SpeechSynthesis.cpp [plain text]
#include "config.h"
#include "SpeechSynthesis.h"
#if ENABLE(SPEECH_SYNTHESIS)
#include "EventNames.h"
#include "PlatformSpeechSynthesisVoice.h"
#include "PlatformSpeechSynthesizer.h"
#include "ScriptController.h"
#include "SpeechSynthesisEvent.h"
#include "SpeechSynthesisUtterance.h"
#include <wtf/CurrentTime.h>
#include <wtf/NeverDestroyed.h>
namespace WebCore {
Ref<SpeechSynthesis> SpeechSynthesis::create()
{
return adoptRef(*new SpeechSynthesis);
}
SpeechSynthesis::SpeechSynthesis()
: m_currentSpeechUtterance(nullptr)
, m_isPaused(false)
#if PLATFORM(IOS)
, m_restrictions(RequireUserGestureForSpeechStartRestriction)
#endif
{
}
void SpeechSynthesis::setPlatformSynthesizer(std::unique_ptr<PlatformSpeechSynthesizer> synthesizer)
{
m_platformSpeechSynthesizer = WTFMove(synthesizer);
m_voiceList.clear();
m_currentSpeechUtterance = nullptr;
m_utteranceQueue.clear();
m_isPaused = false;
}
void SpeechSynthesis::voicesDidChange()
{
m_voiceList.clear();
}
const Vector<Ref<SpeechSynthesisVoice>>& SpeechSynthesis::getVoices()
{
if (m_voiceList.size())
return m_voiceList;
if (!m_platformSpeechSynthesizer)
m_platformSpeechSynthesizer = std::make_unique<PlatformSpeechSynthesizer>(this);
for (auto& voice : m_platformSpeechSynthesizer->voiceList())
m_voiceList.append(SpeechSynthesisVoice::create(*voice));
return m_voiceList;
}
bool SpeechSynthesis::speaking() const
{
return m_currentSpeechUtterance;
}
bool SpeechSynthesis::pending() const
{
return m_utteranceQueue.size() > 1;
}
bool SpeechSynthesis::paused() const
{
return m_isPaused;
}
void SpeechSynthesis::startSpeakingImmediately(SpeechSynthesisUtterance& utterance)
{
ASSERT(!m_currentSpeechUtterance);
utterance.setStartTime(monotonicallyIncreasingTime());
m_currentSpeechUtterance = &utterance;
m_isPaused = false;
if (utterance.text().isEmpty()) {
handleSpeakingCompleted(utterance, false);
return;
}
if (!m_platformSpeechSynthesizer)
m_platformSpeechSynthesizer = std::make_unique<PlatformSpeechSynthesizer>(this);
m_platformSpeechSynthesizer->speak(utterance.platformUtterance());
}
void SpeechSynthesis::speak(SpeechSynthesisUtterance& utterance)
{
#if PLATFORM(IOS)
if (ScriptController::processingUserGesture())
removeBehaviorRestriction(RequireUserGestureForSpeechStartRestriction);
else if (userGestureRequiredForSpeechStart())
return;
#endif
m_utteranceQueue.append(utterance);
if (m_utteranceQueue.size() == 1)
startSpeakingImmediately(m_utteranceQueue.first());
}
void SpeechSynthesis::cancel()
{
RefPtr<SpeechSynthesisUtterance> current = m_currentSpeechUtterance;
m_utteranceQueue.clear();
if (m_platformSpeechSynthesizer)
m_platformSpeechSynthesizer->cancel();
current = nullptr;
ASSERT(!m_currentSpeechUtterance);
}
void SpeechSynthesis::pause()
{
if (!m_isPaused && m_platformSpeechSynthesizer)
m_platformSpeechSynthesizer->pause();
}
void SpeechSynthesis::resume()
{
if (m_currentSpeechUtterance && m_platformSpeechSynthesizer)
m_platformSpeechSynthesizer->resume();
}
void SpeechSynthesis::fireEvent(const AtomicString& type, SpeechSynthesisUtterance& utterance, unsigned long charIndex, const String& name)
{
utterance.dispatchEvent(SpeechSynthesisEvent::create(type, charIndex, (monotonicallyIncreasingTime() - utterance.startTime()), name));
}
void SpeechSynthesis::handleSpeakingCompleted(SpeechSynthesisUtterance& utterance, bool errorOccurred)
{
ASSERT(m_currentSpeechUtterance);
Ref<SpeechSynthesisUtterance> protect(utterance);
m_currentSpeechUtterance = nullptr;
fireEvent(errorOccurred ? eventNames().errorEvent : eventNames().endEvent, utterance, 0, String());
if (m_utteranceQueue.size()) {
Ref<SpeechSynthesisUtterance> firstUtterance = m_utteranceQueue.takeFirst();
ASSERT(&utterance == firstUtterance.ptr());
if (!m_utteranceQueue.isEmpty())
startSpeakingImmediately(m_utteranceQueue.first());
}
}
void SpeechSynthesis::boundaryEventOccurred(PlatformSpeechSynthesisUtterance& utterance, SpeechBoundary boundary, unsigned charIndex)
{
static NeverDestroyed<const String> wordBoundaryString(ASCIILiteral("word"));
static NeverDestroyed<const String> sentenceBoundaryString(ASCIILiteral("sentence"));
ASSERT(utterance.client());
switch (boundary) {
case SpeechWordBoundary:
fireEvent(eventNames().boundaryEvent, static_cast<SpeechSynthesisUtterance&>(*utterance.client()), charIndex, wordBoundaryString);
break;
case SpeechSentenceBoundary:
fireEvent(eventNames().boundaryEvent, static_cast<SpeechSynthesisUtterance&>(*utterance.client()), charIndex, sentenceBoundaryString);
break;
default:
ASSERT_NOT_REACHED();
}
}
void SpeechSynthesis::didStartSpeaking(PlatformSpeechSynthesisUtterance& utterance)
{
if (utterance.client())
fireEvent(eventNames().startEvent, static_cast<SpeechSynthesisUtterance&>(*utterance.client()), 0, String());
}
void SpeechSynthesis::didPauseSpeaking(PlatformSpeechSynthesisUtterance& utterance)
{
m_isPaused = true;
if (utterance.client())
fireEvent(eventNames().pauseEvent, static_cast<SpeechSynthesisUtterance&>(*utterance.client()), 0, String());
}
void SpeechSynthesis::didResumeSpeaking(PlatformSpeechSynthesisUtterance& utterance)
{
m_isPaused = false;
if (utterance.client())
fireEvent(eventNames().resumeEvent, static_cast<SpeechSynthesisUtterance&>(*utterance.client()), 0, String());
}
void SpeechSynthesis::didFinishSpeaking(PlatformSpeechSynthesisUtterance& utterance)
{
if (utterance.client())
handleSpeakingCompleted(static_cast<SpeechSynthesisUtterance&>(*utterance.client()), false);
}
void SpeechSynthesis::speakingErrorOccurred(PlatformSpeechSynthesisUtterance& utterance)
{
if (utterance.client())
handleSpeakingCompleted(static_cast<SpeechSynthesisUtterance&>(*utterance.client()), true);
}
}
#endif // ENABLE(SPEECH_SYNTHESIS)