/*
- * Copyright (C) 2009 Red Hat, Inc.
+ * Copyright © 2009 Red Hat, Inc.
+ * 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-glib.h"
HB_BEGIN_DECLS
+#if !GLIB_CHECK_VERSION(2,29,14)
+static const hb_script_t
+glib_script_to_script[] =
+{
+ HB_SCRIPT_COMMON,
+ HB_SCRIPT_INHERITED,
+ HB_SCRIPT_ARABIC,
+ HB_SCRIPT_ARMENIAN,
+ HB_SCRIPT_BENGALI,
+ HB_SCRIPT_BOPOMOFO,
+ HB_SCRIPT_CHEROKEE,
+ HB_SCRIPT_COPTIC,
+ HB_SCRIPT_CYRILLIC,
+ HB_SCRIPT_DESERET,
+ HB_SCRIPT_DEVANAGARI,
+ HB_SCRIPT_ETHIOPIC,
+ HB_SCRIPT_GEORGIAN,
+ HB_SCRIPT_GOTHIC,
+ HB_SCRIPT_GREEK,
+ HB_SCRIPT_GUJARATI,
+ HB_SCRIPT_GURMUKHI,
+ HB_SCRIPT_HAN,
+ HB_SCRIPT_HANGUL,
+ HB_SCRIPT_HEBREW,
+ HB_SCRIPT_HIRAGANA,
+ HB_SCRIPT_KANNADA,
+ HB_SCRIPT_KATAKANA,
+ HB_SCRIPT_KHMER,
+ HB_SCRIPT_LAO,
+ HB_SCRIPT_LATIN,
+ HB_SCRIPT_MALAYALAM,
+ HB_SCRIPT_MONGOLIAN,
+ HB_SCRIPT_MYANMAR,
+ HB_SCRIPT_OGHAM,
+ HB_SCRIPT_OLD_ITALIC,
+ HB_SCRIPT_ORIYA,
+ HB_SCRIPT_RUNIC,
+ HB_SCRIPT_SINHALA,
+ HB_SCRIPT_SYRIAC,
+ HB_SCRIPT_TAMIL,
+ HB_SCRIPT_TELUGU,
+ HB_SCRIPT_THAANA,
+ HB_SCRIPT_THAI,
+ HB_SCRIPT_TIBETAN,
+ HB_SCRIPT_CANADIAN_ABORIGINAL,
+ HB_SCRIPT_YI,
+ HB_SCRIPT_TAGALOG,
+ HB_SCRIPT_HANUNOO,
+ HB_SCRIPT_BUHID,
+ HB_SCRIPT_TAGBANWA,
+
+ /* Unicode-4.0 additions */
+ HB_SCRIPT_BRAILLE,
+ HB_SCRIPT_CYPRIOT,
+ HB_SCRIPT_LIMBU,
+ HB_SCRIPT_OSMANYA,
+ HB_SCRIPT_SHAVIAN,
+ HB_SCRIPT_LINEAR_B,
+ HB_SCRIPT_TAI_LE,
+ HB_SCRIPT_UGARITIC,
+
+ /* Unicode-4.1 additions */
+ HB_SCRIPT_NEW_TAI_LUE,
+ HB_SCRIPT_BUGINESE,
+ HB_SCRIPT_GLAGOLITIC,
+ HB_SCRIPT_TIFINAGH,
+ HB_SCRIPT_SYLOTI_NAGRI,
+ HB_SCRIPT_OLD_PERSIAN,
+ HB_SCRIPT_KHAROSHTHI,
+
+ /* Unicode-5.0 additions */
+ HB_SCRIPT_UNKNOWN,
+ HB_SCRIPT_BALINESE,
+ HB_SCRIPT_CUNEIFORM,
+ HB_SCRIPT_PHOENICIAN,
+ HB_SCRIPT_PHAGS_PA,
+ HB_SCRIPT_NKO,
+
+ /* Unicode-5.1 additions */
+ HB_SCRIPT_KAYAH_LI,
+ HB_SCRIPT_LEPCHA,
+ HB_SCRIPT_REJANG,
+ HB_SCRIPT_SUNDANESE,
+ HB_SCRIPT_SAURASHTRA,
+ HB_SCRIPT_CHAM,
+ HB_SCRIPT_OL_CHIKI,
+ HB_SCRIPT_VAI,
+ HB_SCRIPT_CARIAN,
+ HB_SCRIPT_LYCIAN,
+ HB_SCRIPT_LYDIAN,
+
+ /* Unicode-5.2 additions */
+ HB_SCRIPT_AVESTAN,
+ HB_SCRIPT_BAMUM,
+ HB_SCRIPT_EGYPTIAN_HIEROGLYPHS,
+ HB_SCRIPT_IMPERIAL_ARAMAIC,
+ HB_SCRIPT_INSCRIPTIONAL_PAHLAVI,
+ HB_SCRIPT_INSCRIPTIONAL_PARTHIAN,
+ HB_SCRIPT_JAVANESE,
+ HB_SCRIPT_KAITHI,
+ HB_SCRIPT_TAI_THAM,
+ HB_SCRIPT_LISU,
+ HB_SCRIPT_MEETEI_MAYEK,
+ HB_SCRIPT_OLD_SOUTH_ARABIAN,
+ HB_SCRIPT_OLD_TURKIC,
+ HB_SCRIPT_SAMARITAN,
+ HB_SCRIPT_TAI_VIET,
+
+ /* Unicode-6.0 additions */
+ HB_SCRIPT_BATAK,
+ HB_SCRIPT_BRAHMI,
+ HB_SCRIPT_MANDAIC
+};
+#endif
+
+hb_script_t
+hb_glib_script_to_script (GUnicodeScript script)
+{
+#if GLIB_CHECK_VERSION(2,29,14)
+ return (hb_script_t) g_unicode_script_to_iso15924 (script);
+#else
+ if (likely ((unsigned int) script < ARRAY_LENGTH (glib_script_to_script)))
+ return glib_script_to_script[script];
+
+ if (unlikely (script == G_UNICODE_SCRIPT_INVALID_CODE))
+ return HB_SCRIPT_INVALID;
+
+ return HB_SCRIPT_UNKNOWN;
+#endif
+}
+
+GUnicodeScript
+hb_glib_script_from_script (hb_script_t script)
