Check raw linkage names when bottling

[MikeMcQuaid]

• brew bottle's relocatability checker resolved @rpath and
  @loader_path load commands against the live keg, so
  relocatable-by-construction linkage read back as absolute
  build-prefix paths and wrongly pinned bottles to their build cellar.
• Verified against the fresh abseil rebuild: its googletest helper
  dylibs carry /opt/homebrew/include/gtest/... assertion strings
  that select the files for checking (while the ignores filters
  drop every text match, leaving binary_relocation_files empty),
  after which their @rpath/libabsl_*.dylib load commands resolved
  through LC_RPATH @loader_path to the keg's absolute path and
  pinned the bottle. Only arm64 macOS was affected:
  /usr/local/include/... never byte-matches Intel's
  /usr/local/opt and /usr/local/Cellar search strings and Linux
  performs no linkage check.
• The checker now reads raw load-command names. Raw names lose no
  true positives: a genuinely unplaceholdered name or RPATH contains
  the prefix in its raw bytes and is already caught.
• 127 formulae are pinned on all arm64 macOS tags yet cellar :any
  on x86_64_linux (e.g. abseil, boost, binutils,
  aws-sdk-cpp); once this fix reaches CI they can be re-marked or
  will flip on rebottle.

This change is part of plans/relocatable-bottles.md

---

• Have you followed our Contributing guidelines?
• Have you checked for other open Pull Requests for the same change?
• Have you explained what your changes do? Performance claims (e.g. "this is faster") must include brew benchmark results.
• Have you explained why you'd like these changes included, not just what they do?
• For bug fixes, have you given step-by-step brew commands to reproduce the bug?
• Have you written new tests (excluding integration tests)? Here's an example.
• Have you successfully run brew lgtm (style, typechecking and tests) locally?

---

• I did not use AI/LLM to create this PR, or I disclosed the tool/model below and reviewed its output; I did not attribute commits to AI and will answer maintainer questions and review comments myself without AI/LLM.

Claude Fable 5 with local review and testing.

---

[github-actions]

Thanks for your pull request. This has been closed because it appears to be missing the pull request template, perhaps because this was written by an AI not a human. We require humans to read and fill in these templates.

Please edit this pull request to fill in the current pull request template. This workflow will reopen this pull request automatically once the template is complete. Do not open a new pull request for this.

[Copilot]

Pull request overview

This PR fixes brew bottle’s relocatability checker on arm64 macOS by ensuring Mach-O linkage is evaluated using the raw load-command names (without resolving @rpath / @loader_path against the live keg), preventing relocatable-by-construction linkage from being misinterpreted as absolute build-prefix paths and incorrectly pinning bottles.

Changes:

• Update macOS Keg.file_linked_libraries to call dynamically_linked_libraries(resolve_variable_references: false) so the checker inspects raw load-command names.
• Add an RSpec example verifying raw (unresolved) linkage names are used for the linkage check.
• Update the relocatable-bottles plan and AGENTS.md with clarifications based on the resolved abseil pinning root cause and contribution workflow guidance.

Reviewed changes

Copilot reviewed 4 out of 4 changed files in this pull request and generated 1 comment.

File	Description
Library/Homebrew/extend/os/mac/keg_relocate.rb	Switch linkage checking to raw load-command names to avoid false pinning from resolved @rpath/@loader_path.
Library/Homebrew/test/os/mac/keg_spec.rb	Add a spec to assert linkage checking does not resolve variable references.
Library/Homebrew/plans/relocatable-bottles.md	Document the resolved checker divergence/root cause and update the Phase 1 plan item accordingly.
AGENTS.md	Add a contribution instruction to always use the PR template rather than bypassing it.

---

💡 Add a code-review agent skill for context-aware, tailored reviews. Learn more in the docs.