rust 1.66.0

[chenrui333]

Created with brew bump-formula-pr.

resource blocks may require updates.

[chenrui333]

linux build looks good 👍

[github-actions]

This pull request has been automatically marked as stale because it has not had recent activity. It will be closed if no further activity occurs. To keep this pull request open, add a help wanted or in progress label.

[cho-m]

ARM failures

Rust failures

• gnirehtet - Probably broken since rust 1.64 - gnirehtet: update cargo.lock to fix build with rust 1.64.0 #117139
• kapacitor - Probably broken since rust 1.64. Upstream needs to update to flux >= 0.187.0 to get influxdata/flux@1ab653c - kapacitor: upgrade flux to 0.191.0 #118950
• rover - Fixed in apollographql/rover@f41e089 but not in release yet. - rover: add build patch to build with rust 1.66.0 #118952

Python poetry-backend failures

• aws-sso-util - Python aws-error-utils -> poetry (build backend) -> rapidfuzz -> cmake -> (FAIL) - aws-sso-util: fix build without dependency on cmake or rust #118910
• cyral-gimme-db-token - Python poetry (build backend) -> rapidfuzz -> cmake -> (FAIL) - cyral-gimme-db-token: patch to use latest poetry_core #118940
• theharvester - Python dnspython -> poetry (build backend) -> rapidfuzz -> cmake -> (FAIL). New version should help theharvester 4.2.0 #115839

Other failures

• dotenv-linter - checksum - dotenv-linter: update checksum #118941
• pyflow - upstream API down - pyflow: update test #118942
• lexicon - test failure - lexicon: update dependencies #118944
• mdcat - https://codeberg.org/flausch redirects to https://codeberg.org/swsnr. https://codeberg.org/swsnr/mdcat is archived and has a link to https://github.com/swsnr/mdcat - mdcat 0.30.3 #118903

[cho-m]

Intel-only failures (to be updated)

Ventura

• sslyze - some linking failure to OpenSSL
• watchman - probably the same SIMD difference when building bottle vs build-from-source

[chenrui333]

I just tested gnirehtet is indeed broken with the functionality for 13-arm

$ /opt/homebrew/Cellar/gnirehtet/2.5_1/bin/gnirehtet relay -p 8000
2022-12-25 20:52:43.547 INFO Main: Starting relay server on port 8000...
2022-12-25 20:52:43.548 ERROR Main: Execution error: IO error: Address family not supported by protocol family (os error 47)

[chenrui333]

accidentally clicked the close button.

[chenrui333]

previous workflow run, https://github.com/Homebrew/homebrew-core/actions/runs/3770623497

[chenrui333]

Most of the issues have been addressed, I think we can skip the dependants and quickly get this PR merged. (will address the intel-only failures separately)

[chenrui333]

rebased for the brew style syntax fixes.

[BrewTestBot]

🤖 A scheduled task has triggered a merge.

[chenrui333]

sslyze audit check passed on intel runner (macos-12), bash-3.2$ brew audit --strict sslyze

[chenrui333]

watchman install/test/audit also passed on intel runner (macos-12), bash-3.2$ brew audit --strict watchman

But I did observe watchman needs python@3.11, while sslyze needs python@3.10

[chenrui333]

regression test on lexicon also passed on intel macos-12 runner

==> Installing lexicon
==> Pouring lexicon--3.11.7.monterey.bottle.1.tar.gz
🍺  /usr/local/Cellar/lexicon/3.11.7: 6,335 files, 120.5MB

==> Testing lexicon
==> /usr/local/Cellar/lexicon/3.11.7/bin/lexicon route53 list brew.sh TXT 2>&1
==> /usr/local/Cellar/lexicon/3.11.7/bin/lexicon cloudflare create domain.net TXT --name foo --content bar 2>&1
==> /usr/local/Cellar/lexicon/3.11.7/bin/lexicon --version

[chenrui333]

🎉

[cho-m]

watchman install/test/audit also passed on intel runner (macos-12), bash-3.2$ brew audit --strict watchman

I think watchman failure happens on brew install --build-from-source ... based on CPU intrinsics. I haven't been able to reproduce via Rosetta which could be due to lack of AVX support.

As I recall, brew install --build-from-source has no -march=... on macOS (it uses -march=native on Linux) while brew install --build-bottle has -march=nehalem. It is possible we need some extra logic like:

if !build.bottle? && Tab.for_formula(Formula["folly"]).built_as_bottle && Hardware::CPU.intel?
  ENV["HOMEBREW_OPTFLAGS"] = "-march=#{Hardware.oldest_cpu}"
end

But have no easy way to test this right now and cannot tell if this even works. Trying to directly pass in -march to C*FLAGS will get removed by superenv.

[EricFromCanada]

Security alert: v1.66.1 is out.