This reverts commit c4b8f36de5.
Having InputContentInfo#requestPermission() should not hurt developers,
but we can polish the behavior in a subsequent CL without changing
the API.
Bug: 29450031
Bug: 29892936
Change-Id: I1b43c19417b643d0c269af860db2d309b73a90d5
It turns out that requiring editor authors to call
InputContentInfo#requestPermission() as needed is just confusing and can
cause compatibility issues, because if an editor author forgot to call
that method then there would be no way for IME developers to prevent
permission denial except for relaxing the default permission of the
ContentProvider just because of such an application. This is not what we
want to see.
My conclusion is that the system should automatically call
InputContentInfo#requestPermission() (or do any equivalent operation)
when InputConnection#INPUT_CONTENT_GRANT_READ_URI_PERMISSION is
specified, like we have done in Context#startActivity().
With this CL, the system automatically grants a temporary URI permission
to the target application when the IME calls
InputConnection#commitContent() with
InputConnection#INPUT_CONTENT_GRANT_READ_URI_PERMISSION, and the
temporary permission will be revoked by any of the following events:
- InputContentInfo#releasePermission() is explicitly called by the
target application.
- The target application returned false in
InputConnection#commitContent().
- All the InputContentInfo instances copied from the original one are
GC-ed.
Bug: 29450031
Bug: 29892936
Change-Id: I37fb744e4d3d1c59177fb0a9be4ef5c325c9a39f
Previously the DisplayMetrics passed to a new ResourcesImpl
object would be generated from the default DisplayAdjustments.
We now use the correct DisplayAdjustments for the ResourcesImpl
and make sure to update them for things like rotation changes.
Bug:29619314
Change-Id: If2ba0d7670a4554dcd3fde9766e2337f20a191fd
(cherry picked from commit 8e8d23214a)
Fix a regression where a change in insets would forceLayout on the
view hierarchy but not run the measure/layout as a result. This would
cause layout requests to become stalled until a window-level relayout
event.
Bug 29634368
Change-Id: Ia3f32f5891c8b32c06c13f95ebd0572233572b04
Previously the DisplayMetrics passed to a new ResourcesImpl
object would be generated from the default DisplayAdjustments.
We now use the correct DisplayAdjustments for the ResourcesImpl
and make sure to update them for things like rotation changes.
Bug:29619314
Change-Id: If2ba0d7670a4554dcd3fde9766e2337f20a191fd
It turns out that we can let the system to call
InputMethodService#exposeContent(InputContentInfo, EditorInfo), which
added in my previous CL [1], during the IME is calling
InputConnection#commitContent() as follows.
[IME]
InputContentInfo contentInfo = new InputContentInfo(
contentUri,
new ClipDescription(description, new String[]{mimeType}),
linkUrl);
getCurrentInputConnection().commitContent(
inputContentInfo,
InputConnection.INPUT_CONTENT_GRANT_READ_URI_PERMISSION,
null);
[App]
try {
contentInfo.requestPermission();
// Load inputContentInfo.getContentUri() here.
} finally {
contentInfo.releasePermission();
}
This gives us flexibility to let InputConnection#commitContent() do all
the magic for IME developers like other APIs such as
Context#startActivity(), rather than asking them to call one more API to
grant a temporary URI permission like a scenario where
Context#grantUriPermission() is used.
[1]: I2772889ca01f2ecb2cdeed4e04a9319bdf7bc5a6
25e0813e6e
Bug: 29450031
Change-Id: I99536cd58c9984af30b0bafb4a1dd25a26634a2d
Bug: 29586513
Also gives BackdropFrameRenderer a direct-destroy
of Choreographer since it's hammering on new Threads
and we don't want to wait for the GC to release
FDs.
Change-Id: Id2ec0af2ee4d5304961c4ab87a104ccb92f35fc2
This is a follow up CL to my previous CLs [1][2] that introduced
InputConnection#commitContent(InputContentInfo, Bundle) API to enable
IMEs to send a content to the target application.
With this CL, IME developers are able to temporarily expose
InputContentInfo object to the target package without permanently
granting URI permission. Although calling IMS#exposeContent() is
allowed only for the IME that is currently selected, the client is able
to request a temporary read-only access even after the current IME is
switched to any other IME as long as the client keeps InputContentInfo
object.
Here is a sample code snippet about how to use this mechanism.
[IME]
InputContentInfo contentInfo = new InputContentInfo(
contentUri,
new ClipDescription(description, new String[]{mimeType}),
linkUrl);
exposeContent(contentInfo, getCurrentInputEditorInfo());
getCurrentInputConnection().commitContent(inputContentInfo, null);
[App]
try {
contentInfo.requestPermission();
// Load inputContentInfo.getContentUri() here.
} finally {
contentInfo.releasePermission();
}
[1]: Iaadf934a997ffcd6000a516cc3c1873db56e60ad
152944f490
[2]: Ica1ba3154795c1bf44e140dfe639b299f83cd8af
adebb52588
Bug: 29450031
Change-Id: I2772889ca01f2ecb2cdeed4e04a9319bdf7bc5a6
Bug: 29547000
Instead of shifting the window off screen fall
back to the UI-thread calculated values. This fixes
a race issue around tear-down where we stop
drawing long before our window is actually no longer
visible, as well as fixes a funky issue with
PRESERVE_GEOMETRY during first draw.
Change-Id: I792d1966f5aaee1e703295ae79166c723b97a1dc
Make sure that when our Resources get updated, that DisplayAdjustment
and Display properly reflect the potentially new screen dimensions.
Bug:28388969
Change-Id: I340550ea094ece87abc8790dd46aaa60ab3cedd3
We need to be careful about stripping FLAG_PRESERVE_GEOMETRY
as it breaks render thread sync. Notably we don't need
to strip it in the case of mUpdateWindowNeeded, this just signals
that Window#resized has been called, or something similar, but if there
has been no actual size change, then we can leave the geometry to
the render thread. This particularly manifests in rotation scenarios
where the SurfaceView will receive resize ahead of the app receiving
a configuration change.
Bug: 28823590
Change-Id: Ie2bbe7d9bd9f0a5ab64a08776bae7344eaecb605
Add a rotation mode which does not require freezing
the screen. For situations like Camera where only small
elements move on screen, this allows for seamless changes
of display orientation. This is achieved by transforming the
windows with their current buffer in the same transaction that
we rotate the display. We set things up so the windows are
frozen this way until they submit buffers in the new orientation.
There is a special case in the Camera window itself, and it's use
of NATIVE_WINDOW_TRANSFORM_INVERSE_DISPLAY. In this case the buffer
contents are rotated by SurfaceFlinger and will never resize, for these
windows we just need to update the scaling matrix.
Bug: 28823590
Change-Id: I52dc6a86fcb3c08f736f0977ba3975a24fb8136c
