Michal Rostecki
Michal Rostecki
> QA issues first: > > * You will need to add a `profiles/desc/` entry for the new USE_EXPAND (and socialise it via the mailing list, you can reference this...
@Kangie I rebased the branch and addressed all your remarks except > I think that RUST_SYSROOT works better for the USE_EXPAND name (i.e. rust_sysroot_bpf, rust_sysroot_wasm) I briefly looked at all...
@mgorny @Kangie I fixed it, it works also for Rust 1.88.0 and produces a correct sysroot: ``` >>> /usr/lib/rust/1.88.0/lib/rustlib/bpfel-unknown-none/ >>> /usr/lib/rust/1.88.0/lib/rustlib/bpfel-unknown-none/lib/ >>> /usr/lib/rust/1.88.0/lib/rustlib/bpfel-unknown-none/lib/libcore-f1f74068811ae06b.rlib >>> /usr/lib/rust/1.88.0/lib/rustlib/bpfel-unknown-none/lib/libcompiler_builtins-c799abf5d063ce62.rlib >>> /usr/lib/rust/1.88.0/lib/rustlib/bpfel-unknown-none/lib/liballoc-7c553e4fb8184643.rlib >>> /usr/lib/rust/1.88.0/lib/rustlib/bpfel-unknown-none/lib/librustc_std_workspace_core-cf15e4fcfd49cf5d.rlib ```...
@mgorny @Kangie ping. I couldn't do a revbump of 1.89.0, because it requires creating a branch in https://gitweb.gentoo.org/proj/rust-patches.git/, which I have no rights to do.
I'm afk today, but I could rebase tomorrow if needed. Thanks for approval!
I see. I can't reproduce this error locally, perhaps it's either: * too old clang on Github runners * compiler-rt not being installed * compiler-rt not being picked up for...
First of all, I'm sorry for the late follow-up. @targos > It seems unlikely. We found recently that newer V8 versions require it for `__atomic_compare_exchange ` Just for clarity `__atomic_compare_exchange`...
@targos @thesamesam First of all, sorry for very late follow up! I went ahead with adding a configure option, which also performs a check whether compiler-rt really supports atomics. I...
Closing this PR. Detecting whether compiler-rt can provide all atomics works, but bloats the gyp script. We decided that the way to go is to fix provide an empty stub...
I'm hesitant towards merging it, because: * I don't like the fact that it modifies `c_void` enums in two separate steps (first in `visit_mdnode` and then in `fix_di_compile_units`). That makes...