U+0951..U+0952 all as OT_VD.
* TESTS:
* U+092E,U+0947,U+0952
* U+092E,U+0952,U+0947
* U+092E,U+0947,U+0951
* U+092E,U+0951,U+0947
* */
if (unlikely (hb_in_range<hb_codepoint_t> (u, 0x0951, 0x0954)))
cat = OT_VD;
if (unlikely (u == 0x17D1))
cat = OT_X;
if (cat == OT_X &&
unlikely (hb_in_range<hb_codepoint_t> (u, 0x17CB, 0x17D3))) /* Khmer Various signs */
{
/* These are like Top Matras. */
cat = OT_M;
pos = POS_ABOVE_C;
}
if (u == 0x17C6) /* Khmer Bindu doesn't like to be repositioned. */
cat = OT_N;
if (unlikely (u == 0x17D2)) cat = OT_Coeng; /* Khmer coeng */
else if (unlikely (u == 0x200C)) cat = OT_ZWNJ;
else if (unlikely (u == 0x200D)) cat = OT_ZWJ;
else if (unlikely (u == 0x25CC)) cat = OT_DOTTEDCIRCLE;
else if (unlikely (u == 0x0A71)) cat = OT_SM; /* GURMUKHI ADDAK. More like consonant medial. like 0A75. */
if (cat == OT_Repha) {
/* There are two kinds of characters marked as Repha:
* - The ones that are GenCat=Mn are already positioned visually, ie. after base. (eg. Khmer)
* - The ones that are GenCat=Lo is encoded logically, ie. beginning of syllable. (eg. Malayalam)
*
* We recategorize the first kind to look like a Nukta and attached to the base directly.
*/
if (_hb_glyph_info_get_general_category (&info) == HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK)
cat = OT_N;
}
/*
* Re-assign position.
*/
if ((FLAG (cat) & CONSONANT_FLAGS))
{
pos = consonant_position (u);
if (is_ra (u))
cat = OT_Ra;
}
else if (cat == OT_M)
{
pos = matra_position (u, pos);
}
else if (cat == OT_SM || cat == OT_VD)
{
pos = POS_SMVD;
}
if (unlikely (u == 0x0B01)) pos = POS_BEFORE_SUB; /* Oriya Bindu is BeforeSub in the spec. */
info.indic_category() = cat;
info.indic_position() = pos;
}
/*
* Things above this line should ideally be moved to the Indic table itself.
*/
/*
* Global Indic shaper options.

2
src/hb-ot-shape-complex-myanmar.cc
View File

@ -28,7 +28,7 @@
/* buffer var allocations */
#define myanmar_category() complex_var_u8_0() /* myanmar_category_t */
#define myanmar_position() complex_var_u8_1() /* myanmar_matra_category_t */
#define myanmar_position() complex_var_u8_1() /* myanmar_position_t */
/*

