Merge branch 'development' into build

src/compiledata.cpp

@@ -44,6 +44,7 @@
 extern void FindPhonemesUsed(void);
 extern void DisplayErrorFile(const char *fname);
 extern int utf8_out(unsigned int c, char *buf);
+extern void DrawEnvelopes();
 char path_source[sizeof(path_home)+20];
 
 
@@ -3503,9 +3504,9 @@ fprintf(f_errors,"\nRefs %d,  Reused %d\n",count_references,duplicate_references
 
 	CompileReport();
 	report_dict = CompileAllDictionaries();
-#ifdef deleted
+//#ifdef deleted
 	DrawEnvelopes();
-#endif
+//#endif
 
 	if(gui_flag)
 	{

src/compiledict.cpp

@@ -879,6 +879,18 @@ static int group3_ix;
 
 
 
+int isHexDigit(int c)
+{
+	if((c >= '0') && (c <= '9'))
+		return(c - '0');
+	if((c >= 'a') && (c <= 'f'))
+		return(c - 'a' + 10);
+	if((c >= 'A') && (c <= 'F'))
+		return(c - 'A' + 10);
+	return(-1);
+}
+
+
 static void copy_rule_string(char *string, int &state)
 {//===================================================
 // state 0: conditional, 1=pre, 2=match, 3=post, 4=phonemes
@@ -889,9 +901,11 @@ static void copy_rule_string(char *string, int &state)
 	int ix;
 	int len;
 	char c;
+	int c2, c3;
 	int  sxflags;
 	int  value;
 	int  literal;
+	int  hexdigit_input = 0;
 	MNEM_TAB *mr;
 
 	if(string[0] == 0) return;
@@ -911,18 +925,38 @@ static void copy_rule_string(char *string, int &state)
 	{
 		literal = 0;
 		c = *p++;
+		if((c == '0') && (p[0] == 'x') && (isHexDigit(p[1]) >= 0) && (isHexDigit(p[2]) >= 0))
+		{
+			hexdigit_input = 1;
+			c = p[1];
+			p+= 2;
+		}
 		if(c == '\\')
 		{
 			c = *p++;   // treat next character literally
+//#ifdef deleted
 			if((c >= '0') && (c <= '3') && (p[0] >= '0') && (p[0] <= '7') && (p[1] >= '0') && (p[1] <= '7'))
 			{
 				// character code given by 3 digit octal value;
 				c = (c-'0')*64 + (p[0]-'0')*8 + (p[1]-'0');
 				p += 2;
 			}
+//endif
 			literal = 1;
 		}
-
+		if(hexdigit_input)
+		{
+			if(((c2 = isHexDigit(c)) >= 0) && ((c3 = isHexDigit(p[0])) >= 0))
+			{
+				c = c2 * 16 + c3;
+				literal = 1;
+				p++;
+			}
+			else
+			{
+				hexdigit_input = 0;
+			}
+		}
 		if((state==1) || (state==3))
 		{
 			// replace special characters (note: 'E' is reserved for a replaced silent 'e')

src/fifo.cpp

@@ -60,7 +60,7 @@ static void* say_thread(void*);
 
 static espeak_ERROR push(t_espeak_command* the_command);
 static t_espeak_command* pop();
-static void init();
+static void init(int process_parameters);
 static int node_counter=0;
 enum {MAX_NODE_COUNTER=400,
       INACTIVITY_TIMEOUT=50, // in ms, check that the stream is inactive
@@ -75,7 +75,7 @@ void fifo_init()
 
   // security
   pthread_mutex_init( &my_mutex, (const pthread_mutexattr_t *)NULL);
-  init();
+  init(0);
 
   assert(-1 != sem_init(&my_sem_start_is_required, 0, 0));
   assert(-1 != sem_init(&my_sem_stop_is_acknowledged, 0, 0));
@@ -457,7 +457,7 @@ static void* say_thread(void*)
 	{ 
 	  // no mutex required since the stop command is synchronous
 	  // and waiting for my_sem_stop_is_acknowledged
-	  init();
+	  init(1);
 
