Switch to GitHub

If the goal of #26728 is reduce friction for new contributors, moving primary contribution flows (PRs + issues) to GitHub is a direct, high-leverage change because it aligns GHC with where (a) the Haskell ecosystem already collaborates and (b) most potential contributors already have accounts and muscle memory.

1. We should meet contributors where they already are

   • GitHub is already home of most of all haskell packages and related projects like ghc-proposals, the core library committee and many others
   • Most software engineers already have GitHub accounts, ssh keys configured, and familiarity with PR review norms. For small, first-time patches, lowering activation energy is often the difference between a contribution landing vs. never being attempted.

2. Reliability

   The GHC GitLab is down very often (last time yesterday)

3. Opportunity cost: GHC time shouldn’t go to commodity hosting

   Right now some of the limited GHC developers time is spent on maintaining, upgrading and hosting GitLab. We can get all this for free by GitHub. Focussing on improving GHC instead of working on hosting related stuff.

4. Other major compiler run on GitHub

   Rust, nodejs, LLVM - they all run on GitHub.

--

If we want GHC to remain healthy long-term, we need to optimize for the contributors we could have, not just the ones we already do.