28
src/hb-ot-shape-complex-private.hh
View File

@ -52,6 +52,7 @@ enum hb_ot_shape_zero_width_marks_type_t {
HB_COMPLEX_SHAPER_IMPLEMENT (arabic) \
HB_COMPLEX_SHAPER_IMPLEMENT (indic) \
HB_COMPLEX_SHAPER_IMPLEMENT (myanmar) \
HB_COMPLEX_SHAPER_IMPLEMENT (sea) \
HB_COMPLEX_SHAPER_IMPLEMENT (thai) \
/* ^--- Add new shapers here */
@ -262,13 +263,11 @@ hb_ot_shape_complex_categorize (const hb_ot_shape_planner_t *planner)
/* Unicode-4.1 additions */
case HB_SCRIPT_BUGINESE:
case HB_SCRIPT_NEW_TAI_LUE:
/* Unicode-5.0 additions */
case HB_SCRIPT_BALINESE:
/* Unicode-5.1 additions */
case HB_SCRIPT_CHAM:
case HB_SCRIPT_LEPCHA:
case HB_SCRIPT_REJANG:
case HB_SCRIPT_SUNDANESE:
@ -277,8 +276,8 @@ hb_ot_shape_complex_categorize (const hb_ot_shape_planner_t *planner)
case HB_SCRIPT_JAVANESE:
case HB_SCRIPT_KAITHI:
case HB_SCRIPT_MEETEI_MAYEK:
case HB_SCRIPT_TAI_THAM:
/* Unicode-6.0 additions */
/* Unicode-6.1 additions */
case HB_SCRIPT_CHAKMA:
@ -287,9 +286,8 @@ hb_ot_shape_complex_categorize (const hb_ot_shape_planner_t *planner)
/* If the designer designed the font for the 'DFLT' script,
* use the default shaper. Otherwise, use the Indic shaper.
* Note that for some scripts, like New Tai Lue, there may
* not be *any* GSUB/GPOS needed, so there may be no scripts
* found! */
* Note that for some simple scripts, there may not be *any*
* GSUB/GPOS needed, so there may be no scripts found! */
if (planner->map.chosen_script[0] == HB_TAG ('D','F','L','T'))
return &_hb_ot_complex_shaper_default;
else
@ -318,6 +316,24 @@ hb_ot_shape_complex_categorize (const hb_ot_shape_planner_t *planner)
return &_hb_ot_complex_shaper_myanmar;
else
return &_hb_ot_complex_shaper_default;
/* Unicode-4.1 additions */
case HB_SCRIPT_NEW_TAI_LUE:
/* Unicode-5.1 additions */
case HB_SCRIPT_CHAM:
/* Unicode-5.2 additions */
case HB_SCRIPT_TAI_THAM:
/* If the designer designed the font for the 'DFLT' script,
* use the default shaper. Otherwise, use the Indic shaper.
* Note that for some simple scripts, there may not be *any*
* GSUB/GPOS needed, so there may be no scripts found! */
if (planner->map.chosen_script[0] == HB_TAG ('D','F','L','T'))
return &_hb_ot_complex_shaper_default;
else
return &_hb_ot_complex_shaper_sea;
}
}

102
src/hb-ot-shape-complex-sea-machine.rl Normal file
View File

@ -0,0 +1,102 @@
/*
* Copyright © 2011,2012,2013 Google, Inc.
*
* This is part of HarfBuzz, a text shaping library.
*
* Permission is hereby granted, without written agreement and without
* license or royalty fees, to use, copy, modify, and distribute this
* software and its documentation for any purpose, provided that the
* above copyright notice and the following two paragraphs appear in
* all copies of this software.
*
* IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
* DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
* ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
* IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
* DAMAGE.
*
* THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
* BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
* FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
* ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
* PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
*
* Google Author(s): Behdad Esfahbod
*/
#ifndef HB_OT_SHAPE_COMPLEX_SEA_MACHINE_HH
#define HB_OT_SHAPE_COMPLEX_SEA_MACHINE_HH
#include "hb-private.hh"
%%{
machine sea_syllable_machine;
alphtype unsigned char;
write data;
}%%
%%{
# Same order as enum sea_category_t. Not sure how to avoid duplication.
C = 1;
GB = 12; # Generic Base
H = 4; # Halant
IV = 2; # Independent Vowel
MR = 22; # Medial Ra
CM = 17; # Consonant Medial
VAbv = 26;
VBlw = 27;
VPre = 28;
VPst = 29;
T = 3; # Tone Marks
A = 10; # Anusvara
syllable_tail = (VPre|VAbv|VBlw|VPst|H.C|CM|MR|T|A)*;
consonant_syllable = (C|IV|GB) syllable_tail;
broken_cluster = syllable_tail;
other = any;
main := |*
consonant_syllable => { found_syllable (consonant_syllable); };
broken_cluster => { found_syllable (broken_cluster); };
other => { found_syllable (non_sea_cluster); };
*|;
}%%
#define found_syllable(syllable_type) \
HB_STMT_START { \
if (0) fprintf (stderr, "syllable %d..%d %s\n", last, p+1, #syllable_type); \
for (unsigned int i = last; i < p+1; i++) \
info[i].syllable() = (syllable_serial << 4) | syllable_type; \
last = p+1; \
syllable_serial++; \
if (unlikely (syllable_serial == 16)) syllable_serial = 1; \
} HB_STMT_END
static void
find_syllables (hb_buffer_t *buffer)
{
unsigned int p, pe, eof, ts HB_UNUSED, te, act;
int cs;
hb_glyph_info_t *info = buffer->info;
%%{
write init;
getkey info[p].sea_category();
}%%
p = 0;
pe = eof = buffer->len;
unsigned int last = 0;
unsigned int syllable_serial = 1;
%%{
write exec;
}%%
}
#undef found_syllable
#endif /* HB_OT_SHAPE_COMPLEX_SEA_MACHINE_HH */

