Replaced the API setOption with content observer to read/get informed of
the CEC settings.
Bug: 16855247
Change-Id: Ieff2399bbfe83f05af4448e7337f91ef40d1e24d
* This allows flash mode=TORCH to keep working if only a single
request is submitted and the device goes to the idle state.
Bug: 16305644
Change-Id: Id2e3451f85dbf9b545d3a70977b8fcfb0922a0eb
Now MHL device information is also a part of HdmiDeviceInfo (previously
HdmiCecDeviceInfo). This change adjusts method/variable names across our code
base to reflect that.
Bug: 16215362
Change-Id: I7a6692329f39842a4aec9a4ff036eca5474cfd33
This change adds a script that renames only font at a time. This helps
in writing better Makefile rules and passes the job for parallelization
to the make rather than doing it in python which doesn't support ctrl-c
properly.
Change-Id: Icd380e3cc5d744260d6a9f497aa65a1a6c829c2e
The copy will be modified to support renaming only one font at a time.
This is done in two steps to improve the diffs.
Change-Id: I47283745c25882283f0464c17e603396a8c9141e
When an app without INSTALL permission attempts to commit a session,
we involve user to confirm permissions. We currently point at the
base APK, which defines all permissions for an app, handling the case
where a session may only be adding splits.
Add failure codes to represent rejection. Fix bug by ignoring stages
during initial boot scan.
Bug: 16515814
Change-Id: I702bb72445216817bcc62b79c83980c1c2bb0120
Invoking aapt after merging resources from a library project
may yield a different ordering to styleable arrays, so have
the indices be non-final too.
Bug:16842410
Change-Id: I0432bea03dc4312d5908a770fc70a11f0a1596ae
Switching from GCM-only inet condition reports to using our network
validation (captive portal check).
Note that currently the GCM signal is disconnected. Next step is to
make the bad-network report API trigger a re-evaluation of the network
and get negative reports from the NetworkMonitor.
Change-Id: Ie2ebab1e5c04775e3c4d6738f656a6c8157dba76
Issue #16655104: Android system used 21% of battery
We weren't clearing the current process state when we now retain
process objects across resets.
Issue #16875776: abnormal outliers in screen-off battery life in go/batterystats
The checkin data since charged was using the since unplugged data
for the amount of drain with the screen on/off.
Also added a new version tag in the checkin output containing
the platform build number and internal version format number, to
start moving away from the wasteful version code on every line.
And fixed a bug in figuring out when things have changed over
discharge levels.
Change-Id: I89b89513a748016aacc848f14bbbc8b2400d92ca
Also hide the inspector whenever the keyguard changes state,
to make sure we don't leave any notification guts lying
around when you turn the screen on and off or unlock.
Bug: 16208321
Change-Id: I47cf6a0cb82e765eb6b50984972f57616b534ed3
Another check in the ActivityManager to stop us from getting into an
invalid state by a fat-fingered 'am switch' etc.
@bug 16814252
Change-Id: Id65ab647d5a535cc2f965139a1b717cf65aa7e69
