Our current top crasher is an unclear error when ProcessManager fails to resolve an executable path. To fix this, we'd like to being adjusting the process resolution logic and adding more instrumentation to track failures. In order to begin the process, the ProcessManager has been folded back into the flutter tool
= gets escaped into %3D which seems to be tripping up cmake on windows since % is a control character. Switch to base64 encoding, since this does not have % nor , in the output character set.
This change is not trivially cherry pickable, and isn't tested on windows aside from my local, manual tests due to the planned CI work not being complete yet.
Fixes#75017Fixes#74705
Overhaul of flutter drive in order to deliver a better experience, namely:
* flutter run and flutter drive now share more flags, so code paths that were previously only testable on run are now testable on drive.
* Removes web-initialize-platform as this is no longer used
* flutter drive correctly sets up a logger that shows native exceptions, by connecting to the vm service.
* VM service connection now provides access to memory info without launching devtools (only for debug/profile mode)
Web changes
* Passes on the one test in the repo, otherwise the webdriver code has been isolated as much as possible
Additional NNBD related bug fixes:
No longer passes --enable-experiment to the test script. (FYI @blasten ). earlier we might have assumed that the flutter gallery benchmarks would be migrated along side the app and flutter driver, but only the app under test needs to be migrated. The test scripts should never be run with the experiment.
Overhaul of flutter drive in order to deliver a better experience, namely:
flutter run and flutter drive now share more flags, so code paths that were previously only testable on run are now testable on drive.
Removes web-initialize-platform as this is no longer used
flutter drive correctly sets up a logger that shows native exceptions, by connecting to the vm service.
VM service connection now provides access to memory info without launching devtools (only for debug/profile mode)
Web changes
Passes on the one test in the repo, otherwise the webdriver code has been isolated as much as possible
Additional NNBD related bug fixes:
No longer passes --enable-experiment to the test script. (FYI @blasten ). earlier we might have assumed that the flutter gallery benchmarks would be migrated along side the app and flutter driver, but only the app under test needs to be migrated. The test scripts should never be run with the experiment.
Remove devicelab specific code for shutting down gradle daemon, add --android-gradle-daemon option to build/run/drive`. Avoids need for un-tested devicelab specific handler. There are also some feature requests for this, so 2 birds one stone.
Example:
flutter build apk --no-android-gradle-daemon will pass --no-daemon on to gradle
Remove devicelab specific code for shutting down gradle daemon, add --android-gradle-daemon option to build/run/drive`. Avoids need for un-tested devicelab specific handler. There are also some feature requests for this, so 2 birds one stone.
Example:
flutter build apk --no-android-gradle-daemon will pass --no-daemon on to gradle
Android Studio 4.1 moved the location of the .home file on Windows which is used to located the install directory. This functionality is important because it is how we locate and discover the Android SDK functionality, as well as the appropriate JRE.
fixes#67986
Work towards #39925
Currently flutter run will uninstall and reinstall if the initial install fails and the APK was previously installed. Allow drive to share this same logic by moving it into installApp and out of startApp.
This should reduce the occurrence of the error in the devicelab.
Cleans up some undocumented classes and re-organizes the AndroidDevices class to avoid the need for the static testing only member. Adds a script for tracking globals.
Refactors the desktop devices and workflow to remove unnecessary usage of global variables. This should make it easier to test and continue enhancing the desktop functionality of the tooling
#47161
In cases where the Intellij/AS plugins are not located, display links to where they can be downloaded but do not surface an error. This should generally reduce confusion about whether the plugins are required for every installed IDE. For example, frequently users may only install AS so that they can install the Android SDK - or they may have multiple copies of Intellij installed.
For example: #66762
This changes the name of the generated output files on android in case a flavor with uppercase letters is used.
Previously, the lowercased flavor name would be used for the apk/aab file. Now, the flavor name is used as-is.
The flutter tool is currently unable to detect missing permissions in gradle/gradle.bat that would cause a gradle build to fail via process exception. Rather than crashing and exiting, we can display the exception as an error message and tool exit.
While linux/macOS are able to add the +x bit, this is not possible on windows with our current file system/OS API. These crashes represent a substantial amount of crash reporting, but are otherwise not actionable on our end.
* Reland "Re-enable the Dart Development Service (DDS) (#64671)"
This reverts commit 2ae25cc2d7.
* Fix MDNS building Observatory URI with port 0 instead of forwarding the device port
* Added MDNS test
This change re-enables DDS and outputs the DDS URI in place of the VM
service URI on the console. If --disable-dds is not provided,
--host-vmservice-port will be used to determine the port for DDS rather
than the host port for the VM service, which will instead be randomly
chosen.
