[1.29.15] Fix compiler warning, conversion from string constant to 'char *'.

espeakedit: Change compile_report file to include the phoneme name as well as phoneme table which uses each data file.


git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@95 d46cf337-b52f-0410-862d-fd96e6ae7743

dictsource/dict_phonemes

@@ -68,26 +68,22 @@ Dictionary en_dict
 
 0    3    3:   @    @-   @2   @5   @L   
 a    a2   A:   A@   aa   aI   aI@  aU   
-aU@  E    e@   eI   I    i    I2   i:   
-i@   IR   O    O:   o:   O@   o@   OI   
-oU   O~   U    u:   U@   V    VR   
+aU@  E    e@   eI   I    I2   i:   i@   
+IR   O    O:   o:   O@   o@   OI   oU   
+O~   U    u:   U@   V    VR   
 
 :    ;    b    C    d    D    dZ   f    
-g    g-   h    j    k    l    L    m    
-n    N    p    Q    r    r-   s    S    
-t    T    t2   tS   v    w    x    z    
-Z    
+g    h    j    k    l    m    n    N    
+p    Q    r    r-   s    S    t    T    
+t2   tS   v    w    x    z    Z    
 
 
 Dictionary eo_dict
 
-@-   a    aI   aU   e    eI   eU   i    
-o    OI   u    uI   
+@-   a    aI   aU   e    i    o    u    
 
-*    b    d    dZ   f    g    h    j    
-k    l    m    n    p    r    R    s    
-S    t    T    tS   ts   v    w    x    
-z    Z    
+*    b    d    k    l    m    n    p    
+r    R    s    t    ts   v    z    Z    
 
 
 Dictionary es_dict
@@ -346,3 +342,21 @@ I    o    r-   u
 k    k^   l    l^   m    n    n^   p    
 r    R    s    S    t    tS   ts   v    
 x    z    Z    
+
+
+Dictionary zh_dict
+
+
+
+Dictionary zhh_dict
+
+&?   0    0?   @    @?   @r   @u   A    
+A?   E    E~   i    i&?  i.   i0   i;   
+i?   i@   i@?  i[   iA   iA?  iE   iE~  
+io   io?  iu   i~   o    u    u&?  u?   
+u@   u@?  uA   uA?  uE   uE~  ui   u~   
+y    y?   y@   y@?  ya   y~   
+
+c;   f    h    k    kh   l    m    n    
+N    p    ph   s    s.   t    tc;  tc;  
+th   ts   ts.  ts.  tsh  v    z.   

dictsource/sv_list

@@ -16,10 +16,10 @@ d	de:
 e	e:
 f	Ef
 g	ge:
-h	hO:
+h	ho:
 i	i:
 j	ji:
-k	kO:
+k	ko:
 l	El
 m	Em
 n	En
@@ -77,6 +77,9 @@ _0M3	bIlj'u:nER
 _1M3	bIlj'u:n
 _dpt    p'8Nkt
 
+(1 : a)  f'Ws.ta
+
+
 // abbreviations
 //**************
 

phsource/ph_greek_ancient

@@ -3,12 +3,43 @@
 //  Ancient Greek  - based on base2
 //====================================================
 
+phoneme : //  Lengthen previous vowel by "length"
+  virtual
+  length 100
+endphoneme
+
+
 phoneme y
   vowel starttype (i) endtype (i)
   length 160
   formants vowel/yy_4
 endphoneme
 
+phoneme e
+  vowel starttype (e) endtype (e)
+  length 140
+  formants vowel/e
+endphoneme
+
+phoneme E
+  vowel starttype (e) endtype (e)
+  length 150
+  formants vowel/ee_1
+endphoneme
+
+phoneme o
+  vowel starttype (o) endtype (o)
+  length 140
+  formants vowel/o
+endphoneme
+
+phoneme O
+  vowel starttype (o) endtype (o)
+  length 150
+  formants vowel/0
+endphoneme
+
+
 phoneme EU
   vowel starttype (e) endtype (u)
   length 230

phsource/phonemes

@@ -1236,6 +1236,13 @@ include ph_vi
 phonemetable zhy base
 include ph_zh_yue
 
+phonemetable zh base
+include ph_zh
+
+phonemetable zhh zh
+include ph_zh_huai
+
+
 phonemetable sw base
 include ph_swahili
 

src/compiledata.cpp

@@ -80,13 +80,11 @@ extern int CompileDictionary(const char *dsource, const char *dict_name, FILE *l
 
 static int markers_used[8];
 
-#define N_USED_BY  12
 typedef struct {
 	void *link;
 	int value;
-	short n_uses;
-	short n_used_by;
-	unsigned char used_by[N_USED_BY];
+	int ph_mnemonic;
+	short ph_table;
 	char string[1];
 } REF_HASH_TAB;
 
@@ -906,12 +904,17 @@ int Compile::LoadDataFile(const char *path, int control)
 
 	FILE *f;
 	int id;
-	int ix;
 	int hash;
+	int addr = 0;
 	int type_code=' ';
 	REF_HASH_TAB *p, *p2;
 	char buf[sizeof(path_source)+120];
 
+	if(strcmp(path,"NULL")==0)
+		return(0);
+	if(strcmp(path,"DFT")==0)
+		return(1);
+
 	count_references++;
 
 	hash = Hash8(path);
@@ -920,102 +923,80 @@ int Compile::LoadDataFile(const char *path, int control)
 	{
 		if(strcmp(path,p->string)==0)
 		{
-			int found = 0;
-
 			duplicate_references++;
-			p->n_uses++;
-
-			// add the current phoneme table to a list of users of this data file, if not already listed
-			for(ix=0; (ix < p->n_used_by) && (ix < N_USED_BY); ix++)
-			{
-				if(p->used_by[ix] == n_phoneme_tabs)
-				{
-					found = 1;
-					break;
-				}
-			}
-			if(found == 0)
-			{
-				if(ix < N_USED_BY)
-				{
-					p->used_by[ix] = n_phoneme_tabs;
-				}
-				p->n_used_by++;
-			}
-
-			return(p->value);    // already loaded this data
+			addr = p->value;    // already loaded this data
+			break;
 		}
 		p = (REF_HASH_TAB *)p->link;
 	}
 
