static INT parse_vowel_syllable(LPCWSTR input, INT cChar, INT start,
INT next, lexical_function lex)
{
if ((next < cChar) && lex(input[next]) == lex_Nukta)
next++;
if ((next < cChar) && is_joiner( lex(input[next]) ) && lex(input[next+1])==lex_Halant && is_consonant( lex(input[next+2]) ))
next+=3;
else if ((next < cChar) && lex(input[next])==lex_Halant && is_consonant( lex(input[next+1]) ))
next+=2;
else if ((next < cChar) && lex(input[next])==lex_ZWJ && is_consonant( lex(input[next+1]) ))
next+=2;
if ((next < cChar) && is_matra( lex(input[next]) ))
{
while((next < cChar) && is_matra( lex(input[next]) ))
next++;
if ((next < cChar) && lex(input[next]) == lex_Nukta)
next++;
if ((next < cChar) && lex(input[next]) == lex_Halant)
next++;
}
if ((next < cChar) && lex(input[next]) == lex_Modifier)
next++;
if ((next < cChar) && lex(input[next]) == lex_Vedic)
next++;
return next;
}
static int
ends_double_consonant(wchar_t *word, int len) {
/* Word must contain at least two characters in order to pass this
* test */
if(len < 2) {
return 0;
}
/* Run the test */
return (towlower(word[len-1]) == towlower(word[len-2]) &&
is_consonant(word, len-1) &&
is_consonant(word, len-2));
}
CFSWString syllabify2(CFSWString s) {
CFSWString res;
for (INTPTR i = 0; i < s.GetLength(); i++) {
CFSWString c = s.GetAt(i);
if (is_consonant(c) && is_vowel(s.GetAt(i - 1)) && is_vowel(s.GetAt(i + 1)))
res += d;
if (is_vowel(c) && is_vowel(s.GetAt(i - 1)) && is_vowel(s.GetAt(i + 1)) && c.ToLower() == s.GetAt(i + 1))
res += d;
if (is_consonant(c) && is_consonant(s.GetAt(i - 1)) && is_vowel(s.GetAt(i + 1)) && has_vowel(res)) //küsitav
res += d;
res += c;
}
return res;
}
/* Computes the measure of a string of characters. If we denote a
* vowel to be v, a consonant to be c, a list of consecutive vowels to
* be V and a list of consecutive consonants to be C, then the measure
* m of a string is computed as [C](VC)^(m)[V] */
static int
compute_measure(wchar_t* word, int len) {
int measure, state, is_vowel;
int i;
measure = 0;
/* state is 1 for vowel and 0 for consonant */
state = 0;
for(i=0; i<len; i++) {
/* Bit of error checking. Just ensure that we've not been passed
* sentence rather than a word */
is_vowel = !is_consonant(word, i);
if(is_vowel < 0) {
return -1;
}
if( is_vowel ) {
state = 1;
} else {
if(state == 1) {
measure++;
}
state = 0;
}
}
return measure;
}
/* Checks if the character at index idx is a consonant. Here, the
* definition of a consonant is a letter other than A, E, I, O, U and
* other than Y preceded by a consonant */
static int
is_consonant(wchar_t *word, int idx) {
int i;
if(idx < 0) {
return 0;
}
/* Make sure we haven't been passed a character that is not a
* letter*/
if(!isalpha(word[idx])) {
return -1;
}
/* First check the non-conditional vowels */
for( i=0; i<wcslen(vowels); i++ ) {
if( towlower(word[idx]) == vowels[i] ) {
return 0;
}
}
/* Check Y, our special case character */
if( towlower(word[idx]) == 'y' && is_consonant(word, --idx) ) {
return 0;
}
return 1;
}
static int FindBaseConsonant(HDC hdc, SCRIPT_ANALYSIS *psa, ScriptCache* psc, LPCWSTR input, IndicSyllable *s, lexical_function lex, BOOL modern)
