/src/mozilla-central/dom/media/webspeech/synth/SpeechSynthesis.cpp
Line | Count | Source (jump to first uncovered line) |
1 | | /* -*- Mode: C++; tab-width: 2; indent-tabs-mode: nil; c-basic-offset: 2 -*- */ |
2 | | /* vim:set ts=2 sw=2 sts=2 et cindent: */ |
3 | | /* This Source Code Form is subject to the terms of the Mozilla Public |
4 | | * License, v. 2.0. If a copy of the MPL was not distributed with this |
5 | | * file, You can obtain one at http://mozilla.org/MPL/2.0/. */ |
6 | | |
7 | | #include "nsISupportsPrimitives.h" |
8 | | #include "nsSpeechTask.h" |
9 | | #include "mozilla/Logging.h" |
10 | | |
11 | | #include "mozilla/dom/ContentChild.h" |
12 | | #include "mozilla/dom/Element.h" |
13 | | |
14 | | #include "mozilla/dom/SpeechSynthesisBinding.h" |
15 | | #include "SpeechSynthesis.h" |
16 | | #include "nsContentUtils.h" |
17 | | #include "nsSynthVoiceRegistry.h" |
18 | | #include "nsIDocument.h" |
19 | | #include "nsIDocShell.h" |
20 | | |
21 | | #undef LOG |
22 | | mozilla::LogModule* |
23 | | GetSpeechSynthLog() |
24 | 0 | { |
25 | 0 | static mozilla::LazyLogModule sLog("SpeechSynthesis"); |
26 | 0 |
|
27 | 0 | return sLog; |
28 | 0 | } |
29 | 0 | #define LOG(type, msg) MOZ_LOG(GetSpeechSynthLog(), type, msg) |
30 | | |
31 | | namespace mozilla { |
32 | | namespace dom { |
33 | | |
34 | | NS_IMPL_CYCLE_COLLECTION_CLASS(SpeechSynthesis) |
35 | | |
36 | 0 | NS_IMPL_CYCLE_COLLECTION_UNLINK_BEGIN_INHERITED(SpeechSynthesis, DOMEventTargetHelper) |
37 | 0 | NS_IMPL_CYCLE_COLLECTION_UNLINK(mCurrentTask) |
38 | 0 | NS_IMPL_CYCLE_COLLECTION_UNLINK(mSpeechQueue) |
39 | 0 | tmp->mVoiceCache.Clear(); |
40 | 0 | NS_IMPL_CYCLE_COLLECTION_UNLINK_END |
41 | | |
42 | 0 | NS_IMPL_CYCLE_COLLECTION_TRAVERSE_BEGIN_INHERITED(SpeechSynthesis, DOMEventTargetHelper) |
43 | 0 | NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mCurrentTask) |
44 | 0 | NS_IMPL_CYCLE_COLLECTION_TRAVERSE(mSpeechQueue) |
45 | 0 | for (auto iter = tmp->mVoiceCache.Iter(); !iter.Done(); iter.Next()) { |
46 | 0 | SpeechSynthesisVoice* voice = iter.UserData(); |
47 | 0 | cb.NoteXPCOMChild(voice); |
48 | 0 | } |
49 | 0 | NS_IMPL_CYCLE_COLLECTION_TRAVERSE_END |
50 | | |
51 | 0 | NS_INTERFACE_MAP_BEGIN_CYCLE_COLLECTION(SpeechSynthesis) |
52 | 0 | NS_INTERFACE_MAP_ENTRY(nsIObserver) |
53 | 0 | NS_INTERFACE_MAP_ENTRY(nsISupportsWeakReference) |
54 | 0 | NS_INTERFACE_MAP_END_INHERITING(DOMEventTargetHelper) |
55 | | |
56 | | NS_IMPL_ADDREF_INHERITED(SpeechSynthesis, DOMEventTargetHelper) |
57 | | NS_IMPL_RELEASE_INHERITED(SpeechSynthesis, DOMEventTargetHelper) |
58 | | |
59 | | SpeechSynthesis::SpeechSynthesis(nsPIDOMWindowInner* aParent) |
60 | | : DOMEventTargetHelper(aParent) |
61 | | , mHoldQueue(false) |
62 | | , mInnerID(aParent->WindowID()) |
63 | 0 | { |
64 | 0 | MOZ_ASSERT(NS_IsMainThread()); |
65 | 0 |
|
66 | 0 | nsCOMPtr<nsIObserverService> obs = mozilla::services::GetObserverService(); |
67 | 0 | if (obs) { |
68 | 0 | obs->AddObserver(this, "inner-window-destroyed", true); |