-	sprintf(buf,"%s%s",path_source,path);
-
-	if(strcmp(path,"NULL")==0)
-		return(0);
-	if(strcmp(path,"DFT")==0)
-		return(1);
-
-	if((f = fopen(buf,"rb")) == NULL)
+	if(addr == 0)
 	{
-		sprintf(buf,"%s%s.wav",path_source,path);
+		sprintf(buf,"%s%s",path_source,path);
+	
 		if((f = fopen(buf,"rb")) == NULL)
 		{
-			Error("Can't read file",path);
-			return(0);
+			sprintf(buf,"%s%s.wav",path_source,path);
+			if((f = fopen(buf,"rb")) == NULL)
+			{
+				Error("Can't read file",path);
+				return(0);
+			}
+		}
+
+		id = Read4Bytes(f);
+		rewind(f);
+	
+		if(id == 0x43455053)
+		{
+			addr = LoadSpect(path, control);
+			type_code = 'S';
+		}
+		else
+		if(id == 0x46464952)
+		{
+			addr = LoadWavefile(f,path);
+			type_code = 'W';
+		}
+		else
+		if(id == 0x43544950)
+		{
+			addr = LoadEnvelope(f,path);
+			type_code = 'E';
+		}
+		else
+		if(id == 0x45564E45)
+		{
+			addr = LoadEnvelope2(f,path);
+			type_code = 'E';
+		}
+		else
+		{
+			Error("File not SPEC or RIFF",path);
+			addr = -1;
+		}
+		fclose(f);
+	
+		if(addr > 0)
+		{
+			fprintf(f_phcontents,"%c  0x%.5x  %s\n",type_code,addr & 0x7fffff,path);
 		}
 	}
 
-	id = Read4Bytes(f);
-//	fread(&id,1,4,f);
-	rewind(f);
-
-	if(id == 0x43455053)
-	{
-		ix = LoadSpect(path, control);
-		type_code = 'S';
-	}
-	else
-	if(id == 0x46464952)
-	{
-		ix = LoadWavefile(f,path);
-		type_code = 'W';
-	}
-	else
-	if(id == 0x43544950)
-	{
-		ix = LoadEnvelope(f,path);
-		type_code = 'E';
-	}
-	else
-	if(id == 0x45564E45)
-	{
-		ix = LoadEnvelope2(f,path);
-		type_code = 'E';
-	}
-	else
-	{
-		Error("File not SPEC or RIFF",path);
-		ix = -1;
-	}
-	fclose(f);
-
-	if(ix > 0)
-	{
-		fprintf(f_phcontents,"%c  0x%.5x  %s\n",type_code,ix & 0x7fffff,path);
-	}
-
 	// add this item to the hash table
-	p = ref_hash_tab[hash];
-	p2 = (REF_HASH_TAB *)malloc(sizeof(REF_HASH_TAB)+strlen(path)+1);
-	p2->value = ix;
-	p2->n_uses = 1;
-	p2->n_used_by = 1;
-	strcpy(p2->string,path);
-	p2->used_by[0] = n_phoneme_tabs;
-	p2->link = (char *)p;
-	ref_hash_tab[hash] = p2;
+	if(addr > 0)
+	{
+		p = ref_hash_tab[hash];
+		p2 = (REF_HASH_TAB *)malloc(sizeof(REF_HASH_TAB)+strlen(path)+1);
+		p2->value = addr;
+		p2->ph_mnemonic = ph->mnemonic;   // phoneme which uses this file
+		p2->ph_table = n_phoneme_tabs-1;
+		strcpy(p2->string,path);
+		p2->link = (char *)p;
+		ref_hash_tab[hash] = p2;
+	}
 
-	return(ix);
+	return(addr);
 }   //  end of Compile::LoadDataFile
 
 
@@ -1892,12 +1873,21 @@ int Compile::LoadEnvelope2(FILE *f, const char *fname)
 
 static int ref_sorter(char **a, char **b)
 {//======================================
-	
-	REF_HASH_TAB *p1 = (REF_HASH_TAB *)(*a);
- 	REF_HASH_TAB *p2 = (REF_HASH_TAB *)(*b);
+	int ix;
 
-  return(strcoll(p1->string,p2->string));
-}   /* end of strcmp2 */
+	REF_HASH_TAB *p1 = (REF_HASH_TAB *)(*a);
+	REF_HASH_TAB *p2 = (REF_HASH_TAB *)(*b);
+
+	ix = strcoll(p1->string,p2->string);
+	if(ix != 0)
+		return ix;
+
+	ix = p1->ph_table - p2->ph_table;
+	if(ix != 0)
+		return ix;
+
+	return(p1->ph_mnemonic - p2->ph_mnemonic);
+}   /* end of ref_sorter */
 
 
 
@@ -1909,10 +1899,13 @@ void Compile::Report(void)
 	REF_HASH_TAB *p;
 	REF_HASH_TAB **list;
 	FILE *f_report;
+	const char *data_path;
+	int prev_table;
+	int prev_mnemonic;
 	char fname[sizeof(path_source)+20];
 
 	// make a list of all the references and sort it
-	list = (REF_HASH_TAB **)malloc(count_references * sizeof(REF_HASH_TAB *));
+	list = (REF_HASH_TAB **)malloc((count_references)* sizeof(REF_HASH_TAB *));
 	if(list == NULL)
 		return;
 
@@ -1932,6 +1925,7 @@ void Compile::Report(void)
 	}
 	fputc('\n',f_report);
 
+	fprintf(f_report,"Data file      Used by\n");
 	ix = 0;
 	for(hash=0; (hash < 256) && (ix < count_references); hash++)
 	{
@@ -1945,25 +1939,35 @@ void Compile::Report(void)
 	n = ix;
 	qsort((void *)list,n,sizeof(REF_HASH_TAB *),(int (*)(const void *,const void *))ref_sorter);
 	
