Various improvements (in particular a new painting fuzzer) to the text manual test.
Some additional documentation.
A fix to Stack to remove an LTR bias: make unpositioned children apply "alignment".
Some more debugging information on RichText and Text.
A fix to the flutter tool to not crash when an RPC call throws an exception.
Previously, the usage analytics would generate a write to the user's
HOME directory during the Fuchsia build. We're tightening down the
environment in which we run the Fuchsia build, and these writes are now
more obvious.
This patch removes the usage analytics during the Fuchsia build,
avoiding the write to the user's HOME directory.
* fix updrade script; upgrade to the latest package versions
* exclude special dependencies from transitive closure
* fix stack trace handling in flutter_test due to stack_trace change
* change type on _emptyStackTrace
The libimobiledevice suite of tools do not include version information.
A simple way to verify they meet our version requirements is to run
idevice_id -l, which will fail when older versions are run against iOS
devices with newer versions of iOS installed.
Unfortunately, idevice_id -l will also fail when libimobiledevice is up
to date, but the attached devices have never been paired with the host
machine in Xcode.
This patch updates the error message to help guide the user in such
situations.
On iOS 11.0.x ideviceinfo fails with the following message
ERROR: Could not connect to lockdownd, error code -3
This workaround should also work for #12330
This patch also makes it possible to create FLX files with no extension.
Previously, the zip tool would add the ".zip" extension if the output file
lacked an extension.
* Restructure hot mode test so it runs interactively.
This allows to add a benchmark for hot reload after actual source code change.
* Add curly braces, refactory copyRecursive
This reverts commit 5e7bcbacf80875730b2002973b5d82cf5331c2a3.
`flutter run --benchmark` was triggering a different quick bailout path in the VM than `flutter run`. The failure has been fixed upstream.
* Fix restart flow for preview-dart-2 mode.
Restart in preview-dart-2 needs to use kernel file and it has to be complete, rather than incremental kernel file.
* Add curly braces
* Do full compile on restart
* Roll engine to pick up changes to hot reload for preview-dart-2
The initial loading happens on the host, which was building a script snapshot and allowing imports of dart:mirrors. Hot reload happens on the device, which then notices the imports and issues a compile-time error. This change causes programs with imports of dart:mirrors to be rejected during the initial load.
Fixes https://github.com/flutter/flutter/issues/12440
* Add hotReloadInitialDevFSSyncMilliseconds to track how long user have to wait before being able to do first reload.
This stat is significantly different between existing and preview-dart-2 setting (for the latter, this stat is ~3x slower: 13s vs ~4s).
* Remove ws
* Cleanup timer-related code
This reverts commit 90028813a89a3de8154144e6e0f1edbe90dc2e4f.
This change caused a few bots to fail with 'JSON-RPC error 110: Extension error', which is odd because _getUnusedChangesInLastReload is not an extension.
This CL introduces 2 hidden options to 'flutter build aot' and 'flutter run' for passing arbitrary arguments to front-end server and to gen_snapshot tool when building and running flutter app in --profile or --release modes.
The ability to pass arbitrary options simplifies various experiments, as it removes the need to change defaults and rebuild flutter engine for every tested configuration.
This class lives in the Context and allows callers to "inject"
flag values, where flag values are first extracted from the
command arguments, then from the global arguments as a fallback.
This makes command validation happen as part of `verifyThenRunCommand()`,
using a newly introduced protected method (`validateCommand()`) rather than
a `commandValidator` property (that subclasses were responsible for manually
invoking).