400
src/hb-ot-shape-complex-sea.cc Normal file
View File

@ -0,0 +1,400 @@
/*
* Copyright © 2011,2012,2013 Google, Inc.
*
* This is part of HarfBuzz, a text shaping library.
*
* Permission is hereby granted, without written agreement and without
* license or royalty fees, to use, copy, modify, and distribute this
* software and its documentation for any purpose, provided that the
* above copyright notice and the following two paragraphs appear in
* all copies of this software.
*
* IN NO EVENT SHALL THE COPYRIGHT HOLDER BE LIABLE TO ANY PARTY FOR
* DIRECT, INDIRECT, SPECIAL, INCIDENTAL, OR CONSEQUENTIAL DAMAGES
* ARISING OUT OF THE USE OF THIS SOFTWARE AND ITS DOCUMENTATION, EVEN
* IF THE COPYRIGHT HOLDER HAS BEEN ADVISED OF THE POSSIBILITY OF SUCH
* DAMAGE.
*
* THE COPYRIGHT HOLDER SPECIFICALLY DISCLAIMS ANY WARRANTIES, INCLUDING,
* BUT NOT LIMITED TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND
* FITNESS FOR A PARTICULAR PURPOSE. THE SOFTWARE PROVIDED HEREUNDER IS
* ON AN "AS IS" BASIS, AND THE COPYRIGHT HOLDER HAS NO OBLIGATION TO
* PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
*
* Google Author(s): Behdad Esfahbod
*/
#include "hb-ot-shape-complex-indic-private.hh"
/* buffer var allocations */
#define sea_category() complex_var_u8_0() /* indic_category_t */
#define sea_position() complex_var_u8_1() /* indic_position_t */
/*
* South-East Asian shaper.
* Loosely based on the Myanmar spec / shaper.
* There is no OpenType spec for this.
*/
static const hb_tag_t
sea_features[] =
{
/*
* Basic features.
* These features are applied in order, one at a time, after initial_reordering.
*/
HB_TAG('p','r','e','f'),
HB_TAG('a','b','v','f'),
HB_TAG('b','l','w','f'),
HB_TAG('p','s','t','f'),
/*
* Other features.
* These features are applied all at once, after final_reordering.
*/
HB_TAG('p','r','e','s'),
HB_TAG('a','b','v','s'),
HB_TAG('b','l','w','s'),
HB_TAG('p','s','t','s'),
/* Positioning features, though we don't care about the types. */
HB_TAG('d','i','s','t'),
};
/*
* Must be in the same order as the sea_features array.
*/
enum {
_PREF,
_ABVF,
_BLWF,
_PSTF,
_PRES,
_ABVS,
_BLWS,
_PSTS,
_DIST,
SEA_NUM_FEATURES,
SEA_BASIC_FEATURES = _PRES /* Don't forget to update this! */
};
static void
setup_syllables (const hb_ot_shape_plan_t *plan,
hb_font_t *font,
hb_buffer_t *buffer);
static void
initial_reordering (const hb_ot_shape_plan_t *plan,
hb_font_t *font,
hb_buffer_t *buffer);
static void
final_reordering (const hb_ot_shape_plan_t *plan,
hb_font_t *font,
hb_buffer_t *buffer);
static void
collect_features_sea (hb_ot_shape_planner_t *plan)
{
hb_ot_map_builder_t *map = &plan->map;
/* Do this before any lookups have been applied. */
map->add_gsub_pause (setup_syllables);
map->add_bool_feature (HB_TAG('l','o','c','l'));
/* The Indic specs do not require ccmp, but we apply it here since if
* there is a use of it, it's typically at the beginning. */
map->add_bool_feature (HB_TAG('c','c','m','p'));
unsigned int i = 0;
map->add_gsub_pause (initial_reordering);
for (; i < SEA_BASIC_FEATURES; i++) {
map->add_bool_feature (sea_features[i]);
map->add_gsub_pause (NULL);
}
map->add_gsub_pause (final_reordering);
for (; i < SEA_NUM_FEATURES; i++) {
map->add_bool_feature (sea_features[i]);
}
}
static void
override_features_sea (hb_ot_shape_planner_t *plan)
{
plan->map.add_feature (HB_TAG('l','i','g','a'), 0, true);
}
enum syllable_type_t {
