OSDN Git Service

android-x86/frameworks-base.git
10 years agocamera2: Add capture sequences and capture failures
Igor Murashkin [Thu, 5 Sep 2013 19:22:00 +0000 (12:22 -0700)]
camera2: Add capture sequences and capture failures

- CaptureResult#getRequest is used to tie a result to a request (for convenience)
- Add new CaptureFailure class to describe capture failure
- Results/frame numbers also return frame numbers, sequence ids
- Captures now all return the sequence id
- A sequence id onComplete is available in the CaptureListener

Bug: 10360518
Change-Id: I9ebaa45698c718a1185b5ae920b7975925fe2f60

10 years agoMerge "Fix for "Can't place cursor after combining accent"" into klp-dev
Raph Levien [Fri, 20 Sep 2013 23:27:57 +0000 (23:27 +0000)]
Merge "Fix for "Can't place cursor after combining accent"" into klp-dev

10 years agoMerge "camera2: Add more camera device states, make #openCamera async." into klp-dev
Igor Murashkin [Fri, 20 Sep 2013 23:21:50 +0000 (23:21 +0000)]
Merge "camera2: Add more camera device states, make #openCamera async." into klp-dev

10 years agoMerge "Ensure access to the map of geofence callbacks is synchronized in all cases...
destradaa [Fri, 20 Sep 2013 23:18:14 +0000 (23:18 +0000)]
Merge "Ensure access to the map of geofence callbacks is synchronized in all cases Fix Java reference leak in the JNI layer" into klp-dev

10 years agoMerge "Surfaces ParcelFileDescriptor#parseMode() as API" into klp-dev
Adam Lesinski [Fri, 20 Sep 2013 22:59:12 +0000 (22:59 +0000)]
Merge "Surfaces ParcelFileDescriptor#parseMode() as API" into klp-dev

10 years agoMerge "Delegate mkdirs() to vold when lacking perms." into klp-dev
Jeff Sharkey [Fri, 20 Sep 2013 22:56:07 +0000 (22:56 +0000)]
Merge "Delegate mkdirs() to vold when lacking perms." into klp-dev

10 years agoMerge "Revert "Workaround for slow AudioRecord destructor"" into klp-dev
Glenn Kasten [Fri, 20 Sep 2013 22:38:18 +0000 (22:38 +0000)]
Merge "Revert "Workaround for slow AudioRecord destructor"" into klp-dev

10 years agoFix for "Can't place cursor after combining accent"
Raph Levien [Fri, 20 Sep 2013 22:11:52 +0000 (15:11 -0700)]
Fix for "Can't place cursor after combining accent"

This patch fixes behavior where attempting to place the cursor at
the end of a string where the last character is a combining accent
actually placed it before that accent. This was especially annoying
for editing Thai text, because it made it difficult to delete a
trailing tone mark.

Fixes bug 8947569 "Can't place cursor after combining accent" and bug
10398332 "[Thailand] Thai Language(IME) -- Can't delete Thai tone mark
while writing message"

Change-Id: Ida1933c8f0ab6cdb0200db39891e9389e4bdba86

10 years agoEnsure access to the map of geofence callbacks is synchronized in all cases
destradaa [Fri, 20 Sep 2013 20:08:41 +0000 (13:08 -0700)]
Ensure access to the map of geofence callbacks is synchronized in all cases
Fix Java reference leak in the JNI layer

Bug: 10624887
Change-Id: I656973653d82c73a7ba2b51f22989ffed144c5df

10 years agoSurfaces ParcelFileDescriptor#parseMode() as API
Adam Lesinski [Fri, 20 Sep 2013 21:08:25 +0000 (14:08 -0700)]
Surfaces ParcelFileDescriptor#parseMode() as API

Moves ContentResolver#modeToMode() to ParcelFileDescriptor#parseMode()

bug:10711271
Change-Id: I28a5de4ca70adfbec4caf739863b77f0fd7fd9a4

10 years agoDelegate mkdirs() to vold when lacking perms.
Jeff Sharkey [Wed, 18 Sep 2013 00:30:33 +0000 (17:30 -0700)]
Delegate mkdirs() to vold when lacking perms.

Apps without sdcard_r or sdcard_rw are still able to write to
their package-specific directory, but someone needs to first make
that directory on their behalf.  This change will delegate the
mkdirs() call through to vold when an app fails to create directly.

MountService validates that the path belongs to the calling user, and
that it's actually on external storage, before passing to vold.

Update Environment to make app-vs-vold paths clearer.

Bug: 10577808
Change-Id: I43b4a77fd6d2b9af2a0d899790da8d9d89386776

10 years agoMerge "Remove over-aggressive stack movement." into klp-dev
Craig Mautner [Fri, 20 Sep 2013 21:28:14 +0000 (21:28 +0000)]
Merge "Remove over-aggressive stack movement." into klp-dev

10 years agoMerge "PrintDocumentAdapter contract not followed on print." into klp-dev
Svetoslav [Fri, 20 Sep 2013 21:12:54 +0000 (21:12 +0000)]
Merge "PrintDocumentAdapter contract not followed on print." into klp-dev

10 years agoRemove over-aggressive stack movement.
Craig Mautner [Fri, 20 Sep 2013 20:42:49 +0000 (13:42 -0700)]
Remove over-aggressive stack movement.

