minor: Update dependency minor versions

[lf-]

bors r+
🤖

[lf-]

damn, my git-sin script had a bug in it 🤣
bors r+

[lnicola]

Are these intentional?

[lf-]

cargo-update generated them.

[lf-]

OOps

[lf-]

bors r-

[bors]

Canceled.

[lnicola]

bors r-

cargo-update generated them.

Ugh, that's strange. Do you have any local Cargo.toml changes?

And in the meanwhile, can you pin libc to 0.2.99 in crates/rust-analyzer/Cargo.toml (I hope that works)? I wanted to open a PR, but then I saw yours. It's needed because newer versions don't build on FreeBSD.

[lf-]

bors r-

cargo-update generated them.

Ugh, that's strange. Do you have any local Cargo.toml changes?

And in the meanwhile, can you pin libc to 0.2.99 in crates/rust-analyzer/Cargo.toml (I hope that works)? I wanted to open a PR, but then I saw yours. It's needed because newer versions don't build on FreeBSD.

Done.

[lnicola]

bors r+

[bors]

Build succeeded:

• Rust (macos-latest)
• Rust (ubuntu-latest)
• Rust (windows-latest)
• TypeScript (ubuntu-latest)
• TypeScript (windows-latest)

[matklad]

FWIW, I've extended my gbors script to give an ooportunity to prep a PR description out of concatenated commit messages: https://github.com/matklad/config/blob/8ccb77cba70ad052ce121c4cfd9a57c8ee39dd93/tool/src/gbors.rs

[lf-]

FWIW, I've extended my gbors script to give an ooportunity to prep a PR description out of concatenated commit messages: https://github.com/matklad/config/blob/8ccb77cba70ad052ce121c4cfd9a57c8ee39dd93/tool/src/gbors.rs

Ah yes, so I can own myself harder and faster 🤣

[matklad]

Yeah; julia has known startup latency problems (known as TTFP -- time to first plot), and, after careful benchmarking and taking Amdahl's law into consideration, I came to the conclusion that gbors performance is the main limiting factor for my productivity, hence the Rust rewrite.

[bjorn3]

Why not make gbors a self-compiling rust script? :) https://github.com/bjorn3/rustc_codegen_cranelift/blob/master/y.rs#L1-L8

[matklad]

That's lovely! I do want to use dependencies though (xshell and anyhow), and I want to have a busy-box style single binary which dispatches based on argv[0].

Also, I wonder why we don't have an RFC for rustc run main.rs? That feels like a feature we should have...

[matklad]

(btw, the stuff is sort-of self-linking at least: https://github.com/matklad/config/blob/7371bb2e043ce37be872e25f0871baacba033747/tool/src/main.rs#L30)