consonant_syllable,
broken_cluster,
non_sea_cluster,
};
#include "hb-ot-shape-complex-sea-machine.hh"
/* Note: This enum is duplicated in the -machine.rl source file.
* Not sure how to avoid duplication. */
enum sea_category_t {
// OT_C = 1,
OT_GB = 12, /* Generic Base XXX DOTTED CIRCLE only for now */
// OT_H = 4, /* Halant */
OT_IV = 2, /* Independent Vowel */
OT_MR = 22, /* Medial Ra */
// OT_CM = 17, /* Consonant Medial */
OT_VAbv = 26,
OT_VBlw = 27,
OT_VPre = 28,
OT_VPst = 29,
OT_T = 3, /* Tone Marks */
// OT_A = 10, /* Anusvara */
};
static inline void
set_sea_properties (hb_glyph_info_t &info)
{
hb_codepoint_t u = info.codepoint;
unsigned int type = hb_indic_get_categories (u);
indic_category_t cat = (indic_category_t) (type & 0x7F);
indic_position_t pos = (indic_position_t) (type >> 8);
/* Medial Ra */
if (u == 0x1A55 || u == 0xAA34)
cat = (indic_category_t) OT_MR;
if (cat == OT_M)
{
switch ((int) pos)
{
case POS_PRE_C: cat = (indic_category_t) OT_VPre; break;
case POS_ABOVE_C: cat = (indic_category_t) OT_VAbv; break;
case POS_BELOW_C: cat = (indic_category_t) OT_VBlw; break;
case POS_POST_C: cat = (indic_category_t) OT_VPst; break;
}
}
info.sea_category() = (sea_category_t) cat;
info.sea_position() = pos;
}
static void
setup_masks_sea (const hb_ot_shape_plan_t *plan HB_UNUSED,
hb_buffer_t *buffer,
hb_font_t *font HB_UNUSED)
{
HB_BUFFER_ALLOCATE_VAR (buffer, sea_category);
HB_BUFFER_ALLOCATE_VAR (buffer, sea_position);
/* We cannot setup masks here. We save information about characters
* and setup masks later on in a pause-callback. */
unsigned int count = buffer->len;
for (unsigned int i = 0; i < count; i++)
set_sea_properties (buffer->info[i]);
}
static void
setup_syllables (const hb_ot_shape_plan_t *plan HB_UNUSED,
hb_font_t *font HB_UNUSED,
hb_buffer_t *buffer)
{
find_syllables (buffer);
}
static int
compare_sea_order (const hb_glyph_info_t *pa, const hb_glyph_info_t *pb)
{
int a = pa->sea_position();
int b = pb->sea_position();
return a < b ? -1 : a == b ? 0 : +1;
}
static void
initial_reordering_consonant_syllable (const hb_ot_shape_plan_t *plan,
hb_face_t *face,
hb_buffer_t *buffer,
unsigned int start, unsigned int end)
{
hb_glyph_info_t *info = buffer->info;
unsigned int base = start;
/* Reorder! */
unsigned int i = start;
for (; i < base; i++)
info[i].sea_position() = POS_PRE_C;
if (i < end)
{
info[i].sea_position() = POS_BASE_C;
i++;
}
for (; i < end; i++)
{
if (info[i].sea_category() == OT_MR) /* Pre-base reordering */
{
info[i].sea_position() = POS_PRE_C;
continue;
}
if (info[i].sea_position() < POS_BASE_C) /* Left matra */
{
continue;
}
info[i].sea_position() = POS_AFTER_MAIN;
}
buffer->merge_clusters (start, end);
/* Sit tight, rock 'n roll! */
hb_bubble_sort (info + start, end - start, compare_sea_order);
}
static void
initial_reordering_broken_cluster (const hb_ot_shape_plan_t *plan,
hb_face_t *face,
hb_buffer_t *buffer,
unsigned int start, unsigned int end)
{
/* We already inserted dotted-circles, so just call the consonant_syllable. */
initial_reordering_consonant_syllable (plan, face, buffer, start, end);
}
static void
initial_reordering_non_sea_cluster (const hb_ot_shape_plan_t *plan HB_UNUSED,
hb_face_t *face HB_UNUSED,
hb_buffer_t *buffer HB_UNUSED,
unsigned int start HB_UNUSED, unsigned int end HB_UNUSED)
{
/* Nothing to do right now. If we ever switch to using the output
* buffer in the reordering process, we'd need to next_glyph() here. */