+	data_path = "";
+	prev_mnemonic = 0;
+	prev_table = 0;
 	for(ix=0; ix<n; ix++)
 	{
-		int j, ph_tab_num;
+		int j = 0;
 
-		fprintf(f_report,"%3d  %s",list[ix]->n_uses, list[ix]->string);
-		for(j = strlen(list[ix]->string); j < 14; j++)
+		if(strcmp(list[ix]->string, data_path) != 0)
+		{
+			data_path = list[ix]->string;
+			j = strlen(data_path);
+			fprintf(f_report,"%s",data_path);
+		}
+		else
+		{
+			if((list[ix]->ph_table == prev_table) && (list[ix]->ph_mnemonic == prev_mnemonic))
+				continue;   // same phoneme, don't list twice
+		}
+
+		while(j < 14)
 		{
 			fputc(' ',f_report);  // pad filename with spaces
+			j++;
 		}
-		for(j=0; (j < list[ix]->n_used_by) && (j<N_USED_BY); j++)
-		{
-			ph_tab_num = list[ix]->used_by[j];
-			fprintf(f_report," %s",phoneme_tab_list2[ph_tab_num-1].name);
-		}
-		if(j < list[ix]->n_used_by)
-		{
-			fprintf(f_report," ++");
-		}
+
+		fprintf(f_report,"  [%s] %s",WordToString(prev_mnemonic = list[ix]->ph_mnemonic), phoneme_tab_list2[prev_table = list[ix]->ph_table].name);
 		fputc('\n',f_report);
+
+		free(list[ix]);
 	}
 	free(list);
 	fclose(f_report);

src/compiledict.cpp

@@ -163,9 +163,9 @@ int compile_line(char *linebuf, char *dict_line, int *hash)
 	char *p;
 	char *word;
 	char *phonetic;
-	int  ix;
+	unsigned int  ix;
 	int  step;
-	int  n_flag_codes = 0;
+	unsigned int  n_flag_codes = 0;
 	int  flag_offset;
 	int  length;
 	int  multiple_words = 0;
@@ -181,10 +181,11 @@ int compile_line(char *linebuf, char *dict_line, int *hash)
 	unsigned char flag_codes[100];
 	char encoded_ph[200];
 	unsigned char bad_phoneme[4];
+static char nullstring[] = {0};
 
 	comment = NULL;
 	text_not_phonemes = 0;
-	phonetic = word = "";
+	phonetic = word = nullstring;
 
 	p = linebuf;
 //	while(isspace2(*p)) p++;
@@ -1219,7 +1220,7 @@ static int compile_dictrules(FILE *f_in, FILE *f_out, char *fname_temp)
 	int n_rules=0;
 	int count=0;
 	int different;
-	char *prev_rgroup_name;
+	const char *prev_rgroup_name;
 	unsigned int char_code;
 	int compile_mode=0;
 	char *buf;
@@ -1247,16 +1248,6 @@ static int compile_dictrules(FILE *f_in, FILE *f_out, char *fname_temp)
 			if(buf[0] == '\r') buf++;  // ignore extra \r in \r\n 
 		}
 
-		if((buf != NULL) && (memcmp(buf,".L",2)==0))
-		{
-			if(compile_lettergroup(&buf[2], f_out) != 0)
-			{
-				fprintf(f_log,"%5d: Bad lettergroup\n",linenum);
-				error_count++;
-			}
-			continue;
-		}
-
 		if((buf == NULL) || (buf[0] == '.'))
 		{
 			// next .group or end of file, write out the previous group
@@ -1277,10 +1268,21 @@ static int compile_dictrules(FILE *f_in, FILE *f_out, char *fname_temp)
 			{
 				// end of the character replacements section
 				fwrite(&n_rules,1,4,f_out);   // write a zero word to terminate the replacemenmt list
+				compile_mode = 0;
 			}
 
 			if(buf == NULL) break;   // end of file
 
+			if(memcmp(buf,".L",2)==0)
+			{
+				if(compile_lettergroup(&buf[2], f_out) != 0)
+				{
+					fprintf(f_log,"%5d: Bad lettergroup\n",linenum);
+					error_count++;
+				}
+				continue;
+			}
+
 			if(memcmp(buf,".replace",8)==0)
 			{
 				compile_mode = 2;

src/dictionary.cpp

@@ -230,7 +230,7 @@ void Translator::InitGroups(void)
 
 		if(p[0] == RULE_REPLACEMENTS)
 		{
-			pw = (unsigned int *)(((int)p+4) & ~3);  // advance to next word boundary
+			pw = (unsigned int *)(((long)p+4) & ~3);  // advance to next word boundary
 			langopts.replace_chars = pw;
 			while(pw[0] != 0)
 			{
@@ -617,8 +617,8 @@ return(0);
 
 
 
-int Translator::IsLetterGroup(char *word, int group)
-{//=================================================
+int Translator::IsLetterGroup(char *word, int group, int pre)
+{//==========================================================
 	// match the word against a list of utf-8 strings
 	char *p;
 	char *w;
@@ -1484,7 +1484,7 @@ char *Translator::DecodeRule(const char *group, char *rule)
 		
 		if(rb == RULE_ENDING)
 		{
-			static char *flag_chars = "ei vtfq t";
+			static const char *flag_chars = "ei vtfq t";
 			flags = ((rule[0] & 0x7f)<< 8) + (rule[1] & 0x7f);
 			suffix_char = 'S';
 			if(flags & (SUFX_P >> 8))
@@ -1802,9 +1802,9 @@ void Translator::MatchRule(char *word[], const char *group, char *rule, MatchRec
 						failed = 1;
 					break;
 
