Merge pull request #865

CHANGELOG.md

 *  ia (Interlingua) -- nesrad
 *  it (Italian) -- Christian Leo
 *  ja (Japanese) -- fukuen
+*  jbo (Lojban) -- Juho Hiltunen, xunsku
 *  mi (Māori) -- boracasli98
 *  shn (Shan Tay Yai) -- ronaldaug
 *  tr (Turkish) -- boracasli98

dictsource/jbo_list

 _dpt	pi_
 
 
-_a	abu
 b	b@
 c	S@
 d	d@
-_e	ebu
 f	f@
 g	g@
-_i	ibu
 j	Z@
 k	k@
-l	l@
-m	m@	// should "words" l,m,n,r say the syllable consonant?
-n	n@
-_o	obu
 p	p@
-r	R@
 s	s@
 t	t@
-_u	ubu
 v	v@
 x	x@
-y	'@bu
 z	z@
 
 h	@h'@bu
 
 
 // stressed cmavo
-cai	S'aI_!
+cai	S'aI__!
 cu'i	S'uhi_!
-pei	p'eI_!
+pei	p'eI__!
 ru'e	R'uhe_!
-sai	s'aI_!
-nai	n'aI_!
-na	n'a_!
+sai	s'aI__!
+nai	n'aI__!
+na	n'a__!
 ja'a	Z'aha_!
 
+// question cmavo
+ma	m'a__ // a longer pause to avoid resegmentation
+mo	m'o__
+xo	x'o__
+cu'e	S'uhe_!
+
+// emphasizing prefixes
+ba'e $u+
+za'e $u+
+
 
 // unstressed words. Note single-syllable cmavo are all unstressed in jbo_rules
-e'o	$u+	// request
 po'e	$u+	// of
 zo'e	$u+	// pronoun
+jo'u	$u+ $brk	// and
+ce'u	$u+ $brk	// lambda-it
+ke'a	$u+ $brk	// relative-it
 
 
-i	i_:	$u	// sentence break (recognised by eSpeak program). Try a short pause after ".i"
-
 // end-of-clause [_;_] before these
 noi	_;_noI	$u
 poi	_;_poI	$u
 po'u	_;_pohu	$u
 goi	_;_goI	$u
 
-ija	_;_iZa		// should this series be unstressed?
-ijanai	_;_iZan'aI
-ije	_;_iZe
-ije'i	_;_iZehi
-ijenai	_;_iZenaI
-ijo	_;_iZo
-ijonai	_;_iZon'aI
-inaja	_;_inaZa
+
+// sentence connectives
+i	__;__i
+ija	_;__iZa
+ijanai	_;__iZan'aI
+ije	_;__iZe
+ije'i	_;__iZehi
+ijenai	_;__iZenaI
+ijo	_;__iZo
+ijonai	_;__iZon'aI
+inaja	_;__inaZa
+
+
+// connectives
+a	$brk
+e	$brk
+o	$brk
+u	$brk
+ja	$brk
+je	$brk
+ji	$brk
+jo	$brk
+ju	$brk
+joi	$brk
+
+
+// terminators and openers
+cu	$brk
+gi	$pause
 
 to	_::to	$u	// start parenthesis, pause but don't raise intonation
-
-// terminators (include a pause after)
-ku	ku_::	$u
-boi	boI_::	$u
+lu	lu_:	$u $brk
+tu'e t,uhe_:	$u $brk
+
+ku	ku_:	$u
+boi	boI_:	$u
+
+ge'u	g,ehu_:	$u $brk
+lu'u	l,uhu_:	$u $brk
+
+li'u	l,ihu_::	$u $brk
+tu'u	t,uhu_::	$u $brk
 vau	vaU_::	$u
 kei	keI_::	$u
+ku'o	k,uho_::	$u
 toi	toI_::	$u
 

dictsource/jbo_rules

 	'	h
      _) ' (_    '@h@
 
+	D) . (D	||p%i||,	// decimal point
 	.	_!	// dot
      _)	. (@P1	_!      // remove . prefix
 	. (_S1  _!	// remove . suffix

src/libespeak-ng/readclause.c

 static int ungot_char2 = 0;
 espeak_ng_TEXT_DECODER *p_decoder = NULL;
 static int ungot_char;
-static const char *ungot_word = NULL;
 
 static bool ignore_text = false; // set during <sub> ... </sub>  to ignore text which has been replaced by an alias
 static bool audio_text = false; // set during <audio> ... </audio>
 	*tone_type = 0;
 	*voice_change = 0;
 
-	if (ungot_word != NULL) {
-		strcpy(buf, ungot_word);
-		ix += strlen(ungot_word);
-		ungot_word = NULL;
-	}
-
 	if (ungot_char2 != 0)
 		c2 = ungot_char2;
 	else
 			if (c1 == 0xf0b)
 				c1 = ' '; // Tibet inter-syllabic mark, ?? replace by space ??
 