}
static void
initial_reordering_syllable (const hb_ot_shape_plan_t *plan,
hb_face_t *face,
hb_buffer_t *buffer,
unsigned int start, unsigned int end)
{
syllable_type_t syllable_type = (syllable_type_t) (buffer->info[start].syllable() & 0x0F);
switch (syllable_type) {
case consonant_syllable: initial_reordering_consonant_syllable (plan, face, buffer, start, end); return;
case broken_cluster: initial_reordering_broken_cluster (plan, face, buffer, start, end); return;
case non_sea_cluster: initial_reordering_non_sea_cluster (plan, face, buffer, start, end); return;
}
}
static inline void
insert_dotted_circles (const hb_ot_shape_plan_t *plan HB_UNUSED,
hb_font_t *font,
hb_buffer_t *buffer)
{
/* Note: This loop is extra overhead, but should not be measurable. */
bool has_broken_syllables = false;
unsigned int count = buffer->len;
for (unsigned int i = 0; i < count; i++)
if ((buffer->info[i].syllable() & 0x0F) == broken_cluster) {
has_broken_syllables = true;
break;
}
if (likely (!has_broken_syllables))
return;
hb_codepoint_t dottedcircle_glyph;
if (!font->get_glyph (0x25CC, 0, &dottedcircle_glyph))
return;
hb_glyph_info_t dottedcircle = {0};
dottedcircle.codepoint = 0x25CC;
set_sea_properties (dottedcircle);
dottedcircle.codepoint = dottedcircle_glyph;
buffer->clear_output ();
buffer->idx = 0;
unsigned int last_syllable = 0;
while (buffer->idx < buffer->len)
{
unsigned int syllable = buffer->cur().syllable();
syllable_type_t syllable_type = (syllable_type_t) (syllable & 0x0F);
if (unlikely (last_syllable != syllable && syllable_type == broken_cluster))
{
last_syllable = syllable;
hb_glyph_info_t info = dottedcircle;
info.cluster = buffer->cur().cluster;
info.mask = buffer->cur().mask;
info.syllable() = buffer->cur().syllable();
buffer->output_info (info);
}
else
buffer->next_glyph ();
}
buffer->swap_buffers ();
}
static void
initial_reordering (const hb_ot_shape_plan_t *plan,
hb_font_t *font,
hb_buffer_t *buffer)
{
insert_dotted_circles (plan, font, buffer);
hb_glyph_info_t *info = buffer->info;
unsigned int count = buffer->len;
if (unlikely (!count)) return;
unsigned int last = 0;
unsigned int last_syllable = info[0].syllable();
for (unsigned int i = 1; i < count; i++)
if (last_syllable != info[i].syllable()) {
initial_reordering_syllable (plan, font->face, buffer, last, i);
last = i;
last_syllable = info[last].syllable();
}
initial_reordering_syllable (plan, font->face, buffer, last, count);
}
static void
final_reordering (const hb_ot_shape_plan_t *plan,
hb_font_t *font HB_UNUSED,
hb_buffer_t *buffer)
{
hb_glyph_info_t *info = buffer->info;
unsigned int count = buffer->len;
/* Zero syllables now... */
for (unsigned int i = 0; i < count; i++)
info[i].syllable() = 0;
HB_BUFFER_DEALLOCATE_VAR (buffer, sea_category);
HB_BUFFER_DEALLOCATE_VAR (buffer, sea_position);
}
static hb_ot_shape_normalization_mode_t
normalization_preference_sea (const hb_segment_properties_t *props HB_UNUSED)
{
return HB_OT_SHAPE_NORMALIZATION_MODE_COMPOSED_DIACRITICS_NO_SHORT_CIRCUIT;
}
const hb_ot_complex_shaper_t _hb_ot_complex_shaper_sea =
{
"sea",
collect_features_sea,
override_features_sea,
NULL, /* data_create */
NULL, /* data_destroy */
NULL, /* preprocess_text */
normalization_preference_sea,
NULL, /* decompose */
NULL, /* compose */
setup_masks_sea,
HB_OT_SHAPE_ZERO_WIDTH_MARKS_BY_GDEF,
false, /* fallback_position */
};