{
int i;
BOOL blwf = FALSE;
BOOL pref = FALSE;
/* remove ralf from consideration */
if (Consonent_is_ralf(hdc, psa, psc, input, s, lex))
{
s->ralf = s->start;
s->start+=2;
}
/* try to find a base consonant */
if (!is_consonant( lex(input[s->base]) ))
{
for (i = s->end; i >= s->start; i--)
if (is_consonant( lex(input[i]) ))
{
s->base = i;
break;
}
}
while ((blwf = Consonent_is_below_base_form(hdc, psa, psc, input, s, lex, modern)) || Consonent_is_post_base_form(hdc, psa, psc, input, s, lex, modern) || (pref = Consonent_is_pre_base_form(hdc, psa, psc, input, s, lex, modern)))
{
if (blwf && s->blwf == -1)
s->blwf = s->base - 1;
if (pref && s->pref == -1)
s->pref = s->base - 1;
for (i = s->base-1; i >= s->start; i--)
if (is_consonant( lex(input[i]) ))
{
s->base = i;
break;
}
}
if (s->ralf >= 0)
s->start = s->ralf;
if (s->ralf == s->base)
s->ralf = -1;
return s->base;
}
int main()
{
//check if it is a vowel
printf("%d == %d\n", is_vowel('a'), true);
printf("%d == %d\n", is_vowel('h'), false);
//check if it is a lettere
printf("%d == %d\n", is_letter('a'), true);
printf("%d == %d\n", is_letter('#'), false);
//checks if it is a consonant
printf("%d == %d\n", is_consonant('a'), false);
printf("%d == %d\n", is_consonant('c'), true);
printf("%d == %d\n", is_consonant('Z'),true);
printf("%d == %d\n", is_consonant('1'), false);
return false;
}
static int
contains_vowel(wchar_t *word, int len) {
int i;
for(i=0; i<len; i++) {
if(is_consonant(word, i) == 0) {
return 1;
}
}
return 0;
}
static int
contains_compound_consonant(char *x)
{
char *p, *q, *s;
if (strlen(x) < 2) {
return 0;
}
p = x;
do {
q = p + 1;
s = q + 1;
if (is_consonant(*p)) {
if (is_consonant(*q)) return 1;
if ((*q == 'y') && (is_consonant(*s))) return 1;
}
p++;
} while (*p);
return 0;
}
static void
pango_indic_engine_shape (PangoFont * font,
const char *text,
int length,
PangoAnalysis * analysis,
PangoGlyphString * glyphs)
{
PangoXSubfont subfont;
int n_chars, n_glyph;
int lvl;
int i;
gunichar *wc;
int sb;
int n_syls;
gunichar **syls = g_new (gunichar *, 2);
g_return_if_fail (font != NULL);
g_return_if_fail (text != NULL);
g_return_if_fail (length >= 0);
g_return_if_fail (analysis != NULL);
n_chars = n_glyph = g_utf8_strlen (text, length);
lvl = pango_x_find_first_subfont (font, &default_charset, 1, &subfont);
if (!lvl)
{
pango_x_fallback_shape (font, glyphs, text, n_chars);
return;
}
pango_indic_split_out_characters (&script, text, n_chars, &wc, &n_glyph,
glyphs);
pango_indic_convert_vowels (&script, TRUE, &n_glyph, wc,
pango_x_has_glyph (font,
PANGO_X_MAKE_GLYPH(subfont, 0xc93e)));
n_syls = 1;
syls[0] = wc;
sb = glyphs->log_clusters[0];
for (i = 0; i < n_chars; i++)
{
if (i && (is_consonant (wc[i]) | is_ind_vowel (wc[i]))
&& wc[i - 1] != VIRAMA)
{
syls = g_renew (gunichar *, syls, n_syls + 2);
syls[n_syls] = wc + i;
n_syls++;
sb = glyphs->log_clusters[i];
}
glyphs->log_clusters[i] = sb;
}
static INT consonant_header(LPCWSTR input, INT cChar, INT start, INT next,
lexical_function lex)
{
if (!is_consonant( lex(input[next]) )) return -1;
next++;