-			if (iswspace(c1)) {
-				char *p_word;
-
-				if (tr->translator_name == 0x6a626f) {
-					// language jbo : lojban
-					// treat "i" or ".i" as end-of-sentence
-					p_word = &buf[ix-1];
-					if (p_word[0] == 'i') {
-						if (p_word[-1] == '.')
-							p_word--;
-						if (p_word[-1] == ' ') {
-							ungot_word = "i ";
-							UngetC(c2);
-							p_word[0] = 0;
-							return CLAUSE_PERIOD;
-						}
-					}
-				}
-			}
-
 			if (c1 == 0xd4d) {
 				// Malayalam virama, check if next character is Zero-width-joiner
 				if (c2 == 0x200d)

src/libespeak-ng/tr_languages.c

 		tr->langopts.stress_rule = STRESSPOSN_2R;
 		tr->langopts.vowel_pause = 0x20c; // pause before a word which starts with a vowel, or after a word which ends in a consonant
 		tr->punct_within_word = jbo_punct_within_word;
-		tr->langopts.param[LOPT_CAPS_IN_WORD] = 2; // capitals indicate stressed syllables
+		tr->langopts.param[LOPT_CAPS_IN_WORD] = 1; // capitals indicate stressed syllables
 		SetLetterVowel(tr, 'y');
 		tr->langopts.max_lengthmod = 368;
 	}

src/libespeak-ng/translate.c

 				if (iswupper(c)) {
 					c = towlower2(c, tr);
 
-					if ((j = tr->langopts.param[LOPT_CAPS_IN_WORD]) > 0) {
-						if ((j == 2) && (syllable_marked == false)) {
+					if (tr->langopts.param[LOPT_CAPS_IN_WORD]) {
+						if (syllable_marked == false) {
 							char_inserted = c;
 							c = 0x2c8; // stress marker
 							syllable_marked = true;

src/libespeak-ng/translate.h

 // change [t] when followed by unstressed vowel
 #define LOPT_REDUCE_T 12
 
-// 1 = allow capitals inside a word
-// 2 = stressed syllable is indicated by capitals
+// 1 = stressed syllable is indicated by capitals
 #define LOPT_CAPS_IN_WORD 13
 
 // bit 0=Italian "syntactic doubling" of consoants in the word after a word marked with $double attribute

tests/language-pronunciation.test

 k,ep\\uko'ete
 ,asak,ijumem'is\\i
 w,eCimos'esuN\\" "イロハニホヘト. チリヌルヲ. ワカヨタレソ. ツネナラム. ウヰノオクヤマ. ケフコエテ. アサキユメミシ. ヱヒモセスン."
-test_phonemes jbo "Latn" "R,o R'emna S,u s,e Z'inzi S,o z'ifRe Z,e simd'uhi b,e l,e R@ n,ilsels'iha _!'eleI R@ s'elS_!Ru
-i_: R@ s,e m'enli _;_g'ihe s,e sezm'aRde
-i_: _!,eI Z,esekih'ubo R@ s,im@z'uhe t'ahi l,e t'unba" "ro remna cu se jinzi co zifre je simdu'i be le ry. nilselsi'a .elei ry. selcru .i ry. se menli gi'e se sezmarde .i .ei jeseki'ubo ry. simyzu'e ta'i le tunba"
+test_phonemes jbo "Latn" "l,o R,o R'emna_! S,u s,e Z'inzi l,o k,a_! S,ehu z'ifRe_! Z,e l,o k,a S_!k,in@duns'ihu g,e l,o k,a_! S,ehu_! S,ehu s,e n,ilsels'inma_:_: g,i l,o k,a_! S,ehu_! S,ehu s,e kR'ali _!__;__'i R,a z'aNkaI l,o k,a s'uho d,a_! S,ehu m'enli_! Z,e l,o k,a s'uho d,a_! S,ehu m'aRde keI_:: _;_g'ihe b'ilga l,o k,a_! S,ehu_! S,ehu f'aRtRa s'uho Z,aI s,e t'asmi b,e l,o n,u R,a t,unb@s'ihu" "lo ro remna cu se jinzi lo ka ce'u zifre je lo ka ckinydunsi'u ge lo ka ce'u ce'u se nilselsinma gi lo ka ce'u ce'u se krali .i ra zankai lo ka su'o da ce'u menli je lo ka su'o da ce'u marde kei gi'e bilga lo ka ce'u ce'u fartra su'o jai se tasmi be lo nu ra tunbysi'u"
+test_phonemes jbo "Latn" "s'uho S'ilSe _!R'atat,aR_! S,u tolS'anSi _!__;__'i m,i Rivl'iha f'aha l,a _!S_!f,aRtsen'ek_!" "su'o cilce .RAtatar. cu tolcanci .i mi rivli'a fa'a la .cfartseNEK."
+
+
+
 test_phonemes ky "Cyrl" "bard'Iq adamd'ar 'Oz b,edelind'e dZan'a ,uquqtarInd'a erk'in dZan'a t'eN ,uquqt'u: bol'up dZaral'at
 alard'In 'aNs,ezim'i men'en ab,ijir'i b'ar dZan'a bir'ib,irin'e b'ir t,u:Gand'Iq mamil'e q,Ilu:G'a tij'iS" "Бардык адамдар өз беделинде жана укуктарында эркин жана тең укуктуу болуп жаралат. Алардын аң-сезими менен абийири бар жана бири-бирине бир туугандык мамиле кылууга тийиш."
 test_phonemes kl "Latn" "inu'itS; tam'armik in,uNN'orput nammin,E:rs,inn,a:ssusEq'arl#utsik ,assiQ,i:mm'il#u at,aqqinassusEq,arl#uts'il#u pis,inn,a:tsit,a:ffEq'arl#utsik