From cf4809a1e573ece70038cd11e25973a6d69619e1 Mon Sep 17 00:00:00 2001
From: Valdis Vitolins <valdis.vitolins@odo.lv>
Date: Fri, 10 Apr 2020 22:55:58 +0300
Subject: [PATCH] Fix bug #706: pronunciation error with abbreviations followed
 by 's

---
 src/libespeak-ng/translate.c | 42 +++++++++++++++++++++---------------
 tests/translate.test         |  6 ++++++
 2 files changed, 31 insertions(+), 17 deletions(-)

diff --git a/src/libespeak-ng/translate.c b/src/libespeak-ng/translate.c
index 4802cc71..9862a9bc 100644
--- a/src/libespeak-ng/translate.c
+++ b/src/libespeak-ng/translate.c
@@ -570,12 +570,6 @@ static int TranslateWord3(Translator *tr, char *word_start, WORD_TAB *wtab, char
 	int add_suffix_phonemes = 0;
 	WORD_TAB wtab_null[8];
 
-	// translate these to get pronunciations of plural 's' suffix (different forms depending on
-	// the preceding letter
-	static char word_zz[4] = { 0, 'z', 'z', 0 };
-	static char word_iz[4] = { 0, 'i', 'z', 0 };
-	static char word_ss[4] = { 0, 's', 's', 0 };
-
 	if (wtab == NULL) {
 		memset(wtab_null, 0, sizeof(wtab_null));
 		wtab = wtab_null;
@@ -738,6 +732,8 @@ static int TranslateWord3(Translator *tr, char *word_start, WORD_TAB *wtab, char
 		strcpy(word_phonemes, phonemes);
 		if (wflags & FLAG_TRANSLATOR2)
 			return 0;
+
+		addPluralSuffixes(wflags, tr, last_char, &word_phonemes);
 		return dictionary_flags[0] & FLAG_SKIPWORDS; // for "b.c.d"
 	} else if (found == false) {
 		// word's pronunciation is not given in the dictionary list, although
@@ -1010,16 +1006,7 @@ static int TranslateWord3(Translator *tr, char *word_start, WORD_TAB *wtab, char
 		}
 	}
 
-	if (wflags & FLAG_HAS_PLURAL) {
-		// s or 's suffix, append [s], [z] or [Iz] depending on previous letter
-		if (last_char == 'f')
-			TranslateRules(tr, &word_ss[1], phonemes, N_WORD_PHONEMES, NULL, 0, NULL);
-		else if ((last_char == 0) || (strchr_w("hsx", last_char) == NULL))
-			TranslateRules(tr, &word_zz[1], phonemes, N_WORD_PHONEMES, NULL, 0, NULL);
-		else
-			TranslateRules(tr, &word_iz[1], phonemes, N_WORD_PHONEMES, NULL, 0, NULL);
-	}
-
+	addPluralSuffixes(wflags, tr, last_char, &word_phonemes);
 	wflags |= emphasize_allcaps;
 
 	// determine stress pattern for this word
@@ -1154,6 +1141,26 @@ static int TranslateWord3(Translator *tr, char *word_start, WORD_TAB *wtab, char
 	return dictionary_flags[0];
 }
 
+// append plural suffixes depending on preceding letter
+void addPluralSuffixes(int flags, Translator *tr, char last_char, char *word_phonemes)
+{
+	char word_zz[4] = { 0, 'z', 'z', 0 };
+	char word_iz[4] = { 0, 'i', 'z', 0 };
+	char word_ss[4] = { 0, 's', 's', 0 };
+	if (flags & FLAG_HAS_PLURAL) {
+		// s or 's suffix, append [s], [z] or [Iz] depending on previous letter
+		if (last_char == 'f')
+			TranslateRules(tr, &word_ss[1], word_phonemes, N_WORD_PHONEMES,
+			NULL, 0, NULL);
+		else if ((last_char == 0) || (strchr_w("hsx", last_char) == NULL))
+			TranslateRules(tr, &word_zz[1], word_phonemes, N_WORD_PHONEMES,
+			NULL, 0, NULL);
+		else
+			TranslateRules(tr, &word_iz[1], word_phonemes, N_WORD_PHONEMES,
+			NULL, 0, NULL);
+	}
+}
+
 int TranslateWord(Translator *tr, char *word_start, WORD_TAB *wtab, char *word_out)
 {
 	char words_phonemes[N_WORD_PHONEMES]; // a word translated into phoneme codes
@@ -2348,7 +2355,8 @@ void TranslateClause(Translator *tr, int *tone_out, char **voice_change)
 					}
 				} else {
 					if ((all_upper_case) && (letter_count > 2)) {
-						if ((c == 's') && (next_in == ' ')) {
+						// Flag as plural only English
+						if (tr->translator_name == L('e', 'n') && (c == 's') && (next_in == ' ')) {
 							c = ' ';
 							all_upper_case |= FLAG_HAS_PLURAL;
 
diff --git a/tests/translate.test b/tests/translate.test
index 729e445e..75a93978 100755
--- a/tests/translate.test
+++ b/tests/translate.test
@@ -27,6 +27,12 @@ test_phonemes lv \
 test_phonemes en "d'eIbju:tI2d" "débuted"
 test_phonemes en-US "d'eIbju:t#I#d" "débuted"
 
+# s at the end of abbreviations
+# https://github.com/espeak-ng/espeak-ng/issues/706
+test_phonemes en ",aIb,i:;'Em m'It 'Ib@mz m'Its ,aIb,i:;'Em ,Em,aIt'i:; ,eIp,i:;'eItS s,i:;,i:;'Es ,aIt,i:;'Eks ,aIb,i:;'Emz ,Em,aIt'i:z ,eIp,i:;'eItSIz s,i:;,i:;'EsIz ,aIt,i:;'EksIz" "ibm mit ibms mits IBM MIT APH CES ITX IBMs MIT's APHs CES's ITXs"
+test_phonemes lv "'ibm m'it 'ibm-s m'its 'ibm m'it 'aph ts'Es 'it_ks 'ibm-s m'its 'aphs ts'Ess 'it_kss" "ibm mit ibms mits IBM MIT APH CES ITX IBMs MIT's APHs CES's ITXs"
+test_phonemes ru "(en),aIb,i:;'Em m'It 'Ib@mz m'Its ,aIb,i:;'Em ,Em,aIt'i:; ,eIp,i:;'eItS s,i:;,i:;'Es ,aIt,i:;'Eks 'Ib@mz m'Its 'afz s'EI2zI2z 'ItEksz(ru)" "ibm mit ibms mits IBM MIT APH CES ITX IBMs MIT's APHs CES's ITXs"
+
 # bug: https://github.com/nvaccess/nvda/issues/7740
 test_phonemes ta "'il." "ள்"
 test_phonemes my "kon'i" "7"