Merge branch 'development'

src/numbers.cpp

@@ -986,6 +986,18 @@ static const char *M_Variant(int value)
 {//====================================
 	// returns M, or perhaps MA for some cases
 	
+	if(translator->translator_name == L('l','t'))
+	{
+		// Lithuanian
+		if((value % 10) == 1)
+			return("0MA");
+		if(((value % 10) == 0) || ((value % 100) > 10) && ((value % 100) < 20))
+		{
+			return("0MB");
+		}
+		return("0M");
+	}
+
 	if((translator->langopts.numbers2 & 0x100) && (value >= 2) && (value <= 4))
 		return("0MA");  // Czech, Slovak
 	else

src/readclause.cpp

@@ -69,12 +69,6 @@ static int ssml_ignore_l_angle = 0;
 //static const char *tone_punct_on = "\0016T";  // add reverberation, lower pitch
 //static const char *tone_punct_off = "\001T\001P";
 
-// ignore these characters
-static const unsigned short chars_ignore[] = {
-  0x200c,  // zero width non-joiner
-  0x200d,  // zero width joiner
-  0 };
-
 // punctuations symbols that can end a clause
 static const unsigned short punct_chars[] = {',','.','?','!',':',';',
   0x2013,  // en-dash
@@ -2392,10 +2386,14 @@ f_input = f_in;  // for GetC etc
 				c2 = ' ';
 			}
 
-			if(lookupwchar(chars_ignore,c1))
+			if((j = lookupwchar2(tr->chars_ignore,c1)) != 0)
 			{
-				// ignore this character (eg. zero-width-non-joiner U+200C)
-				continue;
+				if(j == 1)
+				{
+					// ignore this character (eg. zero-width-non-joiner U+200C)
+					continue;
+				}
+				c1 = j;   // replace the character
 			}
 
 			if(c1 == 0xf0b)

src/synthdata.cpp

@@ -35,7 +35,7 @@
 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.46.19  27.Jun.12";
+const char *version_string = "1.46.20  15.Jul.12";
 const int version_phdata  = 0x014600;
 
 int option_device_number = -1;

src/tr_languages.cpp

@@ -111,6 +111,18 @@ static void SetLetterBitsRange(Translator *tr, int group, int first, int last)
 	}
 }
 
+// ignore these characters
+static const unsigned short chars_ignore_default[] = {
+	0x200c,  1, // zero width non-joiner
+	0x200d,  1, // zero width joiner
+	0, 0 };
+
+// alternatively, ignore characters but allow zero-width-non-joiner (lang-fa)
+static const unsigned short chars_ignore_fa[] = {
+	0x200c,  0x0605, // zero width non-joiner, replace with not-used Arabic character code
+	0x200d,  1, // zero width joiner
+	0, 0 };
+
 
 static Translator* NewTranslator(void)
 {//===================================
@@ -122,6 +134,7 @@ static Translator* NewTranslator(void)
 	static const wchar_t punct_in_word[2] = {'\'', 0};  // allow hyphen within words
 	static const unsigned char default_tunes[6] = {0, 1, 2, 3, 0, 0};
 
+
 	tr = (Translator *)Alloc(sizeof(Translator));
 	if(tr == NULL)
 		return(NULL);
@@ -155,6 +168,7 @@ static Translator* NewTranslator(void)
 
 	tr->char_plus_apostrophe = empty_wstring;
 	tr->punct_within_word = punct_in_word;
+	tr->chars_ignore = chars_ignore_default;
 
 	for(ix=0; ix<8; ix++)
 	{
@@ -647,9 +661,13 @@ Translator *SelectTranslator(const char *name)
 
 	case L('f','a'):   // Farsi
 		{
+			static const char fa_ZWNJ[] = {0x05, 0};  // use letter group G for ZWNJ U+200c
 			tr->letter_bits_offset = OFFSET_ARABIC;
 			tr->langopts.numbers = NUM_AND_UNITS | NUM_HUNDRED_AND;
 			tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1;   // disable check for unpronouncable words
+
+			tr->chars_ignore = chars_ignore_fa;
+			SetLetterBits(tr,LETTERGP_G,(char *)fa_ZWNJ);
 		}
 		break;
 
@@ -972,6 +990,20 @@ SetLengthMods(tr,3);  // all equal
 		}
 		break;
 
+	case L('l','t'): // Lithuanian
+		{
+			tr->charset_a0 = charsets[4]; // ISO-8859-4
+			tr->langopts.stress_rule = STRESSPOSN_2R;
+			tr->langopts.stress_flags = 0x20;
+			tr->langopts.unstressed_wd1 = 0;
+			tr->langopts.unstressed_wd2 = 2;
+			tr->langopts.param[LOPT_DIERESES] = 1;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_4 | NUM_ORDINAL_DOT;
+			tr->langopts.numbers2 = 0x100;
+			tr->langopts.max_roman = 5000;
+		}
+		break;
+
 	case L('l','v'):  // latvian
 		{
 			static const unsigned char stress_amps_lv[8] = {17,13, 20,20, 20,22, 22,21 };

src/translate.cpp

@@ -392,6 +392,9 @@ int IsAlpha(unsigned int c)
 		return(0);
 	}
 
+	if(c == 0x0605)
+		return(1);
+
 	if((c >= 0x64b)  && (c <= 0x65e))
 		return(1);   // arabic vowel marks
 
@@ -467,6 +470,22 @@ int lookupwchar(const unsigned short *list,int c)
 	return(0);
 }
 
+
+int lookupwchar2(const unsigned short *list,int c)
+{//==============================================
+// Replace character c by another character.
+// Returns 0 = not found, 1 = delete character
+	int ix;
+
+	for(ix=0; list[ix] != 0; ix+=2)
+	{
+		if(list[ix] == c)
+			return(list[ix+1]);
+	}
+	return(0);
+}
+
+
 int IsBracket(int c)
 {//=================
 	if((c >= 0x2014) && (c <= 0x201f))

src/translate.h

@@ -29,7 +29,7 @@
 #define N_RULE_GROUP2    120          // max num of two-letter rule chains
 #define N_HASH_DICT     1024
 #define N_CHARSETS        20
-#define N_LETTER_GROUPS   26
+#define N_LETTER_GROUPS   95          // maximum is 127-32 
 
 
 /* dictionary flags, word 1 */
@@ -520,6 +520,7 @@ typedef struct
 	const unsigned short *charset_a0;   // unicodes for characters 0xa0 to oxff
 	const wchar_t *char_plus_apostrophe;  // single chars + apostrophe treated as words
 	const wchar_t *punct_within_word;   // allow these punctuation characters within words
+	const unsigned short *chars_ignore;
 
 // holds properties of characters: vowel, consonant, etc for pronunciation rules
 	unsigned char letter_bits[256];
@@ -632,6 +633,7 @@ int utf8_in2(int *c, const char *buf, int backwards);
 int utf8_out(unsigned int c, char *buf);
 int utf8_nbytes(const char *buf);
 int lookupwchar(const unsigned short *list,int c);
+int lookupwchar2(const unsigned short *list,int c);
 int Eof(void);
 char *strchr_w(const char *s, int c);
 int IsBracket(int c);