Moving a task to the bottom was also moving its stack to the back.
Since chrome always finishes by calling moveTaskToBack this put
the home stack in front of the application stack and would cause
bizarre window layering effects.

Definitely fixes 10764463.
Maybe fixes 10678010.

Change-Id: Ic202ae4fad44b36a8444255764d2135fccd4743d

10 years agoMerge "Clear names from externally supplied WorkSources to WifiManager (b/10733757...
David Christie [Fri, 20 Sep 2013 20:04:04 +0000 (20:04 +0000)]
Merge "Clear names from externally supplied WorkSources to WifiManager (b/10733757)" into klp-dev

10 years agoMerge "Fix incorrect exception being thrown from WifiConfiguration" into klp-dev
Vinit Deshapnde [Fri, 20 Sep 2013 19:30:26 +0000 (19:30 +0000)]
Merge "Fix incorrect exception being thrown from WifiConfiguration" into klp-dev

10 years agoPrintDocumentAdapter contract not followed on print.
Svetoslav [Fri, 20 Sep 2013 18:58:55 +0000 (11:58 -0700)]
PrintDocumentAdapter contract not followed on print.

1. Layout was not called after pressing the print button if the
   print attributes did not change. This is not correct since the
   previous layout calls were for preview purposes and the one
   after pressing print is not for preview. Hence, we always have
   to do this layout.

2. After layout we decide whether to ask the app to write some
   pages. We ask for a write if we do not have the pages selected
   by the user or the document changed (if the page count changed,
   the document type changed, or the app told us that the content
   changed). We were not computing correctly whether the document
   changed since we compared the size but the document info the
   app passes in after a layout does not have the size yet. We set
   the size after a write. So for layout purposes we should ignore
   the size. We only care if the page count, document type, or
   content changed where the latter is reported by the app in
   the layout callback.

3. We were not updating the PrintJob after setting the data
   size of the printed document.

4. Disabled debugging.

bug:10835370

Change-Id: Ic3b2871b4e954cdf610f8cf806de5fc6588a6bec

10 years agoMerge "Null not handled when parsing media size resource id attribute." into klp-dev
Svetoslav [Fri, 20 Sep 2013 18:59:39 +0000 (18:59 +0000)]
Merge "Null not handled when parsing media size resource id attribute." into klp-dev

10 years agoam 53b2a957: am 8ea8ff81: am 78ec732f: Doc change: Update Google Play Services refere...
Dirk Dougherty [Fri, 20 Sep 2013 18:51:37 +0000 (11:51 -0700)]
am 53b2a957: am 8ea8ff81: am 78ec732f: Doc change: Update Google Play Services reference docs.

* commit '53b2a957d19f21cd377d5f93225820d63c6b51d9':
  Doc change: Update Google Play Services reference docs.

10 years agoam 8ea8ff81: am 78ec732f: Doc change: Update Google Play Services reference docs.
Dirk Dougherty [Fri, 20 Sep 2013 18:48:52 +0000 (11:48 -0700)]
am 8ea8ff81: am 78ec732f: Doc change: Update Google Play Services reference docs.

* commit '8ea8ff81a93f367af59e328bb365ca9560e264ec':
  Doc change: Update Google Play Services reference docs.

10 years agoam 78ec732f: Doc change: Update Google Play Services reference docs.
Dirk Dougherty [Fri, 20 Sep 2013 18:46:23 +0000 (11:46 -0700)]
am 78ec732f: Doc change: Update Google Play Services reference docs.

* commit '78ec732f73aedc6afd79c4f924345a5adb61dc54':
  Doc change: Update Google Play Services reference docs.

10 years agoClear names from externally supplied WorkSources to WifiManager (b/10733757)
David Christie [Fri, 20 Sep 2013 17:44:01 +0000 (10:44 -0700)]
Clear names from externally supplied WorkSources to WifiManager (b/10733757)

Change-Id: I36102f13962df2093c1e35fb40081ead647eff32

10 years agoDoc change: Update Google Play Services reference docs.
Dirk Dougherty [Mon, 16 Sep 2013 17:37:40 +0000 (10:37 -0700)]
Doc change: Update Google Play Services reference docs.

Change-Id: Iaed2e01a18263000509f243e66a6b6b33037d6b9

10 years agoMerge "Fix issue #10310128: CTS: android.security.cts.ServicePermissionsTest" into...
Dianne Hackborn [Fri, 20 Sep 2013 18:23:07 +0000 (18:23 +0000)]
Merge "Fix issue #10310128: CTS: android.security.cts.ServicePermissionsTest" into klp-dev

10 years agoMerge "Add equals() and hashcode() to SpannableString" into klp-dev
Chet Haase [Fri, 20 Sep 2013 18:22:54 +0000 (18:22 +0000)]
Merge "Add equals() and hashcode() to SpannableString" into klp-dev

10 years agoMerge "Ensures that CompatibilityInfo is always set in Resources" into klp-dev
Adam Lesinski [Fri, 20 Sep 2013 18:22:17 +0000 (18:22 +0000)]
Merge "Ensures that CompatibilityInfo is always set in Resources" into klp-dev

10 years agoRevert "Workaround for slow AudioRecord destructor"
Glenn Kasten [Fri, 20 Sep 2013 18:10:28 +0000 (18:10 +0000)]
Revert "Workaround for slow AudioRecord destructor"

This reverts commit 8d87d7ba113ad90685f1d2669e65251a9b422a5a.

