* changes:
Fixes 1844680. Interrupted scrolling could lead to rendering artifacts. The ListView was invalidating itself in opaque mode but before the drawing could happen, it was getting rid of its scrolling cache, thus becoming translucent again.
* changes:
Fixes#1846038. DrawableContainer was wrongly returning its opacity by ignoring the visibility of the currently selected layer. This change simply reports a TRANSPARENT opacity if there is no currently selected layer of if the selected layer is not visible. Otherwise it reports the opacity computed by the state class.
* changes:
* update all metrics data when updating density. * Keyboard should use DisplayMetrics from Resource rather than getting it from WindowManager as the display metrics can differ under compatibility mode.
* Keyboard should use DisplayMetrics from Resource rather than getting it from WindowManager as
the display metrics can differ under compatibility mode.
* changes:
Fixes#1596240. Optimize invalidate/draw passes by marking opaque views and avoiding drawing them. Whenever a View requests an invalidate its parent check whether the view is opaque or not. When the view is not opaque, the framework behaves as it used to. However, when a view is opaque, the parent marks itself as being dirty because of an opaque view. Its parent then does the same, and so on. When the framework then starts drawing the views, it does not draw views marked as dirty opaque. If a view is dirty opaque and receives an invalidate request from a non-opaque view, it then clears the dirty opaque flag and behaves as before.
Issue 1837610 Adding a Widget before Running the Associated App Causes a Force Close
We were not retrieving the shared libraries of an application when deliving a
broadcast to an explicit component.
Issue 1753079 loading class path of instrumented app into instrumentation may load wrong path when instrumented app shares process with other apps:
We were using the ApplicationInfo that was used to originally create the process, not the one that the
instrumentation is against.
* changes:
Make in-app search have a different background resource which looks like a slide-out drawer, to provide app context and because it is not full-screen.
like a slide-out drawer, to provide app context and because it is not
full-screen.
Also fix a problem where bringing up global search sometimes still did
not size the dropdown correctly, because the keyboard was not yet showing.
This adds new attributes for specifying a targetSdkVersion and maxSdkVersion.
There is a new ApplicationInfo flag that is set if the application has set
its targetSdkVersion to the current platform or later. Also you can now
use a string for minSdkVersion and targetSdkVerion, to indicate you are
building against a development tree instead of an official platform.
with respect to the soft keyboard. getMaxAvailableHeight in
PopupWindow relies on View's getWindowVisibleDisplayFrame, which
always takes into account the IME, whereas in this case, because
we want the window to extend down to the bottom of the display
when it is being interacted with by the user, I added a new
boolean option 'ignoreBottomDecorations' to getMaxAvailableHeight.
If true, the method returns the height which will extend the
PopupWindow down to the bottom of the screen, rather than just
down to the IME.
Then in AutoCompleteTextView, I set this to true when the
dropDownAlwaysVisible attribute is true and the PopupWindow's
input method mode is INPUT_METHOD_NOT_NEEDED (which happens when
the list is dragged or otherwise interacted with).
Then, because the dropdown height is only calculated when we call
showDropDown(), I had to replace a couple calls to mPopup.update().
There are still a few remaining bugs here I'm tracking Still to do:
* Fix a strange re-placement of the window which happens
intermittently.
* Strange behavior when using the directional pad.
* In a few cases, the list does not correctly size itself to the IME.
This seems to be because the available height is calculated before
the IME is yet on screen, and thus is calculated as the entire screen.
This introduces a new class in the base platform for performing a clean
shutdown (which was copied from the classes in the policies). It
includes new features to send a shutdown broadcast for applications
to do cleanup, and ot have the activity manager pause the current
activity before proceeding with the shutdown. These facilities are
also use to write at the most recent stat files for sync, battery
and user activity.
* changes:
* update density correctly when the configuration is changed. * Turns private sLcdDensity to public DEVICE_DENSITY to use it in ActivityThread
* changes:
Fixes regression introduced in Cupcake: a couple of adapters were using setTag() to store their own data, which could interfere with the use of setTag() made by the applications. The fix is quite simple an rely on a WeakHashMap to store the holders within the adapters; using the Views as the keys of the map.
Given a package name, the Backup Manager schedules a *full* (i.e. non-
incremental) backup pass for that package. Also added the state-file
handling for distinguishing to the target between the full and incremental
backup requests.
- Removes the mSearchable field which was only
for communication between the constructor and
getActivityMetaData().
- Removes the badge and query rewriting fields,
since their values can be efficiently computed
on the fly.
- Makes all the other public fields private and adds getters
for them.
- Makes all fields final, except mActionKeys.
- Removes the DBG_INHIBIT_SUGGESTIONS_CONSTANT.
I don't see why we would every want that, and it
complicated making the fields final.
- Makes all fields in ActionKeyInfo final.
- Makes all fields in ActionKeyInfo private and adds getters.
- Removes the use of ActioKeyInfo.mKeyCode for failure
signalling. Uses IllegalArgumentException instead.
- Replaces the ad hoc linked list for looking up
action keys by a HashMap. This is needed to
make the fields in ActionkeyInfo final, and also avoids O(N)
lookup in the (unlikely) case that an activity
has lots of action keys.
- Don't throw exceptions when reading searchable
meta-data, since that could crash SearchManagerService.
- Adds debug logging.
The previous implementation used a database for storing all of its state, which could cause
a significant amount of IO activity as its tables were updated through the stages of a sync.
This new implementation replaces that in-memory data structures, with hand-written code
for writing them to persistent storage.
There are now 4 files associated with this class, holding various pieces of its state that
should be consistent. These are everything from a main XML file of account information that
must always be retained, to a binary file of per-day statistics that can be thrown away at
any time. Writes of these files as scheduled at various times based on their importance of
the frequency at which they change.
Because the database no longer exists, there needs to be a new explicit interface for
interacting with the sync manager database. This is provided by new APIs on IContentService,
with a hidden method on ContentResolver to retrieve the IContentService so that various
system entities can use it. Other changes in other projects are required to update to the
new API.
The goal here is to have as little an impact on the code and functionality outside of
SyncStorageEngine, though due to the necessary change in API it is still somewhat extensive.
Author: Christopher Tate <ctate@google.com>
Date: Mon May 4 16:38:11 2009 -0700
IBackupService now passes ParcelFileDescriptors rather than int fds
The outlines of backup state file / data file handling are now in place as well
in the BackupManagerService.
Author: Christopher Tate <ctate@google.com>
Date: Thu Apr 30 12:40:19 2009 -0700
Hide the backup stuff for now
Also adjust based on comments:
+ changed service intent string to conform to usage guidelines
+ only publish the IBackupService binder when invoked with the right
intent action
+ docs tweaks
* changes:
Fixes 1830181. Tapping disabled items in a ListView would cause the framework to take a code path that wasn't resetting the mMotionCorrection variable if it had been previously set. This would force ListView to apply a scroll to its children even though it was unnecessary. This simple fix prevents the issue by resetting mMotionCorrection appropriately.
