What is new in patchlevel/event-sourcing in version 3.7

We are happy to announce the release of our php library patchlevel/event-sourcing in version 3.7.0. This release contains a small feature for the InMemorySubscriptionStore, a performance optimization for the subscriptions and a bigger new feature which enable to create small or micro aggregates. So let's get started.

InMemorySubscriptionStore

Next, is a small DX feature regarding testing. The InMemorySubscriptionStore has now a new method clear to reset the internal subscription state. This means that after calling InMemorySubscriptionStore::clear all information on the subscriptions like position and state are removed.

Subscriptions: New Failed Status

Subscriptions can now have a new status called failed. This indicates, that a subscription could not overcome the error state with the retry strategy chosen. This happens for example when the max retries where reached. Subscriptions with status failed are ignored by the subscription engine, which leads to better performance since the subscription engine can now load less events based of the position of the remaining subscriptions.

This new status is technically a BC-Break as we added a new case in our Status enum. We decided to ship this nevertheless in a minor version since it should be really rare that this enum was used by our users. On the other hand this results in a better performance and therefore a better experience when using the subscription system.

If you want to know about the subscription statuses you should read our docs for the subscriptions.

New Stream Attribute to enable Micro Aggregates

Now to the biggest change in this version, a new attribute called #[Stream]. With this we can, when using the StreamStore, configure in which stream the events should be saved and loaded from. Previously, the stream was structured based on the aggregate as follows: {aggregateName}-{aggregateId}. Now, using the new attribute you can decide freely which stream should be used. Also {id} is a placeholder that can be used for the aggregate ID when defining the stream.

use Patchlevel\EventSourcing\Aggregate\BasicAggregateRoot;
use Patchlevel\EventSourcing\Attribute\Aggregate;
use Patchlevel\EventSourcing\Attribute\Stream;

#[Aggregate('profile')]
#[Stream('profile-{id}')]
final class Profile extends BasicAggregateRoot
{
    // ...
}

You can also use the same stream as another aggregate. This would cause both aggregates to read and write to the same stream. The ability to specify the stream of another aggregate is added of better DX by providing the FQCN of the aggregate.

use Patchlevel\EventSourcing\Aggregate\BasicAggregateRoot;
use Patchlevel\EventSourcing\Attribute\Aggregate;

#[Aggregate('meeting')]
final class Meeting extends BasicAggregateRoot
{
    // ...
}

use Patchlevel\EventSourcing\Aggregate\BasicAggregateRoot;
use Patchlevel\EventSourcing\Attribute\Aggregate;

#[Aggregate('guest_list')]
#[Stream(Meeting::class)]
final class GuestList extends BasicAggregateRoot
{
    // ...
}

This feature is a step directed to smaller or even Micro Aggregates and for Dynamic Consistency Boundaries. This feature comes from input we got after giving talks at a symfony user group at cologne. For more information about this topic you can have a look at our docs and this GitHub issue where we listed some resources. We are also planning a dedicated blog post about this topic, so stay tuned!