/*
* convert to the correct display level of THAI vowels and marks.
*/
static HB_Bool HB_ThaiConvertStringToGlyphIndices (HB_ShaperItem *item)
{
char s[128];
char *cstr = s;
const HB_UChar16 *string = item->string + item->item.pos;
const hb_uint32 len = item->item.length;
unsigned short *logClusters = item->log_clusters;
hb_uint32 i = 0, slen = 0;
if (!init_libthai())
return HB_BasicShape (item);
if (len >= 128)
cstr = (char *)malloc(len*sizeof(char) + 1);
if (!cstr)
return HB_BasicShape (item);
to_tis620(string, len, cstr);
/* Get font type */
static ThaiFontType font_type;
static HB_Font itemFont;
if (itemFont != item->font) {
font_type = getThaiFontType (item);
itemFont = item->font;
}
/* allocate temporary glyphs buffers */
HB_STACKARRAY (HB_UChar16, glyphString, (item->item.length * 2));
while (i < item->item.length) {
struct thcell_t tis_cell;
unsigned char rglyphs[4];
int cell_length;
int lgn = 0;
HB_Bool haveSaraAm = false;
cell_length = th_next_cell ((const unsigned char *)cstr + i, len - i, &tis_cell, true); /* !item->fixedPitch); */
haveSaraAm = (cstr[i + cell_length - 1] == (char)0xd3);
/* set shaper item's log_clusters */
logClusters[i] = slen;
for (int j = 1; j < cell_length; j++) {
logClusters[i + j] = logClusters[i];
}
/* Find Logical Glyphs by font type */
switch (font_type) {
case TIS:
lgn = th_render_cell_tis (tis_cell, rglyphs, sizeof(rglyphs) / sizeof(rglyphs[0]), true);
break;
case WIN:
lgn = th_render_cell_mac (tis_cell, rglyphs, sizeof(rglyphs) / sizeof(rglyphs[0]), true);
break;
case MAC:
lgn = th_render_cell_win (tis_cell, rglyphs, sizeof(rglyphs) / sizeof(rglyphs[0]), true);
break;
}
/* Add glyphs to glyphs string and setting some attributes */
for (int lgi = 0; lgi < lgn; lgi++) {
if ( rglyphs[lgi] == 0xdd/*TH_BLANK_BASE_GLYPH*/ ) {
glyphString[slen++] = C_DOTTED_CIRCLE;
} else if ((unsigned char)cstr[i] == (unsigned char)~0) {
// The only glyphs that should be passed to this function that cannot be mapped to
// tis620 are the ones of type Inherited class. Pass these glyphs untouched.
glyphString[slen++] = string[i];
if (string[i] == 0x200D || string[i] == 0x200C) {
// Check that we do not run out of bounds when setting item->attributes. If we do
// run out of bounds then this function will return false, the necessary amount of
// memory is reallocated, and this function will then be called again.
if (slen <= item->num_glyphs)
item->attributes[slen-1].dontPrint = true; // Hide ZWJ and ZWNJ characters
}
} else {
glyphString[slen++] = (HB_UChar16) thai_get_glyph_index (font_type, rglyphs[lgi]);
}
}
/* Special case to handle U+0E33 (SARA AM, ำ): SARA AM is normally written at the end of a
* word with a base character and an optional top character before it. For example, U+0E0B
* (base), U+0E49 (top), U+0E33 (SARA AM). The sequence should be converted to 4 glyphs:
* base, hilo (the little circle in the top left part of SARA AM, NIKHAHIT), top, then the
* right part of SARA AM (SARA AA).
*
* The painting process finds out the starting glyph and ending glyph of a character
* sequence by checking the logClusters array. In this case, logClusters array should
* ideally be [ 0, 1, 3 ] so that glyphsStart = 0 and glyphsEnd = 3 (slen - 1) to paint out
* all the glyphs generated.
*
* A special case in this special case is when we have no base character. When an isolated
* SARA AM is processed (cell_length = 1), libthai will produce 3 glyphs: dotted circle
* (indicates that the base is empty), NIKHAHIT then SARA AA. If logClusters[0] = 1, it will
* paint from the second glyph in the glyphs array. So in this case logClusters[0] should
* point to the first glyph it produces, aka. the dotted circle. */
if (haveSaraAm) {
logClusters[i + cell_length - 1] = cell_length == 1 ? slen - 3 : slen - 1;
if (tis_cell.top != 0) {
if (cell_length > 1) {
/* set the logClusters[top character] to slen - 2 as it points to the second to
* lastglyph (slen - 2) */
logClusters[i + cell_length - 2] = slen - 2;
}
}
/* check for overflow */
if (logClusters[i + cell_length - 1] > slen)
logClusters[i + cell_length - 1] = 0;
}
i += cell_length;
}
glyphString[slen] = (HB_UChar16) '\0';
/* for check, should reallocate space or not */
HB_Bool spaceOK = (item->num_glyphs >= slen);
/* Convert to Glyph indices */
HB_Bool haveGlyphs = item->font->klass->convertStringToGlyphIndices (
item->font,
glyphString, slen,
item->glyphs, &item->num_glyphs,
item->shaperFlags);
HB_FREE_STACKARRAY (glyphString);
if (len >= 128)
free(cstr);
return (haveGlyphs && spaceOK);
}
/*
* Thai Attributes: computes Word Break, Word Boundary and Char stop for THAI.
