Re-sync libstd dependencies with backtrace

[glandium]

3ec9d8d bumped addr2line to 0.22 for some reason, while backtrace uses 0.21.

[rustbot]

r? @Mark-Simulacrum

rustbot has assigned @Mark-Simulacrum.
They will have a look at your PR within the next two weeks and either review your PR or reassign to another reviewer.

Use r? to explicitly pick a reviewer

[rustbot]

These commits modify the Cargo.lock file. Unintentional changes to Cargo.lock can be introduced when switching branches and rebasing PRs.

If this was unintentional then you should revert the changes before this PR is merged.
Otherwise, you can ignore this comment.

[glandium]

Ideally, I'd like this fixed on beta too.

[workingjubilee]

? huh, sorry about making that mistake.

[workingjubilee]

wait, what?

[workingjubilee]

@glandium https://github.com/rust-lang/backtrace-rs/blob/72265bea210891ae47bbe6d4f17b493ef0606619/Cargo.toml pardon?

[glandium]

@glandium https://github.com/rust-lang/backtrace-rs/blob/72265bea210891ae47bbe6d4f17b493ef0606619/Cargo.toml pardon?

that's 0.3.73, you upgraded to 0.3.72.

[glandium]

Actually, Cargo.lock says 0.3.71.
https://github.com/rust-lang/rust/blob/master/Cargo.lock#L308
It looks like #125636 didn't land in full. Even if it did, though, that would still have brought two gimlis because other crates use gimli 0.28 (unwinding and thorin-dwp)

[workingjubilee]

https://docs.rs/crate/backtrace/0.3.72/source/Cargo.toml ?

[workingjubilee]

@glandium That's a version of backtrace used in color-eyre which is used in the compiler and the test suite, independently of std, it has nothing to do with the dependencies of std.

[workingjubilee]

I realize this result may be confusing... it certainly annoys me a bit... but I don't intend to block our release cadence of libstd dependency updates on making sure that all the dependencies of the compiler and its tools, which are compiled separately, perfectly agree, unless it is absolutely imperative that they do. This "for some reason" was because backtrace is transcluded as a module into libstd, thus meaning that the Cargo.lock is utterly irrelevant for it per se because it is then compiled as a module and not a separate crate!

rust/library/std/src/lib.rs

Lines 670 to 672 in ebcb862

	#[path = "../../backtrace/src/lib.rs"]
	#[allow(dead_code, unused_attributes, fuzzy_provenance_casts)]
	mod backtrace_rs;

And if you think this is a silly state of affairs, I agree!

[glandium]

Oh boy, I'm in between a rock and a hard place aren't I. Well, this mess breaks forward ports of rust-lang/cargo#8834 and #78790 ... :(

[workingjubilee]

...Hmm, that sucks!

@glandium I will try to see if this can get settled in rust 1.81.0 / backtrace 0.3.74, can you open an issue describing the problem in more detail?

[glandium]

well, libstd does get two gimlis for target_os = "xous" via unwind depending on unwinding which depends on gimli 0.28. I guess that's really the only problem I have.

[workingjubilee]

I opened nbdd0121/unwinding#30

[glandium]

Oh boy, I'm in between a rock and a hard place aren't I. Well, this mess breaks forward ports of rust-lang/cargo#8834 and #78790 ... :(

Well, since this isn't the first time this happens (last time was duplicated cfg-if), I just went ahead and finally figured out what was missing to make those patches work with multiple crates with the same name but different versions.