strongswan: fix Linux build

[carlocab]

• 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 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 brew install --build-from-source <formula>)? If this is a new formula, does it pass brew audit --new <formula>?

---

These changes are based on feedback from upstream in
strongswan/strongswan#683.

[tobiasbrunner]

The resolve plugin (--enable-resolve) could also be enabled on Linux, it installs DNS servers received from the server and serves a similar purpose as the osx-attr plugin does on macOS.

[carlocab]

Hmm, the Linux build fails with

==> /home/linuxbrew/.linuxbrew/Cellar/strongswan/5.9.4/bin/charon-cmd --version
/home/linuxbrew/.linuxbrew/Cellar/strongswan/5.9.4/bin/charon-cmd: symbol lookup error: /home/linuxbrew/.linuxbrew/Cellar/strongswan/5.9.4/lib/ipsec/libstrongswan.so.0: undefined symbol: XXX

I wonder if we're missing dependencies on Linux.

[tobiasbrunner]

It literally says the missing symbol is XXX?

[carlocab]

Unfortunately, yes. Also, I misspoke. The build succeeded. What failed is trying to run charon-cmd --version. Very odd.

[tobiasbrunner]

Is it possible to build with clang on Linux? The only reference to a similar issue I found is this, and the solution apparently was to build with clang. There is more here: NixOS/nixpkgs#5045 NixOS/patchelf#45

[carlocab]

Yes, it is possible. Though, let's try using a newer GCC first (it pulls in a smaller dep tree). Thanks for the hint!

[carlocab]

Oh wait, the Nix issues probably explain it. We do rely heavily on patchelf on Linux. Let's try building with one of a newer GCC or Clang first. If that doesn't work I'll ping one of our Linux maintainers to have a look.

[SMillerDev]

@Homebrew/linux ideas?

[carlocab]

Looks like patchelf breaks this.

[dawidd6]

==> brew test --verbose strongswan
==> FAILED
==> Testing strongswan
Error: strongswan: failed
An exception occurred within a child process:
  CompilerSelectionError: strongswan cannot be built with any available compilers.
Install Clang or run `brew install gcc`.

Umm. Wut

[carlocab]

Oh right, forgot about that bug. Look at the PR runs from the earlier commits.

[tobiasbrunner]

Yeah, seems to be an issue with our choice of naming a global variable in libstrongswan lib (which is referenced in all binaries of strongSwan), an RPATH that ends with lib, the linker trying to optimize by letting the symbol point at the end of that path (which is also stored in .dynstr) instead of a separate string, and finally patchelf replacing the strings RPATH points to with X. Apparently, it does so without considering that a symbol could point to the replaced content.

[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.