*/
static void HB_ThaiAssignAttributes(const HB_UChar16 *string, hb_uint32 len, HB_CharAttributes *attributes)
{
char s[128];
char *cstr = s;
int *break_positions = 0;
int brp[128];
int brp_size = 0;
hb_uint32 numbreaks, i, j, cell_length;
struct thcell_t tis_cell;
if (!init_libthai())
return ;
if (len >= 128)
cstr = (char *)malloc(len*sizeof(char) + 1);
to_tis620(string, len, cstr);
for (i = 0; i < len; ++i) {
attributes[i].wordBreak = FALSE;
attributes[i].wordStart = FALSE;
attributes[i].wordEnd = FALSE;
attributes[i].lineBreak = FALSE;
}
if (len > 128) {
break_positions = (int*) malloc (sizeof(int) * len);
memset (break_positions, 0, sizeof(int) * len);
brp_size = len;
}
else {
break_positions = brp;
brp_size = 128;
}
if (break_positions) {
attributes[0].wordBreak = TRUE;
attributes[0].wordStart = TRUE;
attributes[0].wordEnd = FALSE;
numbreaks = th_brk((const unsigned char *)cstr, break_positions, brp_size);
for (i = 0; i < numbreaks; ++i) {
attributes[break_positions[i]].wordBreak = TRUE;
attributes[break_positions[i]].wordStart = TRUE;
attributes[break_positions[i]].wordEnd = TRUE;
attributes[break_positions[i]].lineBreak = TRUE;
}
if (numbreaks > 0)
attributes[break_positions[numbreaks - 1]].wordStart = FALSE;
if (break_positions != brp)
free(break_positions);
}
/* manage grapheme boundaries */
i = 0;
while (i < len) {
cell_length = th_next_cell((const unsigned char *)cstr + i, len - i, &tis_cell, true);
attributes[i].graphemeBoundary = true;
for (j = 1; j < cell_length; j++)
attributes[i + j].graphemeBoundary = false;
/* Set graphemeBoundary for SARA AM */
if (cstr[i + cell_length - 1] == (char)0xd3)
attributes[i + cell_length - 1].graphemeBoundary = true;
i += cell_length;
}
if (len >= 128)
free(cstr);
}
/*
* convert to the correct display level of THAI vowels and marks.
*/
static HB_Bool HB_ThaiConvertStringToGlyphIndices (HB_ShaperItem *item)
{
char s[128];
char *cstr = s;
const HB_UChar16 *string = item->string + item->item.pos;
const hb_uint32 len = item->item.length;
unsigned short *logClusters = item->log_clusters;
hb_uint32 i = 0, slen = 0;
if (!init_libthai())
return HB_BasicShape (item);
if (len >= 128)
cstr = (char *)malloc(len*sizeof(char) + 1);
if (!cstr)
return HB_BasicShape (item);
to_tis620(string, len, cstr);
/* Get font type */
static ThaiFontType font_type;
static HB_Font itemFont;
if (itemFont != item->font) {
font_type = getThaiFontType (item);
itemFont = item->font;
}
/* allocate temporary glyphs buffers */
HB_STACKARRAY (HB_UChar16, glyphString, (item->item.length * 2));
while (i < item->item.length) {
struct thcell_t tis_cell;
unsigned char rglyphs[4];
int cell_length;
int lgn = 0;
HB_Bool haveSaraAm = false;
cell_length = th_next_cell ((const unsigned char *)cstr + i, len - i, &tis_cell, true); /* !item->fixedPitch); */
haveSaraAm = (cstr[i + cell_length - 1] == (char)0xd3);
/* set shaper item's log_clusters */
logClusters[i] = slen;
for (int j = 1; j < cell_length; j++) {
logClusters[i + j] = logClusters[i];
}
/* Find Logical Glyphs by font type */
switch (font_type) {
case TIS: lgn = th_render_cell_tis (tis_cell, rglyphs, sizeof(rglyphs) / sizeof(rglyphs[0]), true); break;
case WIN: lgn = th_render_cell_mac (tis_cell, rglyphs, sizeof(rglyphs) / sizeof(rglyphs[0]), true); break;
case MAC: lgn = th_render_cell_win (tis_cell, rglyphs, sizeof(rglyphs) / sizeof(rglyphs[0]), true); break;
}
/* Add glyphs to glyphs string and setting some attributes */
for (int lgi = 0; lgi < lgn; lgi++) {
if ( rglyphs[lgi] == 0xdd/*TH_BLANK_BASE_GLYPH*/ ) {
//if ( !item->fixedPitch ) {
glyphString[slen++] = C_DOTTED_CIRCLE;
item->attributes[slen-1].dontPrint = true; // FIXME this will hide all dotted circle
//}
}
else {
glyphString[slen++] = (HB_UChar16) thai_get_glyph_index (font_type, rglyphs[lgi]);
}
}
if (haveSaraAm) {
logClusters[i + cell_length - 1] = slen - 1; // Set logClusters before NIKAHIT
if (tis_cell.top != 0)
logClusters[i + cell_length - 2] = slen - 2; // Set logClusters before NIKAHIT when tis_cell has top
if (logClusters[i + cell_length - 1] > slen)
logClusters[i + cell_length - 1] = 0;
}
i += cell_length;
}
glyphString[slen] = (HB_UChar16) '\0';
/* for check, should reallocate space or not */
HB_Bool spaceOK = (item->num_glyphs >= slen);
/* Convert to Glyph indices */
HB_Bool haveGlyphs = item->font->klass->convertStringToGlyphIndices (
item->font,
glyphString, slen,
item->glyphs, &item->num_glyphs,
item->shaperFlags);
HB_FREE_STACKARRAY (glyphString);
if (len >= 128)
free(cstr);
return (haveGlyphs && spaceOK);
}