-				case RULE_LETTERGP2:   // match against a list of utf-t strings
+				case RULE_LETTERGP2:   // match against a list of utf-8 strings
 					letter_group = *rule++ - 'A';
-					if((n_bytes = IsLetterGroup(post_ptr-1,letter_group)) >0)
+					if((n_bytes = IsLetterGroup(post_ptr-1,letter_group,0)) >0)
 					{
 						match.points += (20-distance_right);
 						post_ptr += (n_bytes-1);
@@ -1983,6 +1983,17 @@ void Translator::MatchRule(char *word[], const char *group, char *rule, MatchRec
 						failed = 1;
 					break;
 
+				case RULE_LETTERGP2:   // match against a list of utf-8 strings
+					letter_group = *rule++ - 'A';
+					if((n_bytes = IsLetterGroup(pre_ptr-letter_xbytes,letter_group,1)) >0)
+					{
+						match.points += (20-distance_right);
+						pre_ptr -= (n_bytes-1);
+					}
+					else
+						failed =1;
+					break;
+
 				case RULE_NOTVOWEL:
 					if(!IsLetter(letter_w,0))
 					{
@@ -2345,7 +2356,7 @@ int Translator::TranslateRules(char *p_start, char *phonemes, int ph_size, char
 				}
 			}
 		}
-	
+
 		if(match1.phonemes == NULL)
 			match1.phonemes = "";
 	
@@ -2567,10 +2578,10 @@ char *print_dflags(int flags)
 
 
 
-int Translator::LookupDict2(char *word, char *word2, char *phonetic, unsigned int *flags, int end_flags)
-//======================================================================================================
+const char *Translator::LookupDict2(const char *word, const char *word2, char *phonetic, unsigned int *flags, int end_flags)
+//====================================================================================================================
 /* Find an entry in the word_dict file for a specified word.
-   Returns 1 if an entry is found
+   Returns NULL if no match, else returns 'word_end'
 
 	word   zero terminated word to match
 	word2  pointer to next word(s) in the input text (terminated by space)
@@ -2590,8 +2601,8 @@ int Translator::LookupDict2(char *word, char *word2, char *phonetic, unsigned in
 	int  condition_failed=0;
 	int  n_chars;
 	int  no_phonemes;
-	char *word_end;
-	char *word1;
+	const char *word_end;
+	const char *word1;
 	char word_buf[N_WORD_BYTES];
 
 	word1 = word;
@@ -2801,7 +2812,7 @@ int Translator::LookupDict2(char *word, char *word2, char *phonetic, unsigned in
 				fprintf(f_trans,"Found: %s [%s]  %s\n",word1,ph_decoded,print_dflags(flags1));
 			}
 		}
-		return(1);
+		return(word_end);
 
 	}
 	return(0);
@@ -2818,12 +2829,13 @@ int Translator::LookupDictList(char **wordptr, char *ph_out, unsigned int *flags
 */
 {
 	int  length;
-	int  found;
-	char *word1;
-	char *word2;
+	const char *found;
+	const char *word1;
+	const char *word2;
 	unsigned char c;
 	int  nbytes;
 	int  c2;
+	int  len;
 	char word[N_WORD_BYTES];
 	static char word_replacement[N_WORD_BYTES];
 
@@ -2896,10 +2908,14 @@ int Translator::LookupDictList(char **wordptr, char *ph_out, unsigned int *flags
 				word_replacement[0] = 0;
 				word_replacement[1] = ' ';
 				strcpy(&word_replacement[2],ph_out);   // replacement word, preceded by zerochar and space
+
+				word1 = *wordptr;
 				*wordptr = &word_replacement[2];
 
 				if(option_phonemes == 2)
 				{
+					len = found - word1;
+					memcpy(word,word1,len);   // include multiple matching words
 					fprintf(f_trans,"Replace: %s  %s\n",word,*wordptr);
 				}
 			}
@@ -2917,10 +2933,11 @@ int Translator::LookupDictList(char **wordptr, char *ph_out, unsigned int *flags
 
 
 
-int Translator::Lookup(char *word, char *ph_out)
-{//=============================================
+int Translator::Lookup(const char *word, char *ph_out)
+{//===================================================
 	unsigned int flags;
-	return(LookupDictList(&word,ph_out,&flags,0));
+	char *word1 = (char *)word;
+	return(LookupDictList(&word1,ph_out,&flags,0));
 }
 
 

src/espeak_command.cpp

@@ -490,22 +490,22 @@ int delete_espeak_command( t_espeak_command* the_command)
 	  
 	case ET_VOICE_SPEC:
 	  {
-	    espeak_VOICE* data = &(the_command->u.my_voice_spec);
+		espeak_VOICE* data = &(the_command->u.my_voice_spec);
 
-	    if (data->name)
-	      {
-		free(data->name);
-	      }
+		if (data->name)
+		{
+			free((void *)data->name);
+		}
 
-	    if (data->languages)
-	      {
-		free(data->languages);
-	      }
+		if (data->languages)
+		{
+			free((void *)data->languages);
+		}
 
-	    if (data->identifier)
-	      {
-		free(data->identifier);
-	      }
+		if (data->identifier)
+		{
+			free((void *)data->identifier);
+		}
 	  }
 	  break;
 

src/espeakedit.cpp

@@ -42,7 +42,7 @@
 #include "prosodydisplay.h"
 
 
-static char *about_string = "espeakedit: %s\nAuthor: Jonathan Duddington (c) 2007\n\n"
+static const char *about_string = "espeakedit: %s\nAuthor: Jonathan Duddington (c) 2007\n\n"
 "Licensed under GNU General Public License version 3\n"
 "http://espeak.sourceforge.net/";
 

src/extras.cpp

@@ -308,7 +308,7 @@ void Lexicon_De()
 	char phonemes2[80];
 	WORD_TAB winfo;
 
-	static char *vowels = "aeiouyAEIOUY29@";
+	static const char *vowels = "aeiouyAEIOUY29@";
 
 	wxString fname = wxFileSelector(_T("German Lexicon"),path_dir1,_T(""),_T(""),_T("*"),wxOPEN);
 
