Maestro the Second

dev-master / 1.0.x-dev 2021-01-07 08:30 UTC

This package is auto-updated.

Last update: 2024-04-07 15:56:06 UTC



Maestro2 is a Repository Manager System for PHP.

Think of it like Ansible for repositories.

  • Perform upgrades on multiple packages.
  • Standardize configuration accross repositories.
  • Run CI
  • Conduct surveys (discover latest tags, CI status etc)

This project is a work-in-progress.

How It Works

Maestro is essentially a concurrent task runner which is conveniently adapted for working with repositories.

It is your job to create a pipeline class. This class will be instantiated and passed the configuration node, and it must return a task. This task can be an aggregate of many tasks.



This is the main configuration file, which can look something like:

    "core.inventory": [
    "core.templatePath": "example/templates",
    "core.workspacePath": "var",
    "core.concurrency": 10

The inventory files are where we define our repositories and variables:

    "vars": {
        "jobs": [
        "defaultBranch": "master"
    "repositories": [
            "name": "maestro",
            "url": "git@github.com:dantleech/maestro2",
            "vars": {
                "jobs": [
            "name": "worse-reflection",
            "url": "git@github.com:phpactor/worse-reflection"

The inventory files will be merged and cast into configuration nodes which can be used by pipelines.


Create a pipeline and ensure that it is autoloadable, for example:

// example/EmptyPipeline.php

namespace Maestro\Examples\Pipeline;

use Maestro\Core\Inventory\MainNode;
use Maestro\Core\Pipeline\Pipeline;
use Maestro\Core\Task\NullTask;
use Maestro\Core\Task\SequentialTask;
use Maestro\Core\Task\Task;

class EmptyPipeline implements Pipeline
    public function build(MainNode $mainNode): Task
        return new SequentialTask([
            new NullTask(),

You can then run it with:

$ ./vendor/bin/maestro run pipeline/EmptyPipeline.php

This does very little - in fact it will do nothing as all we did was specify a reporting group for subsequent tasks, followed by the NullTask.

For working examples see the example directory in this project.

There are many tasks which you can add (including aggregate tasks to run more tasks sequentially or in parallel). These can be found as implementations of Maestro\Core\Task\Task (and also found in the same namespace).


You will be able to read the documentation here.