On Tue, Dec 12, 2023 at 4:51 PM Alex Kiernan via lists.openembedded.org <alex.kiernan=gmail....@lists.openembedded.org> wrote: > > On Tue, Dec 12, 2023 at 4:27 PM Randy MacLeod via > lists.openembedded.org > <randy.macleod=windriver....@lists.openembedded.org> wrote: > > > > On 2023-12-12 11:25 a.m., Randy MacLeod wrote: > > > > On 2023-12-12 5:41 a.m., Alex Kiernan via lists.openembedded.org wrote: > > > > On Sat, Dec 9, 2023 at 2:06 PM Alex Kiernan via lists.openembedded.org > > <alex.kiernan=gmail....@lists.openembedded.org> wrote: > > > > On Sat, Dec 9, 2023 at 12:18 PM Alexandre Belloni > > <alexandre.bell...@bootlin.com> wrote: > > > > Hello Alex, > > > > rust 1.74.0 is not reproducible: > > https://autobuilder.yocto.io/pub/repro-fail/oe-reproducible-20231208-rfbproj4/packages/diff-html/ > > > > Thanks, I think :) > > > > One observation, which might be useful, these failures are all inside > > the same module src/llvm-project/compiler-rt/lib/builtins which are > > just a bunch of intrinsics. And as far as I can see it looks like it's > > just the object name mangling that's not reproducible. > > > > FYI: > > > > In the weekly YP tech call, we talked about Rust and Richard said that he'd > > merge > > 1.71 to master since it's progress and it unblocks some ptest works that > > Tim was working on. > > > > We can then just jump directly from 1.71 to 1.74 if ( when! ) we can make > > that reproducible. > > > > Or even to rust-1.74.1: > > > > https://blog.rust-lang.org/2023/12/07/Rust-1.74.1.html > > > > Yeah, I did 1.74.1 earlier today > (https://github.com/akiernan/poky/commit/f408e74e37ac8cd6d0e4d9715317f69a039ad50f) > - can post a patch for that if we want to give it a whirl? There's > definitely some reproducibility fixes in there (e.g. > https://github.com/rust-lang/rust/issues/112586) but I'm not convinced > any of them are what we're seeing (which appear to really be in pieces > of LLVM). >
I just realised that because we now have two consecutive versions, diffing the 1.71.1...1.72.0 trees was feasible... for the llvm-project submodule, the changes are tiny (10 files, mostly RISCV specific or tests, nothing that looks even slightly like the issue we're seeing). So I suspect wherever the problem comes from its not actually inside here. -- Alex Kiernan
-=-=-=-=-=-=-=-=-=-=-=- Links: You receive all messages sent to this group. View/Reply Online (#192240): https://lists.openembedded.org/g/openembedded-core/message/192240 Mute This Topic: https://lists.openembedded.org/mt/103055169/21656 Group Owner: openembedded-core+ow...@lists.openembedded.org Unsubscribe: https://lists.openembedded.org/g/openembedded-core/unsub [arch...@mail-archive.com] -=-=-=-=-=-=-=-=-=-=-=-
