On Fri, Oct 21, 2022 at 10:51:59AM +0200, Ludovic Courtès wrote: > Hello, > > Efraim Flashner skribis: > > > I'm not sure there is a bug report, I didn't see it either. It looks > > like when I bumped rust-bootstrap from 1.39 to 1.54 we lost aarch64 > > support. I've bumped mrustc on staging and successfully performed a > > qemu-binfmt build of rust-bootstrap for aarch64 on my x86_64 box. I was > > then able to use that to build rust-1.55 on actual aarch64 hardware, so > > I assume it's good, I just don't have the hardware to build > > rust-bootstrap for aarch64 natively. > > So the presumed fix involves bumping rust-bootstrap from 1.54 to 1.55, > is that right? Not quite. We keep rust-bootstrap at 1.54, but we bump the mrustc commit we use from the v0.10 tag to its current master (about 50 commits). Among the commits there is one that doesn't get aarch64 to try to spit out assembly, (or illegal assembly or something) and then we just have to build out again. > That means we’ll have to rebuild on all architectures. This is > happening here: > > https://ci.guix.gnu.org/eval/739823 > > Could you monitor it? If things go well, can we aim for a merge by next > Thursday? > > Thanks, > Ludo’. rust-1.60 built just fine on x86_64, and it looks like of the ~6700 packages built 567 failed, with another ~3400 waiting to be built. I also haven't seen any movement on aarch64 in the build farm. -- Efraim Flashner אפרים פלשנר GPG key = A28B F40C 3E55 1372 662D 14F7 41AA E7DC CA3D 8351 Confidentiality cannot be guaranteed on emails sent or received unencrypted