Ian Jackson

Thanks! Yes, this worked for me. First, using my existing rustup, but with that rune, and a troublesome symlink at `.rustup/downloads`: ``` rustcargo@zealot:~$ RUSTUP_UPDATE_ROOT='https://dev-static.rust-lang.org/rustup' rustup update info: syncing channel updates...

> I lean towards giving users control over this rather than setting policy for how programs are tested. I agree with the idea of giving users control. I don't think...

FTAOD this implements, but only for `Clone`, these proposals from #16 * "Always using `match`, unifying enum and struct code" * "Always using braced syntax" I haven't tackled "Similarity between...

Oh, rustfmt. I will rewrite the branch.

"CI / Test 1.60 on ubuntu-latest () (pull_request) Failing after 20s" CI failure seems to be unrelated to MR contents. IMO the root cause is not committing a suitable MSRV-compatible...

@magiclen do you like the rough shape of this? If so I can proceed with making similar changes to the other impls.

Our ticket https://gitlab.torproject.org/tpo/core/arti/-/issues/1507

Hi. Thanks for looking. Maybe we can blunder through the forest together ? :-) I ran such a test in our CI (using the gitlab feature to set an env...

We have discovered that cached data (`target/` directories) in our CI would sometimes make this job succeed, when it would otherwise fail. Apparently not everything about the inputs is captured...

I am very late to this party, but: Has consideration been given to using the file access time (on Unix at least) to assist with avoiding deleting things that older...