search

sangria-graphql / sangria

Scala GraphQL implementation
https://sangria-graphql.github.io
Apache License 2.0
1.96k stars 226 forks source link

Moving forward with Sangria #446

Closed yanns closed 3 years ago

yanns commented 5 years ago

Please read the following before: https://github.com/sangria-graphql/sangria/issues/445

With great sadness, the Sangria community will have to organize itself and continue development without Oleg Ilyenko in the future. This thread exists to discuss how to move forward after we’ve processed the sad news. There is no reason to rush this, please take your time. We at commercetools, where Oleg has worked since 2013, are using Sangria extensively and are invested in its future.

Let's open the discussion on how to maintain Sangria in the future.

yasserkaddour commented 5 years ago

I am the only remaining member in sangria-graphql organization, unfortunately I only have write access for 2 repos https://github.com/sangria-graphql/sangria-playground and https://github.com/sangria-graphql/sangria-relay-playground . Your colleague @cneijenhuis told me he already contacted GitHub on how to regain full access to the organization and apparently it's a lengthy process. I would be happy to help appoint worthy maintainers for Sangria to continue the work, like Oleg would have liked.

yanns commented 5 years ago

For information, we are in contact with Github to check if we can gain write access to the whole organization and the related projects. I'll keep you informed.

yanns commented 5 years ago

I guess there are other parts we have to check:

Is there something else?

dave-handy commented 5 years ago

@yanns do you know anything about the "scaldi" org and projects? I don't think these are nearly as popular or active as sangria, but they do need some basic care and feeding to keep up with dependency and scala language updates. My (small) company is heavily invested in the scaldi core library, with great success, and might be willing to take on some of the responsibility.

Any tips? I am thinking about making a similar issue to this one in the scaldi repo and see if someone chimes in with helpful information. However I'm not expecting much in response since it is much less active.

I'm assuming Oleg developed scaldi during his tenure at your company, but you don't use it there any longer?

yanns commented 5 years ago

@dave-handy we are in the process of having all rights on all projects, included scaldi. This is the first step. In a second step, I'd like to make a call for people who are willing to maintain those, and give them enough rights to do so.

You can open a similar issue on scaldi. If you already volunteer to maintain the project, please mention it also.(it'd be very welcome ❤️ )

vmandrychenko commented 5 years ago

Hello everyone,

We (Garner) don't have the resources to maintain the sangria repo full time but we will certainly contribute with PRs, new features, bug fixes and an increase of the code coverage, especially if it aligns with our feature/bugfix road map.

paulpdaniels commented 5 years ago

I know the team at LeadIQ would very much like to contribute to Sangria's continued success. I personally would be interested in helping maintaining Sangria, but I'm not sure how much use I would be since I haven't ever really messed with the internals of the library itself.

Perhaps we could start by going down the list of top contributors and seeing if they would be interested in being a co-maintainer? Then we can start refining the contribution process from there?

anagrius commented 5 years ago

We Humio are also very invested in Sangria and would like to help out as possible.

jregistr commented 5 years ago

Not sure if this has ever been raised in the history of Sangria, but could Scala Center or Lightbend be the folks we are looking for? It's looking so far that even though we have many contributors, we still lack a someone with the resources to be maintainer. With Sangria being THE GraphQl implementation in Scala, it'd make sense for one of these large organizations to lead it. Or is that not how that works?

dave-handy commented 5 years ago

@yanns has there been any progress on getting the needed permissions? I see that there hasn't been any new commits in sangria master either.

yanns commented 5 years ago

Some news:

So we'll try that and keep you in touch. But all in all, github is not very cooperative on that and making everything complex.

dave-handy commented 5 years ago

Thanks for the update! Protenus may start a fork of scaldi if we decide to start our scala 2.13 process soon.

siderakis commented 5 years ago

I started hacking around with the idea of converting a schema created using Sangria to a GraphQL-Java schema. The end result would be similar to kotlin-graphql.

For me this seems like a fine option but I'm wondering what others think. I'm a little biased because I work on rejoiner.io which is also based on GraphQL-Java.

travisbrown commented 5 years ago

@yanns Is there anything any of us can do to help with this process?

yanns commented 5 years ago

@travisbrown If you know some intern at github, maybe this person could help our case. Otherwise I still have to take the time to prepare new papers with the family.

schrepfler commented 5 years ago

I’ve pinged Nat Friedman on Twitter if they can take a look. https://twitter.com/schrepfler/status/1159262893919145984?s=21

TJSomething commented 5 years ago

Would it make sense to do work on a temporary fork while the process is being worked out?

travisbrown commented 5 years ago

@TJSomething Publishing to Maven Central via Sonatype also depends on access to the GitHub organization (although it's possible that @yasserkaddour could get verified now via sangria-playground).

