[1.41.10]

Fixes for ordinal numbers (lang=hu).


git-svn-id: https://espeak.svn.sourceforge.net/svnroot/espeak/trunk@225 d46cf337-b52f-0410-862d-fd96e6ae7743

dictsource/dict_phonemes

@@ -589,3 +589,16 @@ a    e    i    o    u
 g    h    j    J    k    l    m    n    
 N    n^   p    q    R    s    S    S;   
 t    tS   v    w    x    z    
+
+
+Dictionary pa_dict
+
+a    a~   e    E    e~   E~   i    I    
+i:   i~   I~   o    O    o:   O~   o~   
+r-   U    u    u~   U~   V    V~   
+
+:    b    bh   c    ch   d    d.   d.h  
+dh   f    g    gh   H    j    J    Jh   
+k    kh   l    l.   m    n    N    n.   
+n^   p    ph   Q    R    s    S    t    
+t.   t.h  th   v    x    z    

dictsource/en_list

@@ -262,8 +262,8 @@ U+32f	n,0nsI2l'abI2k
 
 // numeric
 
-//_0    zero $text  // TEST
-_0	z'i@roU
+_0    zero $text  // TEST
+//_0	z'i@roU
 _1	w'0n
 ?6 _1	w'Vn
 _2	t'u:
@@ -323,6 +323,7 @@ _70o	s'Ev@nti@
 _80o	'eIti@
 _90o	n'aInti@
 _0Co	h'Vndr@dT
+_0M1o	T'aUz@ndT
 
 
 // ABBREVIATIONS

dictsource/hu_list

@@ -54,8 +54,9 @@ _0Z4	ti:zEzR2Ed
 // ordinal numbers
 
 _ord	Edik
+_0o	n'ullAdik
 _1ox	ElSY:     // number = '1' only
-_2ox	ma:Sodik  // number = '2' only
+_2ox	ma:Sodik // number = '2' only
 _1o	EJ:Edik
 _2o	kEt:Edik
 _3o	hAR2_mAdik
@@ -72,6 +73,9 @@ _60o	hAtvAnAdik
 _80o	n^oltsvAnAdik
 _0Co	sa:zAdik
 _2Co	ke:tsa:zAdik
+_0M1o	EzR2Edik
+_1M1o	EzR2Edik
+_2M1o	ke:tEzR2Edik
 
 // These words mean that a dot after number, immediately preceding, does not mean an ordinal number
 január		$alt
@@ -100,6 +104,15 @@ okt	$alt
 now	$alt
 dec	$alt
 
+//this following 7 lines means the 7 hungarian day name
+hétfő	$alt
+kedd	$alt
+szerda	$alt
+csütörtök	$alt
+péntek	$alt
+szombat	$alt
+vasárnap	$alt
+
 // accent names
 _lig	ligAtu:R2A
 _acu	e:lES
@@ -129,7 +142,7 @@ mm      milime:tER2
 a.m	An^n^i||m'int  $dot
 dr	doktoR2  $dot
 gpu	$abbrev
-mvgyosz $abbrev
+mvgyosz Emve:Je:oEs
 OTP $abbrev
 id	idY:SEb:  $dot
 ill	illEtvE  $dot

dictsource/hu_rules

@@ -21,7 +21,7 @@ _) alattvaló _!'AlAtvAlo:
 
 .group b
         b        b
-//        bb       b:
+bb b:       
 
         biz (A   b'iz,       // bizalmas  etc.
 	biedermeier  bi:dER2ma:jER2
@@ -32,7 +32,8 @@ _) alattvaló _!'AlAtvAlo:
         ccs      tS:
      !) cz       ts         // in names which start with a capital letter
 ar) csz (e ts
-  anar) ch (i    C
+harmin) c ts
+anar) ch (i    C
 hierar) ch (i    C
     me) ch (a    C
     te) ch (n    C
@@ -71,13 +72,16 @@ pá) c ts
 
 .group d
         d        d
+dd d:
 dts tS:
 dt t:
 a) dsz ts:
 cselé) d d
 enge) dsz ts:
 engedelmeske) dsz ts:
+hazu) dsz ts:
 mara) dsz  ts:
+tu) dsz ts:
 kére) dz ts
 //      dd       d:
         dz       dz
@@ -98,6 +102,7 @@ kére) dz ts
 dj J:
 min) dny n^
 
+
 .group e
         e        E
   D_-_) es (_    %ES
@@ -115,7 +120,7 @@ _) egyezség _!'EJ:ESSe:g
 
 .group g
         g        g
-// gg       g:
+gg       g:
         gy       J
         ggy      J:
 
@@ -125,7 +130,7 @@ _) egyezség _!'EJ:ESSe:g
      A) gysz (A  Js:
      C) gyj      J
      A) gyj (A   J:
-
+ha) gyj J:
     _e) gy (es   J:
     _e) gy (et_  J:
     _e) gy (etlen  J:
@@ -175,6 +180,7 @@ _) kétség  ke:tS:e:g
 
 .group l
         l        l
+lj j
         ly       j
         lly      jj
 
@@ -182,9 +188,12 @@ _) kétség  ke:tS:e:g
 á) ll (j j
 beszé) lj  jj
 bére) lj jj
+fáj) l l
+fájla) l (j jj
 gondo) lj jjj
 ke) lj jj
-sajná) lj jj
+sajná) lj jjj
+llj jjj
 C) ly (_    li
  _kéth) ly       li
 szamue) lly	 lli
@@ -259,7 +268,7 @@ r        R2
         sz       s
         ssz      ss2
 szts stS:
-
+föld) s (ánc S
 s (színű S
    hel) s (inki  z
     ki) s (ebb   SS
@@ -296,18 +305,27 @@ pénze) s S
         tc       ts:
         tt       t:
         tt (C    tt
+ttn t:n
         ty       c
 //     s) ty       c:    //
 //     z) ty       c:    // keztyű
         ttj      tc:
         tty      c:
+apá) ts (ág tS:
+becsüle) t t
+cson) t t
+cson) tj c
+ezüs) t t
 éle) t t
 
+felej) ts tS
+já) tsz (ani ts:
 tsz ts:
 szorí) ts (a tS:
 ürí) ts (e tS:
 mula) ts tS:
 nemze) ts tS:
+néme) ts (ég tS:
 min) t t
 tse tSE
 tso tSo
@@ -315,6 +333,7 @@ tsa tSA
 tsá  tSa:
 tsd  tSd
 tsé tSe:
+t (cs t
 ttsé tS:e:
 C) tj       c
      A) tj (A    c:
@@ -324,12 +343,14 @@ C) tj       c
    _ka) ty (n    ti
 
 vörösmar) ty     ti
+á) t (sza t
 á) t (jár t
 á) t (sző t
 á) t (szú t
 á) t (szű t
 bizo) tts (ág tS:
 állí) ts tS:
+állapo) t (sor t
 bará) ts tS:
 dön) ts (ön tS
 elhivato) tts (ág tS:
@@ -337,10 +358,16 @@ folyama) t (jel t
 kiál) ts tS
 köve) ts (ég  tS:
 kür) t t
+kür) tj c
+korlátozo) tts tS:
 neve) ts (ég tS:
+on) ts (u tS
 szen)t (szék t
+szé) t t
 szöve) ts (ég  tS:
+pillan) ts tS
 romlo) tts tS:
+tar) ts tS
 á) t (sor t
 á) t (sé t
 ne) t (c t
@@ -358,6 +385,7 @@ lé) t (szám t
 ké) t (száz t
 ö) t (száz t
 ha) t (száz t
+hé) t (száz t
 vé) ts (ég tS:
 
 
@@ -399,6 +427,7 @@ befeje) z (te s
 bi) z (tons s    
 bi) z (tos s
 bron) z (sz z
+csontvá) z z
 e) z  (t s
 ho) z (ta s
 ho) z (tá s

dictsource/ta_list

@@ -1,4 +1,5 @@
- 
+//_xx  விழுக்காடு   $text  // TESTING doesn't reduce vowels
+
 // This file is UTF8 encoded 
 // Spelling to phoneme words and exceptions for Tamil
 
@@ -76,8 +77,8 @@ _!	a:ctS:Vr,ijVkk,URi
 U+bf9	ru:ba:j
 
 // abbreviations
-ரூ	ru:ba:j   $dot
-Rs	ru:ba:j   $dot
+ரூ	ru:ba:j  $dot
+Rs	ru:pi:z  $dot
 
 // numbers
 _0	suz.ijVm    // சுழியம்
@@ -139,20 +140,23 @@ _1M1x	a:jirVm
 _0M1	a:jirVttU
 _1M1	a:jirVttU
 
