/*
- * Copyright (C) 2009 Red Hat, Inc.
- * Copyright (C) 2009 Keith Stribley
+ * Copyright © 2009 Red Hat, Inc.
+ * Copyright © 2009 Keith Stribley
+ * Copyright © 2011 Google, Inc.
*
* This is part of HarfBuzz, a text shaping library.
*
* PROVIDE MAINTENANCE, SUPPORT, UPDATES, ENHANCEMENTS, OR MODIFICATIONS.
*
* Red Hat Author(s): Behdad Esfahbod
+ * Google Author(s): Behdad Esfahbod
*/
-#include "hb-private.h"
+#include "hb-private.hh"
#include "hb-icu.h"
#include <unicode/uversion.h>
#include <unicode/uchar.h>
+#include <unicode/unorm.h>
+#include <unicode/ustring.h>
-HB_BEGIN_DECLS
hb_script_t
hb_icu_script_to_script (UScriptCode script)
{
- return hb_script_from_string (uscript_getShortName (script));
+ if (unlikely (script == USCRIPT_INVALID_CODE))
+ return HB_SCRIPT_INVALID;
+
+ return hb_script_from_string (uscript_getShortName (script), -1);
}
UScriptCode
hb_icu_script_from_script (hb_script_t script)
{
- switch ((int) script)
- {
-#define CHECK_ICU_VERSION(major, minor) \
- U_ICU_VERSION_MAJOR_NUM > (major) || (U_ICU_VERSION_MAJOR_NUM == (major) && U_ICU_VERSION_MINOR_NUM >= (minor))
-#define MATCH_SCRIPT(C) case HB_SCRIPT_##C: return USCRIPT_##C
-#define MATCH_SCRIPT2(C1, C2) case HB_SCRIPT_##C2: return USCRIPT_##C1
-
- MATCH_SCRIPT2(INVALID_CODE, INVALID);
-
- MATCH_SCRIPT (COMMON);
- MATCH_SCRIPT (INHERITED);
- MATCH_SCRIPT (ARABIC);
- MATCH_SCRIPT (ARMENIAN);
- MATCH_SCRIPT (BENGALI);
- MATCH_SCRIPT (BOPOMOFO);
- MATCH_SCRIPT (CHEROKEE);
- MATCH_SCRIPT (COPTIC);
- MATCH_SCRIPT (CYRILLIC);
- MATCH_SCRIPT (DESERET);
- MATCH_SCRIPT (DEVANAGARI);
- MATCH_SCRIPT (ETHIOPIC);
- MATCH_SCRIPT (GEORGIAN);
- MATCH_SCRIPT (GOTHIC);
- MATCH_SCRIPT (GREEK);
- MATCH_SCRIPT (GUJARATI);
- MATCH_SCRIPT (GURMUKHI);
- MATCH_SCRIPT (HAN);
- MATCH_SCRIPT (HANGUL);
- MATCH_SCRIPT (HEBREW);
- MATCH_SCRIPT (HIRAGANA);
- MATCH_SCRIPT (KANNADA);
- MATCH_SCRIPT (KATAKANA);
- MATCH_SCRIPT (KHMER);
- MATCH_SCRIPT (LAO);
- MATCH_SCRIPT (LATIN);
- MATCH_SCRIPT (MALAYALAM);
- MATCH_SCRIPT (MONGOLIAN);
- MATCH_SCRIPT (MYANMAR);
- MATCH_SCRIPT (OGHAM);
- MATCH_SCRIPT (OLD_ITALIC);
- MATCH_SCRIPT (ORIYA);
- MATCH_SCRIPT (RUNIC);
- MATCH_SCRIPT (SINHALA);
- MATCH_SCRIPT (SYRIAC);
- MATCH_SCRIPT (TAMIL);
- MATCH_SCRIPT (TELUGU);
- MATCH_SCRIPT (THAANA);
- MATCH_SCRIPT (THAI);
- MATCH_SCRIPT (TIBETAN);
- MATCH_SCRIPT (CANADIAN_ABORIGINAL);
- MATCH_SCRIPT (YI);
- MATCH_SCRIPT (TAGALOG);
- MATCH_SCRIPT (HANUNOO);
- MATCH_SCRIPT (BUHID);
- MATCH_SCRIPT (TAGBANWA);
-
- /* Unicode-4.0 additions */
- MATCH_SCRIPT (BRAILLE);
- MATCH_SCRIPT (CYPRIOT);
- MATCH_SCRIPT (LIMBU);
- MATCH_SCRIPT (OSMANYA);
- MATCH_SCRIPT (SHAVIAN);
- MATCH_SCRIPT (LINEAR_B);
- MATCH_SCRIPT (TAI_LE);
- MATCH_SCRIPT (UGARITIC);
-
- /* Unicode-4.1 additions */
- MATCH_SCRIPT (NEW_TAI_LUE);
- MATCH_SCRIPT (BUGINESE);
- MATCH_SCRIPT (GLAGOLITIC);
- MATCH_SCRIPT (TIFINAGH);
- MATCH_SCRIPT (SYLOTI_NAGRI);
- MATCH_SCRIPT (OLD_PERSIAN);
- MATCH_SCRIPT (KHAROSHTHI);
-
- /* Unicode-5.0 additions */
- MATCH_SCRIPT (UNKNOWN);
- MATCH_SCRIPT (BALINESE);
- MATCH_SCRIPT (CUNEIFORM);
- MATCH_SCRIPT (PHOENICIAN);
- MATCH_SCRIPT (PHAGS_PA);
- MATCH_SCRIPT (NKO);
-
- /* Unicode-5.1 additions */
- MATCH_SCRIPT (KAYAH_LI);
- MATCH_SCRIPT (LEPCHA);
- MATCH_SCRIPT (REJANG);
- MATCH_SCRIPT (SUNDANESE);
- MATCH_SCRIPT (SAURASHTRA);
