Apps that run components in separate processes, especially that
host providers in separate processes, can hit a race condition
where two processes simultaneously discover that the files/cache
dir must be created, then each calls mkdirs(). One of these will
fail not because the dir couldn't be created, but because it lost
the race and mkdirs() returned false to signal that it already
existed -- and this was assumed to be a hard failure.
We now recheck existence after a mkdirs() failure to discern this
case and proceed appropriately.
Bug 10515463
Change-Id: I13fbdd838921223f75ab11faa47291c82b21c650
The scale gesture detector has a new behavior to make onScale
callbacks during swipe after a double tap. The screen magnification
is triggered after a trippe tap so if the user trippe taps and
holds to do a temporary magnification and tries to change the
zoom level with another finger, erroneous results are obtained.
The ScaleGestureDetector uses an APIs level check for the new
behavior but the ScreenMagnifier is a platform feature. We now
explicitly ask for the old behavior.
bug:11033376
Change-Id: I0dfb14dd3abcaa34ad1f40447c631b4203797378
Bug: 10918599
Reduce the number of recent tasks to 10 on lowram devices
Use RGB_565 on low ram devices for thumbnails instead of ARGB_8888
Combined this saves ~9MB across system_process and systemui
Change-Id: Ieddcb512c7341a90097bc7cbc72d7355a775b416
Only use the stack bounds for the DimLayer bounds if there is more
than one stack on a layer. Otherwise cover the entire screen (and
more) with the DimLayer. This way there are no exposed regions when
rotating.
Fixes bug 10428085.
Change-Id: I7bfff12b69d59e86610621c498dab35cf0db8eb5
Newly added private flags were being masked in the public flag variable
as opposed to the correct privateFlags variable.
bug:11033280
bug:11043194
Change-Id: Idda3a70a083457f3f1b7d4b46d231f4a7e704cf0
Instead of keeping a single global system decor rect around
in WindowManagerService, calculate and store policy-defined
system-decor frame for each window.
The per-window decor rect is useful for smooth transitions, since it
determines window cropping during transition animations.
Bug:10938001
Change-Id: Ice6652aa5946027c45c0b7ab4e46473a0f8e3f90
The dump code in PrintSpoolerService was relying on being
called only as a result of "adb shell dumpsys print" which
is apparently wrong. As a result the code was making wrong
assumption about the input arguments.
bug:11046234
Change-Id: Ie38f3cc5f17cac98b808fef6d6bbeaca22a62ef0
First, startTime's default value is -1 and not 0, so use it to test
if the startTime is set. Also, startTime isn't being set when connecting
legacy VPN; this change fixes that too.
Bug: 10985174
Change-Id: Ibf52b8f489b18c96efba501b76aa73e9a9e0e3e8
The activity manager can create a new task under certain
circumstances in resetTargetTaskIfNeededLocked(). In that case the
window manager was not told to create a new task. This fix creates the
task in window manager when it recognizes that there is a new task
required.
Fixes bug 10412588.
Change-Id: Ife9ecfeb7f53f9df5950322c9359c836adc258a2
All background work is going through AsyncTasks, which uses a shared
thread pool. Even with the new ContentProviderClient logic to detect
ANRs, the UI can still appear to be unresponsive for 20 seconds, even
if the user attempted to switch to a different backend. In the worst
case, a backlog of thumbnail requests would end up wedging Loaders
for a long time, since they all share the same THREAD_POOL_EXECUTOR.
This change isolates calls to each provider onto their own thread,
which they're free to wedge and recover from over time.
It also means we no longer need a dedicated thread pool for recents
loading, and can use a simpler Semaphore instead.
Disables thumbnails in recents on svelte devices.
Bug: 10993301, 11014856
Change-Id: I7f8a5bbb5f64437e006cb2c48b7e854136d5c38c
The main problem here was a mistake when turning a single process
structure to a multi-package-process structure with a common
process. When we cloned the original process state, if there were
any services already created for the process for that package, they
would be left with their process pointer still referencing the
original now common process instead of the package-specific process,
allowing the active counts to get bad. Now we switch any of those
processes over to the new package-specific process.
There was also another smaller issue with how ServiceRecord is
associated with a ServiceState -- we could be waiting for an
old ServiceRecord to be destroyed while at the same time creating
a new ServiceRecord for that same service class. These would share
the same ServiceState, so when the old record finally finished
destroying itself it would trample over whatever the new service
is doing.
This is fixed by changing the model to instead of using an "active"
reference count, we have an object identifying the current owner
of the ServiceState. Then when the old ServiceRecord is cleaning
up, we know if it is still the owner at that point.
Also some other small things along the way -- new Log.wtfStack()
method that is convenient, new suite of Slog.wtf methods, fixed
some services to use Slog.wtf when catching exceptions being
returned to the caller so that we actually know about them.
Change-Id: I75674ce38050b6423fd3c6f43d1be172b470741f
1. Sometimes selecting a printer from the full printer list does not
work if the selected printer was not in the initial drop down list
of the print dialog.
The reason was that there is a race. We use loaders in the print
dialog activity and the all printers one. When these loaders are
stopped we stop printer discovery since we do not want to keep this
potentially expensive process on going if the activity is paused
because say the user decide to press home and start playing his
favorite game while the printers dialog is up. As a result the
loader does not get printer updates until it is started.
The loader of the print dialog activity is stopped while the user is
selecting a printer from the all printers activity whose loader is
getting discovered recent printers. Now when the user selects a printer
the loader of the print dialog activity is started but may not get
the latest printers by the time onActivityResult is called with the
selected printer. Now we cache the selected printer id and if the
loader reports it we select that printer.
2. In the print dialog we show only a few of the discovered printers.
If the user selects a printer from the all printers activity that
is not in the initial list we shuffle the adapter data to make sure
the selected printer is in the shown subset. Now if the printers
change, i.e. the printers loader reports new result we were not
respecting the reshuffling made before so the short list of
printers changes yet again.
bug:11034216
Change-Id: I54fe3619e3328b65839d9f4b02309699eae7f8eb
