Guess you won't get much far with a build system which has such limited
portability. But it's Google, I'm not surprised anything outside their
own usecase doesn't really work.
I obviously can't speak for the upstream Bazel team, or for their corporate sponsor, but I can give you my perspective on this. Bazel is a relatively new build system and it is still in very heavy development. It makes sense to me that the team would focus their initial efforts on popular 64-bit architectures (e.g. amd64 and arm64) for which they have sufficient in-house expertise. The beauty of FLOSS is that people with expertise in different architectures can help enable Bazel to build, and run, correctly there. That is why I specifically reached out to the RISC-V experts here instead of just excluding the architecture from the list.
I'm personally very impressed with what Bazel can do even now, and I think it has the potential to be very useful in many FLOSS projects, including distributions such as Debian. I'm happy that Google chose to Open Source the code instead of keeping it as an internal tool. That's a trend I would like to see more of. So, I'm asking you to please help honor the steps that the upstream Bazel team has taken to make their software universally available by helping in the areas where neither they (nor the Debian packagers) have sufficient experience.
Thanks in advance!
-Olek