* Read core platform kernel file during Dart initialization.
Currently service isolate is initialized from the source code parsed by VM.
This CL changes it so service isolate created during Dart initialization
is created from the kernel platform.dill file if it is present in the application
bundle. Then this platform kernel file is kept in dart_init module and reused
for application sciprt isolates.
* Reformat and merge
* Use accessor method
* Avoid passing running_from_kernel param. Add TODO for cleanup. Rename param.
* Revert "Pass option to reuse an existing runtime controller (#4253)"
This reverts commit 57f361dc6abcd8441c444882c5e8bb99519532f3.
* Revert "Document the Porter-Duff operators. (#4250)"
This reverts commit b15fa9c64e1028187ec9592289919d77a6843802.
* Revert "Move channel and codec related files to common (#4225)"
This reverts commit 674a3f5869d3ae0e7ea3583a2d0942e184817903.
This reverts commit d43d35347594c893687dec7402a525d83f57db10.
Clear regression to the flutter_gallery_ios__transition_perf benchmarks:
* average_frame_build_time_millis: 1.4x
* missed_frame_build_budget_count: 1.2x
No regression seen for other iOS benchmarks (e.g. scrolling).
When checking whether render buffer size has changed, apply the device
scale factor used to convert points to pixels to ensure we're not
needlessly re-allocating render buffer storage.
Adds trace + debug log to make this easier to detect in future.
Originally introduced in 2d530daeca438ca24562d290616c0aec5ab76b88, but
rolled back in 0a6e415b60581ef7934032df20401da1fae9ea9f to test whether
it was responsible for regressions to
flutter_gallery_ios__transition_perf benchmarks:
* average_frame_build_time_millis
* missed_frame_build_budget_count
This reverts commit 2d530daeca438ca24562d290616c0aec5ab76b88 as it looks
like a potential cause of performance benchmark regressions on
https://flutter-dashboard.appspot.com/benchmarks.html:
- flutter_gallery_ios__transition_perf average_frame_build_time_millis
- flutter_gallery_ios__transition_perf missed_frame_build_budget_count
* Don't crash if there is only a single SemanticsNode
Previously, the code assumed that the root SemanticsNode will allways have a child. This is not true as can be seen in the hello_world example app, which would crash when a11y is turned on.
* review comment
By adding these events the new developer centric timeline will be able
to identify and correlate events related to the different stages of the
drawing pipeline.
Roll Dart to 62045a4590a333ae557f8f261a909ee75449cd70