Avoid creating AndroidDevice discovery if the SDK cannot be located. Previously the tool would use which/where adb, however this required us to handle the AndroidSdk class being potentially null - which required an additional layer of indirection around all access. Sometimes these were forgotten leading to NPEs.
In general, not much can be done with an Android Device if the actual SDK is not installed.
Reland with fixed code + tests for null SDK + adb in AndroidDeviceDiscovery
Avoid creating AndroidDevice discovery if the SDK cannot be located. Previously the tool would use which/where adb, however this required us to handle the AndroidSdk class being potentially null - which required an additional layer of indirection around all access. Sometimes these were forgotten leading to NPEs.
In general, not much can be done with an Android Device if the actual SDK is not installed.
Adds support for size analysis on iOS, macOS, linux, and Windows - using an uncompressed directory based approach. The output format is not currently specified.
Adds support for size analysis on android on windows, switching to package:archive
Updates the console format to display as a tree, allowing longer paths. Increases the number of dart libraries shown (to avoid only ever printing the flutter/dart:ui libraries, which dominate the size)
Attempt to simplify the Android SDK interface ahead of refactoring it. The locateAndroidSdk static method is called at startup to locate the android SDK, returning null if it cannot be found. These helper methods attempted to first look up the AndroidSDK if it was already null - which could only cover the case where someone installed the Android SDK while flutter was running (possibly through an IDE)
Add feature flags for android, ios, and fuchsia (on by default). After updating the g3 rollers, the fuchsia feature will be turned off by default. Creates a simpler base type of feature flags for g3 to extend.
Updates android, ios, fuchsia workflows to use feature flags check.
Removes concept of stable artifacts and checks on flutter version.
Fixes#58999#52859#12768
The global packages path could cause tests to fail when it would be overriden to unexpected (in test setup) values. Remove most usage and make it a configuration on buildInfo, along with most other build information. Cleanup the asset builder to require the .packages path and the resident runners to no longer require it, since they already have the information in build_info.
It needs to stick around for the fuchsia deps we do not control.
Filled #60232 for remaining work.
instead of restricting profile/release mode based on whether the tool thinks the device is an emulator, restrict based on the device target architecture and the requested build mode. Notably, this enables release mode on x86_64 Android emulators, but not x86 emulators since we do not support that as an AOT target.
This does not add release mode support for simulators, since this requires us to build and upload artifacts for simulator/x86_64
Also combines experiments into extraGenSnapshot/ExtraFrontEndOptions. Allows providing --no-sound-null-safety to allow out of order migration and running.
Ensure iOS and android builds can be correctly cached. Use the performance-measurement-file to verify that all targets were skipped on the second invocation. This is only run on the flutter_gallery build.
Support bundling SkSL shaders into an android APK or appbundle via the --bundle-sksl-path command line options. If provided, these are validated for platform engine revision and then placed in flutter_assets/io.flutter.shaders.json
The resident runner does not check if the ApplicationPackage is null when trying to stop the app. Update AndroidDevice.stopApp to handle this case by returning false.
The package will be null when flutter attach is used.
This will allow experimenting with the remove to string transformer before we're ready to turn it on by default. This doesn't work for web yet since we use dart2js instead of the frontend_server for producing kernel
When the AndroidDevice checks for the current API version, it isn't handling a null response. While we can refuse to run on API levels that are unsupported, we should probably assume that a null response = lowest possible API level and allow these to run.
Previously the AdbLogReader did async setup in the StreamController.onListen callback, specifically it would query the api version and start the adb process. If the log subscription was cancelled before this setup completed, then the log output could (haven't confirmed) get added to a closed controller, causing the above state error.
* Update project.pbxproj files to say Flutter rather than Chromium
Also, the templates now have an empty organization so that we don't cause people to give their apps a Flutter copyright.
* Update the copyright notice checker to require a standard notice on all files
* Update copyrights on Dart files. (This was a mechanical commit.)
* Fix weird license headers on Dart files that deviate from our conventions; relicense Shrine.
Some were already marked "The Flutter Authors", not clear why. Their
dates have been normalized. Some were missing the blank line after the
license. Some were randomly different in trivial ways for no apparent
reason (e.g. missing the trailing period).
* Clean up the copyrights in non-Dart files. (Manual edits.)
Also, make sure templates don't have copyrights.
* Fix some more ORGANIZATIONNAMEs
* Generate projects using the new Android embedding
* Add comment about usesNewEmbedding:true
* Feedback
* Rework way to detect new embedding in new apps