-_0M2x	lVd.tSVm
+_0M2x	lVd.tSVm	// 100,000
 _1M2x	orUlVd.tSVm
 _0M2	lVd.tSVttU
 _1M2	orUlVd.tSVttU
 
-_0M3x	ko:d.i
+_0M3x	ko:d.i		// 10,000,000
 _1M3x	orUko:d.i
 _0M3	ko:d.ie:
 _1M3	orUko:d.ie:
 
-_0M4	nu:RUko:d.i   // not correct, but should be understandable
-_1M4	nu:RUko:d.i
-_0M5	patta:jiRUmko:d.i
-_1M5	patta:jiRUmko:d.i
+_0M4x	a:jirVmko:d.i	// 10,000,000,000
+_1M4x	a:jirVmko:d.i
+_0M4	a:jirVmko:d.ie:
+_1M4	a:jirVmko:d.ie:
+
+_0M5	a:jirVma:jirVmko:d.i	// 10,000,000,000,000
+_1M5	a:jirVma:jirVmko:d.i
 	
 
 _dpt	_pul.l.i

dictsource/ta_rules

@@ -16,8 +16,11 @@
 ௭	7
 ௮	8
 ௯	9
+ொ	ொ
+ோ	ோ
+ௌ	ௌ
+
 
-.
 .group 0xe0ae   // characters which start with UTF-8 bytes: [e0 ae]
 
 	ஂ		// anusvara
@@ -39,7 +42,7 @@
 	எ	e
      _) எ	;e     // add a short [j] sound at start of word ?
 
-	ஏ	e::
+	ஏ	e:
      _) ஏ	;e::     // add a short [j] sound at start of word ?
 
 	ஐ	aI
@@ -65,12 +68,14 @@
 	ங	NV
 	ங (B	N
 
-	ச	zV
-	ச (B	z
+	ச	sV    // ?? [z]
+	ச (B	s
      _) ச	sa
      _) ச (B	s
 	ச்ச	tS:V
 	ச்ச (B	tS:
+	ற்ச	tS:
+	ற்ச (B	tS:
      ட்) ச	tSV
      ட்) ச (B	tS
      ஞ்) ச	dZV
@@ -112,7 +117,7 @@
 	ப்ப	ppV
 	ப்ப (B	pp
      ட்)	ப	pV
-     ட்)	ப (B	pV
+     ட்)	ப (B	p
      ற்) ப	pV
      ற்) ப (B	p
      ஃ) ப	fV
@@ -184,6 +189,7 @@
 
 	ௌ	aU
 
+
 	்		// virama
 
 	ௗ	:	// aU length mark

phsource/compile_report

@@ -1,4 +1,4 @@
-60 phoneme tables
+61 phoneme tables
           new total
     base 103  103
    base2  26  124
@@ -17,8 +17,8 @@
       fi  40  134
       fr  55  141
    fr_ca  11  141
-      hi  60  149
-      ta  20  152
+      hi  62  151
+      ta  20  154
       hu  23  119
       lv  29  126
       nl  28  126
@@ -53,13 +53,14 @@
       hy  24  119
       da  21  118
       rw  15  131
-      ml  13  151
-      kn  15  151
-      bn  59  155
-      ne  18  157
-      mr  12  149
+      ml  13  153
+      kn  15  153
+      bn  59  157
+      ne  18  159
+      mr  12  151
       eu   6  125
       mn  15  114
+      pa  12  152
 
 Data file      Used by
 b/b             [b] base
@@ -1341,7 +1342,8 @@ vnasal/e_n      [e~] af
 vnasal/i_n      [i~] pt
                 [i~] bn
                 [i:~] bn
-vnasal/i_n2     [i~] hi
+vnasal/i_n2     [I~] hi
+                [i~] hi
 vnasal/m-       [m-] sw
 vnasal/n-       [n-] sw
 vnasal/nn-      [N-] sw
@@ -1356,7 +1358,8 @@ vnasal/oo_n2    [O~] hi
                 [o] zh
                 [O~] bn
 vnasal/oo_n3    [O~] pl
-vnasal/u_n      [u~] hi
+vnasal/u_n      [U~] hi
+                [u~] hi
                 [u~] pt
                 [u] zh
                 [u~] bn
@@ -1634,6 +1637,7 @@ vowel/aa_9      [a] fi
                 [a:] hi
                 [a:] bn
                 [a] ne
+                [a] pa
 vowel/a_en      [A] fr
 vowel/@_bck     [@] hi
                 [@/] hi
@@ -1641,6 +1645,7 @@ vowel/@_bck     [@] hi
                 [@] bn
                 [V] ne
                 [@/] ne
+                [@] pa
 vowel/e         [e] base2
                 [e:] en
                 [eI] en_n
@@ -1681,6 +1686,7 @@ vowel/e_2       [eI] en_sc
                 [e:] no
                 [e] bn
                 [e:] bn
+                [e] pa
 vowel/e_3       [i] en_n
                 [e:] hu
                 [e] ku
@@ -1707,6 +1713,7 @@ vowel/ee_2      [E] en
                 [E] zh
                 [E#] ku
                 [&] da
+                [E] pa
 vowel/ee#_2     [E-] sv
                 [E#] sq
 vowel/ee_3      [&] af
@@ -1833,6 +1840,7 @@ vowel/i_fnt     [i:] en_wi
                 [i] bn
                 [i:] bn
                 [i:] mr
+                [i] pa
 vowel/ii        [I] en_n
                 [I2] en_n
                 [I] en_rp
@@ -1864,6 +1872,7 @@ vowel/ii_3      [I] cy
                 [I] no
                 [I] tr
                 [I] bn
+                [I] pa
 vowel/ii#_3     [I2] en_us
 vowel/ii_4      [I] en
                 [I2] en
@@ -1901,6 +1910,8 @@ vowel/@_low     [3] en_rp
                 [@/] ne
                 [@] mr
                 [V] mr
+                [@] pa
+                [V] pa
 vowel/@_low2    [@/] en_us
                 [@2] en_us
 vowel/o         [o] base2
@@ -1937,6 +1948,7 @@ vowel/o_2       [o:] cy
                 [o] vi
                 [o] da
                 [o:] bn
+                [o:] pa
 vowel/o-_2      [V] en_n
                 [V] en_wm
 vowel/o_3       [oU] en_sc
@@ -2005,6 +2017,7 @@ vowel/oo_4      [O] base2
                 [O:] hi
                 [O] it
                 [O] bn
+                [O] pa
 vowel/oo_5      [O] pl
                 [O] is
                 [O] sq
@@ -2095,6 +2108,7 @@ vowel/u_bck     [u] base2
                 [U] mr
                 [u:] mr
                 [u] mn
+                [u] pa
 vowel/u_bck2    [u] fr
                 [u:] fr
                 [u:] la
@@ -2120,6 +2134,7 @@ vowel/uu_bck    [U] en_wi
                 [u] zhy
                 [U] bn
                 [U] mn
+                [U] pa
 vowel/V         [3] en_sc
 vowel/V_2       [V] en
                 [a] af
@@ -2130,6 +2145,7 @@ vowel/V_3       [V] en_rp
                 [V] hi
                 [V] ta
                 [V] bn
+                [V] pa
 vowel/V_4       [V] en_sc
                 [V] da
 vowel/V_6       [V] en_us

phsource/ph_hindi

@@ -166,6 +166,12 @@ phoneme i~
   formants vnasal/i_n2
 endphoneme
 
+phoneme I~
+  vowel starttype (i) endtype (i)
+  length 170
+  formants vnasal/i_n2
+endphoneme
+
 phoneme e~
   vowel long starttype (e) endtype (e)
   length 220
@@ -173,7 +179,7 @@ phoneme e~
 endphoneme
 
 phoneme E~ 
-  vowel  starttype (e)  endtype (e)
+  vowel long starttype (e)  endtype (e)
   length 230
   formants vnasal/ee_n2
 endphoneme
@@ -208,6 +214,12 @@ phoneme u~
   formants vnasal/u_n
 endphoneme
 
+phoneme U~
+  vowel starttype (u) endtype (u)
+  length 170
+  formants vnasal/u_n
+endphoneme
+
 
 phoneme r-
   vowel  starttype (@)  endtype (@)

phsource/ph_tamil

@@ -39,7 +39,7 @@ endphoneme
 
 phoneme e:
   vowel  starttype (e)  endtype (e)
-  length 270
+  length 250
   formants vowel/e
 endphoneme
 

phsource/phonemes

@@ -1369,3 +1369,5 @@ include ph_basque
 phonemetable mn base
 include ph_mongolian
 