@@ -477,7 +477,7 @@ void Lexicon_Ru()
 	int input_length;
 
 	int sfx;
-	char *suffix;
+	const char *suffix;
 	int wlen;
 	int len;
 	int check_root;
@@ -495,7 +495,7 @@ void Lexicon_Ru()
 	static char vowels[] = {0xa3,0xc0,0xc1,0xc5,0xc9,0xcf,0xd1,0xd5,0xd9,0xdc,0};
 
 	typedef struct {
-		char *suffix;
+		const char *suffix;
 		int  syllables;
 	} SUFFIX;
 
@@ -1173,7 +1173,7 @@ int x;
 #define TEXT "Hello world."
 #define TEXT_SSML ("<speak>" TEXT "</speak>")
 static void
-speak(char *text)
+speak(const char *text)
 {
   int result;
   result = espeak_Synth(text, strlen(text) + 1, 0, POS_CHARACTER, 0, espeakSSML, NULL, NULL);
@@ -1216,7 +1216,7 @@ result = espeak_SetVoiceByProperties(&voice);
 }
 
 
-char* text1 = "Hello World2. <audio src=\"here\"> Some text</audio>  This is the second sentence";
+const char* text1 = "Hello World2. <audio src=\"here\"> Some text</audio>  This is the second sentence";
 
 void TestTest(int control)
 {//=======================

src/intonation.cpp

@@ -811,10 +811,14 @@ void Translator::CalcPitches_Tone(int clause_tone)
 	PHONEME_LIST *p;
 	int  ix;
 	int  count_stressed=0;
-	int  count_stressed2=0;
 	int  final_stressed=0;
 
 	int  tone_ph;
+	PHONEME_TAB *tph;
+	PHONEME_TAB *prev_tph;   // forget across word boundary
+	PHONEME_TAB *prevw_tph;  // remember across word boundary
+	PHONEME_TAB *prev2_tph;  // 2 tones previous
+	PHONEME_LIST *prev_p;
 
 	int  pitch_adjust = 13;     // pitch gradient through the clause - inital value
 	int  pitch_decrement = 3;   //   decrease by this for each stressed syllable
@@ -845,6 +849,62 @@ void Translator::CalcPitches_Tone(int clause_tone)
 	}
 
 
+	prev_p = p = &phoneme_list[0];
+	prev_tph = prevw_tph = phoneme_tab[phonPAUSE];
+
+	// perform tone sandhi
+	for(ix=0; ix<n_phoneme_list; ix++, p++)
+	{
+		if((p->newword) && ((option_tone1 & 1)==0))
+		{
+			prev_tph = phoneme_tab[phonPAUSE];  // forget across word boundaries
+		}
+
+		if(p->type == phVOWEL)
+		{
+			tone_ph = p->tone_ph;
+			if(tone_ph == 0)
+				p->tone_ph = tone_ph = LookupPh("11");  // default tone 5
+
+			tph = phoneme_tab[tone_ph];
+
+			// Mandarin
+			if(translator_name == L('z','h'))
+			{
+				if(prev_tph->mnemonic == 0x343132)  // [214]
+				{
+					if(tph->mnemonic == 0x343132)   // [214]
+						prev_p->tone_ph = LookupPh("35");
+					else
+						prev_p->tone_ph = LookupPh("21"); 
+				}
+				if((prev_tph->mnemonic == 0x3135)  && (tph->mnemonic == 0x3135))  //  [51] + [51]
+				{
+					prev_p->tone_ph = LookupPh("53");
+				}
+
+				if(tph->mnemonic == 0x3131)  // [11] Tone 5
+				{
+					// tone 5, change its level depending on the previous tone (across word boundaries)
+					if(prevw_tph->mnemonic == 0x3535)
+						p->tone_ph = LookupPh("22");
+					if(prevw_tph->mnemonic == 0x3533)
+						p->tone_ph = LookupPh("33");
+					if(prevw_tph->mnemonic == 0x343132)
+						p->tone_ph = LookupPh("44");
+
+					// tone 5 is unstressed (shorter)
+					p->tone = 1;   // diminished stress
+				}
+			}
+
+			prev_p = p;
+			prev2_tph = prevw_tph;
+			prevw_tph = prev_tph = tph;
+		}
+	}
+
+	// convert tone numbers to pitch
 	p = &phoneme_list[0];
 	for(ix=0; ix<n_phoneme_list; ix++, p++)
 	{
@@ -852,24 +912,18 @@ void Translator::CalcPitches_Tone(int clause_tone)
 		{
 			tone_ph = p->tone_ph;
 
-			if(p->tone >= 2)
+			if(p->tone >= 0)  // TEST, consider all syllables as stressed
 			{
-				// a stressed syllable
-				if(p->tone >= 4)
+				if(ix == final_stressed)
 				{
-
-					count_stressed2++;
-					if(count_stressed2 == count_stressed)
-					{
-						// the last stressed syllable
-						pitch_adjust = pitch_low;
-					}
-					else
-					{
-						pitch_adjust -= pitch_decrement;
-						if(pitch_adjust <= pitch_low)
-							pitch_adjust = pitch_high;
-					}
+					// the last stressed syllable
+					pitch_adjust = pitch_low;
+				}
+				else
+				{
+					pitch_adjust -= pitch_decrement;
+					if(pitch_adjust <= pitch_low)
+						pitch_adjust = pitch_high;
 				}
 
 				if(tone_ph ==0)

src/numbers.cpp

@@ -19,7 +19,6 @@
 
 #include "StdAfx.h"
 
-
 #include <stdio.h>
 #include <ctype.h>
 #include <stdlib.h>
@@ -211,7 +210,7 @@ int Translator::TranslateRoman(char *word, char *ph_out)
 	unsigned int flags;
 	char number_chars[N_WORD_BYTES];
 
-	static char *roman_numbers = "ixcmvld";
+	static const char *roman_numbers = "ixcmvld";
 	static int roman_values[] = {1,10,100,1000,5,50,500};
  
 	acc = 0;
@@ -504,8 +503,8 @@ int Translator::LookupNum3(int value, char *ph_out, int suppress_null, int thous
 
 
 
