[flutter_runner] use dedicated allocation info
This informs the vulkan driver that we're allocating memory
dedicated to a single image. This is required for aemu and
might also allow future drivers to make optimizations currently
not possible.
No change in behavior. Existing tests (modular_tests,
scenic_tests, etc.) are sufficient and regressions will be
prevented by running these tests on aemu.
DX-939 #comment
Test: fx shell run fuchsia-pkg://fuchsia.com/basemgr#meta/basemgr.cmx --base_shell=fuchsia-pkg://fuchsia.com/spinning_cube#meta/spinning_cube.cmx
Change-Id: If4ecd9aaa09f12f94654a68e8e9fe979748f44af
[flutter_runner] use external image create info
DX-939 #progress
MA-394 #progress
This informs the vulkan driver that we're creating an image
that will be backed by external memory. The driver driver can
decide to use different memory requirements based on the if
memory for the image can be exported or not.
No change in behavior. Existing tests (modular_tests,
scenic_tests, etc.) are sufficient and regressions will be
prevented by running these tests on aemu.
Test: fx shell run fuchsia-pkg://fuchsia.com/basemgr#meta/basemgr.cmx --base_shell=fuchsia-pkg://fuchsia.com/spinning_cube#meta/spinning_cube.cmx
Change-Id: I489318c2e31f752f76c80a81245e203861d44d94
Explicitly set |Application|'s |flutter::Settings| |trace_skia| field to
false (it currently defaults to false), in order to slightly simplify
the enabling skia trace events workflow.
PT-145 #comment-patch
Change-Id: Ib40f9ed3dc6f824056465db2cd45309c78b7e3b4
Build rules still reference creating share snapshot data and instructions. This makes the engine to always pass them as empty to the dart vm. To be followed up with a change to alter the build rules to stop referencing the shared snapshots.
This is not being used currently and the fact that the runner will be built outside of the flutter tree means that the apps will not have much to gain via shared snapshots. The rationale behind this change is to partially make migrating the runner out of topaz tree easier.
Change-Id: Ibc4dd6a298d65082416af753522f5a17c88a750a
The earlier design speculated that embedders could affect the same
transformations on the layers post engine compositor presentation but before
final composition.
However, the linked issue points out that this design is not suitable for use
with hardware overlay planes. When rendering to the same, to affect the
transformation before composition, embedders would have to render to an
off-screen render target and then apply the transformation before presentation.
This patch negates the need for that off-screen render pass.
To be clear, the previous architecture is still fully viable. Embedders still
have full control over layer transformations before composition. This is an
optimization for the hardware overlay planes use-case.
Fixes b/139758641
The tests we write must be resilient to the order in which they are run in the
harness. That is, they must not rely on global state set by other tests that
have already run in the process. Also, these tests must themselves be
repeatable. That is, they must correctly clean up after themselves and be able
to run successfully again in the same process.
This patch adds some safeguards against (but does NOT guarantee) the addition of
tests that violate the dictum.
Additionally, test failures must be easily reproducible for folks investigating
the test failure. Also, tests that assert correctness of unrelated code must not
stop progress on the authors patch.
This changes does not hinder reproducibility of test failures because the random
seed is printed in the logs before running each test. Developers attempting to
reproduce the failure locally can do the same via the following invocation
`--gtest_shuffle --gtest_repeat=<the count> --gtest_random_seed=<seed from failing run>`.
This change does introduce potential burden on patch authors that may see
failures in unrelated code as a newly failing shuffle seed is used on their
runs. To ameliorate this, we will formulate guidance for them to aggressively
mark such tests as disabled and file bugs to enable the same.
The test seed is intentionally kept low because it’s purpose is to test that
individual tests are repeatable. It must not be used as a replacement for
fuzzing.
* Do not pass a milliseconds::max() timeout that will overflow when converted
to nanoseconds
* Avoid holding the task_queue_mutex_ while calling glfwWaitEventsTimeout
* Use a signed type to hold the difference between a task's timestamp and
the current engine time
Fixes https://github.com/flutter/flutter/issues/40281
Makes the plugin registration structure consistent with macOS. This will
be used in generated plugin registrant files rather than a specific
implemenation class, so this helps unblock the creation of generated
registrants on Windows and Linux.
Manage resource and onscreen contexts using separate IOSGLContext objects.
FlutterView owns the onscreen context, and PlatformViewIOS owns the resource context.
Additionally, we now use the engine directly as a delegate instead of storing potentially dead runtime_controller.
Unit tests have been updated to include an engine restart check which would fail before the fix.
This fixes https://github.com/flutter/flutter/issues/40303
Creates a starting point for an app delegate. For now it just
incorporates the menu and window renaming functionality that's currently
in the FDE example, but in the future this will also do forwarding of
application lifecycle events to plugins.
Fixes https://github.com/flutter/flutter/issues/32419
Naively embedded platform views on Android were never able to receive
keyboard input, because they were never focusable. So far we've worked
around the limiation by hooking into InputMethodManager and proxying the
InputConnection from a focused window over to the embeded view.
Android Q changed InputMethodManager to be instanced per display instead
of a singleton. Because of this our proxy hook was never being called,
since it was being set up on a different instance of IMM than was being
used in the virtual display.
Update `SingleViewPresentation` to store the IMM from the focused window
and return it whenever there are any calls to `INPUT_METHOD_SERVICE`.
This hooks our proxy back into place for the embedded view in the
virtual display. This restores the functionality of our workaround from
previous versions.
Unfortunately there's still a lot of noisy error logs from IMM here. It
can tell that the IMM has a different displayId than what it's expecting
from the window.
This also updates the unit tests to support SDK=27. SDK 16 doesn't have
DisplayManager, so there were NPEs attempting to instantiate the class
under test.
Manage resource and onscreen contexts using separate IOSGLContext objects. FlutterView owns the onscreen context, and PlatformViewIOS owns the resource context.