 	  // purge start semaphore
 	  SHOW_TIME("say_thread > purge my_sem_start_is_required\n");
@@ -569,14 +569,24 @@ static t_espeak_command* pop()
 }
 
 
-static void init()
+static void init(int process_parameters)
 {
-  ENTER("fifo > init");
-  while (delete_espeak_command( pop() ))
-    {}
-  node_counter = 0;
+	// Changed by Tyler Spivey 30.Nov.2011
+	t_espeak_command *c = NULL;
+	ENTER("fifo > init");
+	c = pop();
+	while (c != NULL) {
+		if (process_parameters && (c->type == ET_PARAMETER || c->type == ET_VOICE_NAME || c->type == ET_VOICE_SPEC))
+		{
+			process_espeak_command(c);
+		}
+		delete_espeak_command(c);
+		c = pop();
+	}
+	node_counter = 0;
 }
 
+
 //>
 //<fifo_init
 void fifo_terminate()
@@ -589,7 +599,7 @@ void fifo_terminate()
   sem_destroy(&my_sem_start_is_required);
   sem_destroy(&my_sem_stop_is_acknowledged);
 
-  init(); // purge fifo
+  init(0); // purge fifo
 }
 
 #endif

src/numbers.cpp

@@ -1862,15 +1862,19 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 
 			case NUM_DFRACTION_1:   // italian, say "hundredths" if leading zero
 			case NUM_DFRACTION_5:   // hungarian, always say "tenths" etc.
+			case NUM_DFRACTION_6:   // kazakh, always say "tenths" etc, before the decimal fraction
 				LookupNum3(tr, atoi(&word[n_digits]), ph_buf, 0,0,0);
-				if((word[n_digits]=='0') || (decimal_mode == NUM_DFRACTION_5))
+				if((word[n_digits]=='0') || (decimal_mode != NUM_DFRACTION_1))
 				{
 					// decimal part has leading zeros, so add a "hundredths" or "thousandths" suffix
 					sprintf(string,"_0Z%d",decimal_count);
 					if(Lookup(tr, string, buf1) == 0)
 						break;   // revert to speaking single digits
 
-					strcat(ph_buf,buf1);
+					if(decimal_mode == NUM_DFRACTION_6)
+						strcat(ph_out, buf1);
+					else
+						strcat(ph_buf, buf1);
 				}
 				strcat(ph_out,ph_buf);
 				n_digits += decimal_count;

src/setlengths.cpp

@@ -715,14 +715,14 @@ if(stress <= 1)
 			if(stress >= 7)
 			{
 				// tonic syllable, include a constant component so it doesn't decrease directly with speed
-				length_mod += 20;
+				length_mod += tr->langopts.lengthen_tonic;
 				if(emphasized)
-					length_mod += 10;
+					length_mod += (tr->langopts.lengthen_tonic/2);
 			}
 			else
 			if(emphasized)
 			{
-				length_mod += 20;
+				length_mod += tr->langopts.lengthen_tonic;
 			}
 
 			if((len = tr->stress_lengths[stress]) == 0)
@@ -739,7 +739,8 @@ if(stress <= 1)
 				}
 			}
 
-			if(end_of_clause == 2)
+
+			if((end_of_clause == 2) && !(tr->langopts.stress_flags & S_NO_EOC_LENGTHEN))
 			{
 				// this is the last syllable in the clause, lengthen it - more for short vowels
 				len = (p->ph->std_length * 2);

src/synthdata.cpp

@@ -35,7 +35,7 @@
 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.46.06  29.Nov.11";
+const char *version_string = "1.46.09  09.Dec.11";
 const int version_phdata  = 0x014600;
 
 int option_device_number = -1;

src/tr_languages.cpp

@@ -164,6 +164,7 @@ static Translator* NewTranslator(void)
 	}
 	memset(&(tr->langopts),0,sizeof(tr->langopts));
 	tr->langopts.max_lengthmod = 500;
+	tr->langopts.lengthen_tonic = 20;
 
 	tr->langopts.stress_rule = STRESSPOSN_2R;
 	tr->langopts.unstressed_wd1 = 1;
@@ -274,11 +275,13 @@ static const unsigned int replace_cyrillic_latin[] =
 	0x45c,0x107,
 0};  // ѓ  ѕ  ќ
 