Change-Id: I78cbe73959d29fc0ff3153268503ae685cd0995c

10 years agoEnsures that CompatibilityInfo is always set in Resources
Adam Lesinski [Fri, 20 Sep 2013 03:33:15 +0000 (20:33 -0700)]
Ensures that CompatibilityInfo is always set in Resources

- Adds null checks when setting mCompatibilityInfo and uses default no-op
  CompatibilityInfo object instead of null
- Removes null checks when calling methods on mCompatibilityInfo since it is never
  null

bug:10204942

Change-Id: I36e88f5cdceb46b8b19fb902270ef74023a81546

10 years agoMerge "Supsend/resume batched scans around dhcp." into klp-dev
Robert Greenwalt [Fri, 20 Sep 2013 17:55:44 +0000 (17:55 +0000)]
Merge "Supsend/resume batched scans around dhcp." into klp-dev

10 years agoFix issue #10310128: CTS: android.security.cts.ServicePermissionsTest
Dianne Hackborn [Fri, 20 Sep 2013 17:54:16 +0000 (10:54 -0700)]
Fix issue #10310128: CTS: android.security.cts.ServicePermissionsTest

Change-Id: Id6432fc1b62db41c6c291ec08450c2fc37ced899

10 years agoMerge "Cleanup the shutdown of BatchedScans." into klp-dev
Robert Greenwalt [Fri, 20 Sep 2013 17:43:29 +0000 (17:43 +0000)]
Merge "Cleanup the shutdown of BatchedScans." into klp-dev

10 years agoNull not handled when parsing media size resource id attribute.
Svetoslav [Fri, 20 Sep 2013 17:43:24 +0000 (10:43 -0700)]
Null not handled when parsing media size resource id attribute.

bug:10835715

Change-Id: I0c9c1120c26e68b4c582599bbda3753d0e443d8c

10 years agoSupsend/resume batched scans around dhcp.
Robert Greenwalt [Fri, 20 Sep 2013 00:30:19 +0000 (17:30 -0700)]
Supsend/resume batched scans around dhcp.

Scanning while dhcp is running breaks dhcp, so stop the batched scans
when we need dhcp and start it up again after.

bug:10691401
Change-Id: Ifdeb6f35cfe4509b90fed1e1e694d0c107f24a7e

10 years agoFix incorrect exception being thrown from WifiConfiguration
Vinit Deshapnde [Fri, 20 Sep 2013 17:12:30 +0000 (10:12 -0700)]
Fix incorrect exception being thrown from WifiConfiguration

Bug: 10817189

Change-Id: Ib2985f50080cb9d983a6de6ce668223e8782c3b6

10 years agoam 8f6ed205: am 208e6079: am 0333302a: Merge "add custom head tag to verify developer...
Scott Main [Fri, 20 Sep 2013 17:10:22 +0000 (10:10 -0700)]
am 8f6ed205: am 208e6079: am 0333302a: Merge "add custom head tag to verify developers@android.com as site owner" into jb-mr2-docs

* commit '8f6ed205a2523a8170cdd3e824170fd77c4fd65a':
  add custom head tag to verify developers@android.com as site owner

10 years agoam 208e6079: am 0333302a: Merge "add custom head tag to verify developers@android...
Scott Main [Fri, 20 Sep 2013 17:07:46 +0000 (10:07 -0700)]
am 208e6079: am 0333302a: Merge "add custom head tag to verify developers@android.com as site owner" into jb-mr2-docs

* commit '208e6079b4ccd161de198553319d285711564ea2':
  add custom head tag to verify developers@android.com as site owner

10 years agoam 0333302a: Merge "add custom head tag to verify developers@android.com as site...
Scott Main [Fri, 20 Sep 2013 17:04:26 +0000 (10:04 -0700)]
am 0333302a: Merge "add custom head tag to verify developers@android.com as site owner" into jb-mr2-docs

* commit '0333302ad7162f66404c08e1c6c5b6b767aa78b3':
  add custom head tag to verify developers@android.com as site owner

10 years agoMerge "add custom head tag to verify developers@android.com as site owner" into jb...
Scott Main [Fri, 20 Sep 2013 16:58:38 +0000 (16:58 +0000)]
Merge "add custom head tag to verify developers@android.com as site owner" into jb-mr2-docs

10 years agoCleanup the shutdown of BatchedScans.
Robert Greenwalt [Fri, 20 Sep 2013 00:27:43 +0000 (17:27 -0700)]
Cleanup the shutdown of BatchedScans.