if ((next < cChar) && lex(input[next]) == lex_Nukta)
next++;
if ((next < cChar) && lex(input[next])==lex_Halant)
{
next++;
if((next < cChar) && is_joiner( lex(input[next]) ))
next++;
if ((next < cChar) && is_consonant( lex(input[next]) ))
return next;
}
else if ((next < cChar) && is_joiner( lex(input[next]) ) && lex(input[next+1])==lex_Halant)
{
next+=2;
if ((next < cChar) && is_consonant( lex(input[next]) ))
return next;
}
return -1;
}
static int
ends_consonant_vowel_consonant(wchar_t *word, int len) {
int i;
static const wchar_t exclude[] = L"wxy";
/* Word must contain at least three characters in order to pass this
* test */
if(len < 3) {
return 0;
}
/* Check if the last character in the string is on our exclusion
* list. If so, then it will fail this test */
for(i=0; i<wcslen(exclude); i++) {
if(word[len-1] == exclude[i]) {
return 0;
}
}
/* Run the test and return the result */
return (is_consonant(word, len-3) &&
!is_consonant(word, len-2) &&
is_consonant(word, len-1));
}
static void
pango_indic_make_ligs (gunichar * start, gunichar * end)
{
int i;
for (i = 0; i < (end-start-2);i++)
{
if (is_consonant (start[i]))
{
if (start[i+2]==RA && start[i+1]==VIRAMA)
{
start[i+1] = 0;
start[i+2] = RA_SUBSCRIPT;
}
}
}
}
static BOOL Consonent_is_pre_base_form(HDC hdc, SCRIPT_ANALYSIS *psa, ScriptCache* psc, LPCWSTR pwChar, IndicSyllable *s, lexical_function lexical, BOOL modern)
{
if (is_consonant(lexical(pwChar[s->base])) && s->base > s->start && lexical(pwChar[s->base-1]) == lex_Halant)
{
if (modern)
return (SHAPE_does_GSUB_feature_apply_to_chars(hdc, psa, psc, &pwChar[s->base-1], 1, 2, "pref") > 0);
else
{
WCHAR cc[2];
cc[0] = pwChar[s->base];
cc[1] = pwChar[s->base-1];
return (SHAPE_does_GSUB_feature_apply_to_chars(hdc, psa, psc, cc, 1, 2, "pref") > 0);
}
}
return FALSE;
}
static INT parse_consonant_syllable(LPCWSTR input, INT cChar, INT start,
INT *main, INT next, lexical_function lex)
{
int check;
int headers = 0;
do
{
check = consonant_header(input,cChar,start,next,lex);
if (check != -1)
{
next = check;
headers++;
}
} while (check != -1);
if (headers || is_consonant( lex(input[next]) ))
{
*main = next;
next++;
}
else
return -1;
if ((next < cChar) && lex(input[next]) == lex_Nukta)
next++;
if ((next < cChar) && lex(input[next]) == lex_Anudatta)
next++;
if ((next < cChar) && lex(input[next]) == lex_Halant)
{
next++;
if((next < cChar) && is_joiner( lex(input[next]) ))
next++;
}
else if (next < cChar)
{
while((next < cChar) && is_matra( lex(input[next]) ))
next++;
if ((next < cChar) && lex(input[next]) == lex_Nukta)
next++;
if ((next < cChar) && lex(input[next]) == lex_Halant)
next++;
}
if ((next < cChar) && lex(input[next]) == lex_Modifier)
next++;
if ((next < cChar) && lex(input[next]) == lex_Vedic)
next++;
return next;
}
static void
pango_indic_make_ligs (gunichar * start, gunichar * end)
{
int num = end - start;
int i;
for (i = 0; i < (end - start); i++)
{
gunichar t0 = pango_indic_get_char (start + i, end);
gunichar t1 = pango_indic_get_char (start + 1 + i, end);
if (is_consonant (t0) && t1 == VIRAMA)
{
start[i+0] = t0 + INTERMEDIATE_HALF_FORM_OFFSET;
