indic-ot-class-tables.c

GTK+-2.0源码之pango-1.15.6.tar.gz

 * FIXME: eyelash RA only for Devanagari??
 */
#define DEVA_SCRIPT_FLAGS (SF_EYELASH_RA | SF_NO_POST_BASE_LIMIT)
#define BENG_SCRIPT_FLAGS (SF_REPH_AFTER_BELOW | SF_NO_POST_BASE_LIMIT)
#define GURU_SCRIPT_FLAGS (SF_NO_POST_BASE_LIMIT)
#define GUJR_SCRIPT_FLAGS (SF_NO_POST_BASE_LIMIT)
#define ORYA_SCRIPT_FLAGS (SF_REPH_AFTER_BELOW | SF_NO_POST_BASE_LIMIT)
#define TAML_SCRIPT_FLAGS (SF_MPRE_FIXUP | SF_NO_POST_BASE_LIMIT)
#define TELU_SCRIPT_FLAGS (SF_MATRAS_AFTER_BASE | 3)
#define KNDA_SCRIPT_FLAGS (SF_MATRAS_AFTER_BASE | 3)
#define MLYM_SCRIPT_FLAGS (SF_MPRE_FIXUP | SF_NO_POST_BASE_LIMIT | SF_PROCESS_ZWJ)
#define SINH_SCRIPT_FLAGS (SF_MPRE_FIXUP | SF_NO_POST_BASE_LIMIT | SF_PROCESS_ZWJ)

/*
 * Indic Class Tables
 */
/* Add a little macro to compute lastChar based on size of the charClasses * table */
#define INDIC_OT_CLASS_TABLE_DEFINE(name, firstChar, worstCaseExpansion, scriptFlags, charClasses, splitMatraTable) \
  const IndicOTClassTable name = {firstChar, firstChar + G_N_ELEMENTS (charClasses) - 1, \
				  worstCaseExpansion, scriptFlags, charClasses, splitMatraTable}
INDIC_OT_CLASS_TABLE_DEFINE (deva_class_table, 0x0900, 2, DEVA_SCRIPT_FLAGS, devaCharClasses, NULL);
INDIC_OT_CLASS_TABLE_DEFINE (beng_class_table, 0x0980, 3, BENG_SCRIPT_FLAGS, bengCharClasses, bengSplitTable);
INDIC_OT_CLASS_TABLE_DEFINE (guru_class_table, 0x0A00, 2, GURU_SCRIPT_FLAGS, guruCharClasses, NULL);
INDIC_OT_CLASS_TABLE_DEFINE (gujr_class_table, 0x0A80, 2, GUJR_SCRIPT_FLAGS, gujrCharClasses, NULL);
INDIC_OT_CLASS_TABLE_DEFINE (orya_class_table, 0x0B00, 3, ORYA_SCRIPT_FLAGS, oryaCharClasses, oryaSplitTable);
INDIC_OT_CLASS_TABLE_DEFINE (taml_class_table, 0x0B80, 3, TAML_SCRIPT_FLAGS, tamlCharClasses, tamlSplitTable);
INDIC_OT_CLASS_TABLE_DEFINE (telu_class_table, 0x0C00, 3, TELU_SCRIPT_FLAGS, teluCharClasses, teluSplitTable);
INDIC_OT_CLASS_TABLE_DEFINE (knda_class_table, 0x0C80, 4, KNDA_SCRIPT_FLAGS, kndaCharClasses, kndaSplitTable);
INDIC_OT_CLASS_TABLE_DEFINE (mlym_class_table, 0x0D00, 3, MLYM_SCRIPT_FLAGS, mlymCharClasses, mlymSplitTable);
INDIC_OT_CLASS_TABLE_DEFINE (sinh_class_table, 0x0D80, 4, SINH_SCRIPT_FLAGS, sinhCharClasses, sinhSplitTable);

const IndicOTSplitMatra *indic_ot_get_split_matra(const IndicOTClassTable *class_table, IndicOTCharClass char_class)
{
    gint32 index = (char_class & CF_INDEX_MASK) >> CF_INDEX_SHIFT;

    return &class_table->splitMatraTable[index - 1];
}

gboolean indic_ot_is_vm_above(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_VM_ABOVE(char_class);
}

gboolean indic_ot_is_vm_post(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_VM_POST(char_class);
}

gboolean indic_ot_is_consonant(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_CONSONANT(char_class);
}

gboolean indic_ot_is_reph(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_REPH(char_class);
}

gboolean indic_ot_is_virama(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return (IS_VIRAMA(char_class) || IS_AL_LAKUNA(char_class));
}

gboolean indic_ot_is_al_lakuna(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_AL_LAKUNA(char_class);
}

gboolean indic_ot_is_nukta(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_NUKTA(char_class);
}

gboolean indic_ot_is_vattu(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_VATTU(char_class);
}

gboolean indic_ot_is_matra(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_MATRA(char_class);
}

gboolean indic_ot_is_split_matra(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_SPLIT_MATRA(char_class);
}

gboolean indic_ot_is_m_pre(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_M_PRE(char_class);
}

gboolean indic_ot_is_m_below(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_M_BELOW(char_class);
}

gboolean indic_ot_is_m_above(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_M_ABOVE(char_class);
}

gboolean indic_ot_is_m_post(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_M_POST(char_class);
}

gboolean indic_ot_is_length_mark(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return IS_LENGTH_MARK(char_class);
}

gboolean indic_ot_has_post_or_below_base_form(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return HAS_POST_OR_BELOW_BASE_FORM(char_class);
}