Fixes binder-death as well as disambiguation when an app has multiple requests in.

bug:10641931
Change-Id: Ib701d531fa51e7e2a027848cfdfcad3b47056923

10 years agoMerge "Include starting windows in the sysui vis calculation." into klp-dev
John Spurlock [Fri, 20 Sep 2013 16:25:30 +0000 (16:25 +0000)]
Merge "Include starting windows in the sysui vis calculation." into klp-dev

10 years agoadd custom head tag to verify developers@android.com as site owner
Scott Main [Fri, 20 Sep 2013 16:10:08 +0000 (09:10 -0700)]
add custom head tag to verify developers@android.com as site owner

Change-Id: I1a79669ab1e30de152354fafe5e397d5d1598f15

10 years agoInclude starting windows in the sysui vis calculation.
John Spurlock [Fri, 20 Sep 2013 16:04:47 +0000 (12:04 -0400)]
Include starting windows in the sysui vis calculation.

Sysui vis needs to be recomputed in the same code path as showing/
hiding the system bar (code path = finishPostLayoutPolicyLw) so
it can perform the new fade in/fade out to transparent modes at
the correct time.

Turns out no new state tracking is required, we already keep track
of this window as mTopFullscreenOpaqueWindowState.

So prefer mFocusedWindow when computing sysui vis as before,
but if null fallback to mTopFullscreenOpaqueWindowState.

Bug:10561554
Change-Id: I492766989a67fdac4f030451dcf00f6741a556c0

10 years agoMerge "UX compliance for CA Cert warning tile in QuickSettings" into klp-dev
Maggie Benthall [Fri, 20 Sep 2013 14:30:29 +0000 (14:30 +0000)]
Merge "UX compliance for CA Cert warning tile in QuickSettings" into klp-dev

10 years agoPut soundpooled sounds in the right place.
Daniel Sandler [Fri, 20 Sep 2013 04:26:29 +0000 (00:26 -0400)]
Put soundpooled sounds in the right place.

Bug: 10846839
Change-Id: I5318c2d54ea0e25fa74bbafc35fb4f7a2883b351

10 years agoFix the build - replace broken assets.
Jim Miller [Fri, 20 Sep 2013 03:01:56 +0000 (20:01 -0700)]
Fix the build - replace broken assets.

Change-Id: If039dee6d7e663d6abacd5239557d4d0a07fc912

10 years agoLoad RootInfo in background, invalidation.
Jeff Sharkey [Thu, 19 Sep 2013 22:25:56 +0000 (15:25 -0700)]
Load RootInfo in background, invalidation.

Move all RootInfo queries to background threads to avoid janking
the UI.  Update passes happen on spawed task, which swaps out updated
cache results when finished.  Support partial updates when only a
single package/authority has changed.  Watch for change notifications
for roots, since flags can change over time.

Ignore stopped packages when in background, but query them for roots
when launching any picker UI.

Optimize management launches by treating as one-shot requests that
don't need to wait for all RootInfo.

Bug: 1060045410745490
Change-Id: Ibc7b15688ef6b41bd7e9dd0d7564b501e60e49a9

10 years agoMerge "Fix default wallpaper dimensions" into klp-dev
Michael Jurka [Fri, 20 Sep 2013 02:33:47 +0000 (02:33 +0000)]
Merge "Fix default wallpaper dimensions" into klp-dev

10 years agoFix default wallpaper dimensions
Michael Jurka [Fri, 20 Sep 2013 02:32:20 +0000 (04:32 +0200)]
Fix default wallpaper dimensions

Bug: 10834711

10 years agoMerge "Don't back up apps in the 'stopped' state" into klp-dev
Christopher Tate [Fri, 20 Sep 2013 02:16:10 +0000 (02:16 +0000)]
Merge "Don't back up apps in the 'stopped' state" into klp-dev

10 years agocamera2: Add more camera device states, make #openCamera async.
Igor Murashkin [Wed, 11 Sep 2013 02:35:24 +0000 (19:35 -0700)]
camera2: Add more camera device states, make #openCamera async.

* Move CAMERA_IN_USE, MAX_CAMERAS_IN_USE to StateListener#onError
* Copy CAMERA_DISABLED to StateListener#onError

Bug: 10360518
Change-Id: Idd8cf42e1511d12682018588b2b413116cf65116

10 years agoMerge "Add camera affordance to navigation bar on phones" into klp-dev
Jim Miller [Fri, 20 Sep 2013 01:33:28 +0000 (01:33 +0000)]
Merge "Add camera affordance to navigation bar on phones" into klp-dev

10 years agoMerge "Move PdfDocument to android.graphics.pdf" into klp-dev
Svetoslav [Fri, 20 Sep 2013 01:14:02 +0000 (01:14 +0000)]
Merge "Move PdfDocument to android.graphics.pdf" into klp-dev

10 years agoMerge "Only show launcher for the bottom activity in a task" into klp-dev
Craig Mautner [Fri, 20 Sep 2013 00:51:20 +0000 (00:51 +0000)]
Merge "Only show launcher for the bottom activity in a task" into klp-dev

10 years agoMerge "Add config setting for default sms app." into klp-dev
David Braun [Fri, 20 Sep 2013 00:26:32 +0000 (00:26 +0000)]
Merge "Add config setting for default sms app." into klp-dev

10 years agoOnly show launcher for the bottom activity in a task
Craig Mautner [Fri, 20 Sep 2013 00:19:51 +0000 (17:19 -0700)]
Only show launcher for the bottom activity in a task

When transitioning from activity-over-launcher to task-over-launcher
ensureActivitiesVisibleLocked() was too aggressive in showing the
launcher. If there were any non-fullscreen activities in a task that
sits over the launcher then the launcher would be shown.

