The app ops mananger service maintains a mapping from UID to
a list of packages where each package is mapped to a list of
non-default app op states (default states are inferred and
not stored). Hence, specifying the app op state for a UID
requires setting the app op for each package in the shared
UID.
This is problematic when installing new packages if there
is a non-default app op policy set for another already
installed package in the same UID as the app op for the new
package has to be updated to be in sync. The package installer
cannot do this as it is in another process and the app op
update will not be atomic. Therefore, the app ops manager
service has to support specifying app op policy on a per
UID basis.
We now have a UID state object that contains the per package
non-default app op states as well as the per uid non-default
app op states. If there is a UID policy specified then it
takes precedence over the per package one. Even further,
changing the uid policy updates the package policies in this
UID if the state is non-default. Changing a package app op
state also updates the app op state for the whole UID if
the per UID policy for this op is non-default. Clearing the
app op state for a package, clears the policy for the UID
as well.
bug:22802981
Change-Id: I78044906d9fcc6066abf07e706c2c88f3397d293
LinearLayout relied on multiple measurement passes to hide a
discrepancy between EXACTLY and non-EXACTLY measurements.
This reverts commit 9cefbda11e.
Bug: 22810327
Change-Id: I515a80749420d00efc5002fa68221b7c236f03df
"Include non-zero dimension views in excess space calculation" and
"Always distribute excess space in LinearLayout measurement" changed
LinearLayout behavior significantly in a way that wasn't covered by
CTS tests.
This reverts commits da2f304409 and
4fabc02158.
Bug: 22862047
Change-Id: I8d37a525ccf295445d3239b80e5cacb10bf3c947
In particular, some clients (*ahem* SystemUI) have been
known to inflate RemoteViews with unusual derived contexts
that may not have valid application contexts. DateTimeView
can now resist this.
Bug: 22852700
Change-Id: I5e91ae0e66859f5f5efd7b19c0ae6dfbc26bcc54
Icon needs the RemoteViews' internal context to reflect not
just the Resources of the originating application, but its
package name as well.
Bug: 22840771
Bug: 22852700
Change-Id: I1164cd38d99283982b00daadcf927c7cf63539b3
The methods startUsingNetworkFeature, stopUsingNetworkFeature and
requestRouteToHost were @removed in all the M preview builds, but
internal and external developers have noted that this imposes
additional burden for applications that need to work across
multiple platform versions because it causes compile-time errors.
We switched from @removed back to @deprecated to avoid these
problems. In order to effectively deprecate these methods, which
are error-prone and insecure, make them throw
UnsupportedOperationException if the app's target SDK is M or
above.
Because there are still one or two places in system code that use
these APIs, exempt Process.SYSTEM_UID and the OMA-DM client from
the check for now.
Bug: 22728205
Change-Id: I790bd32f3aa8067cbb625962a209bb9232f4b58c
These are important values to help diagnose modem related wakeup issues.
Bug:22803686
Bug:22857991
Change-Id: Ie0164ebd194567bc10d1165d671a583674a39829
* commit '964a41214d19f00e8420500ccaa29e0902933c7e':
[Improve docs for AccountManager] Currently, the docs for AccountManager are somewhat misleading and may cause developer errors. To avoid them, we are properly documenting it.
First, getAccounts*() will now return all available accounts depending
on both GET_ACCOUNTS grants and signature matching. This is different
from before where a caller of getAccounts() would need GET_ACCOUNTS to
get any accounts, but if that same caller called getAccountsByType, they
might have gotten back accounts if they shared a signature with the same
developer.
Second, cleaned up some NPEs and javadoc.
This change was motivated by progress on the cts tests.
Change-Id: I2f36226780e074fdf58214b46de3b79d8319ace1
Currently, the docs for AccountManager are somewhat misleading and may
cause developer errors. To avoid them, we are properly documenting it.
Bug: 21924096
Change-Id: If775a54a09219b0f1623d2ff903085b9d12aa863
...of address space causing package manager to fail
Make sure to flush binder commands when done with an async task.
Change-Id: I9f171add7051587ab854226b97a12c7e1844ad5d
changes)
AppOpsManager:
Changed the default operating mode for WRITE_SETTINGS to MODE_DEFAULT from
MODE_ALLOWED.
packages/SettingsProvider:
We no longer do static permission checks for WRITE_SETTINGS in early checks and
defer that to app op when MODE_DEFAULT is returned. For some operations,
checking against WRITE_SECURE_SETTINGS is sufficient.
ActivityManagerService & PowerManagerService:
Incorporated app op checks and handled the MODE_DEFAULT case.
provider/Settings:
Added helper function to do checks on whether app ops protected operations
can be performed by a caller. This includes checks for WRITE_SETTINGS and
SYSTEM_ALERT_WINDOW.
Also added a public API (with javadocs) for apps to query if they can modify
system settings.
Changed the javadocs description for ACTION_MANAGE_WRITE_SETTINGS and
ACTION_MANAGE_OVERLAY_PERMISSION.
Added public API (with javadocs) for apps to query whether they can draw overlays or not,
and also javadocs description on how to use that check.
Change-Id: I7b651fe8af836c2074defdbd6acfec3f32acdbe9