There is ongoing work happening in forks (I'm personally testing local builds for Scala 2.13 now, for example). I'm not a core maintainer, but my own feeling is that the overhead of switching to a temporary "official" fork would be substantial, and we should avoid it if we can, even if it means waiting a little longer for new official releases.

nickhudkins commented 5 years ago

Hi @yanns anything we can do to help out?

yanns commented 5 years ago

@nickhudkins not that I'm aware of, unless you know someone at Github who can help us here. I send them a document, and I'm still waiting for their answer...

nickhudkins commented 5 years ago

Thanks @yanns, I reached out to some folks on twitter in hopes of moving this along, and at least have a party willing to help. I am happy to do the leg work to help you out if you need anything, including just sending emails and talking to people :). Hopefully we can keep Oleg's work alive and well right here in the sangria-graphql org.

nickhudkins commented 5 years ago

We may have an answer sooner thanks to: https://twitter.com/TwitterOSS/status/1171190354642497537 ?

DEGoodmanWilson commented 5 years ago

@yanns GitHub employee here, although not one in a direct position to help. Can you shoot me an email DEGoodmanWilson@github.com ?

mjadczak commented 5 years ago

Given that this has now been a pending issue for nearly 5 months, with no resolution seemingly in sight, should there now be a serious conversation about how long we are prepared to wait for the bureaucratic machine to grind through this issue, and at what point setting up an "official" fork (temporary or otherwise) should be considered?

synedu commented 5 years ago

There is ongoing work happening in forks (I'm personally testing local builds for Scala 2.13 now, for example). I'm not a core maintainer, but my own feeling is that the overhead of switching to a temporary "official" fork would be substantial, and we should avoid it if we can, even if it means waiting a little longer for new official releases.

Do you still feel this way, @travisbrown ?

I agree with @mjadczak that the uncertainty needs to be addressed. Was there any update from github after the documentation they requested, @yanns ?

I think we all agree that sangria is an important and popular project in dire need of maintenance. The longer this issue remains unresolved, the more people will turn away from this library, to the detriment of the community.

@jregistr proposed new custodianship of sangria in June, although it doesn't appear that Scala Center or Lightbend have expressed an interest in doing so. It seems that the largest corporate user of sangria, and therefore the most resourced to maintain it, is twitter. Would it make sense to propose an official twitter fork and direct community efforts there? With twitter OSS behind it, the project would probably be more actively maintained and widely adopted.

yanns commented 5 years ago

There is ongoing work happening in forks (I'm personally testing local builds for Scala 2.13 now, for example). I'm not a core maintainer, but my own feeling is that the overhead of switching to a temporary "official" fork would be substantial, and we should avoid it if we can, even if it means waiting a little longer for new official releases.

Do you still feel this way, @travisbrown ?

I agree with @mjadczak that the uncertainty needs to be addressed. Was there any update from github after the documentation they requested, @yanns ?

Yes we provided the documentation they requested. After some time without news, I asked them about the status, and they answered that they need more documentation.

Github is not really helping here. Every time we provide something, they ask for something more. 😢

So we have to ask the mother, one more time... To be honest, we are feeling demotivated by the lack of any progress compared to the amount of time / energy we put into it.

If an organization or someone is ready to have an official fork, I'd also be ready to help.

synedu commented 5 years ago

Yes we provided the documentation they requested. After some time without news, I asked them about the status, and they answered that they need more documentation.

Github is not really helping here. Every time we provide something, they ask for something more. 😢

So we have to ask the mother, one more time... To be honest, we are feeling demotivated by the lack of any progress compared to the amount of time / energy we put into it.

This is really disappointing from github. It is not fair to expect a grieving mother to keep signing more documents. There must be some way we can send this up the chain.

Thank you for what must be draining work, @yanns .

travisbrown commented 5 years ago

@yasserkaddour Would you be willing to help with the Sonatype verification, so that the publication coordinates don't have to change, even if the GitHub organization does? We'd just need to create a file named OSSRH-48782 in the sangria-playground repo: https://github.com/sangria-graphql/sangria-playground/pull/20

travisbrown commented 5 years ago

I think we should consider moving to a new GitHub organization, at least temporarily. GitHub might be willing to work with us to recover an unused account name (I've had luck with this in the past, and it's been relatively fast—a couple of days, not months), so https://github.com/sangria might be an option, or something like sangria-scala or sangria-graphql-org (both of which are currently available).

If we're able to recover access to the org.sangria-graphql group ID, then we could publish from forks in this organization and there would only be some small overhead for maintainers, but none for adopters.

jregistr commented 5 years ago

Would it make sense to propose an official twitter fork and direct community efforts there? With twitter OSS behind it, the project would probably be more actively maintained and widely adopted.

I honestly agree with @bengraygh here. Has anyone here contacted anyone from or has connection to Twitter OSS? Finatra and all the other projects twitter is behind are well maintained so their involvement in this project would be amazing.

travisbrown commented 5 years ago

@jregistr Yes, we've heard from someone at Twitter:

Thanks for the heads up. We (Sangria-using friends at Twitter) are interested in figuring this out. I'll start some conversations.

…but that was in the context of working with GitHub to get access to the organization here.