This fix adds a test to make sure the launcher will only be shown if
the bottommost activity in such a task is non-fullscreen.

Fixes bug 10840919.

Change-Id: I5dcd63be3fa2865ae38cbb921332937dfa4b5d47

10 years agoAdd equals() and hashcode() to SpannableString
Chet Haase [Wed, 18 Sep 2013 22:12:35 +0000 (15:12 -0700)]
Add equals() and hashcode() to SpannableString

ActionBar uses a transition to animate text changes. This transition
depends on testing the equality of start/end text values in CharSequence
objects. Without equals(), SpannableString will return false for objects
whose references are different, but whose text is exactly the same.

This CL adds the equals() method, and the accompanying hashcode method,
to ensure that two Spanned implementations will always be equal
if their text and span data are equal.

Issue #10760075 Wrong unread count in actionbar

Change-Id: I5e77d40dd302eca035e8c56d40f3cd0aef8e6424

10 years agoAdd camera affordance to navigation bar on phones
Jim Miller [Wed, 11 Sep 2013 01:37:01 +0000 (18:37 -0700)]
Add camera affordance to navigation bar on phones

This adds a camera button on phones that can be used to show
and launch the camera.

- Minor refactoring of touch event dispatch in PagedView.
- Disables usability hints when keyguard loads.
- Only add a touch handler for camera icon once during layout.
- Update after review.
- Updated with latest UX camera and camera background assets

Change-Id: I09cd5cb0e501fd0f4659bea96d00c92b07f805c4

10 years agoMove PdfDocument to android.graphics.pdf
Svetoslav [Wed, 18 Sep 2013 22:58:28 +0000 (15:58 -0700)]
Move PdfDocument to android.graphics.pdf

1. Move PdfDocument to android.graphics.pdf.

2. Changed the PdfDocument as per API concil request.

3. Updated the documentation.

bug:10461180
bug:10552565
bug:10681585
bug:10552336

Change-Id: I08e15b34cf37bb064248c887e6f59808019cafe8

10 years agoDon't back up apps in the 'stopped' state
Christopher Tate [Fri, 20 Sep 2013 00:12:27 +0000 (17:12 -0700)]
Don't back up apps in the 'stopped' state

E.g. right after install from the Play Store, or after a force stop.

Bug 10845317

Change-Id: I7764ce61caa77136ca2758657f111928f9740fc0

10 years agoMerge "Workaround for slow AudioRecord destructor" into klp-dev
Glenn Kasten [Fri, 20 Sep 2013 00:08:58 +0000 (00:08 +0000)]
Merge "Workaround for slow AudioRecord destructor" into klp-dev

10 years agoam 41c3c2b6: am 6e3b21cc: am 72f65780: Merge "docs: Fix for SDK 22.2.1 Windows downlo...
Joe Fernandez [Thu, 19 Sep 2013 23:50:16 +0000 (16:50 -0700)]
am 41c3c2b6: am 6e3b21cc: am 72f65780: Merge "docs: Fix for SDK 22.2.1 Windows download checksum and size" into jb-mr2-docs

* commit '41c3c2b687fa19d10c450a6d4b0b9f9a707a3854':
  docs: Fix for SDK 22.2.1 Windows download checksum and size

10 years agoam 6e3b21cc: am 72f65780: Merge "docs: Fix for SDK 22.2.1 Windows download checksum...
Joe Fernandez [Thu, 19 Sep 2013 23:47:56 +0000 (16:47 -0700)]
am 6e3b21cc: am 72f65780: Merge "docs: Fix for SDK 22.2.1 Windows download checksum and size" into jb-mr2-docs

* commit '6e3b21cc83da2b227704b9a6b0cbad258ea5bd84':
  docs: Fix for SDK 22.2.1 Windows download checksum and size

10 years agoWorkaround for slow AudioRecord destructor
Glenn Kasten [Thu, 19 Sep 2013 23:42:19 +0000 (16:42 -0700)]
Workaround for slow AudioRecord destructor

Bug: 10816117
Change-Id: If59028bc7dc7abb84e1e3642c721c2629653631f

10 years agoam 72f65780: Merge "docs: Fix for SDK 22.2.1 Windows download checksum and size"...
Joe Fernandez [Thu, 19 Sep 2013 23:44:58 +0000 (16:44 -0700)]
am 72f65780: Merge "docs: Fix for SDK 22.2.1 Windows download checksum and size" into jb-mr2-docs

* commit '72f65780c77f9c0c306fcb53efeb62fbeba3899c':
  docs: Fix for SDK 22.2.1 Windows download checksum and size

10 years agoam 354b4fef: am 0d8f0849: am 1d65500d: docs: Google APIs emulator 18 r3 - Platform...
Joe Fernandez [Thu, 19 Sep 2013 23:44:44 +0000 (16:44 -0700)]
am 354b4fef: am 0d8f0849: am 1d65500d: docs: Google APIs emulator 18 r3 - Platform Release Notes