- MATCH_SCRIPT (CHAM);
- MATCH_SCRIPT (OL_CHIKI);
- MATCH_SCRIPT (VAI);
- MATCH_SCRIPT (CARIAN);
- MATCH_SCRIPT (LYCIAN);
- MATCH_SCRIPT (LYDIAN);
-
- /* Unicode-5.2 additions */
- MATCH_SCRIPT (AVESTAN);
-#if CHECK_ICU_VERSION (4, 4)
- MATCH_SCRIPT (BAMUM);
-#endif
- MATCH_SCRIPT (EGYPTIAN_HIEROGLYPHS);
- MATCH_SCRIPT (IMPERIAL_ARAMAIC);
- MATCH_SCRIPT (INSCRIPTIONAL_PAHLAVI);
- MATCH_SCRIPT (INSCRIPTIONAL_PARTHIAN);
- MATCH_SCRIPT (JAVANESE);
- MATCH_SCRIPT (KAITHI);
- MATCH_SCRIPT2(LANNA, TAI_THAM);
-#if CHECK_ICU_VERSION (4, 4)
- MATCH_SCRIPT (LISU);
-#endif
- MATCH_SCRIPT2(MEITEI_MAYEK, MEETEI_MAYEK);
-#if CHECK_ICU_VERSION (4, 4)
- MATCH_SCRIPT (OLD_SOUTH_ARABIAN);
-#endif
- MATCH_SCRIPT2(ORKHON, OLD_TURKIC);
- MATCH_SCRIPT (SAMARITAN);
- MATCH_SCRIPT (TAI_VIET);
-
- /* Unicode-6.0 additions */
- MATCH_SCRIPT (BATAK);
- MATCH_SCRIPT (BRAHMI);
- MATCH_SCRIPT2(MANDAEAN, MANDAIC);
-
-#undef CHECK_ICU_VERSION
-#undef MATCH_SCRIPT
-#undef MATCH_SCRIPT2
- }
+ if (unlikely (script == HB_SCRIPT_INVALID))
+ return USCRIPT_INVALID_CODE;
+
+ for (unsigned int i = 0; i < USCRIPT_CODE_LIMIT; i++)
+ if (unlikely (hb_icu_script_to_script ((UScriptCode) i) == script))
+ return (UScriptCode) i;
return USCRIPT_UNKNOWN;
}
static unsigned int
-hb_icu_get_combining_class (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_icu_unicode_combining_class (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
return u_getCombiningClass (unicode);
}
static unsigned int
-hb_icu_get_eastasian_width (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_icu_unicode_eastasian_width (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
switch (u_getIntPropertyValue(unicode, UCHAR_EAST_ASIAN_WIDTH))
{
}
static hb_unicode_general_category_t
-hb_icu_get_general_category (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_icu_unicode_general_category (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
switch (u_getIntPropertyValue(unicode, UCHAR_GENERAL_CATEGORY))
{
case U_NON_SPACING_MARK: return HB_UNICODE_GENERAL_CATEGORY_NON_SPACING_MARK;
case U_ENCLOSING_MARK: return HB_UNICODE_GENERAL_CATEGORY_ENCLOSING_MARK;
- case U_COMBINING_SPACING_MARK: return HB_UNICODE_GENERAL_CATEGORY_COMBINING_MARK;
+ case U_COMBINING_SPACING_MARK: return HB_UNICODE_GENERAL_CATEGORY_SPACING_MARK;
case U_DECIMAL_DIGIT_NUMBER: return HB_UNICODE_GENERAL_CATEGORY_DECIMAL_NUMBER;
case U_LETTER_NUMBER: return HB_UNICODE_GENERAL_CATEGORY_LETTER_NUMBER;
}
static hb_codepoint_t
-hb_icu_get_mirroring (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_icu_unicode_mirroring (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
return u_charMirror(unicode);
}
static hb_script_t
-hb_icu_get_script (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_icu_unicode_script (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
UErrorCode status = U_ZERO_ERROR;
UScriptCode scriptCode = uscript_getScript(unicode, &status);
+ if (unlikely (U_FAILURE (status)))
+ return HB_SCRIPT_UNKNOWN;
+
return hb_icu_script_to_script (scriptCode);
}
-static hb_unicode_funcs_t icu_ufuncs = {
- HB_REFERENCE_COUNT_INVALID, /* ref_count */
+static hb_bool_t
+hb_icu_unicode_compose (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t a,
+ hb_codepoint_t b,
+ hb_codepoint_t *ab,
+ void *user_data HB_UNUSED)
+{
+ if (!a || !b)
