Le 10/11/2018 à 21:12, Santiago Vila a écrit :
> On Thu, 8 Nov 2018, Sylvestre Ledru wrote:
>
>> Le 08/11/2018 à 21:30, jnq...@gmail.com a écrit :
>>> Package: llvm-7
>>> Version: 1:7.0.1~+rc2-1
>>> Severity: grave
>>>
>>> I've just updated my Sid install and found that building Rust crates
>>> with Cargo now fails with a seg fault.
>>>
>>> Initially I fired a bug report at cargo to kick things off, but I've
>>> now discovered that it relates to the llvm-7 update, as switching llvm7
>>> packages back to testing versions fixes the problem.
>> Do you have more info than "it segfaults"?
> I believe the reported problem is the same that makes all these
> recently uploaded rust packages to FTBFS:
>
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-arc-swap.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-encoding.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-encode-unicode.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-crossbeam-epoch-0.5.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-gobject-sys.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-grep-regex.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-heck.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-iso8601.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-mio-uds.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-rustc-version.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-term.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-stringprep.html
> https://tests.reproducible-builds.org/debian/rb-pkg/unstable/amd64/rust-string-cache-codegen.html

Yeah, this is related to the move of llvm toolchain to be compiled using
clang (stage2).

Reverting llvm to the previous version works.

If someone could help debugging the segfault, this would be great :)

Thanks
S

Reply via email to