69 | 0 | obs->AddObserver(this, "synth-voices-changed", true); |
70 | 0 | } |
71 | 0 |
|
72 | 0 | } |
73 | | |
74 | | SpeechSynthesis::~SpeechSynthesis() |
75 | 0 | { |
76 | 0 | } |
77 | | |
78 | | JSObject* |
79 | | SpeechSynthesis::WrapObject(JSContext* aCx, JS::Handle<JSObject*> aGivenProto) |
80 | 0 | { |
81 | 0 | return SpeechSynthesis_Binding::Wrap(aCx, this, aGivenProto); |
82 | 0 | } |
83 | | |
84 | | bool |
85 | | SpeechSynthesis::Pending() const |
86 | | { |
87 | | switch (mSpeechQueue.Length()) { |
88 | | case 0: |
89 | | return false; |
90 | | |
91 | | case 1: |
92 | | return mSpeechQueue.ElementAt(0)->GetState() == SpeechSynthesisUtterance::STATE_PENDING; |
93 | | |
94 | | default: |
95 | | return true; |
96 | | } |
97 | | } |
98 | | |
99 | | bool |
100 | | SpeechSynthesis::Speaking() const |
101 | 0 | { |
102 | 0 | if (!mSpeechQueue.IsEmpty() && |
103 | 0 | mSpeechQueue.ElementAt(0)->GetState() == SpeechSynthesisUtterance::STATE_SPEAKING) { |
104 | 0 | return true; |
105 | 0 | } |
106 | 0 | |
107 | 0 | // Returns global speaking state if global queue is enabled. Or false. |
108 | 0 | return nsSynthVoiceRegistry::GetInstance()->IsSpeaking(); |
109 | 0 | } |
110 | | |
111 | | bool |
112 | | SpeechSynthesis::Paused() const |
113 | 0 | { |
114 | 0 | return mHoldQueue || (mCurrentTask && mCurrentTask->IsPrePaused()) || |
115 | 0 | (!mSpeechQueue.IsEmpty() && mSpeechQueue.ElementAt(0)->IsPaused()); |
116 | 0 | } |
117 | | |
118 | | bool |
119 | | SpeechSynthesis::HasEmptyQueue() const |
120 | 0 | { |
121 | 0 | return mSpeechQueue.Length() == 0; |
122 | 0 | } |
123 | | |
124 | | bool SpeechSynthesis::HasVoices() const |
125 | 0 | { |
126 | 0 | uint32_t voiceCount = mVoiceCache.Count(); |
127 | 0 | if (voiceCount == 0) { |
128 | 0 | nsresult rv = nsSynthVoiceRegistry::GetInstance()->GetVoiceCount(&voiceCount); |
129 | 0 | if(NS_WARN_IF(NS_FAILED(rv))) { |
130 | 0 | return false; |
131 | 0 | } |
132 | 0 | } |
133 | 0 | |
134 | 0 | return voiceCount != 0; |
135 | 0 | } |
136 | | |
137 | | void |
138 | | SpeechSynthesis::Speak(SpeechSynthesisUtterance& aUtterance) |
139 | 0 | { |
140 | 0 | if (!mInnerID) { |
141 | 0 | return; |
142 | 0 | } |
143 | 0 | |
144 | 0 | if (aUtterance.mState != SpeechSynthesisUtterance::STATE_NONE) { |
145 | 0 | // XXX: Should probably raise an error |
146 | 0 | return; |
147 | 0 | } |
148 | 0 | |
149 | 0 | mSpeechQueue.AppendElement(&aUtterance); |
150 | 0 | aUtterance.mState = SpeechSynthesisUtterance::STATE_PENDING; |
151 | 0 |
|
152 | 0 | // If we only have one item in the queue, we aren't pre-paused, and |
153 | 0 | // we have voices available, speak it. |
154 | 0 | if (mSpeechQueue.Length() == 1 && !mCurrentTask && !mHoldQueue && HasVoices()) { |
155 | 0 | AdvanceQueue(); |
156 | 0 | } |
157 | 0 | } |
158 | | |
159 | | void |
160 | | SpeechSynthesis::AdvanceQueue() |
161 | 0 | { |
162 | 0 | LOG(LogLevel::Debug, |
163 | 0 | ("SpeechSynthesis::AdvanceQueue length=%zu", mSpeechQueue.Length())); |
164 | 0 |
|
