* MicRecorderAPI: allow duration limit to be disabled with zero or negative
* MicRecorderAPI: avoid unnecessary MediaRecorder parameter in RecorderCommandHandler.handle()
* MicRecorderAPI: allow encoder and format to be set by user
All supported encoders can be set with their names. Unless a format
is set explicitly with its corresponding integer, a suitable format
would be chosen automatically for the chosen encoder.
* MicRecorderAPI: allow AudioSource to be set by user
Just in case, so it can only be set with the source's corresponding
integer.
* MicRecorderAPI: allow bit rate, sampling rate and number of channels to be set by user
* MicRecorderAPI: a couple of fixes
Have year before month and day in the default file name for better
sorting. Also fix the extension name in it for automatic format
selection. As we want the file to have extension name by default,
we have to pick a default format/encoder. AAC is chosen because it
is probably the most compatible one among different platforms.
* MicRecorderAPI: replace all getAbsoluteFile() with getAbsolutePath()
None of them were used in places where it should be a File.
* MicRecorderAPI: properly report unlimited MaxDuration
* MicRecorderAPI: make sure the recorder is released when not used
* MicRecorderAPI: remove aac_eld/he_aac/vorbis support
They are at least problematic on my phone. They either fail to work
or crash the API app. Sometimes the crash is even unrecoverable.
* Changed provider to scope to all messages. Future update should be able to provide a flag on which scope the user would like to request. All, Sent, Draft, Inbox, Outbox
* Add ability to specify the type of message to query for
If termux-dialog is called while Termux already has a dialog open, the new dialog fails to open and the command will hang until it is manually terminated. This fix allows multiple dialogs to be open at once, ensuring that termux-dialog will not hang.
* AudioAPI: simplify code
* AudioAPI: remove volume info
Redundunt with the introduction of VolumeAPI
* AudioAPI: remove PROPERTY_SUPPORT_AUDIO_SOURCE_UNPROCESSED
It does not seem to have a point for us to get this property as
it is not very clear what it actually reflects.
We have not been getting PROPERTY_SUPPORT_MIC_NEAR_ULTRASOUND and
PROPERTY_SUPPORT_SPEAKER_NEAR_ULTRASOUND either.
In fact, properties from getProperty() tend to be rather useless.
PROPERTY_OUTPUT_FRAMES_PER_BUFFER and PROPERTY_OUTPUT_SAMPLE_RATE
are kept only to show that, they should NOT be used to optimize
any player in audio output, as they do not change as per the
current sink, not like the values from the AudioTrack methods do.
* AudioAPI: order output
* AudioAPI: check info in PERFORMANCE_MODE_POWER_SAVING
When setPerformanceMode was introduced in Android O, it was
introduced with three possible modes, namely:
PERFORMANCE_MODE_NONE (default)
PERFORMANCE_MODE_LOW_LATENCY
PERFORMANCE_MODE_POWER_SAVING
While PERFORMANCE_MODE_NONE essentially causes tracks to be routed
to the deep buffer mixer path on my phone (as
PERFORMANCE_MODE_POWER_SAVING does), it may not be the case on
every single device.
Therefore, check info with track in PERFORMANCE_MODE_POWER_SAVING,
compare it against that of the default PERFORMANCE_MODE_NONE, and
print it out if they are not the same.
Note: shouldEnablePowerSaving for PERFORMANCE_MODE_NONE always
returns false for SAMPLE_RATE_UNSPECIFIED as of the current
AudioTrack.java.
* DialogAPI Improvements / Enhancements
* New Widgets: Date, Time, Spinner, Text-To-Speech, Confirmation
* Results are returned as JSON including code of button pressed
* Fix typo
* Add BottomSheet InputMethod type
* Change DialogActivity to AppCompatActivity
* Add support library dependency to gradle
* Change theme to be compatible with AppCompatActivity
* Refactor getSpinnerItems to getInputValues so BottomSheet can use
as well
* Add Radio InputMethod
* Better touch outside dialog cancellation behavior
* Prevent speech InputMethod from being cancelled on touch outside
* Add Counter InputMethod
* Add CheckBox InputMethod
* Change FrameLayout to be NestedScrollView to handle scrolling
large dialogs as necessary
* Add custom range for CounterInputMethod
* Set text for ConfirmInputMethod
* Add numeric option / other minor tweaks
* change password type to a boolean option instead of string
* allow multiline password (text only not numeric)
* Allow text hint to be set for SpeechInputMethod
From Android Studio lint:
On versions prior to Android N (24), initializing the WifiManager via
Context#getSystemService can cause a memory leak if the context is not
the application context.
In many cases, it's not obvious from the code where the Context is coming
from (e.g. it might be a parameter to a method, or a field initialized
from various method calls). It's possible that the context being passed in
is the application context, but to be on the safe side, you should consider
changing context.getSystemService(...) to
context.getApplicationContext().getSystemService(...).
Due to https://issuetracker.google.com/issues/37060483:
WifiManager#getScanResults() returns an empty array list
if GPS is turned off
the user needs to enable Location on the device for termux-wifi-scaninfo
to work. So show "Location needs to be enabled on the device" if necessary.
* MediaPlayerAPI: revert commit 299d557
The error message won't be displayed anyway:
$ touch empty
$ dd if=/dev/urandom of=random count=2048
2048+0 records in
2048+0 records out
1048576 bytes (1.0MB) copied, 0.190185 seconds, 5.3MB/s
$ termux-media-player play empty
setDataSourceFD failed.: status=0x80000000
$ termux-media-player play random
Prepare failed.: status=0x1
$
* MediaPlayerAPI: catch NullPointerException for -a play
* MediaPlayerAPI: use getCanonicalPath() instead
* MediaPlayerAPI: use proper values for setVolume()
* MediaPlayerAPI: fix play and stop from pause
* MediaPlayerAPI: add trackName for info and resume
Suggested by Auxilus (#162)
There are several changes in the commit:
1. Use instance methods to get sample rate and buffer size info
on Android N or above only instead of Android M, since both
AudioAttributes.FLAG_LOW_LATENCY (256) and the implicitly used
AudioFormat.SAMPLE_RATE_UNSPECIFIED (0) are available only since
API 24 (N).
2. Use setPerformanceMode() on Android O, as FLAG_LOW_LATENCY is
deprecated since API 26 (O).
3. Create AudioTrack instance of both the fast (LOW_LATENCY) and
deep buffer (POWER_SAVING) mixer path. The two paths can be using
different subdevices on different rate and buffer size (or same
subdevice on same rate but different buffer size). For example on
my phone:
AUDIOTRACK_SAMPLE_RATE: 192000
AUDIOTRACK_BUFFER_SIZE_IN_FRAMES: 23046
AUDIOTRACK_SAMPLE_RATE_LOW_LATENCY: 48000
AUDIOTRACK_BUFFER_SIZE_IN_FRAMES_LOW_LATENCY: 192
4. Get info on sample rate with the getNativeOutputSampleRate()
static method (static in Java's sense; the value is still sink-
based) on builds older than N only. The rate appears to be of the
fast (LOW_LATENCY) mixer path if it is available.
5. A note on the static method getMinBufferSize(). The reason it
is not used even for older devices is, it returns buffer size info
of the deep buffer (POWER_SAVING) mixer path, while NOSR is the
rate of the fast (LOW_LATENCY) mixer path, so the value we can get
with it is not always sensical.
