Command-line tool for managing a release of the Flutter SDK.
Some basic requirements to conduct a release are:
The main entrypoint for the conductor is bin/conductor. For brevity, the rest of this document will assume that this entrypoint is on the shell path.
All available commands can be seen via:
conductor help
Releases are initialized with the start
sub-command, like:
conductor start \ --candidate-branch=flutter-2.2-candidate.10 \ --release-channel=beta \ --framework-mirror=git@github.com:flutter-contributor/flutter.git \ --engine-mirror=git@github.com:flutter-contributor/engine.git \ --engine-cherrypicks=72114dafe28c8700f1d5d629c6ae9d34172ba395 \ --framework-cherrypicks=a3e66b396746f6581b2b7efd1b0d0f0074215128,d8d853436206e86f416236b930e97779b143a100 \ --dart-revision=4511eb2a779a612d9d6b2012123575013e0aef12
For more details on these command line arguments, see conductor help start
. This command will write to disk a state file that will persist until the release is completed. To see the current status of the release (at any time), issue the command:
conductor status
Once initializing the release, the conductor tool will issue instructions for manual steps that must be executed by the user. At any time these instructions can be seen via conductor status
. Once these manual steps have been completed, you can proceed to the next step by using the command:
conductor next
Upon successful completion of the release, the following command will remove the persistent state file:
conductor clean