Release 0.1.43 [auto-version-bump]

VERSION

@@ -1 +1 @@
-0.1.41
+0.1.43

deploy/helm/ai-spel-frontend/Chart.yaml

@@ -2,5 +2,5 @@ apiVersion: v2
 name: ai-spel-frontend
 description: Static web frontend for the Swedish life simulator
 type: application
-version: 0.1.41
+version: 0.1.43
-appVersion: "0.1.41"
+appVersion: "0.1.43"

index.html

@@ -93,7 +93,6 @@
     const voiceSettingsKey = 'ai-spel.voice-settings.v2';
     let voiceSettings = window.GameVoice.loadSettings(localStorage, voiceSettingsKey, 'ai-spel.voice-settings.v1');
     let voiceController;
-    let speechActivatedByGesture = false;
     let audioContext;
     let voiceDisclosureAcknowledged = false;
     let adminOffset = 0;
@@ -228,8 +227,8 @@
       },
     });
     function primeSpeechFromGesture(event) {
-      if (speechActivatedByGesture || !event.isTrusted) return;
+      if (!event.isTrusted) return;
-      speechActivatedByGesture = voiceController.activateFromUserGesture();
+      voiceController.activateFromUserGesture();
     }
     document.addEventListener('pointerdown', primeSpeechFromGesture, true);
     document.addEventListener('keydown', primeSpeechFromGesture, true);
@@ -739,6 +738,10 @@
       refreshMicrophone();
       voiceController.stopListening();
       voiceController.cancelSpeech();
+      // cancel() above may also have discarded Safari's silent speech-channel
+      // primer. Restore it synchronously inside the submit gesture, before
+      // waiting for the API response that will be read aloud.
+      voiceController.activateFromUserGesture();
       messageField.value = '';
       try {
         if (stage === 'biography') await handleBiography(message);

scripts/check-frontend.sh

@@ -28,7 +28,7 @@ grep -q 'Jag förstår – starta mikrofonen' index.html
 grep -q 'supportsRecognition' voice.js
 grep -q 'characterVoiceProfile' voice.js
 grep -q 'utterance.volume = 1' voice.js
-grep -q 'Uppläsningen har startat' voice.js
+grep -q 'Webbläsaren har startat talsyntesen' voice.js
 grep -q 'WebKit versions before Safari 27' voice.js
 grep -q "voiceIdentity: speakingNpc.voice_identity" index.html
 grep -q "role: 'narrator', voiceIdentity: 'neutral'" index.html

tests/voice.test.cjs

@@ -117,17 +117,33 @@ test('recognition watchdog closes a session that never ends and permits retry',
   assert.equal(h.recognitions.length, 2);
 });
 
-test('TTS selects Swedish fallback, applies rate, and cancellation invalidates callbacks', () => {
+test('TTS uses the system default voice unless a named Swedish voice is selected', () => {
   const h = makeHarness();
   h.controller.update({mode: 'game-voice', rate: 1.4});
   assert.equal(h.controller.speak('Hej.'), true);
   const first = h.utterances[0];
   assert.equal(first.lang, 'sv-SE');
   assert.equal(first.rate, 1.4);
-  assert.equal(first.voice.name, 'Svensk');
+  assert.equal(first.voice, null);
   h.controller.cancelSpeech();
   first.onend();
   assert.equal(h.controller.getState(), 'IDLE');
+  assert.equal(h.controller.speak('Namngiven röst.', {voiceName: 'Svensk'}), true);
+  assert.equal(h.utterances.at(-1).voice.name, 'Svensk');
+});
+
+test('a player gesture restores the iOS speech primer after cancelling a queued reply', () => {
+  const h = makeHarness();
+  assert.equal(h.controller.activateFromUserGesture(), true);
+  assert.equal(h.synthesis.spoken[0].text, ' ');
+  h.controller.speak('Tidigare replik');
+  h.controller.cancelSpeech();
+  const countAfterCancel = h.synthesis.spoken.length;
+  assert.equal(h.controller.activateFromUserGesture(), true);
+  assert.equal(h.synthesis.spoken.length, countAfterCancel + 1);
+  assert.equal(h.synthesis.spoken.at(-1).text, ' ');
+  h.controller.speak('Nytt svar');
+  assert.equal(h.synthesis.spoken.at(-1).text, 'Nytt svar');
 });
 
