#Rust sdk archive?
1 messages · Page 1 of 1 (latest)
Hi @hollow tendon. First of all I don't think the rate of changes necessarily correspond to whether a product is useful or not. The Rust SDK is a community lead effort though it is primarily me working on it, and as such I implement what I find useful. The rust SDK is more or less feature complete. The only thing that have changed within the last year or so is special requests from the dagger team to fix potential breaking changes. Otherwise we've been updated to the latest engine version, although I haven't been great at cutting new releases for them.
Dagger module support is also a thing but if I am not mistaken none of the community lead sdks support it atm. I considered building it out, but as my stack is still built around the previous way of doing things I haven't invested the time.
Secondly it has been difficult on boarding other contributors as the dagger stuff has quite a bit of context around it, so we haven't been changing a whole lot as the project during operations just works, and I don't see that changing unless dagger modules become a requirement or the full fledged sdks are gone and done with.
We've haven't gotten any other maintainers ready to pick it up, unless the dagger team themselves want to take a more active role.
The project is still useful for me and others so I don't currently see the need to archive it, so long as there isn't a general consensus that it isn't worth the effort
Yup, I agree with @lilac dock - just because it's stable and doesn't need constant updates doesn't mean it needs to be archived.
We do need to do some work to integrate rust into the release process - that way we can keep doing SDK releases along with the others, that's been sitting on my backlog to properly hack together for the last couple weeks, but there's a lot of other stuff on it as well that's taking longer than expected.
Yep, I've been meaning to do the same thing @jolly root but I've also had other priorities. I guess because the Rust SDK isn't in the main release process, at some point the automatic engine updates stopped, without us catching it.
Based on everything happening in this Discord and with Dagger releases over the last 6 months, Dagger has made a pretty bold play on Modules and the Daggerverse only exists with modules.
While there's still use in the old way, the days are numbered
I've seen this at InfluxDB and Pulumi during my time there, SDKs for languages that fall behind don't just frustrated users; they become a burden on the community and the core team
I'm not saying Rust is that, yet, but if the plan isn't to find a way to keep it up to date then it becomes a net negative for the community
Thanks for the insights @hollow tendon. I think we need a way to not just make the community sdks more visible but also carve out a way to make them available as dagger modules. @jolly root as far as I recall it is still only possible to use the 3 official ones right? I had a look at it a few months ago but the situation might've changed
ah sorry, just saw this 😢
so currently we have module support in go/typescript/python (at mostly roughly the same level of support)
and also some support for elixir and php (i don't think fully support everything but do a pretty good job)