Hopefully it'll be easier to build now. The downside is that it allocates a small vec on each finish, but it's much faster than rust-sha1 anyway, at least on ARM (where I'm running rqbit myself).