1
test/shaping/texts/in-tree/MANIFEST
View File

@ -4,4 +4,5 @@ shaper-default
shaper-hangul
shaper-indic
shaper-myanmar
shaper-sea
shaper-thai

1
test/shaping/texts/in-tree/shaper-indic/south-east-asian/MANIFEST
View File

@ -1,2 +1 @@
script-khmer
script-new-tai-lue

1
test/shaping/texts/in-tree/shaper-myanmar/MANIFEST Normal file
View File

@ -0,0 +1 @@
script-myanmar

3
test/shaping/texts/in-tree/shaper-sea/MANIFEST Normal file
View File

@ -0,0 +1,3 @@
script-cham
script-new-tai-lue
script-tai-tham

0
test/shaping/texts/in-tree/shaper-indic/south-east-asian/script-new-tai-lue/MANIFEST → test/shaping/texts/in-tree/shaper-sea/script-cham/MANIFEST
View File

0
test/shaping/texts/in-tree/shaper-indic/south-east-asian/script-new-tai-lue/misc/MANIFEST → test/shaping/texts/in-tree/shaper-sea/script-cham/misc/MANIFEST
View File

1
test/shaping/texts/in-tree/shaper-sea/script-cham/misc/misc.txt Normal file
View File

@ -0,0 +1 @@
ꩀꨴ

1
test/shaping/texts/in-tree/shaper-sea/script-new-tai-lue/MANIFEST Normal file
View File

@ -0,0 +1 @@
misc

1
test/shaping/texts/in-tree/shaper-sea/script-new-tai-lue/misc/MANIFEST Normal file
View File

@ -0,0 +1 @@
misc.txt

0
test/shaping/texts/in-tree/shaper-indic/south-east-asian/script-new-tai-lue/misc/misc.txt → test/shaping/texts/in-tree/shaper-sea/script-new-tai-lue/misc/misc.txt
View File

1
test/shaping/texts/in-tree/shaper-sea/script-tai-tham/MANIFEST Normal file
View File

@ -0,0 +1 @@
misc

1
test/shaping/texts/in-tree/shaper-sea/script-tai-tham/misc/MANIFEST Normal file
View File

@ -0,0 +1 @@
misc.txt

2
test/shaping/texts/in-tree/shaper-sea/script-tai-tham/misc/misc.txt Normal file
View File

@ -0,0 +1,2 @@
ᨠ᩠ᨠᩮᩕ
ᩋᩫᨶ᩠ᨲᩕᩣ᩠ᨿ