Parts uses notifications to preview custom notification lights.
Since Android O we got NotificationChannels and the internal API uses
the color values of the channel instead of the color set at the notification.
To prevent unexpected breakage in future, introduce a flag to be used for
providing preview colors in a bundle to force said color.
Also introduce flags to control the ON and OFF duration.
Change-Id: Ifbb7995a19d95b6ddb2627c1b14dd201f9dc5430
Signed-off-by: Alexander Martinz <amartinz@shiftphones.com>
* The parameter here only has to have the type you want it to return,
so creating an array of size > 0 is unnecessary and wasteful
Change-Id: I28e490fb6fa3703d7edca21b29d640105072947b
* We can't talk to DisplayTransformManager in C++, so
we must have a fallback in the SDK itself for HWC2
* While we're at it, fix the size of the int array created
in getDisplayColorCalibration() to be 5 rather than 6
Change-Id: Ic524f7d60ae1585efed3fba676cf3301dcaecd7f
* It was moved from Global to System but no migration was in place
to handle upgrades from previous versions.
Change-Id: If76243b0224dbc0138cf3c5dbc8aaeefa69f06a4
* For devices using Qualcomm MPCTL interface as power profile backend,
setting the user's profile preference fails on boot, because we call
perf_lock_acq before MPCTL is ready.
* To fix the issue, add an option to wait for MPCTL service to start
before setting stored power profile.
* We should not only wait for the service to start but also post boot
parsed property to be set, because MPCTL service won't actually
initialize if the property is not set to 1.
Change-Id: Ib749c3eb548dad0ab24868ee0048247a0b80dd6a
* During boot up phase the system is trying to update
LiveDisplay tile status based on the applied mode.
However, with LiveDisplayConfig not yet initialized this results
in a NullPointerException. Address it and fallback to MODE_OFF.
Change-Id: I736da21455e74801aae69e16b9e80c8985b83a5a
* While we're at it, hide the use of getSupportedFeatures(),
since it only works for lineagehw java features, not hidl
* Cleanup some minor styling/organization issues
Change-Id: I905010bc7c6d7d8578eeca501be23f665127aa27
* DISPLAY_GAMMA_CALIBRATION
* LONG_TERM_ORBITS
* SERIAL_NUMBER
* While we're at it, cleanup tests for removal of other features too
Change-Id: I1a572cf3bcc5298d208a3dc234f9d82da1105bf0
* If extphone binder service is invoked after death, phone service
would crash due to android.os.DeadObjectException.
Change-Id: I8dec7bfd709b0443654001ecd67219a2ba8cc134
* The upgradeVersion < newVersion part is never reachable because we
increase upgradeVersion unconditionally. Make it back to stage by
nuking it as no one has experience issues before.
Change-Id: Icac1e18c192292ad3147fd4af4c862e93ef34ba5
* The difference between civil and actual sun{rise,set}
is 6 degrees of movement of the Sun in its path in
the sky, astronomically speaking
* Since the Sun "travels" at 15 degrees/hr in the sky,
that means the difference between civil and actual sun{rise,set}
is about 24 minutes, temporally speaking
* In order to more closely match that temporal difference,
the TWILIGHT_ADJUSTMENT_TIME should be about 30 minutes
Change-Id: I30a600c71ce5dd01fcb96fb4f5108f45fa493b0c
* This makes it clear what the desired behavior is
* In order to make the logic more clear, also start
scaling into night mode before civil sunset so that
we finish by civil sunset -- this makes sure that
by the time there's no light out for practical
purposes, we are in night mode
* While we're at it, add a few comments about what
each return is supposed to do, for the future
Change-Id: I4ec0f8eec6b0129a37e99f9e3a2b3f124aeeb4f2
* Our overlay name now confuses people as Google introduced a screen
burnin protection feature for ambient mode in Oreo. Just re-use AOSP
one and kill this.
This reverts commit 088ddf5aaf.
Change-Id: Ia76b28444cd7b117559062fbd8230602962616b2
After reboot the active WeatherProvider is started (onBootPhase)
too early, resulting in WeatherProvider not syncing anymore.