-static char *M_Variant(int value)
-{//==============================
+static const char *M_Variant(int value)
+{//====================================
 	// returns M, or perhaps MA for some cases
 	
 	if(((value % 100)>20) || ((value % 100)<10))   // but not teens, 10 to 19

src/readclause.cpp

@@ -38,7 +38,7 @@
 #define N_XML_BUF   256
 
 
-char *xmlbase = "";    // base URL from <speak>
+const char *xmlbase = "";    // base URL from <speak>
 
 int namedata_ix=0;
 int n_namedata = 0;
@@ -454,23 +454,19 @@ static int GetC(void)
 static void UngetC(int c)
 {//======================
 	ungot_char = c;
-
-	if((f_input != NULL) && feof(f_input))
-	{
-//		ungetc(' ',f_input);
-	}
 }
 
 
 
-const char *Translator::LookupSpecial(char *string)
-{//================================================
+const char *Translator::LookupSpecial(const char *string)
+{//======================================================
 	unsigned int flags;
 	char phonemes[55];
 	char phonemes2[55];
 	static char buf[60];
+	char *string1 = (char *)string;
 
-	if(LookupDictList(&string,phonemes,&flags,0))
+	if(LookupDictList(&string1,phonemes,&flags,0))
 	{
 		SetWordStress(phonemes,flags,-1,0);
 		DecodePhonemes(phonemes,phonemes2);
@@ -706,13 +702,13 @@ MNEM_TAB ssmltags[] = {
 
 
 
-static char *VoiceFromStack()
-{//==========================
+static const char *VoiceFromStack()
+{//================================
 // Use the voice properties from the SSML stack to choose a voice, and switch
 // to that voice if it's not the current voice
 	int ix;
 	SSML_STACK *sp;
-	char *v_id;
+	const char *v_id;
 	espeak_VOICE voice_select;
 	char voice_name[40];
 	char language[40];
@@ -1083,7 +1079,7 @@ static int GetVoiceAttributes(wchar_t *pw, int tag_type)
 	wchar_t *name;
 	wchar_t *age;
 	wchar_t *variant;
-	char *new_voice_id;
+	const char *new_voice_id;
 
 	static const MNEM_TAB mnem_gender[] = {
 		{"male", 1},

src/speak_lib.cpp

@@ -365,7 +365,9 @@ static espeak_ERROR Synthesize(unsigned int unique_identifier, const void *text,
 	int length;
 	int finished = 0;
 	int count_buffers = 0;
+#ifdef USE_ASYNC
 	uint32_t a_write_pos=0;
+#endif
 
 #ifdef DEBUG_ENABLED
 	if (text)

src/speak_lib.h

@@ -447,9 +447,9 @@ void espeak_CompileDictionary(const char *path, FILE *log);
 
 // voice table
 typedef struct {
-	char *name;            // a given name for this voice. UTF8 string.
-	char *languages;       // list of pairs of (byte) priority + (string) language (and dialect qualifier)
-	char *identifier;      // the filename for this voice within espeak-data/voices
+	const char *name;      // a given name for this voice. UTF8 string.
+	const char *languages;       // list of pairs of (byte) priority + (string) language (and dialect qualifier)
+	const char *identifier;      // the filename for this voice within espeak-data/voices
 	unsigned char gender;  // 0=none 1=male, 2=female,
 	unsigned char age;     // 0=not specified, or age in years
 	unsigned char variant; // only used when passed as a parameter to espeak_SetVoiceByProperties

src/synthdata.cpp

@@ -35,7 +35,7 @@
 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.29.12  29.Oct.07";
+const char *version_string = "1.29.15  01.Nov.07";
 const int version_phdata  = 0x012901;
 
 int option_device_number = -1;

src/synthesize.cpp

@@ -1541,6 +1541,10 @@ int SpeakNextClause(FILE *f_in, const void *text_in, int control)
 	// entries in the wavegen command queue
 	p_text = translator->TranslateClause(f_text,p_text,&clause_tone,&voice_change);
 
+	translator->CalcPitches(clause_tone);
+	translator->CalcLengths();
+
+	translator->GetTranslatedPhonemeString(translator->phon_out,sizeof(translator->phon_out));
 	if(option_phonemes > 0)
 	{
 		fprintf(f_trans,"%s\n",translator->phon_out);
@@ -1550,8 +1554,6 @@ int SpeakNextClause(FILE *f_in, const void *text_in, int control)
 		phoneme_callback(translator->phon_out);
 	}
 
-	translator->CalcPitches(clause_tone);
-	translator->CalcLengths();
 
 	if(skipping_text)
 	{

src/tr_languages.cpp

@@ -535,7 +535,7 @@ SetLengthMods(tr,3);  // all equal
 		{
 			static int stress_amps_sk[8] = {16,16, 20,20, 20,24, 24,22 };
 			static int stress_lengths_sk[8] = {190,190, 210,210, 0,0, 210,210};
-			static char *sk_voiced = "bdgjlmnrvwzaeiouy";
+			static const char *sk_voiced = "bdgjlmnrvwzaeiouy";
 
 			tr = new Translator();
 			SetupTranslator(tr,stress_lengths_sk,stress_amps_sk);
@@ -647,18 +647,18 @@ SetLengthMods(tr,3);  // all equal
 		break;
 
 	case L('z','h'):
-	case L_qa + 'a':   // Test qaa
 		{
-			static int stress_lengths_qaa[8] = {200,200, 248,248, 248,0, 248,250};
-			static int stress_amps_qaa[] = {16,16, 20,20, 24,24, 24,22 };
+			static int stress_lengths_qaa[8] = {220,150, 230,230, 230,0, 230,250};
+			static int stress_amps_qaa[] = {22,18, 22,22, 22,22, 22,22 };
 			tr = new Translator();
 			SetupTranslator(tr,stress_lengths_qaa,stress_amps_qaa);
 