+phonemetable pa hi
+include ph_punjabi

src/compiledict.cpp

@@ -49,6 +49,7 @@ static int transpose_min;
 static int transpose_max;
 static int text_mode = 0;
 static int debug_flag = 0;
+static int error_need_dictionary = 0;
 
 static int hash_counts[N_HASH_DICT];
 static char *hash_chains[N_HASH_DICT];
@@ -223,7 +224,6 @@ static int compile_line(char *linebuf, char *dict_line, int *hash)
 static char nullstring[] = {0};
 
 	WORD_TAB winfo;
-	char decoded_phonemes[128];
 
 	comment = NULL;
 	text_not_phonemes = 0;
@@ -432,23 +432,37 @@ step=1;  // TEST
 	if(text_mode)
 		text_not_phonemes = 1;
 
-	if(text_not_phonemes != translator->langopts.textmode)
-	{
-		flag_codes[n_flag_codes++] = BITNUM_FLAG_TEXTMODE;
-	}
-
 	if(text_not_phonemes)
 	{
 		if(word[0] == '_')
 		{
 			// This is a special word, used by eSpeak.  Translate this into phonemes now
-//			memset(&winfo,0,sizeof(winfo));
-//			TranslateWord(translator,phonetic,0,&winfo);    // but *_dict is not loaded ?
-//			DecodePhonemes(word_phonemes,decoded_phonemes);
+			memset(&winfo,0,sizeof(winfo));
+			strcat(phonetic, " ");     // need a space to indicate word-boundary
+
+	// PROBLEM  vowel reductions are not applied to the translated phonemes
+	// condition rules are not applied
+			TranslateWord(translator,phonetic,0,&winfo);
+			text_not_phonemes = 0;
+			strncpy0(encoded_ph, word_phonemes, N_WORD_BYTES-4);
+
+			if((word_phonemes[0] == 0) && (error_need_dictionary < 3))
+			{
+				// the dictionary was not loaded, we need a second attempt
+				error_need_dictionary++;
+				fprintf(f_log,"%5d: Need to compile dictionary again\n",linenum);
+			}
+{
+//char decoded_phonemes[128];
+//DecodePhonemes(word_phonemes,decoded_phonemes);
 //printf("Translator %x  %s  [%s] [%s]\n",translator->translator_name,word,phonetic,decoded_phonemes);
+}
+		}
+		else
+		{
+			// this is replacement text, so don't encode as phonemes. Restrict the length of the replacement word
+			strncpy0(encoded_ph,phonetic,N_WORD_BYTES-4);
 		}
-		// this is replacement text, so don't encode as phonemes. Restrict the length of the replacement word
-		strncpy0(encoded_ph,phonetic,N_WORD_BYTES-4);
 	}
 	else
 	{
@@ -473,6 +487,12 @@ step=1;  // TEST
 		}
 	}
 
+	if(text_not_phonemes != translator->langopts.textmode)
+	{
+		flag_codes[n_flag_codes++] = BITNUM_FLAG_TEXTMODE;
+	}
+
+
 	if(sscanf(word,"U+%x",&wc) == 1)
 	{
 		// Character code
@@ -1599,6 +1619,7 @@ int CompileDictionary(const char *dsource, const char *dict_name, FILE *log, cha
 	char path[sizeof(path_home)+40];       // path_dsource+20
 
 	error_count = 0;
+	error_need_dictionary = 0;
 	memset(letterGroupsDefined,0,sizeof(letterGroupsDefined));
 
 	debug_flag = flags & 1;

src/dictionary.cpp

@@ -241,9 +241,6 @@ int LoadDictionary(Translator *tr, const char *name, int no_error)
 
 	strcpy(dictionary_name,name);   // currently loaded dictionary name
 
-	if(no_error)   // don't load dictionary, just set the dictionary_name
-		return(1);
-
 	// Load a pronunciation data file into memory
 	// bytes 0-3:  offset to rules data
 	// bytes 4-7:  number of hash table entries

src/espeakedit.cpp

@@ -45,12 +45,12 @@
 
 
 #ifdef deleted
-static const char *about_string = "espeakedit: %s\nAuthor: Jonathan Duddington (c) 2007\n\n"
+static const char *about_string = "espeakedit: %s\nAuthor: Jonathan Duddington (c) 2009\n\n"
 "Licensed under GNU General Public License version 3\n"
 "http://espeak.sourceforge.net/";
 #endif
 
-static const char *about_string = "<font size=0><b>espeakedit </b> %s<br>Author: Jonathan Duddington (c) 2007<br>"
+static const char *about_string = "<font size=0><b>espeakedit </b> %s<br>Author: Jonathan Duddington (c) 2009<br>"
 "<a href=\"http://espeak.sourceforge.net/\">http://espeak.sourceforge.net</a><br>"
 "Licensed under <a href=\"http://espeak.sourceforge.net/license.html\">GNU General Public License version 3</a></font>";
 
@@ -585,6 +585,7 @@ void MyFrame::OnTools(wxCommandEvent& event)
 		sprintf(fname_log,"%s%s",path_dsource,"dict_log");
 		log = fopen(fname_log,"w");
 
+		LoadDictionary(translator, dictionary_name, 0);
 		if((err = CompileDictionary(path_dsource,dictionary_name,log,err_fname,debug_flag)) < 0)
 		{
 			wxLogError(_T("Can't access file:\n")+wxString(err_fname,wxConvLocal));

src/numbers.cpp

@@ -576,7 +576,7 @@ void LookupLetter(Translator *tr, unsigned int letter, int next_byte, char *ph_b
 	ph_stress[0] = phonSTRESS_P;
 	ph_stress[1] = 0;
 
-	for(p=(unsigned char *)ph_buf3; *p != 0; p++)
+	for(p=(unsigned char *)ph_buf3; (*p != 0) && (phoneme_tab[*p] != NULL); p++)
 	{
 		if(phoneme_tab[*p]->type == phSTRESS)
 			ph_stress[0] = 0;  // stress is already marked
@@ -865,6 +865,7 @@ static const char *M_Variant(int value)
 
 static int LookupThousands(Translator *tr, int value, int thousandplex, int thousands_exact, char *ph_out)
 {//=======================================================================================================
+// thousands_exact:  bit 0  no hundreds,tens,or units,  bit 1  ordinal numberr
 	int found;
 	int found_value=0;
 	char string[12];
@@ -874,11 +875,20 @@ static int LookupThousands(Translator *tr, int value, int thousandplex, int thou
 	ph_of[0] = 0;
 
 	// first look for a match with the exact value of thousands
-	if(thousands_exact)
+	if(thousands_exact & 1)
 	{
-		// is there a different pronunciation if there are no hundreds,tens,or units ? (LANG=ta)
-		sprintf(string,"_%dM%dx",value,thousandplex);
-		found_value = Lookup(tr, string, ph_thousands);
+		if(thousands_exact & 2)
+		{
+			// ordinal number
+			sprintf(string,"_%dM%do",value,thousandplex);
+			found_value = Lookup(tr, string, ph_thousands);
+		}
+		if(!found_value)
+		{
+			// is there a different pronunciation if there are no hundreds,tens,or units ? (LANG=ta)
+			sprintf(string,"_%dM%dx",value,thousandplex);
+			found_value = Lookup(tr, string, ph_thousands);
+		}
 	}
 	if(found_value == 0)
 	{
@@ -894,11 +904,20 @@ static int LookupThousands(Translator *tr, int value, int thousandplex, int thou
 		}
 
 		found = 0;
-		if(thousands_exact)
+		if(thousands_exact & 1)
 		{
-			// is there a different pronunciation if there are no hundreds,tens,or units ?
-			sprintf(string,"_%s%dx",M_Variant(value), thousandplex);
-			found = Lookup(tr, string, ph_thousands);
+			if(thousands_exact & 2)
+			{
+				// ordinal number
+				sprintf(string,"_%s%do",M_Variant(value), thousandplex);
+				found = Lookup(tr, string, ph_thousands);
+			}
+			if(!found)
+			{
+				// is there a different pronunciation if there are no hundreds,tens,or units ?
+				sprintf(string,"_%s%dx",M_Variant(value), thousandplex);
+				found = Lookup(tr, string, ph_thousands);
+			}
 		}
 		if(found == 0)
 		{
@@ -994,7 +1013,7 @@ static int LookupNum2(Translator *tr, int value, int control, char *ph_out)
 		{
 			units = (value % 10);
 
