Doing a factory data reset used to always erase eSIMs. Then a few months
ago we added a default-on checkbox to let users opt out of erasing the
eSIM during this process, but only had it show for some devices (ones
which support the "fastboot oem esim_erase" command) by adding a system
property named masterclear.allow_retain_esim_profiles_after_fdr.
When recently updating the strings shown in the factory data reset
screen and the confirmation dialog, we changed the code so that if that
the checkbox is hidden we'll pass false for the ERASE_ESIMS_EXTRA
parameter sent to the factory data reset confirmation dialog. This had
the unintended side effect of making devices that don't specify true for
masterclear.allow_retain_esim_profiles_after_fdr skip erasing the eSIM.
This CL fixes that by removing the "is the checkbox hidden" check, going
back to the previous behavior of just using the checkbox value, which is
on by default even if hidden.
Fixes: 135284765
Test: make RunSettingsRoboTests
Change-Id: Ia9f335920e4e3c4a90f0a6a49d1722a0c19ea83d
(cherry picked from commit 5f612a4b44)
Doing a factory data reset used to always erase eSIMs. Then a few months
ago we added a default-on checkbox to let users opt out of erasing the
eSIM during this process, but only had it show for some devices (ones
which support the "fastboot oem esim_erase" command) by adding a system
property named masterclear.allow_retain_esim_profiles_after_fdr.
When recently updating the strings shown in the factory data reset
screen and the confirmation dialog, we changed the code so that if that
the checkbox is hidden we'll pass false for the ERASE_ESIMS_EXTRA
parameter sent to the factory data reset confirmation dialog. This had
the unintended side effect of making devices that don't specify true for
masterclear.allow_retain_esim_profiles_after_fdr skip erasing the eSIM.
This CL fixes that by removing the "is the checkbox hidden" check, going
back to the previous behavior of just using the checkbox value, which is
on by default even if hidden.
Fixes: 135284765
Test: make RunSettingsRoboTests
Change-Id: Ia9f335920e4e3c4a90f0a6a49d1722a0c19ea83d
From traces analysis, we found getFreeBytes
was taking a long time to return.
getFreeBytes was used when storage controller
tried to get storage information.
In order to prevent this case, we put the action
which takes too much time in background thread.
Test: I can't reproduce it locally. From code view,
this is a reasonable root cause.
Fixes: 136268875
Change-Id: I78e42cde88553c003f198cffb5747b352055f59a
(cherry picked from commit 0c37f019f6)
From traces analysis, we found getFreeBytes
was taking a long time to return.
getFreeBytes was used when storage controller
tried to get storage information.
In order to prevent this case, we put the action
which takes too much time in background thread.
Test: I can't reproduce it locally. From code view,
this is a reasonable root cause.
Fixes: 136268875
Change-Id: I78e42cde88553c003f198cffb5747b352055f59a
(cherry picked from commit 0c37f019f6)
Add tobiast@ to the OWNERS file as libcore TL.
Remove pszczepaniak@ as he hasn't been on libcore
for a while.
Test: None
Change-Id: I714d3448ed60006cb58a0f49e95b7b834a6aac36
Settings activities were launched and resided on home task
which was originally started by FallbackHome. Adding a
unique task affinity for FallbackHome activity in order to
prevent the task being reused by other Settings activities.
Bug: 135696366
Test: starting Settings activity before FallbackHome finishes
Change-Id: I3fe41dd3b77e37236b11006dbf08727783b6a2ec
Added a module licenses option that lives in Legal information settings.
Clicking that option opens module licenses page, which displays every
module by name, filtered to exclude modules without license files.
Clicking a module in the list opens HTMLViewer.
Created ModuleLicensesProvider, a new ContentProvider that serves as a
redirect for the Uris sent to HTMLViewer so that they open asset files.
In order to provide the redirect, the provider will write the license file
to a file in Settings' cache directory when the license does not exist
in the cache or is outdated. The provider then opens that cached file.
Fixes: 135183006
Test: robotests
Change-Id: I7d69da34780c8c4efb150d0c0411078c12bc80d8
On the SIM details page, the preference leading to a page for
configuring wifi calling will appear based on the results of the
MobileNetworkUtils#isWifiCallingEnabled helper function. That helper
uses the ImsManager to check several conditions, among them both
isWfcEnabledByPlatform and isWfcProvisionedOnDevice.
The page for configuring wifi calling has a tabbed UX, with one tab for
each active subscription that supports it. The WifiCallingSettings class
gets a list of the active subscriptions to determine which tabs to show,
and removes any that don't support wifi calling, but was only using the
isWfcEnabledByPlatform test to do so. This is a problem because the code
for showing the contents inside the tab, in WifiCallingSettingsForSub,
includes a sanity check of isWfcProvisionedOnDevice and calls finish()
if that returns false.
What this meant in practice is that if you happened to have 2
subscriptions where one returns true for both isWfcEnabledByPlatform and
isWfcProvisionedOnDevice, but the other only returned true for
isWfcEnabledByPlatform, then you'd never be able to succesfully use the
wifi calling page at all because the tab for the subscription you
*aren't* trying to configure would always call finish() early.
The right long term solution to this problem is probably to remove the
tabbed UX entirely from this page, since we probably don't need it given
the overall new multi-SIM UX. But there may still be legacy uses and
that is likely a bigger change than we want to make right now.
As a stopgap, this CL just adds a check of isWfcProvisionedOnDevice to
the code for filtering out ineligible subscriptions from the tabbed
interface, which we should have always had anyway.
Fixes: 135591718
Test: make RunSettingsRoboTests
Change-Id: I656c3d3fb30cb6fabcb86685eae38c5f0cd0c6f2