* commit '354b4fef036ef9e57d79bf70a6988be92a9813eb':
  docs: Google APIs emulator 18 r3 - Platform Release Notes

10 years agoMerge "fix typo" into klp-dev
Ed Heyl [Thu, 19 Sep 2013 23:43:25 +0000 (23:43 +0000)]
Merge "fix typo" into klp-dev

10 years agoMerge "docs: Fix for SDK 22.2.1 Windows download checksum and size" into jb-mr2-docs
Joe Fernandez [Thu, 19 Sep 2013 23:41:39 +0000 (23:41 +0000)]
Merge "docs: Fix for SDK 22.2.1 Windows download checksum and size" into jb-mr2-docs

10 years agoam 0d8f0849: am 1d65500d: docs: Google APIs emulator 18 r3 - Platform Release Notes
Joe Fernandez [Thu, 19 Sep 2013 23:41:07 +0000 (16:41 -0700)]
am 0d8f0849: am 1d65500d: docs: Google APIs emulator 18 r3 - Platform Release Notes

* commit '0d8f0849ff842ed7ce16f183207ed7ebbc7bac7e':
  docs: Google APIs emulator 18 r3 - Platform Release Notes

10 years agoam 1d65500d: docs: Google APIs emulator 18 r3 - Platform Release Notes
Joe Fernandez [Thu, 19 Sep 2013 23:38:06 +0000 (16:38 -0700)]
am 1d65500d: docs: Google APIs emulator 18 r3 - Platform Release Notes

* commit '1d65500d6e16e497774dbff3b5d94cfd5256d118':
  docs: Google APIs emulator 18 r3 - Platform Release Notes

10 years agodocs: Fix for SDK 22.2.1 Windows download checksum and size
Joe Fernandez [Thu, 19 Sep 2013 23:31:30 +0000 (16:31 -0700)]
docs: Fix for SDK 22.2.1 Windows download checksum and size

external bug: http://b.android.com/60257

Change-Id: I399d21fd792c7c9a1bfac4402bc76fa23d45415b

10 years agofix typo
Ed Heyl [Thu, 19 Sep 2013 23:25:57 +0000 (16:25 -0700)]
fix typo

Change-Id: I97df15ea7d349e0224e1830d7327dde87e840a51

10 years agoMerge "Skip writing zeroes to java allocated memory." into klp-dev
Leon Scroggins III [Thu, 19 Sep 2013 23:19:12 +0000 (23:19 +0000)]
Merge "Skip writing zeroes to java allocated memory." into klp-dev

10 years agoAdd config setting for default sms app.
David Braun [Thu, 19 Sep 2013 23:13:04 +0000 (16:13 -0700)]
Add config setting for default sms app.

Different devices can have different default SMS apps.

Bug: 10818725 Allow OEM to specify or override a default SMS app (package)
Change-Id: I02cee9a3276ae39838cda181c9911041c81ea399

10 years agoMerge "fix emoji clipping in hw draw path" into klp-dev
Victoria Lease [Thu, 19 Sep 2013 22:46:50 +0000 (22:46 +0000)]
Merge "fix emoji clipping in hw draw path" into klp-dev

10 years agoMerge "Fix a regression in pausing activity immediately on sleep" into klp-dev
Amith Yamasani [Thu, 19 Sep 2013 22:42:50 +0000 (22:42 +0000)]
Merge "Fix a regression in pausing activity immediately on sleep" into klp-dev

10 years agofix emoji clipping in hw draw path
Victoria Lease [Thu, 19 Sep 2013 22:38:21 +0000 (15:38 -0700)]
fix emoji clipping in hw draw path

I guess we don't want to overwrite the last line in every RGBA glyph
with our one-line texture atlas spacer?

Bug: 10841207
Change-Id: Ief85ca58650c731e9d21dbf90942b7b44670abcc

10 years agoMerge "TextureView/GLES20Canvas: Support synchronous GLConsumers" into klp-dev
Eino-Ville Talvala [Thu, 19 Sep 2013 21:56:09 +0000 (21:56 +0000)]
Merge "TextureView/GLES20Canvas: Support synchronous GLConsumers" into klp-dev

10 years agoMerge "Implement issue #10691475: Kill cached processes if about to..." into klp-dev
Dianne Hackborn [Thu, 19 Sep 2013 21:44:20 +0000 (21:44 +0000)]
Merge "Implement issue #10691475: Kill cached processes if about to..." into klp-dev

10 years agoImplement issue #10691475: Kill cached processes if about to...
Dianne Hackborn [Thu, 19 Sep 2013 20:34:35 +0000 (13:34 -0700)]
Implement issue #10691475: Kill cached processes if about to...

...be uncached and too large

When the device is in a low RAM state, when we go to pull a cached
process out to use for some background operation, we can now kill
the current process if we consider its size to be too large.