-			if((control & 1) && ((units == 0) || (tr->langopts.numbers & 0x10)))
+			if((control & 1) && ((units == 0) || (tr->langopts.numbers & NUM_SWAP_TENS)))
 			{
 				sprintf(string,"_%dXo",value / 10);
 				if(Lookup(tr, string, ph_tens) != 0)
@@ -1026,7 +1045,7 @@ static int LookupNum2(Translator *tr, int value, int control, char *ph_out)
 					sprintf(string,"_%df",units);
 					found = Lookup(tr, string, ph_digits);
 				}
-				if((control & 1) && ((tr->langopts.numbers & 0x10) == 0))
+				if((control & 1) && ((tr->langopts.numbers & NUM_SWAP_TENS) == 0))
 				{
 					// ordinal
 					sprintf(string,"_%do",units);
@@ -1046,16 +1065,16 @@ static int LookupNum2(Translator *tr, int value, int control, char *ph_out)
 
 	if((control & 1) && (found_ordinal == 0) && (ph_ordinal[0] == 0))
 	{
-		if((value >= 20) && (((value % 10) == 0) || (tr->langopts.numbers & 0x10)))
+		if((value >= 20) && (((value % 10) == 0) || (tr->langopts.numbers & NUM_SWAP_TENS)))
 			Lookup(tr, "_ord20", ph_ordinal);
 		if(ph_ordinal[0] == 0)
 			Lookup(tr, "_ord", ph_ordinal);
 	}
 
-	if((tr->langopts.numbers & 0x30) && (ph_tens[0] != 0) && (ph_digits[0] != 0))
+	if((tr->langopts.numbers & (NUM_SWAP_TENS | NUM_AND_UNITS)) && (ph_tens[0] != 0) && (ph_digits[0] != 0))
 	{
 		Lookup(tr, "_0and", ph_and);
-		if(tr->langopts.numbers & 0x10)
+		if(tr->langopts.numbers & NUM_SWAP_TENS)
 			sprintf(ph_out,"%s%s%s%s",ph_digits, ph_and, ph_tens, ph_ordinal);
 		else
 			sprintf(ph_out,"%s%s%s%s",ph_tens, ph_and, ph_digits, ph_ordinal);
@@ -1063,7 +1082,7 @@ static int LookupNum2(Translator *tr, int value, int control, char *ph_out)
 	}
 	else
 	{
-		if(tr->langopts.numbers & 0x200)
+		if(tr->langopts.numbers & NUM_SINGLE_VOWEL)
 		{
 			// remove vowel from the end of tens if units starts with a vowel (LANG=Italian)
 			if(((ix = strlen(ph_tens)-1) >= 0) && (ph_digits[0] != 0))
@@ -1078,7 +1097,7 @@ static int LookupNum2(Translator *tr, int value, int control, char *ph_out)
 		sprintf(ph_out,"%s%s%s",ph_tens, ph_digits, ph_ordinal);
 	}
 
-	if(tr->langopts.numbers & 0x100)
+	if(tr->langopts.numbers & NUM_SINGLE_STRESS)
 	{
 		// only one primary stress
 		found = 0;
@@ -1107,6 +1126,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 	int tensunits;
 	int x;
 	int exact;
+	int ordinal;
 	char string[12];  // for looking up entries in **_list
 	char buf1[100];
 	char buf2[100];
@@ -1117,6 +1137,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 	char ph_hundred_and[12];
 	char ph_thousand_and[12];
 	
+	ordinal = control & 2;
 	hundreds = value / 100;
 	tensunits = value % 100;
 	buf1[0] = 0;
@@ -1127,7 +1148,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 		ph_thousand_and[0] = 0;
 
 		found = 0;
-		if((control & 2) && (tensunits == 0))
+		if(ordinal && (tensunits == 0))
 		{
 			// ordinal number, with no tens or units
 			found = Lookup(tr, "_0Co", ph_100);
@@ -1137,7 +1158,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 			Lookup(tr, "_0C", ph_100);
 		}
 
-		if(((tr->langopts.numbers & 0x0800) != 0) && (hundreds == 19))
+		if(((tr->langopts.numbers & NUM_1900) != 0) && (hundreds == 19))
 		{
 			// speak numbers such as 1984 as years: nineteen-eighty-four
 //			ph_100[0] = 0;   // don't say "hundred", we also need to surpess "and"
@@ -1151,7 +1172,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 			if ((value % 1000) == 0)
 				exact = 1;
 
-			if(LookupThousands(tr, hundreds / 10, thousandplex+1, exact, ph_10T) == 0)
+			if(LookupThousands(tr, hundreds / 10, thousandplex+1, exact | ordinal, ph_10T) == 0)
 			{
 				x = 0;
 				if(tr->langopts.numbers2 & (1 << (thousandplex+1)))
@@ -1173,7 +1194,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 		ph_digits[0] = 0;
 		if(hundreds > 0)
 		{
-			if((tr->langopts.numbers & 0x100000) && ((control & 1) || (ph_thousands[0] != 0)))
+			if((tr->langopts.numbers & NUM_AND_HUNDRED) && ((control & 1) || (ph_thousands[0] != 0)))
 			{
 				Lookup(tr, "_0and", ph_thousand_and);
 			}
@@ -1184,8 +1205,18 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 			if(tensunits == 0)
 			{
 				// is there a special pronunciation for exactly n00 ?
-				sprintf(string,"_%dC0",hundreds);
-				found = Lookup(tr, string, ph_digits);
+
+				if(ordinal)
+				{
+					// ordinal number
+					sprintf(string, "_%dCo", hundreds);
+					found = Lookup(tr, string, ph_digits);
+				}
+				if(!found)
+				{
+					sprintf(string,"_%dC0",hundreds);
+					found = Lookup(tr, string, ph_digits);
+				}
 			}
 			if(!found)
 			{
@@ -1210,7 +1241,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 	}
 
 	ph_hundred_and[0] = 0;
-	if((tr->langopts.numbers & 0x40) && (tensunits != 0))
+	if((tr->langopts.numbers & NUM_HUNDRED_AND) && (tensunits != 0))
 	{
 		if((value > 100) || ((control & 1) && (thousandplex==0)))
 		{
@@ -1227,7 +1258,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 		if(thousandplex==0)
 		{
 			x = 2;   // allow "eins" for 1 rather than "ein"
-			if(control & 2)
+			if(ordinal)
 				x = 3;   // ordinal number
 			if((value < 100) && !(control & 1))
 				x |= 4;   // tens and units only, no higher digits
@@ -1240,7 +1271,7 @@ static int LookupNum3(Translator *tr, int value, char *ph_out, int suppress_null
 
 		if(LookupNum2(tr, tensunits, x, buf2) != 0)
 		{
-			if(tr->langopts.numbers & 0x80)
+			if(tr->langopts.numbers & NUM_SINGLE_AND)
 				ph_hundred_and[0] = 0;  // don't put 'and' after 'hundred' if there's 'and' between tens and units
 		}
 	}
@@ -1293,7 +1324,7 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 	value = this_value = atoi(word);
 
 	ph_ordinal2[0] = 0;
-	if((tr->langopts.numbers & 0x10000) && (word[ix] == '.') && !isdigit(word[ix+2]))
+	if((tr->langopts.numbers & NUM_ORDINAL_DOT) && (word[ix] == '.') && !isdigit(word[ix+2]))
 	{
 		// ordinal number is indicated by dot after the number
 		ordinal = 2;
@@ -1352,7 +1383,7 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 		prev_thousands = 1;
 	}
 	else
-	if((tr->langopts.thousands_sep == ' ') || (tr->langopts.numbers & 0x1000))
+	if((tr->langopts.thousands_sep == ' ') || (tr->langopts.numbers & NUM_ALLOW_SPACE))
 	{
 		// thousands groups can be separated by spaces
 		if((n_digits == 3) && isdigit(word[-2]))
@@ -1373,7 +1404,7 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 		}
 	}
 
-	if((tr->langopts.numbers & 0x1000) && (word[n_digits] == ' '))
+	if((tr->langopts.numbers & NUM_ALLOW_SPACE) && (word[n_digits] == ' '))
 		thousands_inc = 1;
 	else
 	if(word[n_digits] == tr->langopts.thousands_sep)
@@ -1462,9 +1493,9 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 			max_decimal_count = 2;
 			switch(decimal_mode = (tr->langopts.numbers & 0xe000))
 			{
-			case 0x8000:
+			case NUM_DFRACTION_4:
 				max_decimal_count = 5;
-			case 0x4000:
+			case NUM_DFRACTION_2:
 				// French/Polish decimal fraction
 				while(word[n_digits] == '0')
 				{
@@ -1481,8 +1512,8 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 				}
 				break;
 
