Git Infrastructure Alternatives Keep Multiplying
Two separate threads flagged frustration with existing git hosting and tooling. The 'I regret migrating to Codeberg' post detailed problems with Codeberg's terms of service evolution and the community dynamics of free software hosting. A separate post covered Game of Trees, a git-compatible version control system from the OpenBSD team, positioned as a simpler and more auditable alternative to git itself. Buz, a fork of Bun using modern Zig for sub-1-second incremental builds, also showed up as a quality-over-features bet against the existing dominant tool.
The pattern: there is a persistent low-level dissatisfaction with the current git tooling stack, from hosting (GitHub, Codeberg) to the git protocol itself. None of these alternatives are threatening GitHub's dominance in the near term. But the energy going into them reflects genuine pain: git is complex, hosting platforms have governance risks, and the tooling layer is fragile.
The Buz thread had a realistic counterpoint: without an ecosystem, a single maintainer rewriting Bun in clean Zig is unlikely to sustain long-term. The project lead said they would not accept human-coded contributions until they deemed the project ready, which commenters noted is a fast path to burnout.
So what?
For founders, the Codeberg situation is a reminder that free hosting with ideological governance strings attached carries real platform risk. If your open source project lives on a host with evolving content policies, have a documented migration plan. For tooling builders, the appetite for quality-focused alternatives to bloated defaults is real, but ecosystem lock-in is a moat that good code alone does not break.