fix: allow steps.get in world-postgres to handle undefined runId

[eersnington]

World-postgres steps.get was always expecting runId, so passing undefined returned 404.
This patch makes it fall back to stepId-only lookups, like the other backends (world-local).

Closes #134

[changeset-bot]

🦋 Changeset detected

Latest commit: 9f2461b

The changes in this PR will be included in the next version bump.

This PR includes changesets to release 1 package

Name	Type
@workflow/world-postgres	Patch

Not sure what this means? Click here to learn what changesets are.

Click here if you're a maintainer who wants to add another changeset to this PR

[vercel]

@eersnington is attempting to deploy a commit to the Vercel Labs Team on Vercel.

A member of the Team first needs to authorize it.

[vercel]

The latest updates on your projects. Learn more about Vercel for GitHub.

Project	Deployment	Preview	Comments	Updated (UTC)
example-nextjs-workflow-turbopack	Ready	Preview	Comment	Oct 30, 2025 8:01pm
example-nextjs-workflow-webpack	Ready	Preview	Comment	Oct 30, 2025 8:01pm
example-workflow	Ready	Preview	Comment	Oct 30, 2025 8:01pm
workbench-nitro-workflow	Ready	Preview	Comment	Oct 30, 2025 8:01pm

[pranaygp]

@eersnington awesome, thank you for the fix! Any chance we can also add a test that would ensure this behaviour? We could probabaly do it in the world-testing package, or in the PG world directly?

[eersnington]

@eersnington awesome, thank you for the fix! Any chance we can also add a test that would ensure this behaviour? We could probabaly do it in the world-testing package, or in the PG world directly?

thank you! I saw that local world has tests for this. I'm leaning towards world-testing since it would catch this across all storage implementations per the interface contract.

But, one problem that I can think of is world-testing shouldn't be too tightly coupled to backend specific details. Maybe a generic contract test + specific edge cases in each world package? (I haven't gone through all the tests in local world, but I think that can act as a starting point)

Would love to hear your thoughts and the team's on this. I'd be happy to take it on too!

[pranaygp]

@eersnington world-testing makes sense but I guess to get this PR through without too much friction, I'd suggest copying that same test over from local for now and shipping it a a world-postgres test

The world spec itself needs a bunch of work - we're discussing this as a team right now to

I think the plan for world-testing is to have a good suite of tests, including but not limited e2e workflow tests, and should include testing backend implementation. For example, things like "You cannot cancel a run that's already complete" and "If you mark a step as completed, the completedAt needs to be updated", etc

The creator of worlds is God Gal (aka @Schniz) - so he'll have more to share on those plans soon. We would love your help on it :)

[VaguelySerious]

Are we good to merge this?

[eersnington]

Are we good to merge this?

@VaguelySerious Hey Peter, I've added better tests (my previous commit, I mocked drizzle client which sucked). Now it run against the docker-compose postgres container. Can you review the current ones and point out anything you'd change? I'll add more tests based on the one written by Gal in local-world/storage.test.ts. Rn, there's tests for createRunsStorage, and createStepsStorage, and there's none for createEventsStorage, and createHooksStorage

[VaguelySerious]

I think this is already really good and we can merge as-is. Later, when we use world-testing for this too and have more comprehensive tests there, it'll likely replace the storage tests anyway.

[pranaygp]

Sorry we needed to revert this because of a license change @eersnington

Can you please reopen the PR on the latest main on this repo? tl;dr - we've changed from MIT to Apache with a DCO. #191

When making a new PR, you'll need to signoff the commit using the -s flag in git commit 🙏

Apart from that, the PR itself looks great. Happy to answer any questions on the licensing if you're curious

[eersnington]

Sorry we needed to revert this because of a license change @eersnington

Can you please reopen the PR on the latest main on this repo? tl;dr - we've changed from MIT to Apache with a DCO. #191

When making a new PR, you'll need to signoff the commit using the -s flag in git commit 🙏

Apart from that, the PR itself looks great. Happy to answer any questions on the licensing if you're curious

Thank you @VaguelySerious!

I've just finished writing all the tests locally for storage.ts basing it of Gal's tests for Embedded world and found 1 problem too within the storage contract for PG World through the tests.

^ Would definitely be worth it to write comprehensive integration tests to find these details across all world implementations

Sorry we needed to revert this because of a license change @eersnington

Can you please reopen the PR on the latest main on this repo? tl;dr - we've changed from MIT to Apache with a DCO. #191

When making a new PR, you'll need to signoff the commit using the -s flag in git commit 🙏

Apart from that, the PR itself looks great. Happy to answer any questions on the licensing if you're curious

No worries @pranaygp and thank you for letting me work on this 🫡
And I'll add sign off by on any new of my commits too