-			case 0x2000:   // italian, say "hundredths" is leading zero
-			case 0xa000:   // hungarian, always say "tenths" etc.
+			case NUM_DFRACTION_1:   // italian, say "hundredths" is leading zero
+			case NUM_DFRACTION_5:   // hungarian, always say "tenths" etc.
 				if(decimal_count <= 4)
 				{
 					LookupNum3(tr, atoi(&word[n_digits]), ph_buf, 0,0,0);
@@ -1500,7 +1531,7 @@ static int TranslateNumber_1(Translator *tr, char *word, char *ph_out, unsigned
 				}
 				break;
 
-			case 0x6000:
+			case NUM_DFRACTION_3:
 				// Romanian decimal fractions
 				if((decimal_count <= 4) && (word[n_digits] != '0'))
 				{
@@ -1560,7 +1591,7 @@ int TranslateNumber(Translator *tr, char *word1, char *ph_out, unsigned int *fla
 	if(option_sayas == SAYAS_DIGITS1)
 		return(0);  // speak digits individually
 
-	if((tr->langopts.numbers & 0x3) == 1)
+	if(tr->langopts.numbers != 0)
 		return(TranslateNumber_1(tr, word1, ph_out, flags, wflags));
 
 	return(0);

src/readclause.cpp

@@ -67,8 +67,10 @@ static const char *punct_stop = ".:!?";    // pitch fall if followed by space
 static const char *punct_close = ")]}>;'\"";  // always pitch fall unless followed by alnum
 
 // alter tone for announce punctuation or capitals
-static const char *tone_punct_on = "\0016T";  // add reverberation, lower pitch
-static const char *tone_punct_off = "\001T";
+//static const char *tone_punct_on = "\0016T";  // add reverberation, lower pitch
+//static const char *tone_punct_off = "\001T\001P";
+static const char *tone_punct_on = "";  // add reverberation, lower pitch  TEST apply no effect
+static const char *tone_punct_off = "";
 
 // ignore these characters
 static const unsigned short chars_ignore[] = {
@@ -1903,6 +1905,7 @@ int ReadClause(Translator *tr, FILE *f_in, char *buf, short *charix, int *charix
 	int c1=' ';  // current character
 	int c2;  // next character
 	int cprev=' ';  // previous character
+	int cprev2=' ';
 	int parag;
 	int ix = 0;
 	int j;
@@ -1916,6 +1919,8 @@ int ReadClause(Translator *tr, FILE *f_in, char *buf, short *charix, int *charix
 	int any_alnum = 0;
 	int self_closing;
 	int punct_data;
+	int is_punctuation;
+	int save_c2;
 	int stressed_word = 0;
 	const char *p;
 	wchar_t xml_buf[N_XML_BUF+1];
@@ -1975,6 +1980,7 @@ f_input = f_in;  // for GetC etc
 			}
 		}
 
+		cprev2 = cprev;
 		cprev = c1;
 		c1 = c2;
 
@@ -2324,9 +2330,9 @@ if(option_ssml) parag=1;
 			if((iswspace(c2) || (punct_data & 0x8000) || IsBracket(c2) || (c2=='?') || (c2=='-') || Eof()))
 			{
 				// note: (c2='?') is for when a smart-quote has been replaced by '?'
-				buf[ix] = ' ';
-				buf[ix+1] = 0;
-	
+
+				is_punctuation = 1;
+
 				if((c1 == '.') && (cprev == '.'))
 				{
 					c1 = 0x2026;
@@ -2334,45 +2340,59 @@ if(option_ssml) parag=1;
 				}
 	
 				nl_count = 0;
-				while(!Eof() && iswspace(c2))
+				save_c2 = c2;
+
+				if(iswspace(c2))
 				{
-					if(c2 == '\n')
-						nl_count++;
-					c2 = GetC();   // skip past space(s)
+					while(!Eof() && iswspace(c2))
+					{
+						if(c2 == '\n')
+							nl_count++;
+						c2 = GetC();   // skip past space(s)
+					}
+					if(!Eof())
+					{
+						UngetC(c2);
+					}
 				}
-				if(!Eof())
-				{
-					UngetC(c2);
-				}
-	
+
 				if((nl_count==0) && (c1 == '.'))
 				{
-					if(iswdigit(cprev) && (tr->langopts.numbers & 0x10000) && islower(c2))
+//					if(iswdigit(cprev) && (tr->langopts.numbers & NUM_ORDINAL_DOT) && islower(c2))
+					if(iswdigit(cprev) && (tr->langopts.numbers & NUM_ORDINAL_DOT))
 					{
 						// dot after a number indicates an ordinal number
-						c2 = '.';
-						continue;
+						is_punctuation = 0;
 					}
+					else
 					if(iswlower(c2))
 					{
-						c2 = ' ';
-						continue;  // next word has no capital letter, this dot is probably from an abbreviation
+						// next word has no capital letter, this dot is probably from an abbreviation
+						c1 = ' ';
+						is_punctuation = 0;
 					}
 					if(any_alnum==0)
 					{
-						c2 = ' ';   // no letters or digits yet, so probably not a sentence terminator
-						continue;
+						c1 = ' ';   // no letters or digits yet, so probably not a sentence terminator
+						is_punctuation = 0;
 					}
 				}
-	
-				punct_data = punct_attributes[punct];
-				if(nl_count > 1)
+				c2 = save_c2;
+
+				if(is_punctuation)
 				{
-					if((punct_data == CLAUSE_QUESTION) || (punct_data == CLAUSE_EXCLAMATION))
-						return(punct_data + 35);   // with a longer pause
-					return(CLAUSE_PARAGRAPH);
+					buf[ix] = ' ';
+					buf[ix+1] = 0;
+		
+					punct_data = punct_attributes[punct];
+					if(nl_count > 1)
+					{
+						if((punct_data == CLAUSE_QUESTION) || (punct_data == CLAUSE_EXCLAMATION))
+							return(punct_data + 35);   // with a longer pause
+						return(CLAUSE_PARAGRAPH);
+					}
+					return(punct_data);   // only recognise punctuation if followed by a blank or bracket/quote
 				}
-				return(punct_data);   // only recognise punctuation if followed by a blank or bracket/quote
 			}
 		}
 

src/synthdata.cpp

@@ -35,7 +35,7 @@
 #include "translate.h"
 #include "wave.h"
 
-const char *version_string = "1.41.08  04.Oct.09";
+const char *version_string = "1.41.11  09.Oct.09";
 const int version_phdata  = 0x014100;
 
 int option_device_number = -1;

src/tr_languages.cpp

@@ -48,6 +48,7 @@
 #define OFFSET_ARMENIAN 0x530
 #define OFFSET_DEVANAGARI  0x900
 #define OFFSET_BENGALI 0x980
+#define OFFSET_GURMUKHI 0xa00
 #define OFFSET_TAMIL  0xb80
 #define OFFSET_KANNADA 0xc80
 #define OFFSET_MALAYALAM 0xd00
@@ -165,6 +166,7 @@ static Translator* NewTranslator(void)
 	tr->langopts.max_roman = 49;
 	tr->langopts.thousands_sep = ',';
 	tr->langopts.decimal_sep = '.';
+	tr->langopts.break_numbers = BREAK_THOUSANDS;   // 1000, 1000,000  1,000,000 etc
 
 	memcpy(tr->punct_to_tone, punctuation_to_tone, sizeof(tr->punct_to_tone));
 
@@ -263,7 +265,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.param[LOPT_PREFIXES] = 1;
 			SetLetterVowel(tr,'y');  // add 'y' to vowels
 		
-			tr->langopts.numbers = 0x8d1 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_SWAP_TENS | NUM_HUNDRED_AND | NUM_SINGLE_AND | NUM_ROMAN | NUM_1900;
 			tr->langopts.accents = 1;
 		}
 		break;
@@ -283,7 +285,7 @@ Translator *SelectTranslator(const char *name)
 			SetLetterBitsRange(tr,LETTERGP_F,0x3e,0x4c);   // vowel signs, but not virama
 
 			tr->langopts.numbers = 0x1;
-			tr->langopts.numbers2 = NUM2_100000;
+			tr->langopts.break_numbers = 0x24924aa8;  // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
 		}
 		break;
 
@@ -305,7 +307,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.unstressed_wd2 = 2;
 			tr->langopts.param[LOPT_SONORANT_MIN] = 120;  // limit the shortening of sonorants before short vowels
 
-			tr->langopts.numbers = 0x401;
+			tr->langopts.numbers = NUM_OMIT_1_HUNDRED;
 