165 | 0 | if (mSpeechQueue.IsEmpty()) { |
166 | 0 | return; |
167 | 0 | } |
168 | 0 | |
169 | 0 | RefPtr<SpeechSynthesisUtterance> utterance = mSpeechQueue.ElementAt(0); |
170 | 0 |
|
171 | 0 | nsAutoString docLang; |
172 | 0 | nsCOMPtr<nsPIDOMWindowInner> window = GetOwner(); |
173 | 0 | nsIDocument* doc = window ? window->GetExtantDoc() : nullptr; |
174 | 0 |
|
175 | 0 | if (doc) { |
176 | 0 | Element* elm = doc->GetHtmlElement(); |
177 | 0 |
|
178 | 0 | if (elm) { |
179 | 0 | elm->GetLang(docLang); |
180 | 0 | } |
181 | 0 | } |
182 | 0 |
|
183 | 0 | mCurrentTask = |
184 | 0 | nsSynthVoiceRegistry::GetInstance()->SpeakUtterance(*utterance, docLang); |
185 | 0 |
|
186 | 0 | if (mCurrentTask) { |
187 | 0 | mCurrentTask->SetSpeechSynthesis(this); |
188 | 0 | } |
189 | 0 | } |
190 | | |
191 | | void |
192 | | SpeechSynthesis::Cancel() |
193 | 0 | { |
194 | 0 | if (!mSpeechQueue.IsEmpty() && |
195 | 0 | mSpeechQueue.ElementAt(0)->GetState() == SpeechSynthesisUtterance::STATE_SPEAKING) { |
196 | 0 | // Remove all queued utterances except for current one, we will remove it |
197 | 0 | // in OnEnd |
198 | 0 | mSpeechQueue.RemoveElementsAt(1, mSpeechQueue.Length() - 1); |
199 | 0 | } else { |
200 | 0 | mSpeechQueue.Clear(); |
201 | 0 | } |
202 | 0 |
|
203 | 0 | if (mCurrentTask) { |
204 | 0 | mCurrentTask->Cancel(); |
205 | 0 | } |
206 | 0 | } |
207 | | |
208 | | void |
209 | | SpeechSynthesis::Pause() |
210 | 0 | { |
211 | 0 | if (Paused()) { |
212 | 0 | return; |
213 | 0 | } |
214 | 0 | |
215 | 0 | if (mCurrentTask && !mSpeechQueue.IsEmpty() && |
216 | 0 | mSpeechQueue.ElementAt(0)->GetState() == SpeechSynthesisUtterance::STATE_SPEAKING) { |
217 | 0 | mCurrentTask->Pause(); |
218 | 0 | } else { |
219 | 0 | mHoldQueue = true; |
220 | 0 | } |
221 | 0 | } |
222 | | |
223 | | void |
224 | | SpeechSynthesis::Resume() |
225 | 0 | { |
226 | 0 | if (!Paused()) { |
227 | 0 | return; |
228 | 0 | } |
229 | 0 | |
230 | 0 | mHoldQueue = false; |
231 | 0 |
|
232 | 0 | if (mCurrentTask) { |
233 | 0 | mCurrentTask->Resume(); |
234 | 0 | } else { |
235 | 0 | AdvanceQueue(); |
236 | 0 | } |
237 | 0 | } |
238 | | |
239 | | void |
240 | | SpeechSynthesis::OnEnd(const nsSpeechTask* aTask) |
241 | 0 | { |
242 | 0 | MOZ_ASSERT(mCurrentTask == aTask); |
243 | 0 |
|
244 | 0 | if (!mSpeechQueue.IsEmpty()) { |
245 | 0 | mSpeechQueue.RemoveElementAt(0); |
246 | 0 | } |
247 | 0 |
|
248 | 0 | mCurrentTask = nullptr; |
249 | 0 | AdvanceQueue(); |
250 | 0 | } |
251 | | |
252 | | void |
253 | | SpeechSynthesis::GetVoices(nsTArray< RefPtr<SpeechSynthesisVoice> >& aResult) |
254 | 0 | { |
255 | 0 | aResult.Clear(); |
256 | 0 | uint32_t voiceCount = 0; |
257 | 0 | nsCOMPtr<nsPIDOMWindowInner> window = GetOwner(); |
258 | 0 | nsCOMPtr<nsIDocShell> docShell = window ? window->GetDocShell() : nullptr; |
259 | 0 |
|
260 | 0 |
|
261 | 0 | if (nsContentUtils::ShouldResistFingerprinting(docShell)) { |
262 | 0 | return; |
263 | 0 | } |
264 | 0 | |
265 | 0 | nsresult rv = nsSynthVoiceRegistry::GetInstance()->GetVoiceCount(&voiceCount); |