-			tr->langopts.stress_rule = 0;
+			tr->langopts.stress_rule = 3;   // stress on final syllable of a "word"
+			tr->langopts.stress_flags = 1;          // don't automatically set diminished stress (may be set in the intonation module)
 			tr->langopts.vowel_pause = 0;
 			tr->langopts.intonation = 1;   // Tone language, use  CalcPitches_Tone() rather than CalcPitches()
 			tr->langopts.length_mods0 = tr->langopts.length_mods;  // don't lengthen vowels in the last syllable
-			tr->langopts.tone_numbers = 9;
+			tr->langopts.tone_numbers = 9;   // a number after letters indicates a tone number (eg. pinyin or jyutping)
 			tr->langopts.ideographs = 1;
 			tr->langopts.word_gap = 0x5;   // length of a final vowel is less dependent on the next consonant, don't merge consonant with next word
 		}

src/translate.cpp

@@ -513,8 +513,8 @@ int utf8_out(unsigned int c, char *buf)
 }  // end of utf8_out
 
 
-int utf8_in(int *c, char *buf, int backwards)
-{//==========================================
+int utf8_in(int *c, const char *buf, int backwards)
+{//================================================
 	int c1;
 	int n_bytes;
 	int ix;
@@ -1116,7 +1116,7 @@ int Translator::TranslateWord2(char *word, WORD_TAB *wtab, int pre_pause, int ne
 	int source_ix;
 	int len;
 	int sylimit;        // max. number of syllables in a word to be combined with a preceding preposition
-	char *new_language;
+	const char *new_language;
 	unsigned char bad_phoneme[4];
 
 	len = wtab->length;
@@ -1769,6 +1769,12 @@ void *Translator::TranslateClause(FILE *f_text, const void *vp_input, int *tone_
 		prev_out2 = prev_out;
 		utf8_in(&prev_out,&sbuf[ix-1],1);   // prev_out = sbuf[ix-1];
 
+		if(langopts.tone_numbers && isdigit(prev_out) && IsAlpha(prev_out2))
+		{
+			// tone numbers can be part of a word, consider them as alphabetic
+			prev_out = 'a';
+		}
+
 		if(prev_in2 != 0)
 		{
 			prev_in = prev_in2;
@@ -2270,7 +2276,6 @@ if((c == '/') && (langopts.testing & 2) && isdigit(next_in) && IsAlpha(prev_out)
 
 	prev_clause_pause = clause_pause;
 
-	GetTranslatedPhonemeString(phon_out,sizeof(phon_out));
 	*tone_out = tone;
 
 	new_sentence = 0;

src/translate.h

@@ -178,6 +178,7 @@
 //     match 1 pre 2 post 0     - use common phoneme string
 //     match 1 pre 2 post 3 0   - empty phoneme string
 
+typedef const char *  constcharptr;
 
 typedef struct {
 	int points;
@@ -351,6 +352,7 @@ public:
 	int LoadDictionary(const char *name, int no_error);
 	virtual void CalcLengths();
 	virtual void CalcPitches(int clause_tone);
+	void GetTranslatedPhonemeString(char *phon_out, int n_phon_out);
 	
 	LANGUAGE_OPTIONS langopts;
 	int translator_name;
@@ -385,7 +387,6 @@ private:
 	int TranslateWord2(char *word, WORD_TAB *wtab, int pre_pause, int next_pause);
 	int TranslateLetter(char *letter, char *phonemes, int control);
 	void SetSpellingStress(char *phonemes, int control);
-	void GetTranslatedPhonemeString(char *phon_out, int n_phon_out);
 	void WriteMnemonic(int *ix, int mnem);
 	void MakePhonemeList(int post_pause, int new_sentence);
 	int SubstitutePhonemes(PHONEME_LIST2 *plist_out);
@@ -393,8 +394,8 @@ private:
 	int ReadClause(FILE *f_in, char *buf, unsigned short *charix, int n_buf);
 	int AnnouncePunctuation(int c1, int c2, char *buf, int ix);
 
-	int LookupDict2(char *word, char *word2, char *phonetic, unsigned int *flags, int end_flags);
-	const char *LookupSpecial(char *string);
+	const char *LookupDict2(const char *word, const char *word2, char *phonetic, unsigned int *flags, int end_flags);
+	const char *LookupSpecial(const char *string);
 	const char *LookupCharName(int c);
 	int LookupNum2(int value, int control, char *ph_out);
 	int LookupNum3(int value, char *ph_out, int suppress_null, int thousandplex, int prev_thousands);
@@ -410,7 +411,7 @@ private:
 	void ApplySpecialAttribute(char *phonemes, int dict_flags);
 
 	int IsLetter(int letter, int group);
-	int IsLetterGroup(char *word, int group);
+	int IsLetterGroup(char *word, int group, int pre);
 
 	void CalcPitches_Tone(int clause_tone);
 
@@ -424,7 +425,7 @@ protected:
 
 	int IsVowel(int letter);
 	int LookupDictList(char **wordptr, char *ph_out, unsigned int *flags, int end_flags);
-	int Lookup(char *word, char *ph_out);
+	int Lookup(const char *word, char *ph_out);
 
 	
 	// groups1 and groups2 are indexes into data_dictrules, set up by InitGroups()
@@ -521,7 +522,7 @@ void LoadConfig(void);
 int PhonemeCode(unsigned int mnem);
 void ChangeWordStress(Translator *tr, char *word, int new_stress);
 int TransposeAlphabet(char *text, int offset, int min, int max);
-int utf8_in(int *c, char *buf, int backwards);
+int utf8_in(int *c, const char *buf, int backwards);
 int utf8_out(unsigned int c, char *buf);
 int lookupwchar(const unsigned short *list,int c);
 int Eof(void);

src/transldlg.cpp

@@ -333,13 +333,14 @@ void TranslDlg::OnCommand(wxCommandEvent& event)
 		while((vp != NULL) && (n_ph_list < N_PH_LIST))
 		{
 			vp = translator->TranslateClause(NULL,vp,&clause_tone,NULL);
+			translator->CalcPitches(clause_tone);
+			translator->CalcLengths();
 