 			SetLetterVowel(tr,'w');  // add letter to vowels and remove from consonants
 			SetLetterVowel(tr,'y');
@@ -319,7 +321,7 @@ Translator *SelectTranslator(const char *name)
 
 			tr->langopts.stress_rule = 0;
 			SetLetterVowel(tr,'y');
-			tr->langopts.numbers = 0x10c59;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_ORDINAL_DOT | NUM_1900;
 		}
 		break;
 
@@ -333,7 +335,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.param[LOPT_PREFIXES] = 1;
 			memcpy(tr->stress_lengths,stress_lengths_de,sizeof(tr->stress_lengths));
 		
-			tr->langopts.numbers = 0x11419 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_OMIT_1_HUNDRED | NUM_ALLOW_SPACE | NUM_ORDINAL_DOT | NUM_ROMAN;
 			SetLetterVowel(tr,'y');
 		}
 		break;
@@ -344,7 +346,7 @@ Translator *SelectTranslator(const char *name)
 			SetupTranslator(tr,stress_lengths_en,NULL);
 
 			tr->langopts.stress_rule = 0;
-			tr->langopts.numbers = 0x841 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_HUNDRED_AND | NUM_ROMAN | NUM_1900;
 			tr->langopts.param[LOPT_COMBINE_WORDS] = 2;       // allow "mc" to cmbine with the following word
 		}
 		break;
@@ -381,7 +383,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.unstressed_wd2 = 2;
 			tr->langopts.param[LOPT_SONORANT_MIN] = 130;  // limit the shortening of sonorants before short vowels
 
-			tr->langopts.numbers = 0x109;
+			tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA; 
 			tr->langopts.numbers2 = 0x2;   // variant form of numbers before thousands
 
 			if(name2 == L_grc)
@@ -410,7 +412,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.unstressed_wd1 = 3;
 			tr->langopts.unstressed_wd2 = 2;
 
-			tr->langopts.numbers = 0x1409 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_ALLOW_SPACE | NUM_ROMAN;
 		}
 		break;
 
@@ -435,7 +437,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.unstressed_wd2 = 2;
 			tr->langopts.param[LOPT_SONORANT_MIN] = 120;  // limit the shortening of sonorants before short vowels
 
-			tr->langopts.numbers = 0x529 + NUM_ROMAN + NUM_ROMAN_AFTER;
+			tr->langopts.numbers = NUM_SINGLE_STRESS | NUM_DECIMAL_COMMA | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_ROMAN | NUM_ROMAN_AFTER;
 
 			if(name2 == L('c','a'))
 			{
@@ -457,7 +459,7 @@ Translator *SelectTranslator(const char *name)
 			static const unsigned char stress_amps_eu[8] = {16,16, 18,18, 18,18, 18,18 };
 			SetupTranslator(tr,stress_lengths_eu,stress_amps_eu);
 			tr->langopts.stress_rule = 1;  // ?? second syllable ??
-			tr->langopts.numbers = 0x569 + NUM_VIGESIMAL;
+			tr->langopts.numbers = NUM_SINGLE_STRESS + NUM_DECIMAL_COMMA | NUM_AND_UNITS | NUM_HUNDRED_AND | NUM_OMIT_1_HUNDRED | NUM_VIGESIMAL;
 		}
 		break;
 
@@ -474,7 +476,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.param[LOPT_IT_DOUBLING] = 1;
 			tr->langopts.long_stop = 130;
 
-			tr->langopts.numbers = 0x1009;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA + NUM_ALLOW_SPACE;
 			SetLetterVowel(tr,'y');
 //			tr->langopts.max_initial_consonants = 2;  // BUT foreign words may have 3
 			tr->langopts.spelling_stress = 1;
@@ -493,7 +495,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.stress_flags = 0x0024;  // don't use secondary stress
 			tr->langopts.param[LOPT_IT_LENGTHEN] = 1;    // remove lengthen indicator from unstressed syllables
 
-			tr->langopts.numbers = 0x1509 + 0x8000 + NUM_NOPAUSE | NUM_ROMAN | NUM_VIGESIMAL;
+			tr->langopts.numbers = NUM_SINGLE_STRESS +  NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_OMIT_1_HUNDRED | NUM_NOPAUSE | NUM_ROMAN | NUM_VIGESIMAL | NUM_DFRACTION_4;
 			SetLetterVowel(tr,'y');
 		}
 		break;
@@ -508,6 +510,7 @@ Translator *SelectTranslator(const char *name)
 
 	case L('h','i'):    // Hindi
 	case L('n','e'):    // Nepali
+	case L('p','a'):    // Punjabi
 		{
 			static const short stress_lengths_hi[8] = {190, 190,  210, 210,  0, 0,  230, 250};
 			static const unsigned char stress_amps_hi[8] = {17,14, 20,19, 20,22, 22,21 };
@@ -518,9 +521,15 @@ Translator *SelectTranslator(const char *name)
 
 			tr->langopts.stress_rule = 6;      // stress on last heaviest syllable, excluding final syllable
 			tr->langopts.stress_flags =  0x10004;   // use 'diminished' for unstressed final syllable
-			tr->langopts.numbers = 0x011;
-			tr->langopts.numbers2 = NUM2_100000;
+			tr->langopts.numbers = NUM_SWAP_TENS;
+			tr->langopts.break_numbers = 0x24924aa8;  // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
 			tr->letter_bits_offset = OFFSET_DEVANAGARI;
+
+			if(name2 == L('p','a'))
+			{
+				tr->langopts.numbers = 0;   // no number rules yet
+				tr->letter_bits_offset = OFFSET_GURMUKHI;
+			}
 			SetIndicLetters(tr);
 		}
 		break;
@@ -547,7 +556,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.spelling_stress = 1;
 			tr->langopts.accents = 1;
 
-			tr->langopts.numbers = 0x140d + 0x4000 + NUM_ROMAN_UC;
+			tr->langopts.numbers = NUM_SINGLE_STRESS + NUM_HUNDRED_AND | NUM_DECIMAL_COMMA | NUM_THOUS_SPACE | NUM_DFRACTION_2 | NUM_ROMAN_UC;
 			tr->langopts.numbers2 = 0x4a;  // variant numbers before thousands,milliards
 			tr->langopts.replace_chars = replace_cyrillic_latin;
 
@@ -573,7 +582,7 @@ Translator *SelectTranslator(const char *name)
 			tr->langopts.param[LOPT_IT_DOUBLING] = 1;
 			tr->langopts.param[LOPT_COMBINE_WORDS] = 99;  // combine some prepositions with the following word
 
-			tr->langopts.numbers = 0x1009 + 0xa000 + NUM_ROMAN + NUM_ROMAN_ORDINAL + NUM_ORDINAL_DOT + NUM_OMIT_1_HUNDRED;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_DFRACTION_5 | NUM_ROMAN | NUM_ROMAN_ORDINAL | NUM_ORDINAL_DOT | NUM_OMIT_1_HUNDRED;
 			SetLetterVowel(tr,'y');
 			tr->langopts.spelling_stress = 1;
 SetLengthMods(tr,3);  // all equal
@@ -595,7 +604,7 @@ SetLengthMods(tr,3);  // all equal
 			SetLetterBits(tr,LETTERGP_A,hy_vowels);
 			SetLetterBits(tr,LETTERGP_C,hy_consonants);
 			tr->langopts.max_initial_consonants = 6;
-			tr->langopts.numbers = 0x409;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED;
 //	tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1;   // disable check for unpronouncable words
 		}
 		break;
@@ -607,7 +616,7 @@ SetLengthMods(tr,3);  // all equal
 
 			SetupTranslator(tr,stress_lengths_id,stress_amps_id);
 			tr->langopts.stress_rule = 2;
-			tr->langopts.numbers = 0x1009 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_ROMAN;
 			tr->langopts.stress_flags =  0x6 | 0x10; 
 			tr->langopts.accents = 2;  // "capital" after letter name
 		}
@@ -629,7 +638,7 @@ SetLengthMods(tr,3);  // all equal
 			SetLetterBits(tr,3,"jvr");    // Letter group H
 			tr->letter_groups[1] = is_lettergroup_B;
 			SetLetterVowel(tr,'y');
-			tr->langopts.numbers = 0x8e9;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SINGLE_AND | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_1900;
 			tr->langopts.numbers2 = 0x2;
 		}
 		break;
@@ -652,7 +661,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.param[LOPT_SONORANT_MIN] = 130;  // limit the shortening of sonorants before short vowels
 			tr->langopts.param[LOPT_REDUCE] = 1;        // reduce vowels even if phonemes are specified in it_list
 			tr->langopts.param[LOPT_ALT] = 2;      // call ApplySpecialAttributes2() if a word has $alt or $alt2
-			tr->langopts.numbers = 0x2709 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_SINGLE_VOWEL | NUM_OMIT_1_HUNDRED |NUM_DECIMAL_COMMA | NUM_ROMAN | NUM_DFRACTION_1;
 			tr->langopts.accents = 2;   // Say "Capital" after the letter.
 			SetLetterVowel(tr,'y');
 		}
@@ -686,7 +695,7 @@ SetLengthMods(tr,3);  // all equal
 
 			tr->langopts.stress_rule = 8;   // ?? 1st syllable if it is heavy, else 2nd syllable
 			tr->langopts.param[LOPT_UNPRONOUNCABLE] = 1;   // disable check for unpronouncable words
-			tr->langopts.numbers = 0x0401;
+			tr->langopts.numbers = NUM_OMIT_1_HUNDRED;
 		}
 		break;
 
