--classpath hasn't been properly supported or used for as long as
git history exists, and the non --runtime-init classpath is useless
without it.
Also, forking and exec' a bare dalvikvm instance from the zygote seems
somewhat pointless since we might as well just execute it directly.
Change-Id: Ib07d822d448e7b431d7564381f9cd11735fd1113
Change 3fe595108f changed our pipe creation to CLOEXEC,
which doesn't work so well for wrapper processes that call exec.
Sending down the PID pre-exec won't work so well because wrappers
are arbitrary and might fork themselves before finally starting
the child application process.
Change-Id: I065f0f2150dc1ff04d2b2e8b25f06b4c754809b9
Drop privileges only if we're root. Zygote.preloadClasses can
be called from a non-root uid / gid in the case of wrapped classes,
so we shouldn't die from the resulting EPERM.
Change-Id: Ie3fbe50a17ec49bcbcb875c3a2ee1e1a3a62e88d
If the content to dump is large no output is produced by dump. Change
back to using a loop and flush.
Bug: 18965342
Change-Id: Ibc33da8bfffd5f998a78df0e2d70ed8e8a0aa137
automerge: a319aa8
* commit 'a319aa800df36d40e1de453b64fd1e14573e6d80':
Fix issue #18942959: Phone getting stuck then restarts after unplugging from Audi
FastPrintWriter needs to have its own lock for each instance, or
else one getting blocked can cause others to block and whacky deadlocks
result.
Also:
- Improve error reporting of SystemConfig to tell you which config
file is the problem.
- Fix CoreSettingsProvider to not spew errors if a setting is not
defined (it should just use a default value).
- Get rid of noisy init output of ConditionProviders.
- Reduce log noise of starting a process; move some of that information
to ProcessRecord to print on demand.
Change-Id: I1032d141ddd449968b74ab7b88ab36f2d228ad1a
...an intent with invalid or null package uri
Also tweak battery stats to record in the history when we shut
down, to understand when restarts are due to clean shutdowns or
crashes.
Change-Id: I6443dafc23e356be9f569906f6081152d4f92d2b
Add API for handling nested pre-processing of accessibility events
similar to nested pre-scroll or pre-fling. This allows custom views to
delegate a nested scroll to a parent via the accessibility system.
Use this functionality to allow opening the ResolverDrawerLayout via
accessibility commands.
Bug 18827274
Change-Id: Icd5a502605b78a861bb03e7b11923841a72eb9ab
WindowDecorActionBar and Chrome's ContentViewCore each have their own ActionMode
reference. ActionModeImpl.finish() nulls WindowDecorActionBar's reference and
calls mCallback.onDestroyActionMode() to null ContentViewCore's reference. But
if the callback is deferred, there is a period when the ActionMode is finished
(and mCallback is null), but ContentViewCore doesn't know. ContentViewCore may
try to invalidate() the ActionMode, which will crash on the null mCallback.
Make ActionModeImpl more permissive so that calling invalidate() during this
period does nothing.
BUG:18758329
Change-Id: I407fa0e0cd3cffa217e165caed83130d44760316
The ShutdownActivity was implemented in the system server
code, which is no longer part of the boot class path. This would
work fine when requesting a shutdown from the primary user, since
in that case the activity would be loaded in to the main system
process which also has the system service code loaded.
However, when executed from a secondary user, we need to spin up
a new system process for it; that system process is just a regular
app process, without the system service code, so *boom*.
To fix this, move ShutdownActivity to the framework jar.
Change-Id: Icb831cd6ed619da5feede5605c45ef177a892fbc
Fixes a regression where Builder.setIcon(Drawable) would get overridden
even when Builder.setIcon(int) had never been called and was still 0.
Fixes attribute id to respect all valid resource IDs (e.g. non-zero).
Updates documentation to reflect the long-standing override behavior.
BUG: 18904762
Change-Id: I905703993a59910555d5a858e0aaecab63221a02
ResolverDrawerLayout could get into a state where it thought a drag
was in progress, but since there was no distance to drag it would
return false from onTouchEvent for ACTION_DOWN. This would cause it to
intercept all future touch events, blocking interaction with
ResolverDrawerLayout child views. This would only happen if none of
the child views wanted the initial ACTION_DOWN event and it ended up
being processed in the ResolverDrawerLayout itself, such as in
response to touching in empty space or on the title text, and if all
available options fit in view without needing to expand the drawer.
Fix this by only setting the dragging bit if we're going to claim the
rest of the touch events after ACTION_DOWN.
Bug 18563698
Change-Id: Iab2ec38528697c9ff36bbb08b1f1210a6879d5ef
UI for ResolverActivity doesn't support last used
activity and work/personal profile button at the
same time. So if work/personal button is present
disable last used.
Bug: 18730438
Change-Id: I7ae7319bbdeb9bcaab9d6d03f97dea17d07c6d7f
Show ResolverActivity even if there is only one target
on the current profile that listens to that intent,
if other profiles can also handle it.
Also, fix crash while trying to access out of bounds
element on the list.
BUG: 18701336,18713763
Change-Id: I3a5e9cc5c0eedb2792b8f6d8a5d4aa8ee9e1256b
Due to popular demand the methods putBoolean(), getBoolean(),
putBooleanArray() and getBooleanArray() have been added to
PersistableBundle.
Fixes bug 18390436.
Change-Id: Id133ba902aca774f98529e36ce560e873b88ad5b