+{
+#if GLIB_CHECK_VERSION(2,29,14)
+ return g_unicode_script_from_iso15924 (script);
+#else
+ unsigned int count = ARRAY_LENGTH (glib_script_to_script);
+ for (unsigned int i = 0; i < count; i++)
+ if (glib_script_to_script[i] == script)
+ return (GUnicodeScript) i;
+
+ if (unlikely (script == HB_SCRIPT_INVALID))
+ return G_UNICODE_SCRIPT_INVALID_CODE;
+
+ return G_UNICODE_SCRIPT_UNKNOWN;
+#endif
+}
+
static unsigned int
-hb_glib_get_combining_class (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_glib_unicode_combining_class (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
return g_unichar_combining_class (unicode);
}
static unsigned int
-hb_glib_get_eastasian_width (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_glib_unicode_eastasian_width (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
return g_unichar_iswide (unicode) ? 2 : 1;
}
static hb_unicode_general_category_t
-hb_glib_get_general_category (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_glib_unicode_general_category (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
/* hb_unicode_general_category_t and GUnicodeType are identical */
}
static hb_codepoint_t
-hb_glib_get_mirroring (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_glib_unicode_mirroring (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
g_unichar_get_mirror_char (unicode, &unicode);
return unicode;
}
static hb_script_t
-hb_glib_get_script (hb_unicode_funcs_t *ufuncs,
- hb_codepoint_t unicode,
- void *user_data)
+hb_glib_unicode_script (hb_unicode_funcs_t *ufuncs HB_UNUSED,
+ hb_codepoint_t unicode,
+ void *user_data HB_UNUSED)
{
- GUnicodeScript script = g_unichar_get_script (unicode);
- switch (script)
- {
-#define MATCH_SCRIPT(C) case G_UNICODE_SCRIPT_##C: return HB_SCRIPT_##C
-#define MATCH_SCRIPT2(C1, C2) case G_UNICODE_SCRIPT_##C1: return HB_SCRIPT_##C2
-
- 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);
+ return hb_glib_script_to_script (g_unichar_get_script (unicode));
+}
- /* 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);
+static hb_bool_t
+hb_glib_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 GLIB_CHECK_VERSION(2,29,12)
+ return g_unichar_compose (a, b, ab);
+#endif
- /* 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);
+ /* We don't ifdef-out the fallback code such that compiler always
+ * sees it and makes sure it's compilable. */
- /* Unicode-5.0 additions */
- MATCH_SCRIPT (UNKNOWN);
- MATCH_SCRIPT (BALINESE);
- MATCH_SCRIPT (CUNEIFORM);
- MATCH_SCRIPT (PHOENICIAN);
- MATCH_SCRIPT (PHAGS_PA);
- MATCH_SCRIPT (NKO);
+ if (!a || !b)
+ return FALSE;
- /* 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);
+ gchar utf8[12];
+ gchar *normalized;
+ gint len;
+ hb_bool_t ret;
- /* Unicode-5.2 additions */
-#if GLIB_CHECK_VERSION(2,26,0)
- MATCH_SCRIPT (AVESTAN);
- MATCH_SCRIPT (BAMUM);
- MATCH_SCRIPT (EGYPTIAN_HIEROGLYPHS);
- MATCH_SCRIPT (IMPERIAL_ARAMAIC);
- MATCH_SCRIPT (INSCRIPTIONAL_PAHLAVI);
- MATCH_SCRIPT (INSCRIPTIONAL_PARTHIAN);
- MATCH_SCRIPT (JAVANESE);
