1. API changes: Moved copies API from PrintAttributes to PrintJobInfo;
Changed the PageRange list to an array in PrintDocumentAdapter#onWrite;
Added onCancelled method to the layout and write callbacks.
2. Refactored the serialization of remote layout and write commands. Now
the commands are serialized by the code in the client instead in the spooler.
The benefit is simple code since the client has to do a serialization to delegate
to the main thread anyway. The increased IPC found is fine since these calls
are quite unfrequent.
3. Removed an unused file: IPrintSpoolerObserver.aidl
4. Added equals and hasCode implementation to PageRange, PrintAttributes,
MediaSize, Resolution, Margins, Tray, PrintDocumentInfo.
5. Added shortcut path for query APIs on PrintJob that return cached values
if the print job is in a uncuttable state, i.e. completed or cancelled. Failed
print jobs can be restarted.
6. PrintJobInfo was not properly serialized.
7. Updated the look of the print dialog to be stable if there is and there isn't
currently selected printer.
8. PrintJobCOnfigActivity now calls onLayout on every print attributes change
but requests a write only on print preview or print button press. Also if the
layout did not change the content and it is already written no subsequent
call is made. Also if the selected pages change and we already have them
no subsequent call to write is made. Also the app is called with print preview
attribute set when performing layout and with it cleared after the print button
is pressed. A lot of changes making sure that only valid actions are enabled
in the activity (looks like a dialog) at a given time frame. The print job config
activity is also hidden after we got all the data, i.e. layout and write are done.
9. The callback from the print spooler to the system are scheduled via messages
to avoid lock being held during the call. It was hard to guarantee that since a
method holding a lock may be calling one that would like to release the lock
at some point to make the callbacks.
10. Print spooler state is persisted only if something changes in a completed
print job, i.e. not one that is being constructed due the print job config dialog.
11. Fixed a potential race in the RemotePrintSpooler where it was possible that
a client that got a handle to the remote spooler calls into an unbound spooler.
E.g: the client gets the remote interface with a lock held, now the client releases
the lock to avoid IPC with a lock, during the IPC scheduling the spooler has
notified the system that it is done and the system unbinds from it, now the
client's IPC is made to a spooler that is disconnected.
Change-Id: Ie9c42255940a27ecaed21a4d326a663a4788ac9d
1. Added support for reporting the old print attributes during layout.
Now we keep track of the old print attributes, so the app can
compute the delta and decide whether re-layout work is needed.
2. Fixed PrintDocumentAdapter callback interleavings. Layout callbacks
were intermixing with write ones - a mess. Now we make an attempt
to cancel layout and write if they respond to cancellation, otherwise
we wait but do not interleave them.
3. Refactored the PrintJobConfigActivity for easier maintenance and
to have a single update UI method that does the minimal amount
of work.
Change-Id: I31ada1a0550882e6185018e6f17f923aed165d15
1. Implemented the persistence and restoring of the print spooler state.
The print spooler state is saved as an XML on every print job change
and is restored when we bind to the spooler. The system does not
unbind from the spooler until the state persistence completes. We
are now storing the entire state, i.e. all print jobs, when a single
one changes. This is not optimal but we are not expecting to have
many such at the same time, so for now we err for simplicity of
implementation.
2. Enforcing a non-empty print job name.
3. Hidden the STATE_CREATED print job state which should never be visible to a
client since this is the state of a print job during construction, i.e. the
print dialog is up and we are doing back and forth with the app.
4. Fixed some PrintAttributes APIs that were incorrectly taking in a PackageManager
instance.
5. Updated the PrintSpooler build file due to splitting the framework into multiple
jars.
Change-Id: I52c88eaa1ec9c64920359cc143c79832a4c3d25b
- Make sure Home activity goes in the correct task and on the correct
stack.
- Do not allow different users to be in the same task.
- Do not set stacks aside for each user.
Fixes bug 9775492.
Change-Id: I0e7954e917aac8482a1015a36923e02914e2b692
Handle audio sample start/stop request in a handler
to avoid race conditions.
If a sample is active, do not ignore playback requests but
schedule a second attempt one second later. This is to avoid
ignoring a playback request if a sample is short or about to finish.
Bug: 8176776.
Change-Id: If2d5f7d469eba345af3f4771067a076a56f4bfed
This API is intended to replace the "WebView.loadUrl("javascript:...")
pattern that is the current mechanism for executing JavaScript in the
context of the current page displayed in the WebView.
The new API is more convenient - it doesn't trigger the normal URL
loading path - and so does not have side effects such as hiding
the keyboard - and allows the caller to specify a callback that will
be invoked once the script execution is complete.
BUG=9814043
Change-Id: I3f27e8ff5371077d9265430090d61381a3a86e76
Since this is an operation that could take a few seconds to run and needs to be
completed even if Settings dies, best to do it in the user manager.
Refactored PIN challenge/setup UI with a field to verify existing pin
when changing to a new one.
Change-Id: I0b7df5b2ccb7f343aa9282a9245d3bc2b577a794
Add FX_KEYPRESS_INVALID constant to AudioManager. This is a sound which should
play if the user tries to click or press on an element which is disabled.
Change-Id: Idd8953abeddd1afb51f7f190f9fb612ef672defa
Since list item clicks are delivered asynchronously, it's possible to
get click callbacks after the activity has already paused.
This should actually be blocked further up the event chain, ideally,
so that the list selection state is not out of sync with the application.
Bug: 9752185
Change-Id: I1058d64cba98e9093b462e2e581bdbccdace1420
There is a flaw in the mechanism used by AudioService
to detect the fact that mediaserver process did crash and restart.
It relies on polling AudioFlinger service until a successful
connection triggers a callback into AudioSystem JNI and then into AudioService.
But if another thread in system_server process, not attached to the
JNI environment, reconnects before AudioService, the callback is called in
a detached context and dropped.
The fix consists in attaching the thread executing the callback
and detaching it if needed.
Bug: 9693068.
Change-Id: I184308b12a3f87653bf818abf0159e0e45a66ef0
