On Devsite, the redirect parser doesn't understand regular
expressions, which breaks ~ 10% of our existing redirects; fixing
them now. Also fixing a few other redirect-related bugs while I've got
the hood open.
See first comment for doc stage location.
bug: 28447522
bug: 28467616
Change-Id: I5f315fd54cb9353a56674b5aac4dcd53875263c9
- This ensures that we perform the same dismiss animation when
dismissing via the button or via swipe.
Bug: 28443410
Change-Id: I309c38253e9c2dc78a5882dc663eec84a11619dd
Force a setAppVisibility if the previous app was translucent, so that
the resumed app is added to the opening app list. Otherwise opening
app list is empty and the transition can't be set up correctly.
bug: 28463843
Change-Id: I7f8e28d0f804c799a429743df6fc501c06d62f21
Bug: 28519669
This trace tag makes shadows look way more expensive
than they actually are, particularly troublesome
when tracing things with large number of shadows.
Change-Id: Ib6558b1388edd4b006ec15127470cb9ab563f954
There are two parts to VD's native allocation:
1) VD's internal data structure (i.e. groups, paths, etc that make
up of the VD tree). This structure can change, when a VD is used
to load a different drawable resource.
2) Two bitmap caches, not both of which will necessarily be allocated
The size of the bitmap cache depends on canvas matrix and drawable
bounds, and therefore can often change.
We need to count the native allocation from the above against Java heap.
Bug: 26269056
Change-Id: If833aedcf7f3efe00ea73a41ddccb1b48066ffd8
Registering with the maximum potential memory made by visible by an
ImageReader/Writer can cause the VM to try to heavily garbage-collect
the rest of the application. This can have significant impact on camera
applications, since they often have large ImageReader/Writer queues.
Bug: 28454727
Change-Id: I2ba43635f93da66655be024165a15631b3b421d3
- When adding a new user, adds its default whitelist apps.
- When removing an user, clears its revoked default whitelist apps.
BUG: 28451218
Change-Id: Ic30a3d6d30c7575620edd3ad32c3221203626f8d
On non-FBE devices, we offer to migrate CE/DE storage of certain
system apps during boot. However, if this migration happens after
app code has started to run, we end up clobbering open files such
as databases.
The right way to solve this is to apply all migration early during
boot (in the PMS constructor) before any app code can start.
Bug: 28512897, 28525109
Change-Id: I17eb9a09e5862d873b86c05ce35874a34386e433
Scrim will fade in after the PIP control's focus losing animation,
and will be gone immediately when PIP control gets focus.
Bug: 28041690
Change-Id: Ibb7fa0ce15ba1af13781a1bd5e87609fefa95cc0
The testing shim should be included as a prebuilt. This allows
us to destroy the signing key. This means that the shim can never
be upgraded in the field; aside from its intended purpose in
a CTS test.
Bug: 28398205
Change-Id: I033835d227416ff8aa6602abfc8b977c74cb666d
am: c10ce54761
* commit 'c10ce54761f4bb1a887561af15a51625c902049b':
Delete all of the tools/ directory, but save the adk files by moving them up. Redirects for all the tools/ files are already in Piper's site-android/en/_redirects.yaml file. The new adk redirects are in this cl: https://critique.corp.google.com/#review/121302608
Change-Id: Ia8bc55aab35cdfadd7b5cecf435e68b4b3b7ba60
am: 41f7ff9150
* commit '41f7ff915033ebc8b2210e306cba99c4173a2285':
Delete all of the tools/ directory, but save the adk files by moving them up. Redirects for all the tools/ files are already in Piper's site-android/en/_redirects.yaml file. The new adk redirects are in this cl: https://critique.corp.google.com/#review/121302608
Change-Id: Ib2c3b9bfe2ac1899005c2ef00c3f73dae471032b
- Don’t adjust focus stack when a stack is left empty because we moved
the last task in it to the stack that will be the new top (focused).
Focus can be temporarily adjusted to another stack which we don’t want.
- Protect against trying to resume an already resumed activity in
ASS.resumeFocusedStackTopActivityLocked. This will just cause a crash
on the client side.
- In ASS.moveTasksToFullscreenStackLocked defer resume activity call
until we are done moving all the tasks to the new stack.
- In ASS.moveTaskToStackUncheckedLocked check top running activity
instead of just the top activity when we are trying to determine if
the task as focus and resumed. It is possible the resumed activity
in the task isn’t the top most because the top most is still
initializing or going away.
Bug: 28219871
Change-Id: I8c1113503f6f83d39983375a4f92155ddc04f20f
If we didn't find a home app, the boot just appears to be wedged
and it's hard to investigate what the root-cause is. Let's log to
make it more obvious.
Bug: 28465914
Change-Id: I9697af83eb27efb54f1f703506d4f148c96425b0
