`DartDevelopmentServiceLauncher` was created to share the DDS launch
logic from flutter_tools with other Dart tooling.
---------
Co-authored-by: Andrew Kolos <andrewrkolos@gmail.com>
Catches gradle error and throws a helpful error message that indicates
an incompatability between Java and AGP versions and how to fix the
issue.
Related issue:
[128524](https://github.com/flutter/flutter/issues/128524)
## Pre-launch Checklist
- [x] I read the [Contributor Guide] and followed the process outlined
there for submitting PRs.
- [x] I read the [Tree Hygiene] wiki page, which explains my
responsibilities.
- [x] I read and followed the [Flutter Style Guide], including [Features
we expect every widget to implement].
- [x] I signed the [CLA].
- [x] I listed at least one issue that this PR fixes in the description
above.
- [x] I updated/added relevant documentation (doc comments with `///`).
- [x] I added new tests to check the change I am making, or this PR is
[test-exempt].
- [x] I followed the [breaking change policy] and added [Data Driven
Fixes] where supported.
- [ ] All existing and new tests are passing.
Native libraries that are contributed by native asset builders can depend on each other. For macOS and iOS, native libraries are repackaged into Frameworks, which renders install names that have been written into dependent libraries invalid.
With this change, a mapping between old and new install names is maintained, and install names in dependent libraries are rewritten as a final step.
Related to https://github.com/dart-lang/native/issues/190
The choice screen is irrelevant when debugging apps locally. `flutter run` creates a separate user profile for testing only. It doesn't touch users' browser settings.
Fixes https://github.com/flutter/flutter/issues/153928
Fixes https://github.com/flutter/flutter/issues/153972 (unless the cause of https://github.com/flutter/flutter/issues/153064#issuecomment-2305662791 happens to also prevent this fix from working).
In this PR, I've looked for all non-test call sites of `ChromeConnection.getTabs` and made sure are all wrapped in `try` blocks that handle `IOException` (`HttpException` is what we see in crash reporting, but I figure any `IOException` might as well be the same for all intents and purposes).
I plan on cherry-picking this the stable branch.
Currently, if creating a symlink on Windows fails due to `ERROR_ACCESS_DENIED`, you'll get an error message like:
```
Error: ERROR_ACCESS_DENIED file system exception thrown while trying to create a symlink from source to dest
```
The `source` and `dest` paths are incorrect.
This will help us debug: https://github.com/flutter/flutter/issues/153758
As of Xcode 16, App Store validation now requires that apps uploaded to the App store bundle dSYM debug information bundles for each Framework they embed.
dSYM bundles are packaged in the FlutterMacOS.xcframework shipped in the `darwin-x64-release` tools archive as of engine patches:
* https://github.com/flutter/engine/pull/54696
This copies the FlutterMacOS.framework.dSYM bundle from the tools cache to the build outputs produced by `flutter build macOS`.
Fixes: https://github.com/flutter/flutter/issues/153879
Re-lands https://github.com/flutter/flutter/pull/136880, fixes https://github.com/flutter/flutter/issues/136879.
Additions to/things that are different from the original PR:
- Adds an entry to `gradle_errors.dart` that tells people when they run into the R8 bug because of using AGP 7.3.0 (https://issuetracker.google.com/issues/242308990).
- Previous PR moved templates off of AGP 7.3.0.
- Packages repo has been moved off AGP 7.3.0 (https://github.com/flutter/packages/pull/7432).
Also, unrelatedly:
- Deletes an entry in `gradle_errors.dart` that informed people to build with `--no-shrink`. This flag [doesn't do anything](https://github.com/flutter/website/pull/11022#issuecomment-2297294421), so it can't be the solution to any error.
- Uniquely lowers the priority of the `incompatibleKotlinVersionHandler`. This is necessary because the ordering of the errors doesn't fully determine the priority of which handler we decide to use, but also the order of the log lines. The kotlin error lines often print before the other error lines, so putting it last in the list of handlers isn't sufficient to lower it to be the lowest priority handler.
I am making an assumption `OutputMode.none` should _really_ mean
`OutputMode.failuresOnly`, that is, if we ever get a non-zero exit code,
we still want to know why. If I've somehow misunderstood that, LMK and
I'm happy to revert this PR or make adjustments.
This fixes the bug where if you were to do:
```sh
git clone https://github.com/myuser/fork-of-flutter
cd fork-of-flutter
./bin/flutter update-packages
```
You now get:
1. An actual error message, versus no output at all.
2. A warning that a common reason is not tracking a remote, with
instructions to fix it.
Closes https://github.com/flutter/flutter/issues/148569.
