When the magnifier content moves relative to the parent surface,
it will call setPosition on the content surface control. This conflicts
with position updates called by the BlastBufferQueue adapter
causing a flicker on screen. Fix this by providing a wrapper surface
to BlastBufferQueue adapater to send buffer updates.
Fixes: 186072574
Test: Select text and see magnifier surface does not flicker or move
around
Change-Id: Idfcc06a5d90f400f69e5cbe91008a0cb59fd4646
Some apps operate on the view text based on layout calculations made on
the transformed text. This can cause issues (even crashes) if the
transformed text is longer than the original, as is often the case with
Translate. A previous temporary fix,
I685ae4a9752c817db81f0d736f1d746b2e4a9839, ellipsized the translated
text. This change fixes it properly by instead padding the original
text.
A followup change will make this controllable by API instead of padding
every TextView.
The en space character, U+2002, is used for the padding here. It seems
to work well in practice for multiple languages. We don't make this
configurable through API as it could make debugging harder (issues
triggered by server-side changes).
Bug: 179693024
Test: atest CtsTranslationTestCases
Test: manual - toggling between original and translated, scrolling, new
views appearing, multiple apps
Change-Id: I9d7db5d2bc200e4042baf3267796f00cfc298f19
Updated spring values to the latest ones agreed in design. In addition,
added a velocity multiplier that only affects after it absorbs an impact
at the scroll boundary. This will help to increase the spring behavior
when we fling the content, but keeping it controlled when we pull and
release.
Bug: 183768138
Test: ran EdgeEffectTest and visual
Change-Id: I4ed7b01dd42a9a5adda637917369c349a6ae2f76
Distance and speed were multiplied by LINEAR_STRETCH_INTENSITY which is
not directly related to speed and and distance. Moreover the velocity
threshold was a little bit high (1/0.016, wich means 62 pixels/second)
and the distance as well (1 is ok if there's no scale but since there's
some stretching the difference was being magnified).
Bug: 184892316
Test: EdgeEffectTests and visual
Change-Id: Ibd7adb16d1d00df0ab6e96132bfe96a4fa4d45dc
Updated stretch intensity to 0.016 and added a new uniform to control
the intensity of the interpoloation
Bug: 183768138
Test: ran EdgeEffectTest and visual
Change-Id: Iac10bee5980802fffe838ed83e02b7c41dee6f71
There are some issues the resume/finish seems not work well. The
debug flag should be enabled manually, but in the most time the
user may forget to enable it before testing. Try to enable it in the
debug rom to help debugging, we use b/182433547 to trace the disable
this check before rom release.
Bug: 182433547
Test: manual
Change-Id: I2f85a0e17f8a2b8a4b93252044cf48d739b4eca2
Refactored error checking logic to ignore
all invalid parameters for vecX/vecY
regardless if NaN parameters were
computed through intermediate math
operations or provided as inputs
Fixes: 183781985
Test: Added CTS test
Change-Id: Ib9f3dd1becc0486dc15ce306660bd5d469b81256
It is one of the last users of deferTransactionUntil which we
are hoping to remove in the server (and gain some CPU usage wins
in the process).
Bug: 168505645
Test: Existing tests pass
Change-Id: I2ba66caf7b0eaf2ed7d4fe896d7dc58f57c6e37f
This is a request of the API Council.
There was a hidden and, AFAICT, un-used constructor with the same
signature, which I have removed. Before replacing the constructor I
checked the whole OS builds and run with the constructor simply removed.
Fix: 182824754
Test: atest CtsWidgetTestCases:RemoteViewsRecyclingTest
Test: atest CtsWidgetTestCases:RemoteViewsTest
Change-Id: I0f12d2fa520cf4f0d54a7e3859337cc9a395ae66
There are 2 changes:
* When a view is scrapped, Autofill IDs in its subtree are reset. This
prevents flaky behavior like translating the wrong views when events are
ordered a certain way.
* Views that are being translated are marked as having transient state
since the system needs to attach the response to it later as well as
deliver UI Translation state changes to it.
Bug: 182491706
Test: atest CtsTranslationTestCases
Test: atest CtsContentCaptureServiceTestCases
Test: atest CtsAutoFillServiceTestCases
Test: manual - check in logs that autofill ids aren't reused on
scrolling or new views appearing
Test: manual - translated views stay translated while on the screen
when other views are scrolled off the screen
Test: manual - translated views stay translated when new views appear
Change-Id: I20a52415e3fd191768442d70614d536e11633dfa
Bug: 184392385
The pull consumption should be related to the width, not the height.
Test: Ran ApiDemos and result is vastly improved
Change-Id: I19e1262c42b4b85920e5038a789e2e934569cac3
Added more error checking logic to
protect against invalid usages of EdgeEffect
that include setSize not being called or
invalid bounds parameters.
Fixes: 183945458
Fixes: 183895581
Fixes: 183781985
Test: Added CTS test to EdgeEffectTests
Change-Id: I1eabd96206216526dbe843d0fb8c0fe6cccd8395
> This allows the scroller to be used in Launcher3 instead of a custom copy
Bug: 183966408
Test: Verified with Launcher3
Change-Id: I1c945774d0323fabcf4bfd5a02c2eaf8e95a218d
This change doesn't contain the virtual view part and the API, it
will be done in the next change.
Bug: 178046780
Test: manual
Test: atest CtsTranslationTestCases
CTS-Coverage-Bug: 177960696
Change-Id: Idba66a882a90168ecdd93423c0d5d054ab040dad
For content insertion from the IME, the platform currently assumes
that it can automatically revoke URI permissions (via
InputContentUriTokenHandler.finalize) any time after the
InputContentInfo instance is garbage collected and
InputContentInfo.releasePermission() has not been called explicitly.
This can cause permissions to be revoked before the receiving code
has had a chance to read the content when using
OnReceiveContentListener, since we did not keep a reference to
InputContentInfo.
This change does the following:
* Sets the InputContentInfo instance in the ContentInfo payload passed
to OnReceiveContentListener.
* Exposes a hidden API on ContentInfo to allow system features to
proactively release permissions for sources that allow this (IME and
drag-and-drop).
* Updates notifications direct-reply code to proactively release
permissions as soon as they are no longer needed.
Bug: 183643556
Test: Manual
Change-Id: I1cbc55d53804810a3cd4b6b97c197f4dd2d7d8e6
This change doesn't contain the virtual view part and the API
dispatchTranslationRequests, it will be done in another change.
Bug: 178046780
Test: manual
Test: atest CtsTranslationTestCases
CTS-Coverage-Bug: 177960696
Change-Id: I65fe1db19c9dff21c0caca425fbb7d08559e730b
Previously, the SpellChecker use the start and end positions as the key
to maintain the SuggestionSpan cache.
However, the actual start/end position of the SuggestionSpan may changes
when the text is changed. Hence we may not be able to find the
corresponding SuggestonSpan for a new suggestion with the changed
start/end positions.
In this fix, we deprecated the cache and use Spanned#getSpans to find
overlapped spans instead.
Change-Id: I1e732f335ad745d85a7620e3ced6858551f18232
Fix: 182982338
Test: atest CtsInputMethodTestCases:SpellCheckerTest
Change-Id: I327e5163499cfc978ff8480664a7e728b6ea7249