+			translator->GetTranslatedPhonemeString(translator->phon_out,sizeof(translator->phon_out));
 			if(clause_count++ > 0)
 				strcat(phon_out," ||");
 			strcat(phon_out,translator->phon_out);
 			t_phonetic->SetValue(wxString(translator->phon_out,wxConvLocal));
-			translator->CalcPitches(clause_tone);
-			translator->CalcLengths();
 
 			if((n_ph_list + n_phoneme_list) >= N_PH_LIST)
 			{

src/voice.h

@@ -67,7 +67,7 @@ extern USHORT voice_pcnt[N_PEAKS+1][3];
 extern voice_t *voice;
 extern int tone_points[10];
 
-char *SelectVoice(espeak_VOICE *voice_select);
+const char *SelectVoice(espeak_VOICE *voice_select);
 voice_t *LoadVoice(const char *voice_name, int control);
 voice_t *LoadVoiceVariant(const char *voice_name, int variant);
 void DoVoiceChange(voice_t *v);

src/voices.cpp

@@ -482,7 +482,7 @@ voice_t *LoadVoice(const char *vname, int control)
 	char translator_name[40];
 	char new_dictionary[40];
 	char phonemes_name[40];
-	char *language_type;
+	const char *language_type;
 	char buf[200];
 	char path_voices[sizeof(path_home)+12];
 	char langname[4];
@@ -1005,7 +1005,7 @@ static int __cdecl VoiceScoreSorter(const void *p1, const void *p2)
 static int ScoreVoice(espeak_VOICE *voice_spec, int spec_n_parts, int spec_lang_len, espeak_VOICE *voice)
 {//======================================================================================================
 	int ix;
-	char *p;
+	const char *p;
 	int c1, c2;
 	int language_priority;
 	int n_parts;
@@ -1146,7 +1146,7 @@ static int SetVoiceScores(espeak_VOICE *voice_select, espeak_VOICE **voices, int
 	int nv;           // number of candidates
 	int n_parts=0;
 	int lang_len=0;
-	char *p;
+	const char *p;
 	espeak_VOICE *vp;
 
 	// count number of parts in the specified language
@@ -1194,7 +1194,7 @@ static espeak_VOICE *SelectVoiceByName(espeak_VOICE **voices, const char *name)
 	int match_fname = -1;
 	int match_fname2 = -1;
 	int match_name = -1;
-	char *id;
+	const char *id;
 	int last_part_len;
 	char last_part[41];
 
@@ -1236,8 +1236,8 @@ static espeak_VOICE *SelectVoiceByName(espeak_VOICE **voices, const char *name)
 
 
 
-char *SelectVoice(espeak_VOICE *voice_select)
-{//==========================================
+char const *SelectVoice(espeak_VOICE *voice_select)
+{//================================================
 // Returns a path within espeak-voices, with a possible +variant suffix
 // variant is an output-only parameter
 	int nv;           // number of candidates
@@ -1525,7 +1525,7 @@ espeak_ERROR SetVoiceByName(const char *name)
 
 espeak_ERROR SetVoiceByProperties(espeak_VOICE *voice_selector)
 {//============================================================
-	char *voice_id;
+	const char *voice_id;
 
 	voice_id = SelectVoice(voice_selector);
 

src/vowelchart.cpp

@@ -47,7 +47,7 @@ extern PHONEME_TAB_LIST phoneme_tab_list[N_PHONEME_TABS];
 
 
 // size of the vowelchart png
-#define WIDTH  1200
+#define WIDTH  1580
 #define HEIGHT 800
 
 
@@ -131,8 +131,8 @@ static double log2a(double x)
 
 static int VowelX(int f2)
 {//======================
-	return(WIDTH - int((log2a(f2) - 9.49)*WIDTH/1.8));
-//	return(1024 - int((log2a(f2) - 9.50)*1020/1.8));
+	return(WIDTH - int((log2a(f2) - 9.40)*WIDTH/1.9));
+//	return(WIDTH - int((log2a(f2) - 9.49)*WIDTH/1.8));
 }
 
 static int VowelY(int f1)
@@ -168,7 +168,7 @@ static void DrawVowel(wxDC *dc, wxString name, int f1, int f2, int f3, int g1, i
 	if(y < 0) y = 0;
 	if(y > (HEIGHT-4)) y= (HEIGHT-4);
 	if(x < 0) x = 0;
-	if(x > (WIDTH-8)) x = (WIDTH-8);
+	if(x > (WIDTH-12)) x = (WIDTH-12);
 
 	r = z;
 	g = 255;
@@ -351,7 +351,7 @@ void VowelChart(int control, char *fname)
 		if((ix % 100) == 0)
 			dc.DrawText(wxString::Format(_T("%d"),ix),1,y);
 	}
-	for(ix=800; ix<=2400; ix+=100)
+	for(ix=700; ix<=2400; ix+=100)
 	{
 		x = VowelX(ix);
 		dc.DrawLine(x,0,x,HEIGHT);

src/wave.cpp

@@ -634,7 +634,7 @@ void wave_init()
 //>
 //<wave_open
 
-void* wave_open(char* the_api)
+void* wave_open(const char* the_api)
 {
   ENTER("wave_open");
   static int once=0;
@@ -1022,7 +1022,7 @@ void *wave_test_get_write_buffer()
 
 
 void wave_init() {}
-void* wave_open(char* the_api) {return (void *)1;}
+void* wave_open(const char* the_api) {return (void *)1;}
 size_t wave_write(void* theHandler, char* theMono16BitsWaveBuffer, size_t theSize) {return theSize;}
 int wave_close(void* theHandler) {return 0;}
 int wave_is_busy(void* theHandler) {return 0;}

src/wave.h

@@ -9,7 +9,7 @@ extern int option_device_number;
 
 extern void wave_init();
 // TBD: the arg could be "alsa", "oss",...
-extern void* wave_open(char* the_api);
+extern void* wave_open(const char* the_api);
 
 extern size_t wave_write(void* theHandler, char* theMono16BitsWaveBuffer, size_t theSize);
 extern int wave_close(void* theHandler);