- MATCH_SCRIPT (KAITHI);
- MATCH_SCRIPT (TAI_THAM);
- MATCH_SCRIPT (LISU);
- MATCH_SCRIPT (MEETEI_MAYEK);
- MATCH_SCRIPT (OLD_SOUTH_ARABIAN);
-#if GLIB_CHECK_VERSION(2,27,92)
- MATCH_SCRIPT (OLD_TURKIC);
-#else
- MATCH_SCRIPT2(OLD_TURKISH, OLD_TURKIC);
-#endif
- MATCH_SCRIPT (SAMARITAN);
- MATCH_SCRIPT (TAI_VIET);
-#endif
+ len = g_unichar_to_utf8 (a, utf8);
+ len += g_unichar_to_utf8 (b, utf8 + len);
+ normalized = g_utf8_normalize (utf8, len, G_NORMALIZE_NFC);
- /* Unicode-6.0 additions */
-#if GLIB_CHECK_VERSION(2,28,0)
- MATCH_SCRIPT (BATAK);
- MATCH_SCRIPT (BRAHMI);
- MATCH_SCRIPT (MANDAIC);
+ len = g_utf8_strlen (normalized, -1);
+ if (len == 1) {
+ *ab = g_utf8_get_char (normalized);
+ ret = TRUE;
+ } else {
+ ret = FALSE;
+ }
+
+ g_free (normalized);
+ return ret;
+}
+
+static hb_bool_t
+hb_glib_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)
+{
+#if GLIB_CHECK_VERSION(2,29,12)
+ return g_unichar_decompose (ab, a, b);
#endif
-#undef MATCH_SCRIPT
-#undef MATCH_SCRIPT2
+ /* We don't ifdef-out the fallback code such that compiler always
+ * sees it and makes sure it's compilable. */
+
+ gchar utf8[6];
+ gchar *normalized;
+ gint len;
+ hb_bool_t ret;
+
+ len = g_unichar_to_utf8 (ab, utf8);
+ normalized = g_utf8_normalize (utf8, len, G_NORMALIZE_NFD);
+
+ len = g_utf8_strlen (normalized, -1);
+ if (len == 1) {
+ *a = g_utf8_get_char (normalized);
+ *b = 0;
+ ret = *a != ab;
+ } else if (len == 2) {
+ *a = g_utf8_get_char (normalized);
+ *b = g_utf8_get_char (g_utf8_next_char (normalized));
+ /* 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 :-(. */
+ gchar *recomposed = g_utf8_normalize (normalized, -1, G_NORMALIZE_NFC);
+ hb_codepoint_t c = g_utf8_get_char (recomposed);
+ if (c != ab && c != *a) {
+ *a = c;
+ *b = 0;
+ }
+ g_free (recomposed);
+ ret = TRUE;
+ } else {
+ /* If decomposed to more than two characters, take the last one,
+ * and recompose the rest to get the first component. */
+ gchar *end = g_utf8_offset_to_pointer (normalized, len - 1);
+ gchar *recomposed;
+ *b = g_utf8_get_char (end);
+ recomposed = g_utf8_normalize (normalized, end - normalized, G_NORMALIZE_NFC);
+ /* We expect that recomposed has exactly one character now. */
+ *a = g_utf8_get_char (recomposed);
+ g_free (recomposed);
+ ret = TRUE;
}
- return HB_SCRIPT_UNKNOWN;
+ g_free (normalized);
+ return ret;
}
-static hb_unicode_funcs_t glib_ufuncs = {
- HB_REFERENCE_COUNT_INVALID, /* ref_count */
+
+extern HB_INTERNAL hb_unicode_funcs_t _hb_unicode_funcs_glib;
+hb_unicode_funcs_t _hb_glib_unicode_funcs = {
+ HB_OBJECT_HEADER_STATIC,
+
NULL, /* parent */
TRUE, /* immutable */
{
- hb_glib_get_combining_class,
- hb_glib_get_eastasian_width,
- hb_glib_get_general_category,
- hb_glib_get_mirroring,
- hb_glib_get_script
+#define HB_UNICODE_FUNC_IMPLEMENT(name) hb_glib_unicode_##name,
+ HB_UNICODE_FUNCS_IMPLEMENT_CALLBACKS
+#undef HB_UNICODE_FUNC_IMPLEMENT
}
};
hb_unicode_funcs_t *
hb_glib_get_unicode_funcs (void)
{
- return &glib_ufuncs;
+ return &_hb_glib_unicode_funcs;
}