@@ -700,7 +709,7 @@ SetLengthMods(tr,3);  // all equal
 
 			tr->langopts.stress_rule = 7;   // stress on the last syllable, before any explicitly unstressed syllable
 
-			tr->langopts.numbers = 0x100461;
+			tr->langopts.numbers = NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_AND_HUNDRED;
 			tr->langopts.max_initial_consonants = 2;
 		}
 		break;
@@ -713,7 +722,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.unstressed_wd1 = 0;
 			tr->langopts.unstressed_wd2 = 2;
 			tr->langopts.param[LOPT_DIERESES] = 1;
-			tr->langopts.numbers = 0x1 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_ROMAN;
 			tr->langopts.max_roman = 5000;
 		}
 		break;
@@ -728,7 +737,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_rule = 0;
 			tr->langopts.spelling_stress = 1;
 			tr->charset_a0 = charsets[4];   // ISO-8859-4
-			tr->langopts.numbers = 0x409 + 0x8000 + 0x10000;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_4 | NUM_ORDINAL_DOT;
 			tr->langopts.stress_flags = 0x16 + 0x40000;
 		}
 		break;
@@ -745,7 +754,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->letter_groups[0] = vowels_cyrillic;
 
 			tr->langopts.stress_rule = 4;   // antipenultimate
-			tr->langopts.numbers = 0x0429 + 0x4000;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_AND_UNITS | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_2;
 			tr->langopts.numbers2 = 0x8a;  // variant numbers before thousands,milliards
 		}
 		break;
@@ -761,7 +770,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.param[LOPT_PREFIXES] = 1;
 			SetLetterVowel(tr,'y');
 		
-			tr->langopts.numbers = 0x11c19;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_SWAP_TENS | NUM_OMIT_1_HUNDRED | NUM_ALLOW_SPACE | NUM_1900 | NUM_ORDINAL_DOT;
 			memcpy(tr->stress_lengths,stress_lengths_nl,sizeof(tr->stress_lengths));
 		}
 		break;
@@ -773,7 +782,7 @@ SetLengthMods(tr,3);  // all equal
 			SetupTranslator(tr,stress_lengths_no,NULL);
 			tr->langopts.stress_rule = 0;
 			SetLetterVowel(tr,'y');
-			tr->langopts.numbers = 0x11849;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_HUNDRED_AND | NUM_ALLOW_SPACE | NUM_1900 + NUM_ORDINAL_DOT;
 		}
 		break;
 
@@ -784,7 +793,7 @@ SetLengthMods(tr,3);  // all equal
 
 			SetupTranslator(tr,stress_lengths_om,stress_amps_om);
 			tr->langopts.stress_rule = 2;
-			tr->langopts.stress_flags = 0x16 + 0x80000;
+			tr->langopts.stress_flags = 2 + NUM_SWAP_TENS | NUM_THOUS_SPACE | NUM_NOPAUSE;  //??
 		}
 		break;
 
@@ -800,8 +809,8 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_flags = 0x6;  // mark unstressed final syllables as diminished
 			tr->langopts.param[LOPT_REGRESSIVE_VOICING] = 0x8;
  			tr->langopts.max_initial_consonants = 7; // for example: wchrzczony :)
- 			tr->langopts.numbers=0x1009 + 0x4000;
-			tr->langopts.numbers2=0x40;
+ 			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_DFRACTION_2;
+			tr->langopts.numbers2 = 0x40;
 			tr->langopts.param[LOPT_COMBINE_WORDS] = 4 + 0x100;  // combine 'nie' (marked with $alt2) with some 1-syllable (and 2-syllable) words (marked with $alt)
 			SetLetterVowel(tr,'y');
 		}
@@ -817,7 +826,7 @@ SetLengthMods(tr,3);  // all equal
 
 			tr->langopts.stress_rule = 3;        // stress on final syllable
 			tr->langopts.stress_flags =  0x6 | 0x10 | 0x20000; 
-			tr->langopts.numbers = 0x069 + 0x4000 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_DFRACTION_2 | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_ROMAN;
 			SetLetterVowel(tr,'y');
 			ResetLetterBits(tr,0x2);
 			SetLetterBits(tr,1,"bcdfgjkmnpqstvxz");      // B  hard consonants, excluding h,l,r,w,y
@@ -835,7 +844,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_flags = 0x100 + 0x6;
 
 			tr->charset_a0 = charsets[2];   // ISO-8859-2
-			tr->langopts.numbers = 0x1029+0x6000 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_DFRACTION_3 | NUM_AND_UNITS | NUM_ROMAN;
 			tr->langopts.numbers2 = 0x1e;  // variant numbers before all thousandplex
 		}
 		break;
@@ -850,7 +859,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_flags = 0x16;
 			tr->langopts.length_mods0 = tr->langopts.length_mods;  // don't lengthen vowels in the last syllable
 
-			tr->langopts.numbers = 0x61 + 0x100000 + 0x4000;
+			tr->langopts.numbers = NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_DFRACTION_2 | NUM_AND_HUNDRED;
 			tr->langopts.numbers2 = 0x200;  // say "thousands" before its number
 		}
 		break;
@@ -870,7 +879,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.spelling_stress = 1;
 			tr->langopts.param[LOPT_COMBINE_WORDS] = 4;  // combine some prepositions with the following word
 
-			tr->langopts.numbers = 0x0401 + 0x4000 + NUM_ROMAN;
+			tr->langopts.numbers = NUM_OMIT_1_HUNDRED | NUM_DFRACTION_2 | NUM_ROMAN;
 			tr->langopts.numbers2 = 0x100;
 			tr->langopts.thousands_sep = 0;   //no thousands separator
 			tr->langopts.decimal_sep = ',';
@@ -897,7 +906,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_rule = 2;
 			tr->langopts.stress_flags =  0x16 + 0x100; 
 			SetLetterVowel(tr,'y');
-			tr->langopts.numbers = 0x69 + 0x8000;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_HUNDRED_AND | NUM_AND_UNITS | NUM_DFRACTION_4;
 			tr->langopts.accents = 2;  // "capital" after letter name
 		}
 		break;
@@ -911,7 +920,7 @@ SetLengthMods(tr,3);  // all equal
 
 			tr->langopts.stress_rule = 0;
 			SetLetterVowel(tr,'y');
-			tr->langopts.numbers = 0x1909;
+			tr->langopts.numbers = NUM_SINGLE_STRESS + NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_1900;
 			tr->langopts.accents = 1;
 		}
 		break;
@@ -928,8 +937,8 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_rule = 2;
 			tr->langopts.stress_flags =  0x6 | 0x10; 
 
-			tr->langopts.numbers = 0x4e1;
-			tr->langopts.numbers2 = NUM2_100000a;
+			tr->langopts.numbers = NUM_AND_UNITS | NUM_HUNDRED_AND | NUM_SINGLE_AND | NUM_OMIT_1_HUNDRED;
+			tr->langopts.break_numbers = 0x49249268;  // for languages which have numbers for 100,000 and 1,000,000
 		}
 		break;
 
@@ -946,7 +955,7 @@ SetLengthMods(tr,3);  // all equal
 
 			tr->langopts.stress_rule = 0;
 			tr->langopts.stress_flags =  0x10004;   // use 'diminished' for unstressed final syllable
-			tr->langopts.numbers2 = NUM2_100000;
+			tr->langopts.break_numbers = 0x24a8;  // 1000, 100,000  10,000,000 
 
 			if(name2 == L('t','a'))
 			{
@@ -1002,7 +1011,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->langopts.stress_rule = 7;   // stress on the last syllable, before any explicitly unstressed syllable
 			tr->langopts.stress_flags = 0x20;  //no automatic secondary stress
 