gboolean indic_ot_has_post_base_form(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return HAS_POST_BASE_FORM(char_class);
}

gboolean indic_ot_has_below_base_form(const IndicOTClassTable *class_table, gunichar ch)
{
  IndicOTCharClass char_class = indic_ot_get_char_class(class_table, ch);

  return HAS_BELOW_BASE_FORM(char_class);
}

IndicOTCharClass indic_ot_get_char_class(const IndicOTClassTable *class_table, gunichar ch)
{
    if (ch == C_SIGN_ZWJ) {
	return CF_CONSONANT | CC_ZERO_WIDTH_MARK;
    }

    if (ch == C_SIGN_ZWNJ) {
	return CC_ZERO_WIDTH_MARK;
    }

    if (ch < class_table->firstChar || ch > class_table->lastChar) {
	return CC_RESERVED;
    }

    return class_table->charClasses[ch - class_table->firstChar];
}

static const gint8 stateTable[][CC_COUNT] =
{
/*   xx  ma  mp  iv  ct  cn  nu  dv  vr  zw  al */
    { 1,  1,  1,  5,  3,  2,  1,  1,  1,  1,  1},
    {-1, -1, -1, -1, -1, -1, -1, -1, -1, -1, -1},
    {-1,  6,  1, -1, -1, -1, -1,  5,  4, -1, -1},
    {-1,  6,  1, -1, -1, -1,  2,  5,  4, 10,  9},
    {-1, -1, -1, -1,  3,  2, -1, -1, -1,  8, -1},
    {-1,  6,  1, -1, -1, -1, -1, -1, -1, -1, -1},
    {-1,  7,  1, -1, -1, -1, -1, -1, -1, -1, -1},
    {-1, -1,  1, -1, -1, -1, -1, -1, -1, -1, -1},
    {-1, -1, -1, -1,  3,  2, -1, -1, -1, -1, -1},
    {-1, -1, -1, -1, -1, -1, -1, -1, -1,  8, -1},
    {-1, -1, -1, -1, -1, -1, -1, -1,  8, -1,  8}

};

glong indic_ot_find_syllable(const IndicOTClassTable *class_table, const gunichar *chars, glong prev, glong char_count)
{
    glong cursor = prev;
    gint8 state = 0;

    while (cursor < char_count) {
	IndicOTCharClass char_class = indic_ot_get_char_class(class_table, chars[cursor]);

	state = stateTable[state][char_class & CF_CLASS_MASK];

	/*for the components of split matra*/
	if ((char_count >= cursor + 3) &&
	    (chars[cursor] == 0x0DD9 && chars[cursor + 1] == 0x0DCF && chars[cursor + 2] == 0x0DCA)) {  /*for 3 split matra of Sinhala*/
	    return cursor + 3;
	}
	else if ((char_count >= cursor + 3) &&
		 (chars[cursor] == 0x0CC6 && chars[cursor + 1] == 0x0CC2 && chars[cursor + 2] == 0x0CD5)) {  /*for 3 split matra of Kannada*/
	    return cursor + 3;
	}
	/*for 2 split matra*/
	else if (char_count >= cursor + 2) {
		/*for Bengali*/
	    if ((chars[cursor] == 0x09C7 && chars[cursor + 1] == 0x09BE) ||
		(chars[cursor] == 0x09C7 && chars[cursor + 1] == 0x09D7) ||
		/*for Oriya*/
		(chars[cursor] == 0x0B47 && chars[cursor + 1] == 0x0B3E) ||
		(chars[cursor] == 0x0B47 && chars[cursor + 1] == 0x0B56) ||
		(chars[cursor] == 0x0B47 && chars[cursor + 1] == 0x0B57) ||
		/*for Tamil*/
		(chars[cursor] == 0x0BC6 && chars[cursor + 1] == 0x0BBE) ||
		(chars[cursor] == 0x0BC6 && chars[cursor + 1] == 0x0BD7) ||
		(chars[cursor] == 0x0BC7 && chars[cursor + 1] == 0x0BBE) ||
		/*for Malayalam*/
		(chars[cursor] == 0x0D46 && chars[cursor + 1] == 0x0D3E) ||
		(chars[cursor] == 0x0D46 && chars[cursor + 1] == 0x0D57) ||
		(chars[cursor] == 0x0D47 && chars[cursor + 1] == 0x0D3E) ||
		/*for Sinhala*/
		(chars[cursor] == 0x0DD9 && chars[cursor + 1] == 0x0DCA) ||
		(chars[cursor] == 0x0DD9 && chars[cursor + 1] == 0x0DCF) ||
		(chars[cursor] == 0x0DD9 && chars[cursor + 1] == 0x0DDF) ||
		(chars[cursor] == 0x0DDC && chars[cursor + 1] == 0x0DCA) ||
		/*for Telugu*/
		(chars[cursor] == 0x0C46 && chars[cursor + 1] == 0x0C56) ||
		/*for Kannada*/
		(chars[cursor] == 0x0CBF && chars[cursor + 1] == 0x0CD5) ||
		(chars[cursor] == 0x0CC6 && chars[cursor + 1] == 0x0CD5) ||
		(chars[cursor] == 0x0CC6 && chars[cursor + 1] == 0x0CD6) ||
		(chars[cursor] == 0x0CC6 && chars[cursor + 1] == 0x0CC2) ||
		(chars[cursor] == 0x0CCA && chars[cursor + 1] == 0x0CD5))
		    return cursor + 2;
	}

	if (state < 0) {
	    break;
	}

	cursor += 1;
    }

    return cursor;
}