I don't know much about how things currently work at Twitter OSS, but I was on the team a few years ago, and I think it's unlikely that Twitter would adopt the project in the sense you're talking about. I also don't think that's necessary, though—there are plenty of contributors and potential maintainers, the issue is just GitHub making this transfer difficult.

sachee commented 5 years ago

@travisbrown @yanns @jregistr Hey! This is Sasha from Twitter (I work on the team that works with GraphQL/Sangria). We would love to set something up and talk about what that general situation with the project is and see what we can do to help.

felipesilva commented 5 years ago

@sachee @travisbrown @yanns I sent you an email so we can get this going.

yanns commented 5 years ago

Status update:

we had a meeting with @felipesilva @sachee @travisbrown and others. It was great to see you!

We still would like to have access to the organization. -> I'm checking for the papers github is asking for. I'll update this issue if I have news.

@travisbrown opened a PR so that we could be allowed to push to sonatype. For the moment, only @yasserkaddour has the permissions to merge it.

travisbrown commented 5 years ago

Thanks for the update, @yanns! As I mentioned in the call, I'm also checking with Sonatype about alternative ways to get access to the org.sangria-graphql group ID, so that we don't have to change the Maven Central coordinates even if we do end up deciding (possibly temporarily) to use "official" forks on GitHub.

yasserkaddour commented 5 years ago

I merged the PR, sorry for the delay

travisbrown commented 5 years ago

@yasserkaddour Thank you! I've commented on the Sonatype issue with a link to the file.

yanns commented 5 years ago

Thx a lot @yasserkaddour!

travisbrown commented 5 years ago

Update

We're now able to publish to the org.sangria-graphql group ID on Maven Central, and we've created a new organization on GitHub that will serve as a temporary home for the official Sangria forks: https://github.com/sangria-graphql-org

This new organization is owned by @yanns, who also has the final sign-off on all PRs and releases. Yann is continuing to work on recovering admin access to the sangria-graphql GitHub organization, and when that happens we will move back here from sangria-graphql-org.

I'm currently in the process of forking the repositories in sangria-graphql to sangria-graphql-org. I'm prioritizing library repos today, but we can also fork the sandbox and example repos as needed.

We've also started publishing new releases introducing Scala 2.13 support from these forks. Right now only two are available, but we're hoping to get the rest of the projects listed here published over the next few days.

Note that while the GitHub repositories have changed, the artifact coordinates on Maven Central are the same as they've always been. If you previously had the following in your build:

libraryDependencies +=
  "org.sangria-graphql" %% "sangria-marshalling-api" % "1.0.3"

…all you need to do is bump the final digit.

Both of the currently available new releases are verified by MiMa to be binary-compatible with the previous versions, and there are no bug fixes or other substantive changes, so unless you need 2.13 support there's no reason to upgrade.

We don't know of a convenient automated way to move pull requests from sangria-graphql to sangria-graphql-org, since we don't have admin access to the former, so we're asking contributors to move their open pull requests manually.

Which organization new issues should be opened against is an open question. Right now we're prioritizing unblocking 2.13 users, and this is one of probably many details we haven't discussed yet. If anyone has thoughts please comment here.

travisbrown commented 5 years ago

We've now published new releases of nine of Sangria's modules with Scala 2.13 support, as well as a first 2.0.0 milestone release of Sangria itself:

I've published GitHub release notes for all ten repos, but some still need CHANGELOG and README update PRs.

There are eight other modules (listed here) that still need 2.13 PRs. I might not be able to get to these for a couple of weeks, so any help would be appreciated.

yanns commented 4 years ago

Small update: we still haven't been granted access by Github. We are still in the process.

yanns commented 4 years ago

I'm now an owner of https://github.com/sangria-graphql 🥳 We now have to replay all changes done on the fork https://github.com/sangria-graphql-org

zvuki commented 4 years ago

@yanns great news! Do you mind sharing the ownership with someone, you know, just in case? p.s. I wish you good health.

yanns commented 4 years ago

@yanns great news! Do you mind sharing the ownership with someone, you know, just in case? p.s. I wish you good health.

of course. We have to avoid being a single point of failure. And I'll need help maintaining all the repositories.

travisbrown commented 4 years ago

Wow, this is great news, @yanns!

yanns commented 4 years ago

update: I've replayed all changes made on the temp organization to this one

In the short term, I'd like to add at least one other owner to be on the safe side. Any volunteer?

travisbrown commented 4 years ago

@yanns I probably won't have much time to help with maintenance in the immediate future, but I'd be happy to join as a backup (and I already have org.sangria-graphql publishing rights on Sonatype).

yanns commented 4 years ago

@yanns I probably won't have much time to help with maintenance in the immediate future, but I'd be happy to join as a backup (and I already have org.sangria-graphql publishing rights on Sonatype).

Thanks you very much for proposing. I added you as backup. Now the project is quite safe from this perspective.

sachee commented 4 years ago

@yanns let me know if we (Twitter) can be of any help here. We have time set aside to help if you need it.

yanns commented 4 years ago

For the moment, I'm maintaining the sangria ecosystem with minimal time investment.

I'm not pushing sangria forward, like checking for new GraphQL features and implementing them. I don't spent much time on opened issues. I review PR as they come.

I will not be able to spend more time on it.

I think there are different kind of help I'd need:

I think checking the website is a top priority to avoid losing the domain.