Revert "workflows/tests: set HOMEBREW_RELOCATABLE_INSTALL_NAMES"

[cho-m]

This reverts commit 83ef417.

• Have you followed the guidelines for contributing?
• Have you ensured that your commits follow the commit style guide?
• Have you checked that there aren't other open pull requests for the same formula update/change?
• Have you built your formula locally with HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source <formula>, where <formula> is the name of the formula you're submitting?
• Is your test running fine brew test <formula>, where <formula> is the name of the formula you're submitting?
• Does your build pass brew audit --strict <formula> (after doing HOMEBREW_NO_INSTALL_FROM_API=1 brew install --build-from-source <formula>)? If this is a new formula, does it pass brew audit --new <formula>?

---

This has broken app bundle creation via CMake and Qt, e.g.

• Broken Qt5/6 macdeployqt: qt5: Output of macdeployqt not working after update to 5.15.10_1 #140930
• Broken CMake cpack/fixup_bundle:
  • https://github.com/orgs/Homebrew/discussions/4735
  • wxmaxima 23.08.0 #139981
  • qt5: Output of macdeployqt not working after update to 5.15.10_1 #140930 (comment)

In current state, breakages to common use cases are worse then advantages from enabling relocatable names. I recall various projects using Homebrew for app bundling which may now be negatively impacted.

Until we figure out a way of resolving reported issues (which may require modifying qt/cmake to support rewrites of these modified RPATHs), we should disable relocatable names and fix regressions.

[MikeMcQuaid]

Could this be unset in those specific formulae instead or in brew test-bot where e.g. qt or cmake are in the dependency tree?

[cho-m]

Could this be unset in those specific formulae instead or in brew test-bot where e.g. qt or cmake are in the dependency tree?

I think the issue can impact any formulae that user wants to link in their CMake project and then distribute as app bundle. For example, https://github.com/orgs/Homebrew/discussions/4735 is for harfbuzz's RPATHs which is meson-based formula. The more formulae we bottle with rewritten RPATHs, the more users will hit this.

As I recall, CMake app bundle copies over dylibs into Framework directory and then RPATH rewrites them while copying over recursive dylibs. Seems like Homebrew's rewrites are interfering with CMake's rewrites.

Not sure yet if there is anything that can be done on our side without checking how the CMake/Qt app bundle logic works.

[MikeMcQuaid]

Not sure yet if there is anything that can be done on our side without checking how the CMake/Qt app bundle logic works.

This is something we can do, it's all open source 😁. That said, you've convinced me we should just revert this for now.