After accepting a connection on a listening socket and
storing the resulting FileDescriptor into a newly created
LocalSocketImpl, the new impl's "descriptor was created
locally and should be closed normally" flag should be set.
Bug: 11805817
(cherry picked from commit 3e7305c6bf6062b5cb1e2ddcec6c6d30b4a8bc0d)
Change-Id: I723d7c5544ee4b6858894c215716cbc32a958df1
Due to an API change in LocalSocket, Zygote must now
manually close the FileDescriptor it created when it
registered a LocalServerSocket. The LocalSocket.close()
routine will no longer do so.
Bug: 12114500
(cherry picked from commit 70ef29b04ee4ef7d7acfec79041dbe800961195e)
Change-Id: Ief23a3c99e007dc4aa6f94dfb47a1b2b6c854dad
Keep track of how many clients are requesting scans and scan
continuously until all of them are gone then explicitly terminate the
scan instead of letting it time out as before.
Suspend wifi display scans while connecting or connected to a remote
display. This is handled by both the display manager and media router
since neither has complete information about what is happening.
Much of this code will no longer be needed once wifi display support
is integrated directly into the media router service.
Ensure that we don't attempt to scan or connect to wifi displays
while the wifi display feature is off.
Infer when a connection attempt fails and unselect the wifi display
route automatically so it doesn't appear to be connecting forever.
Fix issues around correctly canceling and retrying connection attempts.
Often we would cancel but not retry.
Improved connection reliability somewhat. It seems that discovery must
already be in progress in order for a connection attempt to succeed.
Ensure QuickSettings uses exactly the same logic as the MediaRouteButton
to determine when the remote display tile should be made visible.
Bug: 11717053
Change-Id: I18afc977b0e8c26204b8c96adaa79f05225f7b6e
Whoops persistent processes are, well, persistent. Don't remove
services from them. We'll be keeping that process record around.
Change-Id: I29e9fb6f704efdf0caad5e0307a7adbb416eed3b
The init & clear operations are particularly important to ensure
delivery when at all possible, so we retry those periodically
if the transport is unavailable when we first attempt them.
Now with 100% less build break.
Bug 11716868
Change-Id: I2af4e93788068cfac97c0a48d3568c561eefa23d
There is a data ownership issue related to recent documents Cursors,
and the loader and UI threads can end up traversing the same Cursor
at runtime. This change masks OOBE crashes caused by this race until
we have time to fix cleanly.
Bug: 11736668
Change-Id: I2039a5edcf341fcaedacc2f08bde96fc5a271782
