The haskell admin team manages the server https://github.com/haskell-infra/haskell-admins and also controls releases.
There are no significant active contributors at this time (though a fair amount of work was done over last summer during the last gsoc, all of which was merged), and deploying new releases is a careful process done rarely (to ensure stabiility). I know the HF has long term goals of attempting to direct more resources in this directiion.
I will note that it is a bit misleading to point to open issues and PRs and not closed/merged ones as well. I.e. things are not just “piling up” but also (slowly) they do get resolved!
As always with open source, the best thing one can do is familiarize oneself with the codebase, pick up tickets and try to make contributions. Additionally, people with time to dedicate and ops experience/capacity can volunteer to play a role in helping the admin group (but this is a position that entails some long-term commitment and responsibility – ops unfortunately is not an arena suited to “drive by” contributions).
@sclv, thanks for sharing the docs with that info. One question I was unable to find an answer to: is there code for the infrastructure itself? If there is no code for that, are there docs about the infrastructure provisioning and hosting details?
Various pieces of the infra live in repos here https://github.com/haskell-infra
However, a lot of this is “just debian servers with some configs” and those configs are kept private. There is some further private and public documentation, but a fair amount is out of date. It is expected the HF will help systematize this to some degree. The details of the providers of infra and hosting are listed at the bottom of the page.
What can be done for the existing PRs to become merged?
I just checked and there was only one merge-ready PR, which I just merged. The big remaining PR that’s close is one I’ve been working on and then stalled out on at the test phase: https://github.com/haskell/hackage-server/pull/904
So testing that change on a sufficient subset of packages to make sure it doesn’t break anything would be very useful.
Outside of that, everything else is a stalled WIP as far as I know, or only still around for reference, and could probably be tagged a bit better to indicate that.
There are lots of issues that people can pick up to work on, though! The tags haven’t been updated lately, so its good to check which issues are still relevant or important, but the “good first issue” and “help wanted” ones can give some guidance.