Note that the current implementation for killing processes is to
just use the same killUnneededProcessLocked() method that we already
have for other things like too many cached processes.  This is a
little wrong here, though, because in this case we are at the
point where the caller is actually looking for a process to use.
This current code is not actually removing or cleaning up the
process, so we still need to return the now killed ProcessRecord
and let things fall out from there, which typically means the caller
trying to make an IPC on it and failing and falling into its "oh
no the process died unexpectedly" path.  All code using this
*should* be able to handle this correctly, anyway, since processes
really can be killed at any time.

At some point we may to make this implementation cleaner, where it
actually tears down the process right in the call and returns a
null ProcessRecord.  That is very dangerous however (we'd need to
go through all paths into this to make sure they are going to be
okay with process state changing on them like that), and I'm not
sure it is really worthwhile.  This intention is that killing
processes like this is unusual, due to processes being too large,
and anyway as I wrote all of our incoming code paths must already
be able to handle the process being killed at this point and one
could argue this is just another way to excercise those code paths.
Really, the main negative to this is that we will often have spam
in the log with exceptions about processes dying unexpectedly.
If that is the only issue, we could just add some conditions to
quiet that up at in this case.

We don't want to compute the size of the process each time we try
to evaluate it here (it takes 10s or ms to do so), so there is now
a new field associated with the process to give us the last pss
size we computed for it while it was in the cached state.

To be able to have better cached pss data when we now need it, the
timing for computing process pss has been tuned to use a much
shorter delay for the situations when the process has first switch
into a new state.  This may result in us having a fair amount more
pss data overall, which is good, as long as it doesn't cause us to
be computing pss excessively and burning cpu.

Procstats now also has new state to keep track of the number of
times each process has been killed by this new system, along with
the min, avg, max pss of all the times it has happened.  This has
slightly changed the checkin format to include this additional data
at the end of pkgkills/prockills lines.

Other changes here:

- Fixed a problem where GPU RAM was not being seen when dumping
  the full RAM details of a process.  This was because in that
  case the system would ask the process to compute its own MemInfo,
  which it returned, but the process doesn't have permission to
  access the files containing the GPU RAM data.  So now the system
  always computes the MemInfo and hands it to the app.

- Improved broadcast delays to not apply the delay if the next receiver
  of the broadcast is going to run in the same process as the last
  one.  A situation I was seeing was an application that had two
  receivers, one of which started a service; we are better off letting
  the second receiver run while the service is running.

- Changed the alarm manager's TIME_TICK broadcast to be a foreground
  broadcast.  This really should have been anyway (it is supposed to
  go out even minute, on the minute, very accurately, for UI elements
  to update), and is even more important now that we are doing more
  things to delay background broadcasts.

- Reworked how we maintain the LRU process list.  It is now divided
  into the two parts, the top always containing the processes holding
  activities.  This better matches the semantics we want (always try
  to keep those around modulated by the LRU order we interleave with
  other cached processes), and we now know whether a process is being
  moved on the LRU list because of an activity operation so we can
  only change the order of these activity processes when user operations
  happen.  Further, this just makes that common code path a lot simpler
  and gets rid of all the old complexity that doesn't make sense any
  more.

Change-Id: I04933ec3931b96db70b2b6ac109c071698e124eb

10 years agoFix a regression in pausing activity immediately on sleep
Amith Yamasani [Thu, 19 Sep 2013 19:30:32 +0000 (12:30 -0700)]
Fix a regression in pausing activity immediately on sleep

At some point during refactoring of ActivityStack, the code to pause the current
activity got deleted. Added back that line of code. Activity will now pause
as soon as the screen is turned off, rather than after 5 seconds (sleep timeout).

Bug: 10632898

Change-Id: If3cc8708d692d29a13dbd8cfd9edda8883b38c2e

10 years agoMerge "Camera2Tests: create new directory for manual camera2 test apps" into klp-dev
Igor Murashkin [Thu, 19 Sep 2013 21:11:57 +0000 (21:11 +0000)]
Merge "Camera2Tests: create new directory for manual camera2 test apps" into klp-dev

10 years agoMerge "Remove STOPSHIP in WifiWatchdogStateMachine" into klp-dev
Yuhao Zheng [Thu, 19 Sep 2013 21:11:00 +0000 (21:11 +0000)]
Merge "Remove STOPSHIP in WifiWatchdogStateMachine" into klp-dev

10 years agoMerge "Debug for b/10689184." into klp-dev
Craig Mautner [Thu, 19 Sep 2013 21:08:27 +0000 (21:08 +0000)]
Merge "Debug for b/10689184." into klp-dev

10 years agoRemove STOPSHIP in WifiWatchdogStateMachine
Yuhao Zheng [Thu, 19 Sep 2013 21:02:58 +0000 (14:02 -0700)]
Remove STOPSHIP in WifiWatchdogStateMachine

There used to be some STOPSHIP code in WifiWatchdogStateMachine for debug
purposes. We don't need them for the release.