Reverts: flutter/flutter#152487
Initiated by: gmackall
Reason for reverting: I forgot that I need to override the compileSdkVersion in the AGP 8.0 [instance of this test](ef9cd32f5a/dev/devicelab/bin/tasks/android_java17_dependency_smoke_tests.dart (L19))
Original PR Author: gmackall
Reviewed By: {reidbaker}
This change reverts the following previous change:
Updates `compileSdk`, `targetSdk`, and `ndk` versions (former 2 to latest, latter to the version of the ndk we are hosting on CIPD).
Summary of changes:
- Updates mentioned template values
- `compileSdk` 35 requires AGP 8.0+, so updated to 8.1 in many places.
- This also necessitated Gradle upgrades in most places
- This also necessitated moving the `package` xml attribute to the AGP `namespace` field in a couple places (test + template).
- Some tests use the output of `flutter create` but then use intentionally lower AGP versions. [I downgraded the `compileSdk` in these tests.](fee34fd61a)
- [Stopped lockfile generation](82324a2570) script from hitting the `hello_world` example because it uses `.kts` gradle files.
- One test needed [some Gradle options we had already added to templates](6aa187b4b6).
Updates `compileSdk`, `targetSdk`, and `ndk` versions (former 2 to latest, latter to the version of the ndk we are hosting on CIPD).
Summary of changes:
- Updates mentioned template values
- `compileSdk` 35 requires AGP 8.0+, so updated to 8.1 in many places.
- This also necessitated Gradle upgrades in most places
- This also necessitated moving the `package` xml attribute to the AGP `namespace` field in a couple places (test + template).
- Some tests use the output of `flutter create` but then use intentionally lower AGP versions. [I downgraded the `compileSdk` in these tests.](fee34fd61a)
- [Stopped lockfile generation](82324a2570) script from hitting the `hello_world` example because it uses `.kts` gradle files.
- One test needed [some Gradle options we had already added to templates](6aa187b4b6).
The stream subscriptions in the device's VmService are used by other parts of FlutterVmService and other components throughout flutter_tools. Components that listen to streams should not call VmService.streamCancel because that will interfere with other users who still want the events.
See https://github.com/flutter/flutter/issues/153049
See https://github.com/flutter/flutter/issues/153563
The Flutter tool has a bug where removing the last Flutter plugin does not correctly update the CocoaPods integration.
This adds a test to ensure that the generated Swift package is properly updated when the last Flutter plugin is removed.
See: https://github.com/flutter/flutter/issues/11819#issuecomment-2289782626
When DevTools is active HotRunner.attach makes a deferred RPC to a DevTools service. That RPC appears to cause issues if it executes at the wrong time during the hot restart.
Passing the --no-devtools flag works around this by disabling the RPC.
See https://github.com/flutter/flutter/issues/153049
The test should not assume that a stream of bytes delivers them in a
certain chunked order (e.g. all bytes of a line are delivered in one
chunk and the newline comes in another chunk).
Instead it should use a line splitter to obtain individual lines and can
match those against expectations.
This fragile test is broken after [0] which combines writing a string
together with the newline instead of seperately.
[0] https://dart-review.googlesource.com/c/sdk/+/378024
Hide the `--web-renderer` option in the Flutter Tool. The defaults already cover all fully supported modes:
- `flutter build web` provides canvaskit + dart2js
- `flutter build web --wasm` provides skwasm + dart2wasm
We do not want to encourage production usage of any other permutations (e.g. `auto` or `html`), in particular those that simply do not work (e.g. `skwasm` + dart2js).
Fixes https://github.com/flutter/flutter/issues/140096
Fixes https://github.com/flutter/flutter/issues/151786
Reverts: flutter/flutter#152358
Initiated by: zanderso
Reason for reverting: Speculative revert to determine whether this PR is related to https://github.com/flutter/flutter/issues/153026
Original PR Author: andrewkolos
Reviewed By: {christopherfujino}
This change reverts the following previous change:
Contributes to fixing https://github.com/flutter/flutter/issues/137184.
Cleaned up version of earlier PR, https://github.com/flutter/flutter/pull/152187.
This PR guards all the writes to `Process::stdin` by wrapping them with `ProcessUtils.writelnToStdinUnsafe`. This way, if any writes fail, we should at least get a stacktrace in our crash reporting.
Opening a Swift package in Xcode generates `.build` and `.swiftpm` directories. These should be ignored as they contain intermediary build artifacts that aren't meant to be checked-in.
Part of https://github.com/flutter/flutter/issues/148018
Nullable types for values in map patterns require the key to be present.
Since the 'uri' key is not always present in DDS exception responses,
this was causing us to fall back to throwing a StateError.
Fixes https://github.com/flutter/flutter/issues/152684
Sets up tests that verify we can build a fresh counter app across our Gradle/AGP/Kotlin support range.
Post submit only, because the suite takes ~30 minutes to run, and I expect it to be _somewhat_ rare that we break only one of these versions (and therefore it doesn't get caught by existing presubmits).
