regression?: llvm version error

[Mark-Simulacrum]

• https://crater-reports.s3.amazonaws.com/beta-1.51-1/beta-2021-02-14/gh/moparisthebest.sendxmpp-rs/log.txt
• https://crater-reports.s3.amazonaws.com/beta-1.51-1/beta-2021-02-14/reg/sendxmpp-2.0.0/log.txt

potentially related, but if not good to file another issue:

• https://crater-reports.s3.amazonaws.com/beta-1.51-1/beta-2021-02-14/gh/Hoverbear.rust-rosetta/log.txt

[jyn514]

This looks like https://docs.rs/rustc_version/0.3.2/rustc_version/enum.Error.html#variant.LlvmVersionError (note that the crate has a dependency on rustc_version 0.3.2). It was fixed in 0.3.3, so I think this can be closed: djc/rustc-version-rs@v0.3.2...v0.3.3.

[jyn514]

potentially related

[INFO] [stderr] thread 'rustc' panicked at 'called `Result::unwrap()` on an `Err` value: FromUtf8Error { bytes: [128, 0, 2, 8, 0, 12, 5, 4], error: Utf8Error { valid_up_to: 0, error_len: Some(1) } }', compiler/rustc_codegen_llvm/src/metadata.rs:75:22

This looks like a different issue.

[apiraino]

removing the prioritize label @rustbot label -I-prioritize

[apiraino]

Seems that the master branch of this crate compiles on stable, beta and nightly, I guess the crate used in the crater run is not updated to the latest version (or there's no more recent release of xmpp-rs).

Therefore I would close this issue, what do you think @Mark-Simulacrum ?

Perhaps worth mentioning that xmpp-rs fails to compile on nightly+beta (on stable is fine) for another reason, a linting warning on another dependency (xml5ever):

Compiling xml5ever v0.16.1

   Compiling xml5ever v0.16.1
error: field is never read: `ns`
   --> tokio-xmpp/src/xmpp_codec.rs:183:5
    |
183 |     ns: Option<String>,
    |     ^^^^^^^^^^^^^^^^^^
    |
note: the lint level is defined here
   --> tokio-xmpp/src/lib.rs:3:22
    |
3   | #![deny(unsafe_code, unused, missing_docs, bare_trait_objects)]
    |                      ^^^^^^
    = note: `#[deny(dead_code)]` implied by `#[deny(unused)]`

error: aborting due to previous error

error: could not compile `tokio-xmpp`

worth opening an issue?

[jyn514]

Perhaps worth mentioning that xmpp-rs fails to compile on nightly+beta (on stable is fine) for another reason, a linting warning on another dependency (xml4ever):

Is crater running without --cap-lints warn? That seems bad. Anyway, I don't think making lints more accurate should ever be considered a breaking change, the author opted-in to the breakage with deny(unused).

[Mark-Simulacrum]

crater usually caps lints at forbid, as we want to know if we've added regressions in lints too. This seems reasonable to me. (I might even want --force-lints = deny or something).

But yes it sounds like this is not a rust regression at least.