Bug: 10841961
Change-Id: I501d62e9891ace52317e6c1d399b877175099a3c

10 years agoMerge "Increasing visibility of pressed feedback for default ui widget assets" into...
petergng@google.com [Thu, 19 Sep 2013 20:57:02 +0000 (20:57 +0000)]
Merge "Increasing visibility of pressed feedback for default ui widget assets" into klp-dev

10 years agodocs: Google APIs emulator 18 r3 - Platform Release Notes
Joe Fernandez [Thu, 19 Sep 2013 18:09:22 +0000 (11:09 -0700)]
docs: Google APIs emulator 18 r3 - Platform Release Notes

Change-Id: I5c11c8c480aa5dcd515f11ae5a4c40b28197a572

10 years agoTextureView/GLES20Canvas: Support synchronous GLConsumers
Eino-Ville Talvala [Thu, 19 Sep 2013 19:49:13 +0000 (12:49 -0700)]
TextureView/GLES20Canvas: Support synchronous GLConsumers

Always update to the newest available frame from a GLConsumer.
Otherwise, with a synchronous queue, rendering can fall behind and
eventually deadlock with producer.

Bug: 10830400
Change-Id: I7f1d752c80ae5dac892a26d82e86806c27f5d955

10 years agoSkip writing zeroes to java allocated memory.
Leon Scroggins III [Thu, 19 Sep 2013 15:34:06 +0000 (11:34 -0400)]
Skip writing zeroes to java allocated memory.

If pixel memory was just allocated by Java, tell our decoders not
to write 0s, since the memory was initialized to 0. Likewise,
when drawing to a bitmap with memory just allocated by Java, do
not erase to 0.

Depends on a change to external/skia to add the new option on
image decoders:
https://googleplex-android-review.git.corp.google.com/362663

BUG:10016979
Change-Id: I9a3dc969870f8516e7d8495fe96d0a6b8225eda2

10 years agoMerge "Return tasks in correct order." into klp-dev
Craig Mautner [Thu, 19 Sep 2013 20:03:24 +0000 (20:03 +0000)]
Merge "Return tasks in correct order." into klp-dev

10 years agoDebug for b/10689184.
Craig Mautner [Thu, 19 Sep 2013 19:02:09 +0000 (12:02 -0700)]
Debug for b/10689184.

Change-Id: I67b7a4bd7fb902b0d0275af0665625593dc1af00

10 years agoMerge "Support for "not the home app you're looking for" metadata" into klp-dev
Christopher Tate [Thu, 19 Sep 2013 19:46:20 +0000 (19:46 +0000)]
Merge "Support for "not the home app you're looking for" metadata" into klp-dev

10 years agoIncreasing visibility of pressed feedback for default ui widget assets
petergng@google.com [Thu, 19 Sep 2013 18:31:38 +0000 (11:31 -0700)]
Increasing visibility of pressed feedback for default ui widget assets

Change-Id: Ibfab7a6428b9a971a8624d3b8daeb40a503f7a1f

10 years agoSupport for "not the home app you're looking for" metadata
Christopher Tate [Thu, 19 Sep 2013 01:31:59 +0000 (18:31 -0700)]
Support for "not the home app you're looking for" metadata

* Make sure that pm.getHomeActivities() returns the activity metadata
  as well, so that the caller can trace the reference

* Add a public canonical name for that metadata key

Bug 10749961

Change-Id: Ic4d0750d61001ffe5af180398f042afa30eea1ce

10 years agoMerge "Bar gradients now 55% to 0% black." into klp-dev
John Spurlock [Thu, 19 Sep 2013 19:26:25 +0000 (19:26 +0000)]
Merge "Bar gradients now 55% to 0% black." into klp-dev

10 years agoUX compliance for CA Cert warning tile in QuickSettings
Maggie Benthall [Tue, 17 Sep 2013 23:23:59 +0000 (19:23 -0400)]
UX compliance for CA Cert warning tile in QuickSettings

Text not all uppercase
Reduce spacing between icon and text
Break the string in the middle.

Bug: 10785095
Change-Id: If85da55452b06d77d2ae1c40f3d63431c2ae2a9a

10 years agoBar gradients now 55% to 0% black.
John Spurlock [Thu, 19 Sep 2013 19:22:17 +0000 (15:22 -0400)]
Bar gradients now 55% to 0% black.

Bug:10840205
Change-Id: If30e13ab4f61cd65b7fdf7e5e987191fec1f45f2

10 years agoMerge "Use a separate thread for services that do NTP lookup" into klp-dev
Amith Yamasani [Thu, 19 Sep 2013 19:17:50 +0000 (19:17 +0000)]
Merge "Use a separate thread for services that do NTP lookup" into klp-dev

10 years agoMerge "Also catch one more tablet layout NPE." into klp-dev
Jeff Sharkey [Thu, 19 Sep 2013 18:27:51 +0000 (18:27 +0000)]
Merge "Also catch one more tablet layout NPE." into klp-dev

10 years agoAlso catch one more tablet layout NPE.
Jeff Sharkey [Thu, 19 Sep 2013 18:25:56 +0000 (11:25 -0700)]
Also catch one more tablet layout NPE.

Bug: 10837208
Change-Id: I178110431221773658403319b2be7aff22adadc8

10 years agoMerge "Fix tablet layout and resources." into klp-dev
Jeff Sharkey [Thu, 19 Sep 2013 18:22:12 +0000 (18:22 +0000)]
Merge "Fix tablet layout and resources." into klp-dev