+
+static const unsigned char ru_vowels[] = {0x10,0x15,0x31,0x18,0x1e,0x23,0x2b,0x2d,0x2e,0x2f,  0xb9,0xc9,0x91,0x8f,0x36,0};  //also kazakh
+static const unsigned char ru_consonants[] = {0x11,0x12,0x13,0x14,0x16,0x17,0x19,0x1a,0x1b,0x1c,0x1d,0x1f,0x20,0x21,0x22,0x24,0x25,0x26,0x27,0x28,0x29,0x2a,0x2c, 0x73,0x7b,0x83,0x9b,0};
+
 static void SetCyrillicLetters(Translator *tr)
 {//===========================================
 	// character codes offset by 0x420
-	static const char ru_vowels[] = {0x10,0x15,0x31,0x18,0x1e,0x23,0x2b,0x2d,0x2e,0x2f,0};
-	static const char ru_consonants[] = {0x11,0x12,0x13,0x14,0x16,0x17,0x19,0x1a,0x1b,0x1c,0x1d,0x1f,0x20,0x21,0x22,0x24,0x25,0x26,0x27,0x28,0x29,0x2a,0x2c,0};
 	static const char ru_soft[] = {0x2c,0x19,0x27,0x29,0};   // letter group B  [k ts; s;]
 	static const char ru_hard[] = {0x2a,0x16,0x26,0x28,0};   // letter group H  [S Z ts]
 	static const char ru_nothard[] = {0x11,0x12,0x13,0x14,0x17,0x19,0x1a,0x1b,0x1c,0x1d,0x1f,0x20,0x21,0x22,0x24,0x25,0x27,0x29,0x2c,0};
@@ -291,14 +294,14 @@ static void SetCyrillicLetters(Translator *tr)
 
 	tr->letter_bits_offset = OFFSET_CYRILLIC;
 	memset(tr->letter_bits,0,sizeof(tr->letter_bits));
-	SetLetterBits(tr,LETTERGP_A,ru_vowels);
+	SetLetterBits(tr,LETTERGP_A,(char *)ru_vowels);
 	SetLetterBits(tr,1,ru_soft);
-	SetLetterBits(tr,2,ru_consonants);
+	SetLetterBits(tr,2,(char *)ru_consonants);
 	SetLetterBits(tr,3,ru_hard);
 	SetLetterBits(tr,4,ru_nothard);
 	SetLetterBits(tr,5,ru_voiced);
 	SetLetterBits(tr,6,ru_ivowels);
-	SetLetterBits(tr,LETTERGP_VOWEL2,ru_vowels);
+	SetLetterBits(tr,LETTERGP_VOWEL2,(char *)ru_vowels);
 }  // end of SetCyrillicLetters
 
 
@@ -893,6 +896,30 @@ SetLengthMods(tr,3);  // all equal
 		}
 		break;
 
+	case L('k','k'):   // Kazakh
+		{
+			static const unsigned char stress_amps_tr[8] = {18,16, 20,21, 20,21, 21,20 };
+			static const short stress_lengths_tr[8] = {190,180, 230,230, 0,0, 250,250};
+ 
+			tr->letter_bits_offset = OFFSET_CYRILLIC;
+			memset(tr->letter_bits,0,sizeof(tr->letter_bits));
+			SetLetterBits(tr,LETTERGP_A,(char *)ru_vowels);
+			SetLetterBits(tr,LETTERGP_C,(char *)ru_consonants);
+			SetLetterBits(tr,LETTERGP_VOWEL2,(char *)ru_vowels);
+
+			SetupTranslator(tr,stress_lengths_tr,stress_amps_tr);
+
+			tr->langopts.stress_rule = 7;   // stress on the last syllable, before any explicitly unstressed syllable
+			tr->langopts.stress_flags = S_NO_AUTO_2 + S_NO_EOC_LENGTHEN;  //no automatic secondary stress, don't lengthen at end-of-clause
+			tr->langopts.lengthen_tonic = 0;
+			tr->langopts.param[LOPT_SUFFIX] = 1;
+
+			tr->langopts.numbers =  NUM_OMIT_1_HUNDRED | NUM_DFRACTION_6 ;
+			tr->langopts.max_initial_consonants = 2;
+SetLengthMods(tr,3);  // all equal
+		}
+		break;
+
 	case L('k','o'):   // Korean, TEST
 		{
 			static const char ko_ivowels[] = {0x63,0x64,0x67,0x68,0x6d,0x72,0x74,0x75,0};  // y and i vowels

