* Do not close the text editing connection when an input text element is blurred (upon clicking another element on the page) on a desktop browser. keep the current behaviour for mobile browsers
* change the unit tests for blur
* refocus after blur so that the user can keep inputing text to the TextFormField
* skipping failing firefox check. active element didn't get updated in firefox in the automated test. manually checks working
A recent change to the analyzer package introduced an API incompatibility with
the build_resolvers package. The analyzer package needs to be pinned to an
older version until this is resolved.
SkImage references get held by both our Image and Picture objects. The GC has no idea about this, and so sometimes pictures look very small when they're actually holding a reference to a large chunk of memory. This patch helps make sure that the GC can more adequately catch the real size impact of Picture objects, and combined with an upstream patch in Dart allows for much more aggressive collection of unused image related memory.
The underlying _updateNode call requires that all parameters be set
non-null. There's a single call site in the framework in
lib/src/semantics/semantics.dart in SemanticsNode.updateWith(). At that
call site, all parameters are either asserted non-null in the
constructor of SemanticsData or defaulted to null, with the sole
exception of textDirection.
The ergonomics of this method are currently pretty ugly and we should
consider migrating most of the defaulting and assertions that we apply
at the call site up to the definition in dart:ui. That work is filed as
https://github.com/flutter/flutter/issues/57720.
* runs ios unit tests, if the simulator is already booted.
* address reviewer comments
* changing comments for the Safari desktop tests
* addressing reviewer comments. removing the timeout.
* change mobile browsers ver name for safari.dart
* removing then calback on browser exit, only leaving the exception callback
* try to run integration tests on luci
* adds extra logs to check if chrome is on lUCI
* use the cipd chrome package for integration tests in LUCI
* add driver version for chrome 83.(cipd package has 83 now)
* no headless mode didn't worked on the bots. try the other option
* On LUCI also use chrome driver as a cipd package
* change the directory name to fit the cpid package
* adding blacklist functionality
* remove logs added for troubleshooting. remove the check that blocks int-test runs on LUCI
* add more comments to blacklists
* also use CIPD package chrome for unit tests
* addressing reviewer comments
* run integration tests first, upon reviewer request
* fix bug. keep running integration tests after unit tests
* update the logs for LUCI
* fix todo comments