-			tr->langopts.numbers = 0x1509 + 0x4000;
+			tr->langopts.numbers = NUM_SINGLE_STRESS + NUM_DECIMAL_COMMA | NUM_ALLOW_SPACE | NUM_OMIT_1_HUNDRED | NUM_DFRACTION_2;
 			tr->langopts.max_initial_consonants = 2;
 		}
 		break;
@@ -1034,7 +1043,7 @@ SetLengthMods(tr,3);  // all equal
 			tr->letter_groups[0] = vowels_vi;
 			tr->langopts.tone_language = 1;   // Tone language, use  CalcPitches_Tone() rather than CalcPitches()
 			tr->langopts.unstressed_wd1 = 2;
-			tr->langopts.numbers = 0x0049 + 0x8000;
+			tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_HUNDRED_AND | NUM_DFRACTION_4;
 
 		}
 		break;
@@ -1070,13 +1079,13 @@ SetLengthMods(tr,3);  // all equal
 
 	tr->translator_name = name2;
 
-	if(tr->langopts.numbers & 0x8)
+	if(tr->langopts.numbers & NUM_DECIMAL_COMMA)
 	{
 		// use . and ; for thousands and decimal separators
 		tr->langopts.thousands_sep = '.';
 		tr->langopts.decimal_sep = ',';
 	}
-	if(tr->langopts.numbers & 0x4)
+	if(tr->langopts.numbers & NUM_THOUS_SPACE)
 	{
 		tr->langopts.thousands_sep = 0;   // don't allow thousands separator, except space
 	}
@@ -1128,7 +1137,7 @@ static void Translator_Russian(Translator *tr)
 	tr->langopts.stress_rule = 5;
 	tr->langopts.stress_flags = 0x0020;  // waas 0x1010
 
-	tr->langopts.numbers = 0x0409;
+	tr->langopts.numbers = NUM_DECIMAL_COMMA | NUM_OMIT_1_HUNDRED;
 	tr->langopts.numbers2 = 0xc2;  // variant numbers before thousands
 	tr->langopts.phoneme_change = 1;
 	tr->langopts.testing = 2;

src/translate.cpp

@@ -366,6 +366,10 @@ int IsAlpha(unsigned int c)
 {//========================
 // Replacement for iswalph() which also checks for some in-word symbols
 
+	const unsigned short extra_indic_alphas[] = {
+	0xa70,0xa71,	// Gurmukhi: tippi, addak
+	0 };
+
 	if(iswalpha(c))
 		return(1);
 
@@ -374,6 +378,8 @@ int IsAlpha(unsigned int c)
 		// Indic scripts: Devanagari, Tamil, etc
 		if((c & 0x7f) < 0x64)
 			return(1);
+		if(lookupwchar(extra_indic_alphas, c) != 0)
+			return(1);
 		return(0);
 	}
 
@@ -2570,10 +2576,6 @@ if((c == '/') && (tr->langopts.testing & 2) && IsDigit09(next_in) && IsAlpha(pre
 		int c_temp;
 		char *pn;
 		char *pw;
-		static unsigned int break_numbers1 = 0x49249248;
-		static unsigned int break_numbers2 = 0x24924aa8;  // for languages which have numbers for 100,000 and 100,00,000, eg Hindi
-		static unsigned int break_numbers3 = 0x49249268;  // for languages which have numbers for 100,000 and 1,000,000
-		unsigned int break_numbers;
 		char number_buf[80];
 
 		// start speaking at a specified word position in the text?
@@ -2591,7 +2593,7 @@ if((c == '/') && (tr->langopts.testing & 2) && IsDigit09(next_in) && IsAlpha(pre
 		// digits should have been converted to Latin alphabet ('0' to '9')
 		word = pw = &sbuf[words[ix].start];
 
-		if(iswdigit(word[0]) && (tr->langopts.numbers2 & NUM2_100000))
+		if(iswdigit(word[0]) && (tr->langopts.break_numbers != BREAK_THOUSANDS))
 		{
 			// Languages with 100000 numbers.  Remove thousands separators so that we can insert them again later
 			pn = number_buf;
@@ -2628,34 +2630,26 @@ if((c == '/') && (tr->langopts.testing & 2) && IsDigit09(next_in) && IsAlpha(pre
 			pn = &number_buf[1];
 			nx = n_digits;
 
-			if((tr->langopts.numbers2 & NUM2_100000a) == NUM2_100000a)
-				break_numbers = break_numbers3;
-			else
-			if(tr->langopts.numbers2 & NUM2_100000)
-				break_numbers = break_numbers2;
-			else
-				break_numbers = break_numbers1;
-
 			while(pn < &number_buf[sizeof(number_buf)-3])
 			{
 				if(!isdigit(c = *pw++) && (c != tr->langopts.decimal_sep))
 					break;
 
 				*pn++ = c;
-				if((--nx > 0) && (break_numbers & (1 << nx)))
+				if((--nx > 0) && (tr->langopts.break_numbers & (1 << nx)))
 				{
 					if(tr->langopts.thousands_sep != ' ')
 					{
 						*pn++ = tr->langopts.thousands_sep;
 					}
 					*pn++ = ' ';
-					if(break_numbers & (1 << (nx-1)))
+					if(tr->langopts.break_numbers & (1 << (nx-1)))
 					{
 						// the next group only has 1 digits (i.e. NUM2_10000), make it three
 						*pn++ = '0';
 						*pn++ = '0';
 					}
-					if(break_numbers & (1 << (nx-2)))
+					if(tr->langopts.break_numbers & (1 << (nx-2)))
 					{
 						// the next group only has 2 digits (i.e. NUM2_10000), make it three
 						*pn++ = '0';

src/translate.h

@@ -328,12 +328,27 @@ typedef struct {
 	unsigned char *length_mods;
 	unsigned char *length_mods0;
 
+#define NUM_THOUS_SPACE  0x4
+#define NUM_DECIMAL_COMMA 0x8
+#define NUM_SWAP_TENS    0x10
+#define NUM_AND_UNITS    0x20
+#define NUM_HUNDRED_AND  0x40
+#define NUM_SINGLE_AND   0x80
+#define NUM_SINGLE_STRESS 0x100
+#define NUM_SINGLE_VOWEL 0x200
 #define NUM_OMIT_1_HUNDRED 0x400
-#define NUM_19_HUNDRED   0x800
+#define NUM_1900         0x800
+#define NUM_ALLOW_SPACE  0x1000
+#define NUM_DFRACTION_1  0x2000
+#define NUM_DFRACTION_2  0x4000
+#define NUM_DFRACTION_3  0x6000
+#define NUM_DFRACTION_4  0x8000
+#define NUM_DFRACTION_5  0xa000
 #define NUM_ORDINAL_DOT  0x10000
 #define NUM_ROMAN        0x20000
 #define NUM_ROMAN_UC     0x40000
 #define NUM_NOPAUSE      0x80000
+#define NUM_AND_HUNDRED  0x100000
 #define NUM_ROMAN_AFTER  0x200000
 #define NUM_VIGESIMAL    0x400000
 #define NUM_ROMAN_ORDINAL 0x800000
@@ -361,17 +376,15 @@ typedef struct {
 	// bit23=Roman numbers are ordinal numbers
 	int numbers;
 
-#define NUM2_100000     0x800   // numbers for 100,000 and 10,000,000
-#define NUM2_100000a    0xc00   // numbers for 100,000 and 1,000,000
 	// bits 1-4  use variant form of numbers before thousands,millions,etc.
 	// bit6=(LANG=pl) two forms of plural, M or MA
 	// bit7=(LANG-ru) use MB for 1 thousand, million, etc
 	// bit8=(LANG=cs,sk) two forms of plural, M or MA
 	// bit9=(LANG=rw) say "thousand" and "million" before its number, not after
-   // bit10=(LANG=sw) special word for 100,000 and 1,000,000
-	// bit11=(LANG=hi) special word for 100,000 and 10,000,000
 	int numbers2;
 
+#define BREAK_THOUSANDS   0x49249248
+	int break_numbers;  // which digits to break the number into thousands, millions, etc (Hindi has 100,000 not 1,000,000)
 	int max_roman;
 	int thousands_sep;
 	int decimal_sep;

src/wavegen.cpp

@@ -794,12 +794,15 @@ static void WavegenSetEcho(void)
 		amp = embedded_value[EMBED_H];
 		delay = 130;
 	}
+#ifdef deleted
 	if(embedded_value[EMBED_T] > 0)
 	{
-		// announcing punctuation
+		// announcing punctuation, add a small echo
+// This seems unpopular
 		amp = embedded_value[EMBED_T] * 8;
 		delay = 60;
 	}
+#endif
 
 	if(delay == 0)
 		amp = 0;