Killing the WifiWatchdogService thread from
WifiService can cause messages to be handled on
a dead thread. Quit the thread on the broadcast
instead.
A couple of more fixes:
- Do an asynchronous bring up of Wifi. This will
allow WifiWatchdogServiceThread to be immediately
brought up, instead of relying on an update.
- There is no need to listen on supplicant connection
in wifiwatchdog anymore. We kill the thread when
supplicant connection is no more.
Bug: 2546756
Change-Id: I9beae3eebee9cf0446d1dc031ce8fc1c27943521
We add a variable for saving wifi state
to restore after tethering.
Bring up wifi on boot up if the state indicates so.
Bug: 2537983
Change-Id: I9c6548b93df6fcbc0cec1e6b857f7224dc6d1b2c
Previously we only supported a single range - this was inadequate for
multiple interfaces. Adding a second range so we can support
both usb and wifi tethering.
Also moving out of the zero-conf range as our dhcp client won't
accept ip addrs in that range (no nexus to nexus wifi action).
bug: 2537963
bug: 2533491
bug: 2538303
Change-Id: I600b421343c28c2f9839ed2076122ae3d0ff5d3d
As it appears earlier in the list, .3gpp will be used as the extension for
video/3gpp mime types.
Bug: 2542939
Change-Id: I7d9188b460774af7bc29f4e8f1ebada43bf2ed10
This change forwards to the WebViewClient HTTP authentication credentials that
are supplied to an XHR from JavaScript. This allows the client to store these
credentials for use with later requests.
Currently, the browser only stores credentials that have been entered manually
by the user through the authentication dialog.
Bug: 2544330
Change-Id: I913e35d80b7ad41ff75586092408b86ea9f543d8
Store a list of Handlers in a global ComponentCallbacks instance. Use
WeakReference to avoid leaking BrowserFrame objects.
Bug: 2542935
Change-Id: I67b102288f47f9fa6e855ba6558f2b1d611e8427
Looper.quit() is asynchronous. The looper may still has some
preview callbacks in the queue after quit is called. The preview
callback still uses the camera object (setHasPreviewCallback).
After camera is released, RuntimeException will be thrown from
the method. So we need to join the looper thread here.
This change also fixes a potential race condition.
bug:2521202
Change-Id: If69bbb3125d1d30192563559579b87c20fa5aac8
This is kinda gross, but I can't find a good way to check for the
existence of gettid(), except by finding its syscall number. Then
might as well just use it rather than hope gettid's around, as it's
not in sim-eng.
Change-Id: Ieb7b39426dec08bd715b6fe1a9ab5b2801bdf775
This mimics what we do already for SQLiteDatabase's db_operation and
ContentProvider's content_query_operation and
content_update_operation: over a threshold things are always logged,
and under which they're sub-sampled.
Change-Id: Ia0280b9b07b336ee88b17be2a31a7af0fd7c5770
On keyboardful devices, it is possible to disable the system soft input
method. Something changed in eclair that caused the ime to be re-enabled
on every package manager update (packages added/deleted).
Now keep track of disabled system imes in the settings db and search
in that list before enabling a system IME on package changes.
Every time the user goes to settings to enable/disable imes, the list
is re-created.
Any new system IMEs that may be added via an OTA will get enabled if
they have a different package name.
