Instead of using a series of booleans, create a single flags integer
that contains all of the dexopt options.
Change-Id: Ia8fa968f64b164267f43dd29cea9dc0413058125
Propagate the boot status explicitly to installd so that we do not
have to rely on dev.bootcomplete, which isn't meaningfully set
when the device needs the decryption screen on boot.
Bug: 23898216
(cherry picked from commit 06bb908b78)
Change-Id: I9b34298caf70b1e5d40970cc0d04c469016a80a7
If the id is negative, it is not from Resource.
Passing the negative value to getResourcePackageName makes AssetManager dump unnecessary warning logs.
It is reasonable not to get package/type/entry information for the negative id.
The warning logs :
W/ResourceType( 3711): No known package when getting name for resource
number 0x9b010100
Change-Id: Ic89acb4f32205ba5a2fdac61dc14b00ccf251148
default seinfo values are those for which policy lacks a
matching stanza in the corresponding mac_permissions.xml
file. Prior to this change the null object reference was
used to represent the non-matching state. This is in
contrast to a policy supplied default stanza which will
assign a non null seinfo value. Confused yet?
Basically, two default states were distinguished in the
code to describe the two cases where either a policy stanza
spelled out the base case seinfo label or not. Policy
writers could either supply a default stanza or rely on
the class instantiated value assigned by the ApplicationInfo
object. The hope was that the later assignment could be
used to help distinguish the cases where policy writers
intentionally white listed apps. This change will just use
the hard coded "default" string implicitly to describe all
cases and removes the idea of a policy supplied default stanza.
Change-Id: Ib7b01ee004775f24db9a69340a31784b967ce030
Signed-off-by: rpcraig <rpcraig@tycho.ncsc.mil>
- Add TRACE_TAG_SYSTEM_SERVER for tracing system server code.
- Add a few more tracing code to the code paths related to boot.
BUG: 21739901
BUG: 22207144
Change-Id: I08b424d2f83783f512c3818394099909cd55f327
Document that some devices may not support submitting multiple
outputs in a single reprocess capture requests. Remove the
guaranteed reprocess session configurations that are less common
and may not be supported by all devices.
Bug: 23422291
Change-Id: I1c869612e9a83ef480113ca2628f68511a150b64
Devices that do not support the BURST_CAPTURE capability do not
get the split resolution lists. However, one spot in the stream
configuration map enumeration did not guard against this properly.
As a result, the category of devices that are not-BURST-CAPABLE and
do list resolutions that operate at a 'slow' (<= 20fps) rates cause
an assert to fire from an apparently-inconsistent size list.
Guard the filtering correctly, to correctly support such
devices.
Bug: 23489536
Change-Id: Ib026f6555c69b7fac0cd1ea5027d598644cd8701
This ensures that sizes don't change after being initially inspected,
and/or stale size information isn't reused incorrectly.
Bug: 22092273
Change-Id: I5b26ea53977da4c1412e476d06d94aa791846393
- Correct spelling of quaternion
- LENS_INTRINSIC_CALIBRATION and LENS_RADIAL_DISTORTION were not
linkifying correctly
Bug: 23360503
Change-Id: I9c5acc0087d7115548b75ccfa1bc1eb1866557b2
To help MediaProvider uniquely identify storage devices, pass through
the UUID of the underlying private storage volume.
Bug: 23329611
Change-Id: I22ee7ea98fcb208d7797310acb3396a3f074f09b
Requests without NET_CAPABILITIES_INTERNET and just the default network
capabilities should not be marked restricted. Without this fix apps
can hit permissions exceptions if they inadvertently make requests
without NET_CAPABILITIES_INTERNET.
Bug:23164917
Change-Id: I4c7136821315bcb05dfc42ffbc505a5d4f6109e6
RecoverySystem.handleAftermath() is called to clean up recovery-related
files. It needs to handle non-empty directories as well such as leftover
stashes from failed OTAs.
Bug: 23199081
Change-Id: I33fb326d376d04c793ad7b5948a10e8c1c572812