+ return false;
+
+ UChar utf16[4], normalized[5];
+ int len;
+ hb_bool_t ret, err;
+ UErrorCode icu_err;
+
+ len = 0;
+ err = false;
+ U16_APPEND (utf16, len, ARRAY_LENGTH (utf16), a, err);
+ if (err) return false;
+ U16_APPEND (utf16, len, ARRAY_LENGTH (utf16), b, err);
+ if (err) return false;
+
+ icu_err = U_ZERO_ERROR;
+ len = unorm_normalize (utf16, len, UNORM_NFC, 0, normalized, ARRAY_LENGTH (normalized), &icu_err);
+ if (U_FAILURE (icu_err))
+ return false;
+ if (u_countChar32 (normalized, len) == 1) {
+ U16_GET_UNSAFE (normalized, 0, *ab);
+ ret = true;
+ } else {
+ ret = false;
+ }
+
+ return ret;
+}
+
+static hb_bool_t
+hb_icu_unicode_decompose (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t ab,
+ hb_codepoint_t *a,
+ hb_codepoint_t *b,
+ void *user_data HB_UNUSED)
+{
+ UChar utf16[2], normalized[20];
+ int len;
+ hb_bool_t ret, err;
+ UErrorCode icu_err;
+
+ /* This function is a monster! Maybe it wasn't a good idea adding a
+ * pairwise decompose API... */
+ /* Watchout for the dragons. Err, watchout for macros changing len. */
+
+ len = 0;
+ err = false;
+ U16_APPEND (utf16, len, ARRAY_LENGTH (utf16), ab, err);
+ if (err) return false;
+
+ icu_err = U_ZERO_ERROR;
+ len = unorm_normalize (utf16, len, UNORM_NFD, 0, normalized, ARRAY_LENGTH (normalized), &icu_err);
+ if (U_FAILURE (icu_err))
+ return false;
+
+ len = u_countChar32 (normalized, len);
+
+ if (len == 1) {
+ U16_GET_UNSAFE (normalized, 0, *a);
+ *b = 0;
+ ret = *a != ab;
+ } else if (len == 2) {
+ len =0;
+ U16_NEXT_UNSAFE (normalized, len, *a);
+ U16_NEXT_UNSAFE (normalized, len, *b);
+
+ /* Here's the ugly part: if ab decomposes to a single character and
+ * that character decomposes again, we have to detect that and undo
+ * the second part :-(. */
+ UChar recomposed[20];
+ icu_err = U_ZERO_ERROR;
+ unorm_normalize (normalized, len, UNORM_NFC, 0, recomposed, ARRAY_LENGTH (recomposed), &icu_err);
+ if (U_FAILURE (icu_err))
+ return false;
+ hb_codepoint_t c;
+ U16_GET_UNSAFE (recomposed, 0, c);
+ if (c != *a && c != ab) {
+ *a = c;
+ *b = 0;
+ }
+ ret = true;
+ } else {
+ /* If decomposed to more than two characters, take the last one,
+ * and recompose the rest to get the first component. */
+ U16_PREV_UNSAFE (normalized, len, *b);
+ UChar recomposed[20];
+ icu_err = U_ZERO_ERROR;
+ len = unorm_normalize (normalized, len, UNORM_NFC, 0, recomposed, ARRAY_LENGTH (recomposed), &icu_err);
+ if (U_FAILURE (icu_err))
+ return false;
+ /* We expect that recomposed has exactly one character now. */
+ U16_GET_UNSAFE (recomposed, 0, *a);
+ ret = true;
+ }
+
+ return ret;
+}
+
+
+extern HB_INTERNAL const hb_unicode_funcs_t _hb_icu_unicode_funcs;
+const hb_unicode_funcs_t _hb_icu_unicode_funcs = {
+ HB_OBJECT_HEADER_STATIC,
+
NULL, /* parent */
- TRUE, /* immutable */
+ true, /* immutable */
{
- hb_icu_get_combining_class,
- hb_icu_get_eastasian_width,
- hb_icu_get_general_category,
- hb_icu_get_mirroring,
- hb_icu_get_script
+#define HB_UNICODE_FUNC_IMPLEMENT(name) hb_icu_unicode_##name,
+ HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
+#undef HB_UNICODE_FUNC_IMPLEMENT
}
};
hb_unicode_funcs_t *
hb_icu_get_unicode_funcs (void)
{
- return &icu_ufuncs;
+ return const_cast<hb_unicode_funcs_t *> (&_hb_icu_unicode_funcs);
}
-HB_END_DECLS
projects / framework / uifw / harfbuzz.git / blobdiff