Project Abandoned?

[brandonkal]

I'd really like to see this project succeed. However, it appears that the project has been abandoned.

[stephank]

It's not! It's just difficult to move things forward. I think it's low on everyone's list of priorities? But this is a very good point to raise!

We use Portier at Angry Bytes extensively, self-hosted broker, clients using Node.js and PHP libraries, and it's been rock solid. We don't really need anything more from it, except keeping those components up to date, which is why I get a little less time to invest in Portier at Angry Bytes.

Even if there was more time to spend, I'm not sure what the goal is, really?

I'd personally like to give a shot at Rust async sometime, and use the broker codebase to do that. But that doesn't really drive the project forward.

The more important thing I want is to update the public broker at broker.portier.io, but I don't know how to plan or communicate that to people.

I think communication is probably the most important thing for the project, really. The issues on this repo and on the broker repo all need attention, but for example, changes that affect users are blocked if we can't communicate them. Or another example, packaging is something we can technically solve, but that sort of thing also needs to be published on the site. (Maybe bad communication is also what brought you here? 🙂)

We don't really have good communication channels, and I don't think anyone has given those things thought. I think we used to have google groups? If so, I don't know about their status. Whatever we use, it should probably be documented in this repo. (And probably have low-barrier to entry not just for users, but also maintainers, considering the current time available.)

(EDIT: Oh man, the google group is on the front page. I never pay attention to it, and it looks like some spam has gotten through. 🙈 No clue who has access to that group, though.)

Any way, that's what's on my mind whenever I think about the project.

[onli]

Similar situation for me: I use portier in projects of mine, it works and I have thus not much reason to work on it. But I definitely would, for example creating a new client library when moving to a new stack with a new project, which I'm actually considering.

Though we could definitely move more developers and maintainers deciding to dedicate work for the project.

@stephank Btw communication, I tried to get in touch with you on gitter but did not receive a response. Let's get in touch there or via mail to move the blockers out of the way. I have a bit more time on my hand now/soon to really act on what needs to be done.

[brandonkal]

It is good to hear this hasn't been abandoned. The reason I opened this is because a lot of issues have been open for years and seem to have gone stale.

I'd like to see the spec finished which hasn't been touched for a year. Diagrams would be a nice value-add as well.

Other wish list items:

1. An example of an integration with Dex.
2. An example docker-compose or kubernetes deployment with a sample client app would also help.
3. Better discovery: It should be possible to use MX records to determine if an email provider supplies an OpenID Connect flow. i.e. Google Apps domains shouldn't need to be manually configured.
4. A project logo and branding

[Boscop]

Is portier abandoned now, or still working?

Would you recommend switching to it?

If not, what is the alternative? :)

[stephank]

I actively maintain it and use it for work. I keep on top of dependencies and security issues, but I'm not actively working on new features.

So if the current functionality is acceptable, then it's perfectly fine to use, I'd say.