State of ruby ecosystem in conda-forge

[h-vetinari]

No-one has been touching the ruby ecosystem for a while. There's an open migration to ruby 3.x that started over half a year ago and has immediately run into blockers that have gone unresolved since then.

Ruby <3.0 is also incompatible with OpenSSL 3, which has been our default since January.

Ruby itself has a bunch of downloads, but it seems the other packages are pretty dead.

What - if anything - are we going to do about this? @conda-forge/ruby

CC @conda-forge/core

[traversaro]

Personally I am interested in ruby itself as it is a dependency for some packages I mantain (mainly https://github.com/conda-forge/libsdformat-feedstock/ and https://github.com/conda-forge/gz-tools2-feedstock). While in the long term the dependency will be dropped (see gazebosim/sdformat#274 and gazebosim/gz-tools#7) in the short/medium term I need to depend on it, so if necessary I will be happy to mantain the ruby feedstock. On the other hand, I am not particularly interested in any gem/other ruby-related packages.

Based on conda-forge/ruby-feedstock#122, perhaps @lmrodriguezr could be interested in this discussion?

[wolfv]

I think it would be great to have ruby itself well supported in conda-forge. I am not sure if we have the man-power / interest for the packages from the ruby ecosystem though (although I ran into some compilation problems when using some ruby packages that needed C extensions in the past ... )

[lmrodriguezr]

Indeed, I'd be interested, thanks for tagging @traversaro !

I would agree with @wolfv , ruby itself is probably the only mission-critical part. Any packages could be installed on top of it with not a lot of effort. I've also encountered some issues installing native packages in the past, but recent forge versions seem to be pretty stable

[h-vetinari]

Hey @jdmarques, I see you've got a couple of recent ruby PRs in staged recipes:

• add rb-pdf-core staged-recipes#28685
• add rb-ttfunk staged-recipes#28686
• add rb-polyglot staged-recipes#28744
• add rb-matrix staged-recipes#28746
• add rb-treetop staged-recipes#29182
• add rb-tilt package staged-recipes#29185

Perhaps you want to join the ruby team here and help out on the affected feedstocks? We've been stuck for lack of resources (= interested contributors) for a while on essentially all ruby-related feedstocks (except ruby itself).

[jdmarques]

Hello @h-vetinari, I do have interest in getting some more packages in for Ruby v3.x. Where/How can I help?

P.S.: Full disclosure I'm not a ruby developer or am involved in the eco-system. I'm just a python dev with some dependencies on ruby packages.

[h-vetinari]

Where/How can I help?

Thanks for being willing to help! What you could do is take any of the feedstocks from the long-dead migration, and open a PR that tries to move them to ruby 3.x, ideally by adding

ruby:
  - "3.4"

to recipe/conda_build_config.yaml. Also, ideally you'd add yourself as a maintainer on those feedstocks (note that many of them are probably severely out of date compared to the latest available version, and you can update that too if it helps). Once they're green we can then merge, and once there's a couple of them that are revived, we can try for a proper migration to 3.4

[jdmarques]

and once there's a couple of them that are revived, we can try for a proper migration to 3.4

just see if we can have an alignment of expectations does it have to be the packages in that list? I'm asking because I have an active interest in getting some packages in conda-forge working with ruby 3.4, but I think most of them are not in that list. e.g.

• rb-addressable
• rb-afm
• rb-prawn

My point is if by adding the packages I need it would help in that migration?

[h-vetinari]

New packages can be added subject to the normal review process in staged-recipes, and I'm not trying to gate the packages you want on anything. You don't have to take over maintenance of the existing feedstocks, I'm just apprehensive about adding more ruby packages while we can't even manage to maintain the existing ones (the core of this issue). It would be good to revive feedstocks worth reviving, and archive the ones where no-one's interested in taking over maintenance, and any help in that direction would be much appreciated. 🙃