Start the active Provider when userland apps are ready to fix this,
like on 14.1:
0cee14b03c
Found by meiser@xda
lineage-15.1 ready Weatherproviders for testing:
https://github.com/nvertigo/android_vendor_nlos/tree/nlos-15.1/weatherproviders
Change-Id: I8d3eb7aacaf265c55d5d0b5157731e04068a7c7d
* WiFi service is not considered as a core service so it won't start in
min framework, but ProfileManager accesses to WiFi service's API.
This would lead to a fatal exception and trigger RescueParty.
* We don't really need ProfileManager during such early stage, so
simply disable it to avoid invoking WifiManager.
Change-Id: I2c66217599e0062eaff2e5f4e86ebc092d4a41d5
* Since changing overlays is now guarded with enforceCallingPermission()
instead of enforceCallingOrSelfPermission() we can no longer call
setEnabled() after resetting our identify.
Change-Id: I25df30d0ba83f15ca1124d6fc6f79d3465d052f2
*) On pie, lineage system server is starting before
overlay manager service so mOverlay was being
initialized with null (and crashing).
*) Change to only attempt to get service handles after
boot phase system services ready.
Change-Id: I4965ce6015275a421e2e3864e66e77b8c15477f0
* New devices don't have hardware navigation buttons
Change-Id: I989ec6f7f07fd746c24b3eecb527412e3c5ec0e0
Signed-off-by: Giuseppe Barillari <joe2k01dev@gmail.com>
*) Don't depend on framework libs when building our org.lineageos.platform-res
package (as per framework-res bp).
*) It doesn't work out of the box because our resources refer to android: framework-res
attributes (and compile fails).
*) We can fix this in soong by making org.lineageos.platform-res depend on
framework-res (see topic:pie-sdk-no-framework-libs). This allows us to
build with no_framework_libs true.
Change-Id: I426d44ef2521971adfc6b8ad33a7cc042b04e0e6
* Since we removed onehand from settings
we need to make sure user doesn't have them
set in settings otherwise we will hit
ArrayIndexOutOfBoundsException exception.
* Fixes : BUGBASH-1525
Change-Id: I7a5737ecfe5e70c3f0108c03b3e8745eade80433
(cherry picked from commit eb8fc0648f645a265ce8d1b6daf575795eb69510)
* An alternative to using the getRecentTasksForUser method
that got removed with the following commit:
5fa397533f
Change-Id: I752c2445c3ffc0933dfe6290c99d839a036e4bc3
*) This was initialized from Settings.Global.DEVICE_PROVISIONED and it is no
longer possible to access system settings during provider startup since:
235272a216
*) LINEAGE_SETUP_WIZARD_COMPLETED has been deprecated since 14.1:
797a45aa69
*) We will later cleanup the unused setting during topic:pie-db-upgrade
Change-Id: I185648ca25995fae80362642d297573795d04bfa
Replaces the need for:
32e67e6020
*) frameworks/support are now built with gradle and are included as prebuilts.
*) API level 26 (8.0) introduced PreferenceDataStorage which can be used
to implement what we need without touching the support library.
Change-Id: Ide3417019ccf8bde5dca3b2aec3d2075e7c8b597
* Apparently creating it so early can cause serious breakages
that make StorageManagerService and UsbDeviceManager end up
getting null pointer when trying to get KeyguardManager with
ctx.getSystemService(KeyguardManager.class).
* Moving ctx.getSystemService(KeyguardManager.class) out of
constructor to isKeyguardLocked() solves these issues.
* Also move to getSystemService(Class<T>) while at it.
Change-Id: Ib3f65ba2e726149089ab1fb12030321e0ecdeab1
*) the sdk is now using aapt2
*) aapt2 is unhappy with "lineage:..." name attribute value format
(errors with invalid character).
*) update in a way that aapt2 is happy with
Change-Id: I96ecc14c4a6de00272a5ad4f513b62f21289ac45
