The ANR is caused by SurfaceFlinger waiting for buffers of a removed surface to become availlable.
When it is removed from the current list, a Surface is marked as NO_INIT, which causes SF to return
immediately in the above case. For some reason, the surface here wasn't marked as NO_INIT.
This change makes the code more robust by always (irregadless or errors) setting the NO_INIT status
in all code paths where a surface is removed from the list.
Additionaly added more information in the logs, should this happen again.
We now just don't send out dock broadcasts/launches when the
device is not provisioned. Good enough for our purposes.
Change-Id: Iee6384121675e0e9854745ec1168245e8a23a241
Use a dedicated Connection to process a synchronous request using the main
RequestQueue and on the WebCore thread.
HttpConnection, HttpsConnection, Connection:
No need for ConnectionManager, just pass the proxy host to
HttpsConnection. In Connection.clearPipe, empty should be set to true if the
RequestFeeder does _not_ have a request for the host. I don't think that
broke anything but it is incorrect.
ConnectionThread:
Refactor the proxy selection into ConnectionManager.getConnection
RequestHandle:
Add a new constructor that takes a Connection argument. Add processRequest
to process a single request on that Connection.
RequestQueue:
Add determine host to select between the request's host and the proxy. Add
queueSynchronousRequest to make a new request that operates on a private
connection. Add SyncFeeder to handle RequestFeeder.requeueRequest if the
request fails.
Network: Use the new RequestQueue api to handle a synchronous request.
Bug: 1315832
This is the frameworks/base part of the fix for Bug 2158434: add a new
callback to the OnDialTriggerListener interface, so the RotarySelector can
tell the app about state changes between NOTHING_GRABBED and
LEFT_HANDLE_GRABBED and RIGHT_HANDLE_GRABBED.
BUG=2158434
DRNO=timsullivan
TESTED=regular incoming calls, call waiting calls, answering a call,
rejecting a call. Also verified I didn't affect the
lock screen at all.
AbsListView can reject views from the recycler's scrap heap but when that
happens the rejected view is not detached from the parent. This can be pretty
bad in the case of TextView since it prevents them from unregistering their
OnPreDrawListeners.
this fixes a safari-only bug in which encoded chars were inserted
into the search string.
http://b/issue?id=2031559
Change-Id: Ie60902bafcbb327d488b921f8d3925aaf9ff0566
Since a fullscreen IME covers the entire search dialog, there was
no discoverable way to get back to the search dialog.
Now, on BACK presses when the IME is fullscreen (e.g. in landscape
mode without hard keyboard), we dismiss only the IME.
A second BACK press will dismiss the search dialog.
Fixes http://b/issue?id=2159132
Change-Id: I8e5736d83d7c35936a718929017ea04c6214700b
API council says:
"In reviewing the new triggerSearch API, we are concerned that
applications could use the option to perform a global search to spam
the user with frequent searches. We would like the global search
option to be removed for now (not just hidden, but removed from the
internal IPC API so that nobody can find this and abuse it). The rest
of the API should be fine as long as it is restricted to local
searches."
Fixes http://b/editIssue?id=2158785
Change-Id: Ie69a9c0ab6373cc4427aab50606885bdede40585
Windows with a negative Y position can end up in createSurfaceLocked()
with mFrame containing a negative height, causing SurfaceFlinger to go
crazy when asked to create the surface. This change simply guards
against such a situation by instead asking for a 1x1 surface and relying
or later layout operations to resize the window to the appropriate size.
Change-Id: I66f2058f4cd1cf069b12d3d23e6fd340dc76b74e
* move DaemonHelper out from VpnService to VpnDaemons for better
managing native daemons.
* check connectivity and dns less frequently to save battery.