start[i+1] = 0;
}
}
if (num > 2 && start[0] == INTERMEDIATE_HALF_FORM_OFFSET + RA)
{
for (i=1;i<num;i++)
start[i-1] = start[i];
start[num-1] = RA_SUPERSCRIPT;
}
dev_mini_shuffle (start, end);
for (i = 0; i < (end - start - 1); i++)
if (is_intermediate_form (start[i]))
{
if (start[i+1] == RA)
{
start[i] = nominal_form (start[i]);
start[i+1] = RA_SUBSCRIPT;
}
else if (start[i+1] == (RA + INTERMEDIATE_HALF_FORM_OFFSET))
{
start[i] = nominal_form (start[i]);
start[i+1] = RA_SUBSCRIPT;
shuffle_one_along (start+2, end);
start[i+2] = VIRAMA;
}
}
}
CFSWString shift_pattern(CFSWString s) {
if (s == L'j') return L'j';
else
if (s == L'h') return L'h';
else
if (s == L'v') return L'v';
else
if (s.FindOneOf(L"sS") > -1) return L's';
else
if (s.FindOneOf(L"lmnrLN") > -1) return L'L';
else
if (s.FindOneOf(L"kptfšT") > -1) return L'Q';
else
if (is_vowel(s)) return L'V';
else
if (is_consonant(s)) return L'C';
return s;
}
static int
IC(char c) {
return is_consonant(c);
}
static char *
split_lujvo (char *x)
{
static char result[2048];
int first=1; /* At start, so we can handle r and n hyphens */
result[0] = 0;
while (1) {
/* Advance x to look at the tail of the string. */
int len = strlen(x);
switch (len) {
case 0:
case 1:
case 2:
/* Impossible */
return NULL;
break;
case 3:
case 4:
{
char *comp = lookup_component(x);
if (comp) {
if (!first) strcat(result, "+");
strcat(result,comp);
}
return result;
}
break;
case 5:
if (!first) strcat(result, "+");
strcat(result, x);
return result;
break;
default:
{
if (x[3] == 'y') {
strip_leading_rafsi(&x, 3, 1, first, result);
} else if (x[4] == 'y') {
strip_leading_rafsi(&x, 4, 1, first, result);
} else if (x[5] == 'y') {
/* Handle cultural rafsi of form CCVVCy */
strip_leading_rafsi(&x, 5, 1, first, result);
} else if ((x[6] == 'y') && (x[3] == '\'')) {
/* Handle cultural rafsi of form CCV'VCy */
strip_leading_rafsi(&x, 6, 1, first, result);
} else {
/* Have to decide if there's an r/n hyphen */
if (!is_consonant(x[0])) return NULL; /* can't ever happen */
if (IC(x[1]) && IV(x[2])) { /* CCV at start */
strip_leading_rafsi(&x, 3, 0, first, result);
} else if (IV(x[1]) && IC(x[2])) {
strip_leading_rafsi(&x, 3, 0, first, result);
} else if (IV(x[1])) {
if (x[2] == '\'') {
/* Everything pushed up by 1 */
if (IV(x[3])) {
if (((x[4] == 'r') && (IC(x[5]))) ||
((x[4] == 'n') && (x[5] == 'r'))) {
/* This pattern can only occur if a hyphen is there. */
strip_leading_rafsi(&x, 4, 1, first, result);
} else {
strip_leading_rafsi(&x, 4, 0, first, result);
}
} else {
/* CV'C form, impossible. */
return NULL;
}
} else {
if (IV(x[2])) {
if (((x[3] == 'r') && (IC(x[4]))) ||
((x[3] == 'n') && (x[4] == 'r'))) {
/* This pattern can only occur if a hyphen is there. */
strip_leading_rafsi(&x, 3, 1, first, result);
} else {
strip_leading_rafsi(&x, 3, 0, first, result);
}
} else {
/* CVC form, no r/n hyphen */
strip_leading_rafsi(&x, 3, 0, first, result);
}
}
} else {
/* Impossible */
return NULL;
}
}
}
break;
}
first = 0;
}
}
static int
is_consonantal_form (int q)
{
return (q == PANGO_ZERO_WIDTH_JOINER) || is_consonant (q) || (q >= 0xd000);
}