When user decides to change tts engine, and if it's not system one,
then he sees dialog warning about possibility of sending all kinds of
data to the engine.
If user chooses to not changes engine, radio button still sticks to the
new position.
This change delays all operations regarding changing current engine
after user closes dialog. It also unsets the radio button if user chooses
to cancel it.
Bug: 7628362
Change-Id: I977abe71b3547f2545a971fc0d69179be6fafb44
If engine is not active, its settings icon is disabled in
Settings > Language & input > Text-to-speech output screen.
Currently, settings icons for all TTS engines are shown at
the same opacity. This fix dims settings icons of not-selected
engines.
Make sure that more than one TTS engine are installed on the
target, for e.g., Google Text-to-speech Engine and Classic
Text To Speech Engine (SVOX Classic TTS).
Additionally, since setAlpha() is used in multiple places within
Settings package, moved DISABLED_ALPHA declaration to Utils.java
in order to have single point of reference.
Change-Id: Ifa7de79814a2f4a4aa021cd8621cbfab41655680
Signed-off-by: Shuhrat Dehkanov <uzbmaster@gmail.com>
If TTS requires network for selected language, we will
show dialog informing why we can't synthesize an example
if user have no internet connectivity.
Bug: 7398316
Change-Id: I01b033a9a8b8538b322ed3639b671f90f849ae29
If user added new voice pack, then set it as a default language,
went level back and went to TTS settings again, it ended in crash.
This is because first updateVoiceDetails uses old voice data details
that settings retreived long time ago.
locales preference is hidden until TTs engine returns list of locales.
Change-Id: I33a81ff259c69adb4c051ee693b6c41b93f302d0
Added one BroadcastListener that listens for TextToSpeech.Engine.ACTION_TTS_DATA_INSTALLED intents.
If it gets one, it will call TextToSpeech.Engine.ACTION_CHECK_TTS_DATA to update list of TTS voices.
It expects TTS engine to broadcast TTS_DATA_INSTALLED also when removing voice data.
In order to catch changes done when user sits in Text-to-speech output main activity,
TextToSpeech.Engine.ACTION_CHECK_TTS_DATA is called on start of tts engine config fragment.
Change-Id: Ic1e299ea3ffc45d938083f1ab5b2bf101fe25f26
The issue is that Fragment.setVoiceDataDetails can
be called before Fragment.getView. We guard against this
issue. No thread visibility issues here because both functions
are called on the UI thread.
bug:5884355
Change-Id: Iad91b91c58b04dcb9f34f6b5ff8752f2e8295423
The spec does not specify case sensitivity and support
for 3 letter locales is currently poor.
bug:5885830
Change-Id: I1b8f582bd19b1b0aed220036c5f17904478dba46
This is required to fix the case where an app has instantiated
and is using a TextToSpeech instance, and the user switches to
settings and changes the language for that instance. If no
app was using TTS, then onCreate() would have be called when
TTS is used next, and the correct locale will be loaded.
bug:5823836
Change-Id: I62adf3f7f0f14b67450ba08b8743e0d31e615a7e
(a) The TTS API doesn't mandate that unavailable voices should
be set and null, so deal with it correctly.
(b) The android manifest contained a reference to the old
TTS class.
bug:5525023
Change-Id: I1c16dbc75893288e5fa221c73a7284fca7ab25c5
Re-added if there are voices to install.
DO NOT MERGE : Was already submitted to master.
bug:5511343
Change-Id: Iac95f37c3dda3b213f6ba34c32b7dc725673d64d
This prevents QUEUE_FLUSH from working as intented since
there is a (blocking) setLanguage() call that waits for
the previous synthesis to complete before flushing the queue.
bug:5288327
Change-Id: I00467f0ac1584863ae27bfc076b8413e29e5379b
This makes the sample play faster, and results in a reliable
return value from getLanguage().
bug:5207139
Change-Id: Ic6ca4a2f4aed299d35c320ecc3f327579c66471b