266 | 0 | if(NS_WARN_IF(NS_FAILED(rv))) { |
267 | 0 | return; |
268 | 0 | } |
269 | 0 | |
270 | 0 | nsISupports* voiceParent = NS_ISUPPORTS_CAST(nsIObserver*, this); |
271 | 0 |
|
272 | 0 | for (uint32_t i = 0; i < voiceCount; i++) { |
273 | 0 | nsAutoString uri; |
274 | 0 | rv = nsSynthVoiceRegistry::GetInstance()->GetVoice(i, uri); |
275 | 0 |
|
276 | 0 | if (NS_FAILED(rv)) { |
277 | 0 | NS_WARNING("Failed to retrieve voice from registry"); |
278 | 0 | continue; |
279 | 0 | } |
280 | 0 |
|
281 | 0 | SpeechSynthesisVoice* voice = mVoiceCache.GetWeak(uri); |
282 | 0 |
|
283 | 0 | if (!voice) { |
284 | 0 | voice = new SpeechSynthesisVoice(voiceParent, uri); |
285 | 0 | } |
286 | 0 |
|
287 | 0 | aResult.AppendElement(voice); |
288 | 0 | } |
289 | 0 |
|
290 | 0 | mVoiceCache.Clear(); |
291 | 0 |
|
292 | 0 | for (uint32_t i = 0; i < aResult.Length(); i++) { |
293 | 0 | SpeechSynthesisVoice* voice = aResult[i]; |
294 | 0 | mVoiceCache.Put(voice->mUri, voice); |
295 | 0 | } |
296 | 0 | } |
297 | | |
298 | | // For testing purposes, allows us to cancel the current task that is |
299 | | // misbehaving, and flush the queue. |
300 | | void |
301 | | SpeechSynthesis::ForceEnd() |
302 | 0 | { |
303 | 0 | if (mCurrentTask) { |
304 | 0 | mCurrentTask->ForceEnd(); |
305 | 0 | } |
306 | 0 | } |
307 | | |
308 | | NS_IMETHODIMP |
309 | | SpeechSynthesis::Observe(nsISupports* aSubject, const char* aTopic, |
310 | | const char16_t* aData) |
311 | 0 | { |
312 | 0 | MOZ_ASSERT(NS_IsMainThread()); |
313 | 0 |
|
314 | 0 |
|
315 | 0 | if (strcmp(aTopic, "inner-window-destroyed") == 0) { |
316 | 0 | nsCOMPtr<nsISupportsPRUint64> wrapper = do_QueryInterface(aSubject); |
317 | 0 | NS_ENSURE_TRUE(wrapper, NS_ERROR_FAILURE); |
318 | 0 |
|
319 | 0 | uint64_t innerID; |
320 | 0 | nsresult rv = wrapper->GetData(&innerID); |
321 | 0 | NS_ENSURE_SUCCESS(rv, rv); |
322 | 0 |
|
323 | 0 | if (innerID == mInnerID) { |
324 | 0 | mInnerID = 0; |
325 | 0 | Cancel(); |
326 | 0 |
|
327 | 0 | nsCOMPtr<nsIObserverService> obs = mozilla::services::GetObserverService(); |
328 | 0 | if (obs) { |
329 | 0 | obs->RemoveObserver(this, "inner-window-destroyed"); |
330 | 0 | } |
331 | 0 | } |
332 | 0 | } else if (strcmp(aTopic, "synth-voices-changed") == 0) { |
333 | 0 | LOG(LogLevel::Debug, ("SpeechSynthesis::onvoiceschanged")); |
334 | 0 | nsCOMPtr<nsPIDOMWindowInner> window = GetOwner(); |
335 | 0 | nsCOMPtr<nsIDocShell> docShell = window ? window->GetDocShell() : nullptr; |
336 | 0 |
|
337 | 0 | if (!nsContentUtils::ShouldResistFingerprinting(docShell)) { |
338 | 0 | DispatchTrustedEvent(NS_LITERAL_STRING("voiceschanged")); |
339 | 0 | // If we have a pending item, and voices become available, speak it. |
340 | 0 | if (!mCurrentTask && !mHoldQueue && HasVoices()) { |
341 | 0 | AdvanceQueue(); |
342 | 0 | } |
343 | 0 | } |
344 | 0 | } |
345 | 0 |
|
346 | 0 | return NS_OK; |
347 | 0 | } |
348 | | |
349 | | } // namespace dom |
350 | | } // namespace mozilla |