Dev command leaves containers running when process exits with an error

[TylerAldrich]

When we hit this code:

try {
    await compose_process;
} catch (ex) {
    if (!is_exiting) {
        throw ex;
    }
}

it was possible for the compose_process to end (with an error that we throw) - but the process finishing with an error doesn't guarantee that containers get cleaned up. One container may hard-fail, the process ends, and other containers remain running.

To fix that, we now run:

    try {
      await this.compose_process;
    } catch (ex) {
      if (!this.is_exiting) {
        // Always call `docker compose stop` here - the process died so there's nothing to kill,
        // need to call `docker compose stop` to clean up any remaining running containers.
        console.error(chalk.red(ex));
        console.log(chalk.red('\nError detected - Stopping running containers...'));
        await this.stop();
      }
    }

to call docker compose stop after the compose up process ends to ensure that containers are stopped.

I also moved all the code that manages the docker compose up process into it's own class to make the logic in this file more easily understandable.

• The Dev command class is still handling most of the work
• The UpProcessManager now starts/registers the process interrupts/stops the long-running docker compose up process and when it's finished running, the Dev.run can continue cleanup.

Tested on

• Mac
• WSL2
• (Windows) Powershell
• (Windows) Cmd

[github-actions]

🎉 This PR is included in version 1.25.0 🎉

The release is available on:

• v1.25.0
• npm package (@latest dist-tag)
• GitHub release

Your semantic-release bot 📦🚀