 test('speech is reported as in progress only after the native onstart event', () => {
@@ -137,8 +153,8 @@ test('speech is reported as in progress only after the native onstart event', ()
   assert.match(h.states.at(-1).message, /Startar uppläsningen/);
   h.utterances[0].onstart();
   assert.equal(h.controller.getState(), 'SPEAKING');
-  assert.match(h.states.at(-1).message, /Uppläsningen har startat/);
+  assert.match(h.states.at(-1).message, /Webbläsaren har startat talsyntesen/);
-  assert.match(h.states.at(-1).message, /iPadens volym/);
+  assert.match(h.states.at(-1).message, /Läs senaste repliken/);
 });
 
 test('iOS user gesture primes speech silently and idle speech is not cancelled before speak', () => {
@@ -214,7 +230,7 @@ test('TTS and microphone are mutually exclusive', () => {
   h.controller.speak('Svar');
   assert.equal(h.recognitions[0].aborted, true);
   assert.equal(h.controller.getState(), 'QUEUED');
-  h.utterances[0].onstart();
+  h.utterances.at(-1).onstart();
   assert.equal(h.controller.getState(), 'SPEAKING');
 });
 

voice.js

@@ -100,6 +100,9 @@
       // same turn as cancel(). Do not clear an already-empty native queue.
       if (wasQueued) {
         try { win.speechSynthesis?.cancel(); } catch (_) { /* Best effort. */ }
+        // cancel() also removes the silent user-gesture primer from the native
+        // queue. A later reply must prime again from the next player gesture.
+        speechPrimed = false;
       }
       if (!recognition) setState('IDLE');
     }
@@ -149,6 +152,10 @@
         return false;
       }
       cancelSpeech();
+      // This runs synchronously from the microphone button gesture. If
+      // cancelSpeech() removed the earlier primer, restore it before starting
+      // recognition so a spoken reply can follow the transcript.
+      activateFromUserGesture();
       const token = ++generation;
       let instance;
       try { instance = new Recognition(); }
@@ -226,15 +233,22 @@
       utterance.pitch = isNarrator ? 1 : Math.min(2, Math.max(0, Number(voiceOptions.pitch ?? profile.pitch ?? identityProfile.pitch) || 1));
       const requestedName = voiceOptions.voiceName || profile.voiceName || settings.voiceName;
       const voices = availableVoices();
-      utterance.voice = voices.find(voice => voice.name === requestedName)
+      // An empty selection means the operating system should choose its
-        || voices.find(voice => /^sv(-|$)/i.test(voice.lang))
+      // default voice for the requested language. Forcing the first voice
-        || null;
+      // returned by WebKit can select a voice that is listed but cannot speak.
+      utterance.voice = requestedName
+        ? voices.find(voice => voice.name === requestedName)
+          || voices.find(voice => /^sv(-|$)/i.test(voice.lang))
+          || null
+        : null;
       utterance.onstart = () => {
         if (token !== generation) return;
         if (speechStartTimer !== null) win.clearTimeout(speechStartTimer);
         speechStartTimer = null;
-        setState('SPEAKING', 'Uppläsningen har startat. Om du inte hör ljud, kontrollera iPadens volym och ljudutgång.');
         if (typeof voiceOptions.onStart === 'function') voiceOptions.onStart();
+        // onstart confirms the browser speech engine accepted the utterance;
+        // it cannot confirm that sound reached the device's speakers.
+        setState('SPEAKING', 'Webbläsaren har startat talsyntesen. Om du inte hör ljud, tryck på Läs senaste repliken.');
       };
       utterance.onend = () => {
         if (token !== generation) return;