DNSControl vs Obsidian

DNSControl is mid-way through a v5 refactor being shipped as a visible candidate series — rc1 through rc4 in roughly two weeks, all of them explicitly promising zero user-facing change. The work replaces how DNS record fields are stored internally with a dnsv2.RDATA type, with legacy structs converted bidirectionally so existing providers keep working untouched. Two other threads run alongside: migrating the miekg/dns dependency from GitHub to Codeberg, now complete and no longer dual-imported, and removing old fields to cut memory use. The release notes read as direct appeals to users — naming nine providers nobody has tested and asking for volunteers, and inviting anyone to a monthly community call.

The stated goals are about future velocity rather than present features: making new DNS record types faster to adopt and custom record types easier to implement. That is a bet that the bottleneck on supporting new record types was the internal representation, and it is being paid for with a long candidate series rather than a big-bang release. The provider conversion finished during this window — rc3 declares all providers converted with long-tail work remaining — so the refactor is past its riskiest stretch.

The maintainer states v5.x is on schedule to ship before September, and the untested-provider list is the visible blocker; expect further candidates focused on provider coverage rather than new goals.