Good, that will assist with the builder stuff, although I'm not convinced that it goes far enough. In all honesty, I would wish Mercurial away (and Python before it), but I believe in Codereview unconditionally :-)
I'll try to get a builder going (plan9/386) in my environment, but I expect it will not be very representative (old, too slow and running on VMware ESXi 3.5).
I'd like to know how the community proposes to deal with each fork of the Plan 9 codebase. For my part, I think it's up to us to hide all the fractures from the Go developers: multiple architectures, poorly supported, is already a problem, adding kernel forks would really put the knife in it. On the other hand, it's a bit of that "he who pays the fiddle calls the tune" refrain: "he who hosts the builder, makes the rules?"
Lucio.