It would return an arbitrary selection around current position.
Rely on select all instead.
Bug: 3100750
Change-Id: I73d995e6481d7c230cc9f334c72fbfb7e9828007
This is a race condition exposed after we turned off some STOPSHIP
logging and some other logging. We drop the ACl connection after 2 secs
of inactivity and hence the device didn't exist -> DBUS crash.
Bug: 3097224
Change-Id: I90adbbee2c5793924037685e484027ed5cd2e0d0
Switch to using PBKDF2 for the key generation for OBBs. Any previously
generated OBBs will stop being read correctly. A small pbkdf2gen program
is available to allow generation of appropriate keys with the salts.
Bug: 3059950
Change-Id: If4305c989fd692fd1150eb270dbf751e09c37295
1. Some headsets send an AVCTP connection before an AVDTP connection
When that AVCTP connection fails, we get stuck in IncomingA2DP state
because we don't handle AVCTP signals for now.
We need to handle the signals and fix it better.
2. Also when ACL gets disconnected, reset state i.e
when the low level connection disconnected reset our state.
This is like a fail safe in case we get stuck.
Bug: 3051490
Change-Id: Ibcf188103999ffb1e08c36c2554504c639fb7f50
gingerbread.
(for more detailed comments, please see the dependent CL
Change-Id: Ib90eb9e647f543312c865d3bbf9a06fb867a648b )
2 changes in this CL.
1. add 2 new columns to downloads table in downloads.db:
mediaprovider_uri = downloaded file's content_uri in mediaprovider db
deleted = flag is set to true if a file is to be deleted
2. download UI app shows only those files whose 'deleted' flag is not set
Change-Id: I47c5f72b0a0b354fac3abe3a97752130b8189794
The decay is now in line for the decay for absorb glows.
Fixes b/3094064 Overscroll fade timeout is too long
Change-Id: Ida987c8f6193fcb15f743d72c590a5c1daef056d
We can't reliably detect when the key for an OBB is incorrect, so just
remove the mention in the documentation. It still returns an error, but
just not the most specific error.
Bug: 3091064
Change-Id: I70e506822f2178a0ac5e4617fe545b23ce0026f4
If something kills system_server before it completes its shutdown
action, the runtime will just restart giving the illusion that a reboot
for an OTA or something else has happened.
To prevent this, write a system property containing the reboot reason
before initiating the shutdown with all the services. If the
system_server is killed before it completes, the next time the main
thread of system_server starts up, it will immediately execute the
shutdown action.
Bug: 3022556
Change-Id: I81723bac333430f04205e7a7b799914d96f170eb
This change caused Music APP to throw an NPE the fix
is to test that scheme is not null before using it.
Bug: 3092886
Change-Id: Ib29f944cbd9a9ff628462bdd6f4b68a308468536
- Add the second half of the new NFC API: NfcAdapter, Tag, NdefTag,
RawTagConnection, NdefTagConnection.
- Add implementations for all of the new NFC API.
Change-Id: I8da9fd1d2971ed2c8b76f2a93f207fa9e305b0f6
Signed-off-by: Nick Pelly <npelly@google.com>
Bug: 3027023
1. user is connected to a headset. Headset is turned off.
2. User connects to another device. The priority of the first
headset must be set to ON. We only have 1 device set at AUTO_CONNECT
priority.
1. A remote headset sends an incoming connection. Don't set
AUTO_CONNECT till its gets connected.
2. For outgoing connections, set it to AUTO_CONNECT in Connecting state
since the user initated this connection.
Change-Id: Iaa211bf22b1fd99850875f7eda686c47142173ba
Bug:3085639
Call super.onTouchEvent for ACTION_UP and ACTION_CANCEL.
This removes the old selection as appropriate.
Bug:3085587
Make the cursor visible so that the insertion handle
appears when tapping on the field.
Change-Id: If5ea6219de9ca351667021743b2e63edc6c47250
Rearrange structure of MountService handling of OBBs to be entirely
asynchronous so we don't rely on locking as much. We still need the
locking to support dumpsys which has been improved to output all the
data structures for OBBs.
Added more tests to cover more of the error return codes.
Oh and fix a logic inversion bug.
Change-Id: I34f541192dbbb1903b24825889b8fa8f43e6e2a9
Added the concept of a "trusted" event to distinguish between events from
attached input devices or trusted injectors vs. other applications.
This change enables us to move certain policy decisions out of the
dispatcher and into the policy itself where they can be handled more
systematically.
Change-Id: I4d56fdcdd31aaa675d452088af39a70c4e039970
Previously a tight loop of StrictMode violations would still allocate
Exception objects and populate their stack frames, just to orphan them
later when checking the max-10-violations-per-loop constraint.
With this patch, we do that check _before_ allocating any memory.
Change-Id: Iae96aba33f8fcc6a8ec5838a231aecc08e95122d