src/translate.cpp

@@ -334,16 +334,16 @@ static unsigned char length_mods_en0[100] = {
 
 static unsigned char length_mods_equal[100] = {
 /*  a   ,   t   s   n   d   z   r   N   <- next */
-	110,110,110,110,110,110,110,110,110, 110,  /* a  <- next2 */
-	110,110,110,110,110,110,110,110,110, 110,  /* , */
-	110,110,110,110,110,110,110,110,110, 110,  /* t */
-	110,110,110,110,110,110,110,110,110, 110,  /* s */
-	110,110,110,110,110,110,110,110,110, 110,  /* n */
-	110,110,110,110,110,110,110,110,110, 110,  /* d */
-	110,110,110,110,110,110,110,110,110, 110,  /* z */
-	110,110,110,110,110,110,110,110,110, 110,  /* r */
-	110,110,110,110,110,110,110,110,110, 110,  /* N */
-	110,110,110,110,110,110,110,110,110, 110 }; // SPARE
+	110,120,100,110,110,110,110,110,110, 110,  /* a  <- next2 */
+	110,120,100,110,110,110,110,110,110, 110,  /* , */
+	110,120,100,110,100,110,110,110,100, 110,  /* t */
+	110,120,100,110,110,110,110,110,110, 110,  /* s */
+	110,120,100,110,110,110,110,110,110, 110,  /* n */
+	110,120,100,110,110,110,110,110,110, 110,  /* d */
+	110,120,100,110,110,110,110,110,110, 110,  /* z */
+	110,120,100,110,110,110,110,110,110, 110,  /* r */
+	110,120,100,110,110,110,110,110,110, 110,  /* N */
+	110,120,100,110,110,110,110,110,110, 110 }; // SPARE
 
 
 static unsigned char *length_mod_tabs[6] = {

src/translate.h

@@ -355,6 +355,7 @@ typedef struct {
 #define S_INITIAL_2         0x2000
 #define S_NO_AUTO_DIM       0x10000
 #define S_HYPEN_UNSTRESS    0x100000
+#define S_NO_EOC_LENGTHEN   0x200000
 // bit0=don't stress monosyllables, except at end of clause
 // bit1=don't set diminished stress,
 // bit2=mark unstressed final syllables as diminished
@@ -373,6 +374,7 @@ typedef struct {
 // bit18= don't lengthen short vowels more than long vowels at end-of-clause
 // bit19=stress on final syllable if it has a long vowel, but previous syllable has a short vowel
 // bit20= hyphenated words, 2nd part is unstressed
+// bit21= don't lengthen vowels at end-of-clause
 
 	int stress_flags; 
 	int unstressed_wd1; // stress for $u word of 1 syllable
@@ -398,6 +400,7 @@ typedef struct {
 #define NUM_DFRACTION_3  0x6000
 #define NUM_DFRACTION_4  0x8000
 #define NUM_DFRACTION_5  0xa000
+#define NUM_DFRACTION_6  0xc000
 #define NUM_ORDINAL_DOT   0x10000
 #define NUM_NOPAUSE       0x20000
 #define NUM_AND_HUNDRED   0x40000
@@ -476,7 +479,8 @@ typedef struct {
 	const unsigned int *replace_chars;      // characters to be substitutes
 	const char *ascii_language;  // switch to this language for Latin characters
 	int max_lengthmod;
-	int suffix_add_e;     // replace a suffix (which has the SUFX_E flag) with this character
+	int lengthen_tonic;   // lengthen the tonic syllable
+	int suffix_add_e;      // replace a suffix (which has the SUFX_E flag) with this character
 } LANGUAGE_OPTIONS;