Previously, the TextChange transition didn't handle interruption/
cancellation at all, which made it problematic to use in any real
situation where a transition might get interrupted mid-animation.
Also, the way that it side-effected the text of TextView objects caused
errors in the UI when the transition was interrupted, because it would
not clean up after itself properly as new transitions queried the
current state of the UI.
Also, the prior cancellation logic for all transitions was not quite
correct; we were pausing transitions but resuming the animations, making it
tricky to write transitions that would restore state correctly.
Change-Id: I5a9f3c915e9834ec59ce1e1c3c96a88d11e4aa1b
- Make it more intuitive by enforcing callers to set a listener while
starting recognition
- Fix TODO to resolve the ComponentName if none is provided
- Unbind from the Hotword service if stopRecognition is called
- Get rid of the callback in onStopHotwordRecognition
Change-Id: I8edad0ef0f3671283aeb51eaed1d3b8dab01baa0
* commit '627dd9b584a267b206bace8dc518287fe58b9563':
Fix bug #10515013 CalendarView is always forcing first day of the week to Sunday and not taking into account CLDR data
As per UX, we're removing the EXTRA_DELETE_AFTER_USE flag. Instead,
apps that pick up these intents should automatically delete alarms
and timers, unless the user is taken into the app to explicicitly
specify the time/length.
Change-Id: Ibd582788c04e121427467f6549759c8b7706df8c
Sets contentInvalid for TextViews with errors, sets collection
and item properties in AbsListView, GridView, and ListView.
BUG: 10391326
Change-Id: I91191c8ce1e10a1112f74215e30b26349396344b
Allows applications to temporarily force the NFC controller
to only do tag discovery. This will allow Android applications
to read and interact with devices that employ HCE.
Bug: 10360259
Change-Id: I709ead9a26f8e6ae8582cc295d82bd896e7c5bba
ViewGroup.setLayoutTransition() has logic to cancel a previously-running
transition and remove the ViewGroup's listener from that transition. However,
it is possible for the transition to be set to null on the ViewGroup on the
cancel() call, making the next call to remove the listener crash with an NPE.
The fix is to use a temporary variable to hold the old reference and
not depend on the class instance variable remaining valid/non-null across
the call to cancel().
Issue #10488932 Hangout NPE in ViewGroup.setLayoutTransition on initiating a new conversation
Issue #10509201 ViewGroup.setLayoutTransition(null) causes crash when being called in LayoutTransition.TransitionListener.endTransition()
Change-Id: I7c181b79e6601024c93a8dc246d32b751f78b216
1. Added scale to fit and scale to fill fitting modes.
2. Added APIs for a print app to specify which constraints imposed
by the print attributes were satsified during a layout so they
are not handled by the printer again, e.g. if the content is to
be in landscape the the app generated such content the printer
should rotate the content again.
3. Added some printer statuses.
4. Added a helper class that generates PDF documents with correct
size based on the passed in print attributes.
5. Exposed a ctor for PageRange which was hidden by mistake and apps
could not create page ranges. Added API for changing the printer
status.
Change-Id: If1334a61d2d931027b98075f653018d456b1b768
- Fix additional getInt() path, restores the location settings screen
functionality.
- Should fix "unresolved link" build breakages in
git_klp-dev-plus-aosp-without-vendor, which is much more persnickety than
klp-dev for some reason.
- Add warning that we may add additional location modes in the future.
- Finish fix for b/10461763 and b/10461474
Change-Id: Id7155e3a0d7526a377d446018ef3bdb057bad3a6
Previously, Visibility would determine whether a given view was
worthy of a transition by checking the visibility of that view's
parent hierarchy. This is done to avoid fading every view in
a hierarchy when only the top-most node should be faded.
However, the logic was flawed because it assumed the end scene
parent was stable between scenes, which is not the case with inflated
layout scenes, in which the parent views are not the same, even though they
may represent the same parents (via ids).
The new approach is to check information on both the start and end scene
parents, and to take ids into account when checking the state of these
parents in the start/end scenes. Also, avoid this logic altogether if the
transition is targeted at specific views (don't bother checking the
parents if the transition was told to fade specific views).
Issue #10442447 Transitions: Fades don't happen correctly in some situations
Change-Id: Icea24b892f2eff2d37c6436ccfe4e288aac84178
Integer doesn't seem to like number > 2 Gig; so converting all bits of
samples to long; and handling exceptions that may crash the service
Bug: 10459447
Change-Id: I1291e97ba6d0ab7304dc6a6e8ef819d17491e017
