Fades out the mouse pointer:
- after 15 seconds of inactivity normally
- after 3 seconds of inactivity in lights out mode
- after a non-modifier key down
- after a touch down
Extended the native Looper to support enqueuing time delayed
messages. This is used by the PointerController to control
pointer fade timing.
Change-Id: I87792fea7dbe2d9376c78cf354fe3189a484d9da
bug:3504589
WebView#removeJavascriptInterface was not performing check if
a mWebViewCore instance is present. Added such a check.
Change-Id: I7f6dfe51a08f23dddf2fc94df635fdfa68e9a7cf
Also some fine tuning of zoom scales.
1. Since overview scale will be used to determine min zoom scale, its own
computation shall be free of min zoom scale.
2. make sure mMaxZoomScale is at lease mMinZoomScale.
3. Use default scale in case of non-overview mode, since it reflects the
current zoom density.
issue: 3494868
Change-Id: I4297878b820e437b706bb7e0f143336ff1b39f82
Generates InetAddresses without risking an accidental dns lookup. For use with supposedly
numeric-only ip address strings.
Change-Id: I694f3976ce1c6382854706f6557ea88a289add3a
There was an issue with stale recipient tracking when BinderProxy weak
references had been purged and a new proxy object allocated for a
still-live underlying IBinder. The death recipient bookkeeping has
now been reworked so that it's fundmentally tied to the BinderProxy
instances, not maintained as global state, to prevent this sort of
confusion entirely.
Bug 3499939
Change-Id: I75c5216b6d53b90868ac969e32c9725201e51be3
Bug 3422121
Cherry-picked from master's 95472.
With ellipsize, lines starting with a very long word that does not
fit inside the width were simply ignored. Cut the long word instead.
start - widthStart index offset shift in BiDi.
The original ellipsize-end patch that added '...' after the last
word on end-ellipsized lines has been punted in favor of a true
ellipsize support in I.
I believe the StaticLayout calculateEllipsise is a no-op since textwidth <= avail
by construction: fitWidth and okwidth are < outerWidth. The only exception is the
paraEnd != here case in generate (when not a single character fits in width).
This case is exercised by StaticLayoutTest in cts (width of 8 pixels) and revealed
an offset error in widstart.
All in all, it looks like this code was probably never really tested. I tried some
typical text configuration to make sure these changes improved the situation.
Change-Id: I6c2cb26436a21f0f89078c275a89e891f0f23b92
...at android.app.DialogFragment.dismissInternal(DialogFragment.java:264)
Don't allow a DialogFragment to be dismissed twice.
Change-Id: Id2e9e3be1046b0d7862492c57c36001d8fd44a69
AIOOB exception fix in TabWidget
Bug http://code.google.com/p/android/issues/detail?id=15005
The problem was not specific to the legacy theme. The code that first
measure the tab's width with no contraint was incorrectly using the
mImposedTabsWidth array which could not have the right size if a
child was added.
The first measure after a child is added should indeed crash. Could
be investigated. This fix is sure anyway.
Change-Id: If5015aaa2d5574939fd5d6c6362ed6db94d35d4a
Well, I'm not sure it is right for onCreateDialog() to return a null
dialog, but if it does, let's not crash here.
Change-Id: I5ff49b9b3c326d9005f70a01435c01bfc7307343
This is due to the window doing a relayout after its activity is
stopped, at which point it may need to interact with the adapter
to load data.
The fix here is to tell ViewRoot about an activity being stopped
and, if in this state, hold off on doing any new measurements and
layouts of the hierarchy until it is no longer stopped.
In this case the relayout was happening due to the cursor
being deactivated, with causes the adapter to invalidate
its data. Because this is now in a dialog window, this
allows the window to actually be resized smaller (unlike when
in a full screen activity), and boom.
Change-Id: I26442b4679819b4a4e6bc56289afd3445526750b
